Table of Contents

Planning Committees Page 863

Abstracts

1 e 4 Plenary Session Pages 864-865LB1 Late-Breaking (LB) Abstract Page 8665 e 12 Oral Abstract Session Pages 866-870100 e 358 Poster Presentations Pages 871-970

100 e 202 Definitive Management of Head and Neck Squamous Cell Carcinoma Pages 871-911203 e 217 Epidemiology and Prevention Pages 911-916218 e 247 Imaging Pages 916-928248 e 250 Immunology and Immunotherapy Pages 928-929251 e 265 Management of Recurrent Head and Neck Squamous Cell Carcinoma Pages 929-935266 e 293 Molecular Biology and Therapeutics Pages 935-946294 e 321 Nonsquamous Cell Malignancies of the Head and Neck: Thyroid, Skin,

Salivary Gland, and Sinus Cancers Pages 946-956322 e 358 Survivorship Pages 956-970

Index of Presenting Authors Pages 971-972

All abstracts accepted for presentation at the 2016 Multidisciplinary Head and Neck Cancer Symposium are embargoed until the opening ceremony of the

symposium, Thursday, February 18, 2016, at 8:00 a.m. Mountain Time.

T A R G E T I N G C A N C E R C A R E

Volume 94 � Number 4 � 2016 863

Planning Committees

STEERING COMMITTEE

AHNSDavid W. Eisele, MD, Johns Hopkins Hospital (Symposium Chair)Cherie-Ann O. Nathan, MD, Louisiana State University Health Sciences Center

ASCORobert I. Haddad, MD, Brigham and Women’s Hospital (Symposium Chair-elect)Merrill S. Kies, MD, MD Anderson Cancer Center

ASTRODavid Raben, MD, University of Colorado Cancer Center (Immediate Past Chair)Quynh L. Le, MD, FASTRO, Stanford School of Medicine

PROGRAM COMMITTEE

AHNSD. Gregory Farwell, MD, FACS, UC Davis Health SystemChristine Gail Gourin, MD, Johns Hopkins University School of MedicineBrian Nussenbaum, MD, Washington University School of Medicine

ASCOEthan Argiris, MD, Hygeia HospitalRenato Martins, MD, MPH, University of WashingtonVictoria Meucci Villaflor, MD, University of Chicago

ASTRONancy Y. Lee, MD, Memorial Sloan Kettering Cancer CenterRandall J. Kimple, MD, PhD, University of Wisconsin School of Medicine and Public HealthBrian O’Sullivan, MD, Princess Margaret Hospital

CONTENT ADVISOR

Oral OncologistBetsy K. Davis, DMD, MS, Medical University of South Carolina

International Journal of

Radiation Oncology

biology physics

www.redjournal.org

K. Kian Ang, MD, PhD, FASTRO, Commemorative Plenary Session

1Molecular Profile of Human PapillomavirusePositiveOropharyngeal Squamous Cell Carcinoma Stratified by SmokingStatusJ.P. Zevallos,1 E. Yim,2 P. Brennan,3 A.Y. Liu,4 J.M. Taylor,2 M. Weissler,1

D. Anantharaman,3 B. Abedi-Ardekani,3 A.F. Olshan,4 and N.N. Hayes5;1University of North Carolina Hospitals, Chapel Hill, NC, 2Department

of Otolaryngology/Head and Neck Surgery, University of North

Carolina at Chapel Hill, Chapel Hill, NC, 3International Agency

for Research on Cancer, Lyon, France, 4Gillings School of Global

Public Health, University of North Carolina at Chapel Hill, Chapel

Hill, NC, 5University of North Carolina School of Medicine, Chapel

Hill, NC

Purpose/Objective(s): HPV-positive oropharyngeal squamous cell carci-

noma (OPSCC) among smokers represents a distinct clinical entity with

intermediate prognosis compared to HPV-positive never smokers and

HPV-negative cases. Despite recent advances in head and neck cancer

genomics, the number of HPV-positive cases evaluated to date has been

modest and the interplay between smoking and HPV has not been fully

evaluated. The purpose of this study is to characterize the mutational

profile of HPV-positive OPSCC by smoking status. We hypothesize a

higher frequency of TP53 and CDKN2A mutations in HPV-positive

OPSCC among heavy smokers.

Materials/Methods: Targeted next-generation sequencing of >800 genes

including all commonly mutated genes in cancer was performed in 66

HPV-positive OPSCC cases stratified by smoking status (<10 pack years

vs >10 pack years). Cases were identified from an NC population-based

epidemiologic study conducted from 2001 to 2006 with follow-up for vital

status. Copy number variation was also examined. Mutation frequency was

compared to previously reported frequencies in the Catalogue of Somatic

Mutations in Cancer database.

Results: Sixty-six HPV-positive OPSCC cases were examined,

including 40 HPV+ OPSCC >10 pack-year and 26 HPV+ <10 pack-

year smokers. Disease-free and overall survival were significantly

better in the <10 pack-year history group. The most commonly

mutated genes in both groups were HLA-A, PIK3CA, and MLL-3.

Several differences in mutation frequency and copy number variation

were noted between <10 and >10 pack-year smokers: TP53, CDKN2A,

KRAS, and NOTCH1 mutations were found almost exclusively among

>10 pack-year smokers and were associated with worse survival, while

HLA-A mutations were more common in the <10 pack-year cohort

(73.1% vs 47.5%, PZ.047).

Conclusion: This study provides a molecular basis for the intermediate

prognosis in patients with dual HPV/tobacco exposure. In addition to

expected tobacco-associated mutations in the >10 pack-year group, we

demonstrate a novel immune signature in the <10 pack-year group. If

validated, these findings could further stratify risk in HPV-positive

OPSCC and provide novel mutational parameters for treatment de-

intensification.

Author Disclosure: J.P. Zevallos: None. E. Yim: None. P. Brennan:

None. A.Y. Liu: None. J.M. Taylor: None. M. Weissler: None.

D. Anantharaman: None. B. Abedi-Ardekani: None. A.F. Olshan:

None. N.N. Hayes: None.

2Systemic Immunologic Effects of Definitive Radiation in Head andNeck CancerV. Sridharan,1 D.N. Margalit,2 S.A. Curreri,2 M. Severgnini,2 F.S. Hodi,2

R.I. Haddad,3 R.B. Tishler,2 and J.D. Schoenfeld2; 1Harvard Medical

School, Boston, MA, 2Dana-Farber Cancer Institute / Brigham and

Women’s Hospital, Boston, MA, 3Dana-Farber Cancer Institute,

Harvard Medical School / Brigham and Women’s Hospital, Boston, MA

Purpose/Objective(s): Radiation therapy (RT) may potentiate antitumor

immunity; however, immune effects of RT are complex and can also be

inhibitory. Therefore, studies evaluating RT’s effects on various compo-

nents of the immune system are needed. Of interest are immune effects of

fractionated RT used in the definitive setting for squamous cell carcinomas

of the head and neck (SCCHN), given the activity of anti-PD-1 immune

checkpoint blockade in this disease. Here, we assess temporal changes in

circulating CD4+ and CD8+ T-effector (CD69+), regulatory T (CD25+

CD127lo), myeloid derived suppressor (MDSC, CD14+ HLA-DR-), and

“exhausted” immune cells (CD4+/CD8+ LAG3+, TIM3+, PD-1+) during

definitive RT in SCCHN patients. We also evaluate levels of circulating

cytokines, soluble PD-L1 (sPD-L1), and potential antitumor antibodies.

Materials/Methods:We prospectively obtained blood from 16 consecutive

SCCHN patients undergoing curative-intent RT. Samples were obtained at

the beginning (week 1) and end (week 6-7) of therapy. Flow cytometry was

used to quantify T-cell subsets. Cytokine and sPD-L1 levels were assessed

using a multiplex platform. We performed proteomic analysis of antibody

responses using a profiling application according to manufacturer’s pro-

tocols. We compared changes in immunologic factors using Wilcoxon

signed rank tests.

Results: Patients received a median 70 Gy for disease in the oropharynx

(nZ12, 75%), nasopharynx (nZ2, 12%), larynx (nZ1, 6%), or oral cavity

(nZ1, 6%). The majority had stage IV disease metastatic to regional lymph

nodes (nZ11, 69%), and received concurrent platinum-based chemotherapy

(nZ13, 81%). During RT, circulating CD8+ T-effector cells increased

(PZ.01), as did CD4+ PD1+ (PZ.02), CD8+ LAG3+ (PZ.02), and regu-

latory T cells (PZ.04). Increases in sPD-L1 levels mirrored increases in

CD8+ T cells over the course of therapy (PZ.047). RTalso decreased levels

of CXCL10 (PZ.004), and increasedCXCL16 (PZ.006).An increase in the

diversity of antigens targeted by antibody responses following treatmentwas

seen in all 3 patients in whom this was evaluated.

Conclusion: Preclinical models suggest RT may lead to antitumor immune

responses that may be blunted by systemic immune regulation. We find

evidence supporting these counteracting effects in SCCHN patients

receiving fractionated RT with or without chemotherapy, where systemic

increases in T-effector cells and potential antitumor antibodies were found

in addition to increases in sPD-L1 and exhausted and regulatory T cells

expressing immune checkpoint receptors. Although the extent to which

these systemic changes reflect changes in the tumor microenvironment is

unknown, our findings support complex immunologic effects of fraction-

ated chemoradiation therapy and mechanisms for potential synergy

between chemotherapy, RT, and immunotherapy in SCCHN.

Author Disclosure: V. Sridharan: None. D.N. Margalit: None.

S.A. Curreri: None. M. Severgnini: None. F. Hodi: None. R.I. Haddad:

None. R.B. Tishler: None. J.D. Schoenfeld: None.

Volume 94 � Number 4 � 2016 Oral Abstract Session 865

3Response-Adapted Volume De-escalation (RAVD) in LocallyAdvanced Head and Neck Cancer: Efficacy and HumanPapillomavirusePositive Subgroup AnalysisJ.M. Melotek,1 V.M. Villaflor,1 T.G. Karrison,1 R.J. Brisson,1 E.A. Blair,1

L. Portugal,1 K.M. Stenson,2 J. De Souza,1 E. Cohen,3 A. Langerman,1

M.T. Spiotto,1 T.Y. Seiwert,1 E.E. Vokes,1 and D.J. Haraf1;1University of Chicago, Chicago, IL, 2Rush University, Chicago, IL,3University of California San Diego, La Jolla, CA

Purpose/Objective(s): Efforts to reduce the late toxicity associated

with chemoradiation (CRT) for locally advanced head and neck squa-

mous cell cancer (LA-HNSCC) have focused on radiation therapy (RT)

dose de-escalation in select populations. In this phase 1/randomized 2

trial investigating the addition of everolimus to induction chemotherapy

(IC), we incorporated a novel response-adapted volume de-escalation

(RAVD) approach using IC response to guide the extent of RT volume

reduction in a nonselected population of patients (pts) with LA-

HNSCC.

Materials/Methods: Pts with measurable LA-HNSCC received 2 cycles

of IC (cisplatin 75 mg/m2, paclitaxel 175 mg/m2 day 1, and weekly

cetuximab, with or without everolimus). Pts with “good” response

(GR), defined as �50% reduction in the sum of gross tumor diameters,

received TFHX2 (paclitaxel, fluorouracil, hydroxyurea, and 1.5 Gy

twice daily RT every other week) to 75 Gy with the planning target

volume (PTV1) encompassing exclusively gross disease. Pts with

<50% response (NR) were treated with volumes encompassing PTV1

and the next nodal station at risk (PTV2) to 45 Gy, followed by a

sequential boost to PTV1 to 75 Gy. Survival rates were estimated by the

Kaplan-Meier method and compared between groups using the log-rank

test.

Results: Ninety-four pts were enrolled: median age 57 (range 27-76)

years, 84% male, 63% HPV+ oropharynx (OPX), 54% �10 pack-year

tobacco use, 56% �T3, 88% �N2b. Everolimus was discontinued on

interim analysis after 43 pts due to futility. IC response was evaluable in 89

pts. Thirty-seven (41.6%) had GR, and 52 (58.4%) had NR. Thirty out of

thirty-seven pts with GR had HPV+ OPX SCC. With mean follow-up of 2

years, there were no significant differences in progression-free survival

(PFS; PZ.086) or overall survival (OS; PZ.94) between GR and NR.

Two-year PFS and OS were 86.0% and 83.5% for GR and 68.7% and

85.4% for NR, respectively. Two-year PFS and OS were 93.1% and 92.1%

for HPV+ OPX GR and 74.0% and 95.2% for HPV+ OPX NR, respec-

tively. There was no statistically significant difference in PFS between the

HPV+ OPX GR and NR groups (PZ.10). There were too few deaths to

provide reliable comparisons for OS between the HPV+ OPX GR and NR

groups.

Conclusion: RAVD is a novel treatment approach that uses IC response to

determine the extent of RT volume reduction. In this study, elimination of

elective nodal coverage did not appear to compromise outcomes in the

entire cohort nor specifically in the HPV+ OPX subgroup. Further inves-

tigation is warranted.

Author Disclosure: J.M. Melotek: None. V.M. Villaflor: Research

Grant; Novartis, Celgene. T.G. Karrison: None. R.J. Brisson: None.

E.A. Blair: None. L. Portugal: None. K.M. Stenson: None.

J. De Souza: None. E. Cohen: Consultant; Bayer, Novartis, VentiRx,

Merck, Bristol-Myers Squibb, AstraZeneca, Eisai. Speaker’s Bureau;

Biodesix, Bayer. A. Langerman: None. M.T. Spiotto: None.

T.Y. Seiwert: Research Grant; Genentech/Roche, Boehringer Ingelheim.

Honoraria; Novartis, Bayer/Onyx, Merck. E.E. Vokes: Consultant;

Abbvie, AstraZeneca, Boehringer Ingelheim, Celgene, Clovis Oncology,

GeneCentric, Genentech, Merck, Synta, VentiRx, Eisai, Lilly, Transgene.

D.J. Haraf: None.

4Proton Beam Reirradiation for Recurrent Head and Neck Cancer:Multi-institutional Report on Feasibility and Early OutcomesJ.H.C. Chang,1 P.B. Romesser,2 E.D. Scher,3 O. Cahlon,2 E.B. Hug,4

K. Sine,4 C.J. DeSelm,2 D. Mah,5 and N. Lee2; 1Northwestern Medicine

Chicago Proton Center, Warrenville, IL, 2Memorial Sloan Kettering

Cancer Center, New York, NY, 3Memorial Sloan Kettering Cancer

Center - Radiation Oncology, New York, NY, 4ProCure Proton Therapy

Center New Jersey, Somerset, NJ, 5ProCure Proton Therapy Center,

Somerset, NJ

Purpose/Objective(s): Most head and neck cancers (HNC) present with

locally advanced disease, with radiation as a treatment component. For

recurrent disease, surgery or reirradiation (re-RT) are the only 2 curative

options. Photon re-RT can offer improved disease-free survival and

locoregional control (LRC) compared to chemotherapy alone, but severe

acute and late toxicities often complicate treatment. The physical prop-

erties of proton beam radiation therapy (PBRT) produce an insignificant

exit dose, which can lead to reduced toxicity in a re-RT setting compared

to photons. We report the first multi-institutional clinical experience

using PBRT for re-RT of recurrent HNC.

Materials/Methods: A retrospective analysis of an ongoing prospective

data registry was conducted. Patients with recurrent HNC and at least 1

prior course of definitive intent external beam RT were included. Acute

and late toxicities were assessed by the National Cancer Institute Com-

mon Terminology Criteria for Adverse Events version 4.0 and by the

Radiation Therapy Oncology Group late radiation morbidity scoring

system, respectively. Six-month LRC, freedom from distant metastasis

(FFDM), and overall survival (OS) were calculated using the Kaplan-

Meier method.

Results: There were 90 patients who received re-RT with PBRT between

2011 and 2014, with median follow-up of 7.0 months (interquartile range

[IQR], 4.4e10.3 months). Median time between PBRT and prior RT was

31.4 months (IQR, 13.4e91.9 months). There were 73 patients with 1

prior RT course (81.1%), and 17 with 2 or more courses (18.9%). Median

PBRT dose was 60.2 CGE (IQR, 50.0e66.0 CGE; maximum, 90.0 CGE).

Actuarial 6-month LRC, FFDM, and OS were 86.5%, 92.8%, and 84.1%,

respectively. Grade 3 acute toxicity included mucositis (10.0%), dermatitis

(3.3%), dysphagia (2.2%), and esophagitis (1.1%). There was 1 death

during PBRT secondary to disease progression. Late toxicity data were

available for 57 out of the 74 patients with >90 days follow-up. There

were 3 (5.3%) grade 3 or greater skin toxicities: a small grade 3 neck ulcer

with plans for future closure, a chronic neck wound successfully managed

with hyperbaric oxygen treatments, and pneumocephalus from a skull base

defect in the setting of recurrent/progressive disease. There was 1 grade 5

toxicity secondary to treatment-related nasopharyngeal or parapharyngeal

bleeding.

Conclusion: PBRT is an attractive option for re-RT of recurrent HNC, with

encouraging disease control and survival rates. Due to the finite range of

protons and subsequent reduced dose to normal tissue, to date, we have

achieved markedly improved toxicity rates compared to photon re-RT.

Early outcomes of PBRT re-RT appear to be associated with tolerable

toxicities. This will guide us in creating a prospective trial with specific

dose parameters (combined dose constraints) to use across multiple proton

institutions.

Author Disclosure: J.H. Chang: Stock; Chicago Proton Therapy, Inc.

Research Director; Northwestern Medicine Chicago Proton Center. P.B.

Romesser: None. E.D. Scher: None. O. Cahlon: Stock; Procure. E.B.

Hug: Director of Clinical Operations; ProCure Proton Therapy Centers. K.

Sine: None. C.J. DeSelm: None. D. Mah: Manage physics department at

the proton center; ProCure Proton Therapy Center, NJ. N. Lee: Manage

clinical operations of the center; Memorial Sloan Kettering Cancer Center.

International Journal of Radiation Oncology � Biology � Physics866

Late-Breaking (LB) Abstract

LB1Determinants of Cost in the Treatment of T1-T3 Oropharynx CancerA.D. Pinheiro1 and R.W. Kramar2; 1Mercy Clinic - Head and Neck Surgery

Springfield, MO, 2Mercy Clinic - Springfield, Springfield, MO

Purpose/Objective(s): Current treatment guidelines for T1-T3

oropharynx squamous cell carcinoma (OPC) offer surgical (Surg) or

nonsurgical (NSurg) treatments as equivalent alternatives. In order to make

value decisions, we must understand cost of care. Our goal was to identify

the primary determinants of cost and we hypothesized that surgical

treatment did not increase cost of care in OPC. To this end, we examined

the relationship of cost to tumor stage, AJCC cancer stage, Charlson age-

comorbidity index (CACI), and treatment strategy (Surg vs NSurg).

Materials/Methods: Retrospective review of patient records in EPIC

identified 299 patients with a diagnosis of oropharynx cancer between July

12, 2011, and May 15, 2015. We excluded patients with tumors that

extended to the oral cavity, those with second primaries or distant me-

tastases, and those whose histology was other than squamous cell carci-

noma (SCC). We identified 71 patients staged T1-T3 who received all their

treatment (S, RT, and/or CRT) at our facility. Cost was defined as revenue

collected by the hospital and clinic for a 6-month episode of care that

started with a biopsy positive for OPC.

Results: A total of 72 patients were available for evaluation. Forty-twowere

treated with Surg and 29 were treated with NSurg. Among the Surg patients,

22 received adjuvant treatment. Of those tested for p16, 92.5% (62/67) were

positive and 4 had unknown p16 status. All 5 p16 negative patients were

treated with Surg. Among the 4 patients with unknown p16 status, 1 was

treated with Surg and 3 with NSurg. Median age was 61 and 62 for the Surg

and NSurg groups (tZ-0.16, PZ.8747). There were no differences between

the Surg andNSurg groups in distribution of T stage (c2Z4.83,PZ.0893) or

AJCC Stage (c2Z6.06, PZ.1946). Comorbidity was higher for the Surg

group (CACIZ8.07) relative to the NSurg group (CACIZ7.34) but this did

not reach significance (tZ-1.36, PZ.1792). Cost was lowest for those

treated with surgery only relative to the NSurg group ($38,462 vs $83,222;

tZ2.26, PZ.0298). Surgery followed by adjuvant CRT had similar cost to

primary CRT (respectively, $84,598 and $83,222; tZ�0.06, PZ.9528)

Conclusion: Surgically treated patients with higher CACI, similar age, and

greater proportion of p16-negative tumors had more favorable cost relative

to those treated with primary CRT. Surgical patients who require adjuvant

CRT had similar cost to those treated with primary CRT. The highest

opportunity for cost savings is in those patients who do not require adju-

vant CRT. Starting with a surgical approach does not increase cost even for

those who require adjuvant treatment. Future research should determine

which treatment strategy yields the best value.

Author Disclosure: A.D. Pinheiro: None. R.W. Kramer: None.

Oral Abstract Session

5Phase 2 Trial of Deintensified Chemoradiation Therapy for Low-Risk Human PapillomaviruseAssociated Oropharyngeal SquamousCell CarcinomaB.S. Chera,1 R.J. Amdur,2 J.E. Tepper,3 B. Qaqish,4 R. Green,5

N.N. Hayes,5 J. Weiss,6 J. Grilley-Olson,5 A. Zanation,1 T. Hackman,1

W. Funkhouser,5 N.C. Sheets,7 M. Weissler,1 and W.M. Mendenhall8;1University of North Carolina Hospitals, Chapel Hill, NC, 2University of

Florida Hospitals, Gainesville, FL, 3University of North Carolina

Hospitals, Chapel Hill, NC, United States, 4University of North Carolina,

Chapel Hill, NC, 5University of North Carolina School of Medicine,

Chapel Hill, NC, 6University of North Carolina, Chapel Hill, NC,7Rex/UNC, Raliegh, NC, 8University of Florida Health Proton Therapy

Institute, Jacksonville, FL

Purpose/Objective(s): We performed a prospective multi-institutional

phase 2 study of a substantial decrease in concurrent chemoradiation

therapy (CRT) intensity as primary treatment for favorable-risk, human

papillomavirus (HPV)-associated oropharyngeal squamous cell carcinoma

(OPSCC).

Materials/Methods: The major inclusion criteria were (1) T0-T3, N0-

N2c, M0; (2) HPV or p16 positive; and (3) minimal/remote smoking

history. Treatment was limited to 60 Gy intensity modulated radiation

therapy (IMRT) with concurrent weekly intravenous cisplatinum (30

mg/m2). The primary study endpoint was pathologic complete response

rate (pCR) based on required biopsy of the primary site and dissection

of pretreatment positive lymph node regions, regardless of radiographic

response. Power computations were performed for the null hypothesis

that the pCR rate is 87% and nZ40, resulting in a type I error of

14.2%. Secondary endpoint measures included physician-reported

toxicity (Common Terminology Criteria for Adverse Events [CTCAE]),

patient-reported symptoms (Patient-Reported Outcomes [PRO]-

CTCAE), quality of life (European Organization for Research and

Treatment of Cancer Quality of Life Questionnaires for cancer

[EORTC QLQ-C30] and head and neck patients [H&N35]), and

penetration aspiration scale (PAS) scores for modified barium swallow

studies.

Results: The study population is 43 patients. The pCR rate was 86%

(37/43). All 6 non-pCR cases were limited to microscopic foci of

residual cancer: 1 primary site and 5 nodal. All patients are alive with

no evidence of disease (median follow-up 21.3 months, range 4-41

months). Thirty-eight patients had a follow-up of at least 1 year. The

incidence of acute CTCAE grade 3/4 toxicity and PRO-CTCAE severe/

very severe symptoms were mucositis 34%/45%, pain 5%/48%, nausea

18%/52%, vomiting 5%/34%, dysphagia 39%/55%, and xerostomia

2%/75%. Grade 3/4 hematological toxicities were 11%. Mean pre- and

6-months-post-CRT EORTC QOL scores were as follows: Global 80/71

(lower worse), Pain (mouth, jaw, throat) 19/21 (higher worse), Swal-

lowing 11/16, Coughing 17/26, Dry Mouth 16/68, and Sticky Saliva 6/

49. Six-months-post CRT mean PRO-CTCAE severity scores for

swallowing and dry mouth were mild and moderate, respectively.

Thirty-nine percent of patients required a feeding tube (none perma-

nently) for a median of 15 weeks (range, 5-22 weeks). There were no

significant differences in PAS scores for thin, pureed, and solid foods

before and after CRT.

Conclusion: In conclusion, pCR rate with decreased intensity of therapy

with 60 Gy of IMRT and weekly low-dose cisplatinum is very high in

favorable-risk OPSCC with evidence of decreased toxicity compared to

standard therapies. (ClinicalTrials.gov, NCT01530997)

Author Disclosure: B.S. Chera: None. R.J. Amdur: None. J.E. Tepper:

None. B. Qaqish: None. R. Green: None. N.N. Hayes: None. J. Weiss:

None. J. Grilley-Olson: None. A. Zanation: None. T. Hackman: None.

W. Funkhouser: None. N.C. Sheets: None. M. Weissler: None.

W.M. Mendenhall: None.

6Detection of Recurrence in Human PapillomaviruseAssociatedOropharynx Squamous Cell CarcinomaJ.M. Frakes, A.O. Naghavi, T. Strom, A.R. Giuliano, L.B. Harrison,

A. Trotti, and J.J. Caudell; H. Lee Moffitt Cancer Center and Research

Institute, Tampa, FL

Purpose/Objective(s): Human papillomavirus (HPV)-associated squa-

mous cell carcinomas of the oropharynx are on the rise with a much higher

likelihood of cure. We reviewed our institutional experience to determine

time to recurrence as well as mode of detection of recurrence to further

help guide optimal follow-up.

Materials/Methods: After institutional review board approval, we queried

an institutional database for patients with HPV- or p16-positive non-

metastatic oropharyngeal cancers treated with definitive radiation therapy

(RT) between 2006 and 2014, and 246 cases were identified. Charts were

reviewed and patient, tumor, and treatment factors as well as mode of

Volume 94 � Number 4 � 2016 Oral Abstract Session 867

detection of recurrence were abstracted. Patients either received definitive

RT alone (nZ38, 15.4%) or concurrent systemic therapy and RT (nZ209,

84.6%). Outcomes, including local control (LC), regional control (RC),

locoregional control (LRC), freedom from distant metastases (FFDM), and

overall survival (OS) were calculated according to Kaplan-Meier method

from the end of RT. Patients received a 3-month posttreatment positron

emission tomographic/computed tomographic (PET/CT) scan and were

seen every 3 months in the first year, every 4 months in Year 2, and every 6

months in Years 3 to 5.

Results: Median follow-up of all patients was 36 months. LC was

achieved in 239 of 245 patients, with a 3-year LC rate of 97.8%. Of

the 6 local failures, all were detected by direct visualization (nZ2) or

flexible laryngoscopy (nZ4). Three-year RC was 95.3%, where pa-

tients with �5 nodes or level 4 lymph nodes present were more likely

to suffer regional failure (P<.05). Of the 9 regional recurrences, 89%

(nZ8) were found by symptoms or 3-month posttreatment PET/CT.

Three-year LRC was 94%. Of the 13 patients that suffered a

locoregional failure, 92% (nZ12) presented with either symptoms or

persistent disease on 3-month posttreatment PET/CT. Three-year

FFDM rate was 91.4%, with increased risk of metastases occurring in

patients with a lymph node greater than 6 cm, bilateral lymphade-

nopathy, 5 or more nodes, or if a lymph node was present in level 4

(P<.05). Of the 21 patients who suffered distant recurrence, 71%

(nZ15) were found due to symptoms or 3-month posttreatment im-

aging. Three-year OS was 91% for all patients. Late grade �3 toxicity

occurred in 21 patients (9%), with 19 being grade 3 toxicities and 2

grade 4 toxicities. The majority of toxicity and/or failure

occurred within the first 6 months (64%), with only 4 events beyond 2

years.

Conclusion: HPV-associated oropharyngeal cancer treated with

definitive RT has excellent outcomes. Of the few patients that suf-

fered a local failure, all were identified with physical exam.

Symptoms and/or 3-month posttreatment PET/CT identified 92% of

locoregional failures and 71% of distant failures. Given these find-

ings, if posttreatment of PET/CT is negative, no further imaging is

warranted. Follow-up should include history and physical exam with

direct visualization.

Author Disclosure: J.M. Frakes: None. A.O. Naghavi: None. T. Strom:

None. A.R. Giuliano: None. L.B. Harrison: None. A. Trotti: None.

J.J. Caudell: None.

7Final Results of a Randomized Phase 2 Trial Investigating theAddition of Cetuximab to Induction Chemotherapy and Acceleratedor Hyperfractionated Chemoradiation Therapy for LocoregionallyAdvanced Head and Neck Cancer: HPV-negative Subset AnalysisJ.M. Melotek,1 D.J. Haraf,1 E.A. Blair,1 M.E. Witt,1 R.J. Brisson,1

A. Dekker,1 M. Lingen,1 M. Kocherginsky,1 V.M. Villaflor,1

E. Cohen,2 E.E. Vokes,1 and T.Y. Seiwert1; 1University of

Chicago, Chicago, IL, 2University of California San Diego,

La Jolla, CA

Purpose/Objective(s): Cetuximab improves survival when added to

cytotoxic chemotherapy in the metastatic disease setting and when added

to radiation therapy (RT) in the curative intent setting. In this phase 2

randomized study, we investigated the addition of cetuximab to both in-

duction chemotherapy (IC) and hyperfractionated or accelerated chemo-

radiation therapy (CRT) for patients with locoregionally advanced head

and neck squamous cell cancer (LA-HNSCC). Outcomes of poor prognosis

HPV-negative patients are presented herein.

Materials/Methods: Patients with LA-HNSCC were randomized to

receive 2 cycles of weekly IC (cetuximab, paclitaxel, carboplatin) and

either Cetux-FHX (concurrent cetuximab, 5-FU, hydroxyurea, and 1.5 Gy

twice-daily RT every other week to 75 Gy) or Cetux-PX (cetuximab,

cisplatin, and accelerated RT with delayed concomitant boost to 72 Gy in

42 fractions). All patients were treated with image guided intensity

modulated RT. Progression-free survival (PFS) and overall survival (OS)

rates were estimated by the Kaplan-Meier method and compared between

groups using the log-rank test. Cumulative incidences of locoregional

failure (LRF) and distant failure (DF) were estimated using the competing

risks approach.

Results: One hundred ten patients were enrolled from September 2006 to

April 2010. HPV status was assessed in 103 patients, of whom 56 were

HPV negative. Patient characteristics of this subset include: median age 57

years (range 32-81); 84% male; 66% �10 pack-year tobacco use; 98%

stage IV (61% �T3, 77% �N2b, 54% T4 or N3). Thirty-two patients

received Cetux-FHX, and 24 patients received Cetux-PX. With median

follow-up of 68 months, there were no significant differences in PFS

(PZ.57) or OS (PZ.43) between treatment arms. Two-year PFS was

75.0% for Cetux-FHX and 70.8% for Cetux-PX. Late outcomes for the

entire HPV-negative cohort include 5-year PFS, OS, LRF, and DF rates of

65.9%, 72.5%, 21.7%, and 7.2%, respectively.

Conclusion: The addition of cetuximab to IC and hyperfractionated or

accelerated CRT produced long-term control of LA-HNSCC among

poor-prognosis HPV-negative patients. Further investigation of cetuximab

may be warranted in the neoadjuvant setting and with nonplatinum-based

CRT.

Author Disclosure: J.M. Melotek: None. D.J. Haraf: None. E.A. Blair:

None. M.E. Witt: None. R.J. Brisson: None. A. Dekker: None.

M. Lingen: None. M. Kocherginsky: None. V.M. Villaflor: Research

Grant; Novartis, Celgene. E. Cohen: Consultant; Bayer, Novartis,

VentiRx, Merck, Bristol-Myers Squibb, AstraZeneca, Eisai. Speaker’s

Bureau; Biodesix, Bayer. E.E. Vokes: Consultant; Abbvie, AstraZeneca,

Boehringer Ingelheim, Celgene, Clovis Oncology, GeneCentric,

Genentech, Merck, Synta, VentiRx, Eisai, Lilly, Transgene. T.Y. Seiwert:

Research Grant; Genentech/Roche, Boehringer Ingelheim. Honoraria;

Novartis, Bayer/Onyx, Merck.

8A Phase 1b Study of Neoadjuvant Immune Biomarker ModulationWith Cetuximab and Motolimod in Head and Neck Cancer (HNC)R.L. Ferris,1 B.A. Kansy,2 S.P. Gibson,2 R.M. Srivastava,2 G. Shayan,3

J.K. Bryan,4 and R.M. Hershberg4; 1University of Pittsburgh

Medical Center, Pittsburgh, PA, 2University of Pittsburgh Cancer

Institute, Pittsburgh, PA, 3University of Pittsburgh; University of

Pittsburgh Cancer Institute, Pittsburgh, PA, 4VentiRx Pharmaceuticals,

Seattle, WA

Purpose/Objective(s): Inflamed tumors are known to generate a better

response to immunotherapy. HNC has many immunosuppressive

features, including impaired natural killer cell (NK) activity. The

monoclonal antibody cetuximab (CTX) is effective in a subset of

HNC patients, with clinical activity linked to NK-mediated antibody-

dependent cellular cytotoxicity (ADCC). Recent data show that CTX

increases the frequency of intratumoral CTLA-4+FoxP3+ regulatory T

cells (Treg), which suppress ADCC and are associated with poor

clinical outcome. In ex vivo experiments, this effect could be

attenuated by targeting CTLA-4 on Tregs. It is possible that the

clinical efficacy of CTX may be improved by enhancing tumor

inflammation, activating the immune system, and/or inhibiting the

Treg suppressive effects. Motolimod (VTX-2337), a novel Toll-like

receptor 8 (TLR8) agonist. Preclinical data show enhanced CTX and

NK-mediated lysis of HNC cells and dendritic cross-priming of

EGFR-specific CD8+ T cells. CTX and motolimod in HNC patients

was tolerable and active in a phase 1b study, with enhanced NK cell

activity. The central hypothesis in this study is that NK and mono-

cyte/mDC activation by CTX is enhanced by concomitant adminis-

tration of motolimod, thereby amplifying the innate and adaptive

immune response in the circulation and in the tumor microenviron-

ment (TM).

Materials/Methods: In this prospective phase 1b clinical trial

(NCT02124850) of preoperative treatment with CTX and motolimod,

the primary objective is to evaluate how neoadjuvant CTX plus

International Journal of Radiation Oncology � Biology � Physics868

motolimod modulates innate and adaptive immune biomarkers. An

exploratory objective of the study is to assess whether this modulation

of biomarkers can predict antitumor response. Subjects (nZ12) with

stage IIeIVA HNC received weekly doses of CTX and motolimod

followed by definitive surgical resection. Biomarker modulation in

tumor and blood are correlated with clinical response by CT or MRI.

Tumor apoptosis/proliferation is assessed by biopsy pre- and

posttreatment.

Results: Phenotypic markers of suppression in Treg were reduced when

motolimod was combined with CTX, including lower levels of CTLA-4

(PZ.01), CD73 (PZ.04), and membrane-bound TGF-b (PZ.05). These

mediators were induced in nonresponders in our single-agent cetuximab

therapy study (PZ.005). Myeloid antigenepresenting cells achieved

strong maturation and expression of stimulatory markers in the peripheral

blood.

Conclusion: CTX plus motolimod induces inflammatory changes in the

TM and peripheral blood. Treg-suppressive mediators are reduced with

CTX/motolimod, overcoming a negative prognostic biomarker (Treg

induction) of CTX therapy alone. A pending amendment will assess

the impact of adding checkpoint inhibition to the CTX/motolimod

combination.

Author Disclosure: R.L. Ferris: Research Grant; Bristol Myers

Squibb, AZ/Medimmune, VentiRx Pharmaceuticals. Advisory Board;

Bristol Myers Squibb, AZ/Medimmune, Merck, Celgene. B.A.

Kansy: None. S.P. Gibson: None. R.M. Srivastava: None. G.

Shayan: Visiting Scholar; University of Pittsburgh Cancer Institute.

J.K. Bryan: Employer; VentiRX Pharmaceuticals. Stock Options;

VentiRx Pharmaceuticals. Clinical leadership; VentiRx Pharmaceuti-

cals. R.M. Hershberg: Employer; VentiRx Pharmaceuticals. Stock

Options; VentiRx Pharmaceuticals. Leadership; VentiRx Pharmaceu-

ticals.

9A Phase 1 Trial of Pazopanib Added to Cetuximab in PatientsWith Incurable Head and Neck Squamous Cell Carcinoma(HNSCC)D. Adkins, J. Ley, T. Wildes, K. Trinkaus, M. Siegel, and

L. Michel; Washington University School of Medicine,

St. Louis, MO

Purpose/Objective(s): Up-regulation of angiogenesis is a frequent event

in head and neck squamous cell carcinoma (HNSCC) and a mechanism of

resistance to inhibitors of the epidermal growth factor receptor. Pazopanib

is an inhibitor of angiogenesis through multiple targets, including vascular

endothelial growth factor receptors. The primary objective of this trial was

to determine the safety of adding pazopanib to cetuximab in patients with

incurable HNSCC. A suspension formulation of pazopanib was chosen

due to swallowing problems or gastrostomy tube dependency in these

patients.

Materials/Methods: A phase 1 trial using Fibonacci (3+3) design was

performed to determine the dose-limiting toxicity (DLT) and

maximum tolerated dose (MTD) of pazopanib added to fixed doses of

cetuximab in patients with incurable HNSCC. Pazopanib was admin-

istered daily as an oral suspension in cycles of 8 weeks: dose levels 1

(200 mg/d; starting), 2 (400 mg/d), 3 (600 mg/d), and 4 (800 mg/d;

maximum) with weekly fixed, standard doses of cetuximab. Tumor

response was evaluated using RECIST 1.1 criteria after each 8-week

cycle. Cetuximab naıve and cetuximab refractory patients were

eligible to participate.

Results: Twenty-two patients were enrolled: 3 on dose level 1, 6 on level

2, 7 on level 3, and 6 on level 4. DLT events (all grade 3) occurred at dose

levels 2 (neutropenia, nZ1), 3 (proteinuria, nZ1), and 4 (fatigue, nZ1).

An MTD of pazopanib was not reached at the maximum dose level. Best

overall tumor response: 6 partial (PR, 27%), 11 stable (SD, 50%), and 5

progression (23%). Three of the 6 PRs occurred in patients with cetux-

imab- and platin-resistant disease. Median overall time-to-progression

(TTP) was 112 (range: 21-280) days, and was 140 (112-280) days in

patients with disease control (PR + SD) and 56 (21-56) days in patients

with progressive disease as best overall tumor response. Disease control

occurred in 9 of 10 patients with cetuximab- and platin-resistant disease, 2

of 2 patients with cetuximab-resistant disease, 1 of 4 patients with platin-

resistant disease, and 5 of 6 patients with platin- and cetuximab-naıve

disease.

Conclusion: This trial, the first to evaluate pazopanib in HNSCC, deter-

mined the pazopanib suspension dose recommended for phase 2 trials was

800 mg/day in combination with cetuximab. The tumor response rate of

27%, median overall TTP of 112 days, and disease control rate of 90% in

patients with cetuximab- and platin-resistant disease demonstrate that the

combination of pazopanib and cetuximab have impressive efficacy in

incurable HNSCC.

Author Disclosure: D. Adkins: None. J. Ley: None. T. Wildes: None.

K. Trinkaus: None. M. Siegel: None. L. Michel: None.

10Tumor Biomarker Association With Clinical Outcomes in Recurrentand/or Metastatic Head and Neck Squamous Cell Carcinoma (R/MHNSCC) Patients (Pts) Treated With Afatinib Versus Methotrexate(MTX): LUX-Head & Neck 1 (LHN1)E. Cohen,1 L. Licitra,2 B. Burtness,3 J. Fayette,4 T. Gauler,5

P.M. Clement,6 J.J. Grau,7 J.M. del Campo,8 A. Mailliez,9 R.I. Haddad,10

J.B. Vermorken,11 M. Tahara,12 J. Guigay,13 L. Geoffrois,14

M.C. Merlano,15 X.J. Cong,16 N. Gibson,17 F. Solca,18 E. Ehrnrooth,19

and J.P.H. Machiels20; 1University of California San Diego, La Jolla, CA,2Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy,3Yale University School of Medicine, New Haven, CT, 4Leon Berard Center

and Hospices Civils de Lyon, University of Lyon, Lyon, France,5West German Cancer Center, University Hospital Essen, Essen, Germany,6Department of Oncology, KU Leuven, Leuven, Belgium,7Hospital Clınic de Barcelona, University of Barcelona, Barcelona, Spain,8Medical Oncology Department, Hospital Universitario Vall D’Hebron,

Barcelona, Spain, 9Centre Oscar Lambret, Lille, France,10Department of Medical Oncology, Dana-Farber Cancer Institute,

Harvard Medical School and Department of Medicine, Brigham and

Women’s Hospital, Boston, MA, 11Antwerp University Hospital, Edegem,

Belgium, 12National Cancer Center Hospital East, Kashiwa, Japan,13Gustave Roussy, Villejuif and Centre Antoine Lacassagne, Nice, France,14Institut de Cancerologie de Lorraine, Vandoeuvre-les-Nancy, France,15Azienda Ospedaliera Santa Croce e Carle, Cuneo, Italy,16Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, CT,17Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany,18Boehringer Ingelheim RCV GmbH & Co. KG, Vienna, Austria,19Boehringer Ingelheim, Danmark A/S, Copenhagen, Denmark,20Institut Roi Albert II, Service d’Oncologie Medicale, Cliniques

Universitaires Saint-Luc and Institut de Recherche Clinique et

Experimentale (Pole MIRO), Universite Catholique de Louvain,

Brussels, Belgium

Purpose/Objective(s): In the phase 3 LHN1 trial, afatinib significantly

improved progression-free survival (PFS; median 2.6 vs 1.7 months;

hazard ratio [HR] 0.80; PZ.03) versus MTX as second-line treatment in

R/M HNSCC pts. Because afatinib selectively targets ErbB dysregulations,

we evaluated the association of prespecified ErbB-related biomarkers and

p16 on the antitumor activity of afatinib versus MTX.

Materials/Methods: Optional tumor biomarker assessments, including p16

by immunohistochemistry (IHC), EGFR amplification (FISH), HER3

(IHC), and PTEN (IHC) were conducted in a central laboratory combining

tissue microarray (TMA) cores and slides. Subgroups combining bio-

markers of interest were also analyzed.

Results: Concordance rates for TMAs and slides were 66% to 78% for the

biomarkers tested. Improvements in PFS and overall survival (OS) were

observed with afatinib versus MTX in p16-negative, EGFR-amplified,

HER3-low, and PTEN-high disease (Table 1). Objective response rate

(ORR) was notably improved with afatinib versus MTX in EGFR-tar-

geted therapy-naıve pts with p16-negative tumors (27.5% vs 4.8%) and

pts with both p16-negative and EGFR-amplified disease (15.5% vs 0%).

Abstract 10; Table 1 Efficacy outcomes according to biomarkers (all comparisons are afatinib vs MTX)

Subgroup Median PFS, mos PFS HR (95% CI) Median OS, mos OS HR (95% CI) ORR, %

p16-neg* (H-score <210), nZ199p16-pos* (H-score �210), nZ35

2.7 vs 1.62.0 vs 2.3 0.70 (0.50, 0.97)0.81 (0.39, 1.69)

6.7 vs 6.49.5 vs 13.0 1.22 (0.88, 1.68)2.08 (0.87, 4.99)

14.1 vs 1.60 vs 8.3

EGFR-amplifiedy, nZ101EGFRnot amplified, nZ105

2.7 vs 1.41.7 vs 2.4 0.53 (0.32, 0.88)1.03 (0.67, 1.61)

6.8 vs 4.75.6 vs 6.8 0.81 (0.50, 1.32)1.42 (0.91, 2.20)

11.8 vs 05.7 vs 0

HER3-low (H-score �50), nZ91HER3-high (H-score >50), nZ88

2.9 vs 2.11.6 vs 2.4 0.47 (0.27, 0.80)1.41 (0.83, 2.38)

9.8 vs 6.46.0 vs 7.3 0.95 (0.57, 1.57)1.26 (0.78, 2.04)

10.5 vs 09.7 vs 0

PTEN-high (H-score >150), nZ49PTEN-low (H-score �150), nZ137

2.9 vs 1.42.7 vs 2.5 0.44 (0.21, 0.95)0.86 (0.57, 1.28)

8.9 vs 4.47.1 vs 8.5 0.56 (0.29,1.08)1.07 (0.73, 1.57)

8.6 vs 011.5 vs 0

Abbreviations: HR Z hazard ratio; H-score Z histology score; OS Z overall survival; PFS Z progression-free survival; pos Z positive;

neg Z negative.

*p16 staining was analyzed in tumors from all subsites. p16 status was unknown for 249 patients.y�50% of cells with �4 copies, or �1 cell with

�8 copies.

Volume 94 � Number 4 � 2016 Oral Abstract Session 869

No objective responses were observed in the 35 pts with p16-positive

disease.

Conclusion: In this analysis, more pronounced effects on survival were

observed with afatinib versus MTX in R/M HNSCC pts with p16-negative,

EGFR-amplified, HER3-low, and PTEN-high disease. Future prospective

studies are warranted to provide a more robust readout of clinical outcomes

with afatinib in these pts.

Author Disclosure: E. Cohen: Consultant; Merck, Pfizer, Eisai,

Bayer. L. Licitra: None. B. Burtness: Honoraria; VentiRx, Johnson

and Johnson, Merck, Amgen, Medimmune. Consultant; VentiRx,

Johnson and Johnson, Merck, Amgen, Medimmune. Study investi-

gator; Merck, Genentech, Koltan Pharmaceuticals. J. Fayette: None.

T. Gauler: Honoraria; Boehringer Ingelheim, Merck Serono,

Novartis, BMS, MSD. Consultant; Boehringer Ingelheim, Merck

Serono, Novartis, BMS. Speaker’s Bureau; Boehringer Ingelheim,

Merck Serono, Novartis. Advisory Board; Boehringer Ingelheim,

Merck Serono, Novartis, BMS, MSD. Travel Expenses; Boehringer

Ingelheim, Merck Serono, BMS, MSD. Stock; BAYER AG. P.M.

Clement: None. J.J. Grau: None. J.M. del Campo: None.

A. Mailliez: None. R.I. Haddad: Research Grant; Boehringer

Ingelheim, Merck, BMS, Celgene, AstraZeneca. Consultant; Boeh-

ringer Ingelheim, Merck, BMS, Celgene, AstraZeneca. J.B. Ver-

morken: Advisor; Merck Serono, Boehringer Ingelheim,

Vaccinogen, Synthon, Innate Pharma. Consultant; Merck Serono.

Speaker’s Bureau; Merck Serono, Vaccinogen. Advisory Board;

Merck Serono, Vaccinogen. Travel Expenses; Merck Serono, Vac-

cinogen. M. Tahara: Research Grant; Eisai, Merck Sharp & Dohme,

AstraZeneca. Honoraria; Merck Serono, Bristol-Myers Squibb, Eisai,

Otsuka, Bayer. Advisory Board; Merck Sharp & Dohme. J. Guigay:

Research Grant; BMS, Merck Serono, Boehringer Ingelheim, Chu-

gai, GSK, MSD, Sanofi. Advisory Board; BMS, Merck Serono. L.

Geoffrois: None. M.C. Merlano: None. X.J. Cong: None. N.

Gibson: None. F. Solca: None. E. Ehrnrooth: Clinical Programme

Lead at Boehringer Ingelheim; Boehringer Ingelheim. J.H.

Machiels: Research Grant; Novartis, Bayer, Janssen. Advisory

Board; MSD. Participation in advisory board (non-compensated);

Boehringer Ingelheim.

Abstract 11; Table 1

SOM (Grades 3 and 4)

Comparative historical control data

Matched data set

Placebo datafrom phase 3 trials

(JCO, 2011)

3S. Sonis Le et al Henke et al

# Patients z300 94 94Incidence through 60 Gray 60% 57% 62%Intensity: Grade 4 overall incidence 20% 19% 26%Onset: Median days 28 35 32Duration: Median days 28 26 22

11Phase 1b/2a Trial of Superoxide (SO) Dismutase (SOD)Mimetic GC4419 to Reduce Chemoradiation TherapyeInduced OralMucositis (OM) in Patients With Oral Cavity or OropharyngealCarcinoma (OCC)C.M. Anderson,1 B.G. Allen,1 W. Sun,1 C.M. Lee,2 S. Agarwala,3

M. Venigalla,4 L. Greenberg,5 D. Adkins,6 Y. Chen,7 W. Zhen,8 D. Mould,9

J. Holmlund,10 J. Brill,10 S. Sonis,11 and J. Buatti1; 1University of Iowa

Hospitals & Clinics, Iowa City, IA, 2Cancer Care Northwest, Spokane, WA,3St Luke’s Medical Center, Easton, PA, 4Lakeland Regional Health

Cancer Center, Lakeland, FL, 5Allegheny General Hospital, Pittsburgh, PA,6Washington University School of Medicine, St. Louis, MO,7Wilmot Cancer Institute, University of Rochester, Rochester, NY, 8University

of Nebraska Medical Center, Omaha, NE, 9Projections

Research Inc., Phoenixville, PA, 10Galera Therapeutics, Inc., Malvern, PA,11BioModels, Boston, MA

Purpose/Objective(s): Seventy percent of patients (pts) with OCC

develop severe OM (SOM) after a median 40-Gy radiation therapy (RT)

treatment plus concurrent cisplatin (CDDP). There are no approved in-

terventions to reduce the incidence or severity of OM. RT-induced SO is a

key initiator of OM. GC4419’s (MWZ483) specific mimicry of SOD’s

dismutation of SO to H2O2 mitigated SOM in a translational hamster

model. We hypothesized that GC4419 immediately prior to intensity

modulated RT (IMRT) fractions would reduce incidence, severity, dura-

tion, and delay onset, of SOM.

Materials/Methods: Serial 3 to 6 pt cohorts with locally advanced, non-

metastatic OCC were planned for definitive or post-op IMRT to approxi-

mately 70 Gy total (>50 Gy to � 2 oral sites), 1.8 to 2.2 Gy/fx M-F plus

CDDP; plus escalating doses of GC4419, over 60-min IV, M through F for

3 to 7 weeks, ending <60 min before IMRT. WHO grade OM was assessed

twice weekly. GC4419 was measured in plasma. OM-related circulating

cytokines (Cy) and changes in gene expression were assayed prior to and

during treatment.

Results: Forty-six pts received GC4419; 43 were evaluable for OM at dose

levels/d: 15 to 112 mg x 3 wks (nZ20); 112 mg x 4 to 5+ wks (nZ9); 30

GC4419 treatment

weeks (15-112 mg/d) <6 weeks (15-112 mg/d) 6-7 weeks (30 or 90 mg/d)

20 29 1440% 41% 28%25% 31% 0%>54 >50 >504.5 9 2.5

Abstract 12; Table 1 Adjusted outcomes for head and neck cancer patientsaccording to health insurance status

Insured Medicaid Uninsured

HR 95% CI HR 95% CI HR 95% CI

Overall mortality 1.00 – 1.55 1.49-1.62 1.48 1.38-1.58Cause-specific mortality 1.00 – 1.60 1.55-1.69 1.65 1.52-1.79

Abbreviations: CI Z confidence interval; HR Z hazard ratio.

International Journal of Radiation Oncology � Biology � Physics870

mg x 7 wks (nZ4); 90 mg x 6 weeks (nZ4); 90 mg x 7 wks (nZ6).

Efficacy: SOM duration, incidence through 6 wks/60 Gy, severity, and

onset appeared markedly improved from historical controls (Table 1);

overall SOM incidence (50% vs 70%) was also reduced. GC4419 efficacy

was related to treatment duration (Table 1), but not absolute daily dose.

Safety: A true maximum tolerated dose was not reached. Gr 3 gastroen-

teritis and Gr 3 nausea/vomiting (1 each, 112 mg/d x 3 or 6 wks) were

considered dose-limiting. Dose-related Gr 1-2 peri-infusional facial

tingling was attributable to GC4419. Other AEs were consistent with

known effects of IMRT/CDDP. PK: Dose-related Cmax and AUC; plasma

t1/2 approximately 2 hours. Biomarkers: Cy levels were associated with

GC4419 dose and WHO severity, consistent with known OM pathogenesis.

Conclusion: When GC4419 was given throughout IMRT, SOM was less

frequent, briefer, delayed, and less severe than expected. GC4419-related

toxicity was mild-to-moderate and acceptable in combination with IMRT/

CDDP. Doses of 30 and 90 mg/d were chosen for a future randomized,

placebo-controlled trial.

Author Disclosure: C.M. Anderson: Employee; University of Iowa Hos-

pitals & Clinics. Uncompensated, involved in design of randomized Phase

2 trial of GC4419; Galera Therapeutics, Inc. B.G. Allen: Physician sci-

entist, employed and self-funded through grants; University of Iowa

Hospitals & Clinics. Research Grant; ASTRO. W. Sun: None. C.M. Lee:

Partner; Cancer Care Northwest, Gamma Knife of Spokane. Speaker’s

Bureau; Bayer Pharmaceuticals, Bristal Myers Squibb. Partnership;

Kobold Medical Devices, Provenir Publishing. Committee Member; Sa-

cred Heart Medical Center. S. Agarwala: None. M. Venigalla: None.

L. Greenberg: None. D. Adkins: None. Y. Chen: None. W. Zhen: None.

D. Mould: President; Projections Research Inc. Consultant; Projections

Research Inc. J. Holmlund: Independent Contractor; Galera Therapeutics,

Inc. Consultant; Baxalta, Prometheus Labs. Stock Options; Galera Ther-

apeutics, Inc. Ad-hoc reviewer; Aspire IRB. J. Brill: Employee; Incyte

Corporation. Stock Options; Galera Therapeutics, Inc, Incyte Corporation.

S. Sonis: Senior Surgeon; Brigham and Women’s Hospital. Consultant;

Clinical Assistance Programs. Stock; Inform Genomics. Partnership;

BioInsight Diagnostics. Advisor; Galera. Board Member; Inform Geno-

mics. J. Buatti: Research Grant; NIH.

12The Impact of Health Insurance Status on the Presentation, LocalManagement, and Outcomes of Patients With Head and Neck Cancerin the United StatesT.M. Churilla,1 B. Egleston,1 Y. Dong,2 M. Lango,1 and T.J. Galloway1;1Fox Chase Cancer Center, Philadelphia, PA, 2Fox Chase Cancer Center,

Cheltenham, PA, United States

Purpose/Objective(s): We sought to evaluate the association between

health insurance status and stage, treatment, and outcomes among patients

with head and neck (H&N) cancer. We hypothesized that patients with

Medicaid or lack of health insurance more frequently present with

advanced disease, undergo less surgery or radiation, and have increased

cancer-specific mortality compared to patients with non-Medicaid

insurance.

Materials/Methods: We queried the National Cancer Institute Surveil-

lance, Epidemiology, and End Results database for primary squamous

cell carcinoma of the oral cavity, pharynx, and larynx from 2007 to

2012. We characterized clinical and demographic variables according to

insurance status (insured vs Medicaid vs uninsured). We tested for

associations between patient insurance status and American Joint

Committee on Cancer (AJCC) stage, receipt of cancer-directed surgical

procedure (� wide local excision), and receipt of external beam radi-

ation therapy. We calculated odds ratios (ORs) and computed Pearson

X2 test and used multiple logistic regression analysis to adjust for

clinical and demographic covariates. We evaluated cancer-specific

mortality according to insurance status by the Kaplan-Meier method

and adjusted for demographic and clinical information using Cox

regression.

Results: A total of 53,848 patients were analyzed: 80.1% insured, 15.0%

with Medicaid, and 4.9% uninsured. AJCC stage III or IV disease was

more common among patients with Medicaid (72.9%) and uninsured

patients (75.1%) compared to insured patients (60.1%), P<.001. After

adjustment for site, stage, use of radiation, age, race, location, education,

and income, uninsured patients were less likely to receive cancer-

directed surgery (OR [95% CI] Z 0.86 [0.77-0.97]). Similarly, after

adjustment for site, stage, cancer-directed surgery, age, race, location,

education, and income, patients with Medicaid and uninsured patients

were less likely to receive external beam radiation therapy (OR [95% CI]

Z 0.77 [0.72-0.81] and 0.68 [0.62-0.75], respectively). Patients with

Medicaid or uninsured status had inferior outcomes after adjustment for

surgery, radiation therapy, tumor, and demographic characteristics, as

seen in Table 1.

Conclusion: Patients with Medicaid or uninsured status frequently pre-

sented with advanced-stage H&N cancer in the United States and were

less likely to undergo cancer-directed surgery or radiation therapy.

Overall and cancer-specific mortality were increased among Medicaid

and uninsured patients after adjustment for clinical and treatment

characteristics suggesting additional barriers to care or associated risk

factors.

Author Disclosure: T.M. Churilla: None. B. Egleston: None. Y. Dong:

None. M. Lango: None. T.J. Galloway: None.

Volume 94 � Number 4 � 2016 Posters 871

Posters

100Does Age Matter? Survival Outcomes With the Addition ofConcurrent Chemotherapy for Elderly Head and Neck CancerPatients Undergoing Definitive Radiation Using the NationalCancer Data BaseA. Amini,1 B. Jones,1 J. McDermott,2 A. Jimeno,2 D.W. Bowles,2

D. Raben,1 and S. Karam1; 1Department of Radiation Oncology,

University of Colorado Denver, Aurora, CO, 2Department of Medical

Oncology, University of Colorado Denver, Aurora, CO

Purpose/Objective(s): The addition of chemotherapy to radiation (CRT)

for head and neck squamous cell carcinomas (HNSCC) improves overall

survival (OS) compared to radiation therapy (RT) alone; however, the

Pignon meta-analysis of over 17,000 patients demonstrated no OS benefit

in patients >70 years of age. Because of this, many elderly patients receive

RT alone. Using a nationwide database, this study examines the outcomes

of elderly patients receiving CRT versus RT alone.

Materials/Methods: The National Cancer Data Base (NCDB) was queriedfor patients >70 years of age with nonmetastatic oropharynx, larynx, and

hypopharynx cancers treated from 1998 to 2011. Patients received defin-

itive RT (66-81.6 Gy in 1.2-2.0 Gy fractions); CRT was defined as

chemotherapy started within 14 days of the beginning of RT. Multivariate

(MVA) and propensity score-matched (PSM) analyses were performed to

compare OS outcomes. Recursive partitioning analysis (RPA) based on OS

using age, Charlson comorbidity score, T stage, and N stage was also

performed.

Results: A total of 5,265 patients were included: 3,604 (68%) received RT

alone, and 1,661 (32%) received CRT. Median follow-up was 31 mo (2-181

mo). Median age of patients undergoing RT alone was 77 years (71-90

years); median age of CRT patients was 75 years (71-90 years). When ac-

counting for age, gender, race,median county household income, percentage

without a high school diploma, comorbidity index, facility, tumor site, T

stage, and N stage, CRT improved OS under MVA (hazard ratio [HR], 0.84;

95% confidence interval [CI], 0.76-0.94; PZ.002) and PSM analyses (HR,

0.86; 95% CI, 0.76-0.98; PZ.022) compared to RT alone. On subgroup

analysis, patients ages <79 years (HR, 0.80; PZ.001), those with a co-

morbidity score of 0-1 (HR, 0.84; PZ.002), stage III/IV disease (HR, 0.77;

P<.001), and treatment with intensity modulated RT (HR, 0.76; PZ.002)

had an OS benefit with CRT under MVA. The addition of chemotherapy in

patients aged�79 years (HR, 0.93;PZ.368), thosewith a comorbidity score

of �2 (HR, 1.00; PZ.992), stage I/II disease (HR, 1.09; PZ.448), and

treatment with 3-dimensional RT (HR, 1.02; PZ.923) did not have

improved OS. RPA showed patients aged <79 years presenting with T3-4,

any N, and a comorbidity score of 0 had an OS improvement with CRT

versus RTalone; all others, including patients�79 years of age, did not have

an OS benefit with CRT. Patients�79 years old with a comorbidity score�2

had a trend for worse OS with CRT (HR, 2.36; PZ.080).

Conclusion: Patients <79 years of age with low comorbidity scores with

T3/T4, any N-stage disease, appear to have an improvement in OS with

CRT versus RT alone within the NCDB. CRT had worse OS in those aged

�79 years with poor comorbidity scores. In summary, patients >70 years

of age should not be denied concurrent chemotherapy based solely on age;

additional factors ought to be accounted for.

Author Disclosure: A. Amini: None. B. Jones: None. J. McDermott:

None. A. Jimeno: None. D.W. Bowles: None. D. Raben: None. S.

Karam: None.

101Validation of NRG Oncology/RTOG 0129 Risk Groups for p16-Positive and p16-Negative Oropharyngeal Squamous Cell Cancer(OPSCC)C. Fakhry,1 Q. Zang,2 M. Gillison,3 P.F. Nguyen-Tan,4 D.I. Rosenthal,5

R.S. Weber,5 L. Lambert,6 A. Trotti,7 W.L. Barrett,8 W. Thorstad,9

S.S. Yom,10 S. Wong,11 J.A. Ridge,12 S.S.D. Rao,13 W.K. Huh,14

E. Vigneault,15 D. Raben,16 J. Harris,2 and Q.T. Le17; 1Johns Hopkins,

Baltimore, MD, 2NRG, Philadelphia, PA, 3The Ohio State University,

Columbus, OH, United States, 4Centre, Montreal, DE, Canada, 5The

University of Texas MD Anderson Cancer Center, Houston, TX, 6Hopital

Notre Dame, Montreal, QC, Canada, 7H. Lee Moffitt Cancer Center and

Research Institute, Tampa, FL, 8University of Cincinnati, Cincinnati,

OH, 9Washington University, St Louis, MO, 10University of California

San Francisco, San Francisco, CA, 11Medical College of Wisconsin,

Department of Hematology and Oncology, Milwaukee, WI, 12Fox Chase

Cancer Center, Philadelphia, PA, 13University of California, Davis,

Sacramento, CA, 14University of Alabama, Birmingham, AL, 15Centre

Hospitalier Universitaire de Quebec, Hotel-Dieu de Quebec, Quebec,

QC, Canada, 16Department of Radiation Oncology, University of

Colorado Denver, Aurora, CO, 17Stanford Radiation Oncology,

Stanford, CA

Purpose/Objective(s): Radiation Therapy Oncology Group (RTOG) 0129

risk groups stratified OPSCC patients into low-, intermediate-, and high-

risk groups based on p16 status, clinical stage, and tobacco history. These

risk groups have not been confirmed in other clinical trials. Reduction of

treatment toxicities for OPSCC is of interest; however, whether toxicities

differ by risk groups is unknown. Therefore, we hypothesized that RTOG

0129 risk groups were reproducible in RTOG 0522 and that observed

toxicities differed by risk groups.

Materials/Methods: Patients with stage III/IV OPSCC enrolled in RTOG

0129 or 0522, available p16 status, and tobacco history were eligible for

this retrospective analysis. RTOG 0129 evaluated standard versus

accelerated fractionation (AFX) radiation therapy concurrent with

cisplatin. RTOG 0522 compared cisplatin-AFX � cetuximab. RTOG

0129 risk groups were assigned to RTOG 0522 to estimate overall (OS)

and progression-free (PFS) survival rates by Kaplan-Meier method and

log-rank. Grade 3+ toxicities were compared between risk groups by X2

test.

Results: A total of 260 and 287 patients from RTOG 0129 and RTOG

0522, respectively, were eligible for analysis. Median follow-up for

surviving patients in RTOG 0129 and RTOG 0522 was 7.9 years (range

1.7-9.9) in RTOG 0129 and 4.7 years (0.1-7.0) in RTOG 0522. In

RTOG 0522, 5-year OS for the low-, intermediate-, and high-risk

groups were 88.1%, 69.9%, and 45.1%, respectively (PZ.002 for low

vs intermediate and PZ.004 for intermediate vs high). Five-year PFS

for 0522 was also higher for the low-risk group relative to the inter-

mediate- and high-risk groups (PZ.01 for low vs intermediate and

PZ.07 for intermediate vs high). Data from RTOG 0129 and RTOG

0522 were combined to determine differences in acute and late toxic-

ities. Similar overall rates of acute toxicities were observed by risk

groups (PZ.14). Severe acute anemia and neutropenia were less

common in the low/intermediate groups (PZ.01). However, severe

acute dysphagia, nausea, and mucositis were significantly more com-

mon in the low/intermediate groups (P<.02 for all). Late toxicities were

similar by risk groups.

Conclusion: RTOG 0129 risk groups are reproducible in RTOG 0522.

When treated with standard or intensified chemoradiation, low/interme-

diate risk patients more commonly experience specific acute toxicities, but

not late toxicities.

Acknowledgment(s): This project was supported by National Cancer

Institute and Bristol Myers Squibb grants U10CA21661, U10CA180868,

U10CA180822, and U10CA37422.

Author Disclosure: C. Fakhry: None. Q. Zang: None. M. Gillison: None.

P. Nguyen-Tan: None. D.I. Rosenthal: None. R.S. Weber: None. L.

Lambert: None. A. Trotti: None. W.L. Barrett: None. W. Thorstad:

None. S.S. Yom: None. S. Wong: None. J.A. Ridge: None. S.S. Rao:

None. W.K. Huh: None. E. Vigneault: None. D. Raben: None. J. Harris:

None. Q. Le: None.

International Journal of Radiation Oncology � Biology � Physics872

102Variability in Outcome Based on Type of Concurrent Chemotherapyfor Treatment of Squamous Cell Carcinoma of the OropharynxD.N. Margalit,1 M. Puzanov,2 B. Rawal,3 R.I. Haddad,4 J.D. Schoenfeld,1

G. Rabinowits,3 N. Chau,3 J. Lorch,5 L.A. Goguen,1 D.J. Annino,1

T. Thomas,1 P.J. Catalano,3 and R.B. Tishler1; 1Dana-Farber Cancer

Institute / Brigham and Women’s Hospital, Boston, MA, 2Brigham &

Women’s Hospital, Boston, MA, 3Dana-Farber Cancer Institute, Boston,

MA, 4Department of Medical Oncology, Dana-Farber Cancer Institute,

Harvard Medical School and Department of Medicine, Brigham and

Women’s Hospital, Boston, MA, 5Dana Farber Cancer Institute, Boston,

MA

Purpose/Objective(s): Concurrent carboplatin and taxol (C/T) is

commonly utilized in the management of head and neck squamous cell

carcinoma (HNSCC) based on phase 2 data. At our institution we under-

went a shift in practice over a relatively short period of time from frequent

use of concurrent C/T to predominantly cisplatin (Cis)-based concurrent

therapy. We sought to determine cancer-specific outcomes with Cis-based

versus C/T concurrent therapy.

Materials/Methods: We retrospectively analyzed 336 consecutive patients

(pts) with oropharyngeal HNSCC treated at a single institution with

definitive intensity modulated radiation therapy to a median dose of 70 Gy

(median 2 Gy/fraction) using predominantly a dose-painting technique.

Univariable and multivariable Cox proportional hazard regression models

were used to evaluate the association of concurrent chemotherapy-type

with overall survival (OS), recurrence-free survival (RFS), local-recur-

rence failure (LRF), and distant metastasis (DM). The multivariable model

included clinical predictors that were significant in the univariate model for

survival outcomes with a P value �.05. The Fine & Gray method was used

to adjust for competing risks.

Results: The cohort comprised 336 pts, mostly male (85%), with a median

age of 57.3 years. The most common sites were tonsil (49.4%) and base of

tongue (46.7%). The American Joint Committee on Cancer stage was I

(0.3%), II (3.9%), III (1.1%), IVA (71.5%), and IVB (9.2%). Most pts were

human papillomavirus (HPV)-positive (72.3%), 13.1% were HPV-nega-

tive, and 14.6% had unknown HPV status. Pts received upfront concurrent

chemoradiation (58%) or sequential therapy (38.4%); 3.6% received ra-

diation and no chemotherapy. Eighty pts (23.8%) received Cis-based

concurrent chemotherapy, and 103 pts (30.7%) received weekly C/T

concurrent chemotherapy. With a median follow-up of 3.8 years, the 3-year

OS was 89.3% with CI of LRF of 9.6% and CI of DM of 9.1%. Compared

with Cis-based concurrent chemotherapy, pts that received weekly C/T had

worse OS (hazard ratio [HR] 6.57; 95% confidence interval [CI] 1.54-28.0)

and RFS (HR 3.38; 95% CI 1.29-8.87) and higher rates of LRF (HR 9.42;

95% CI 1.26-70.5). The significantly worse OS seen with C/T concurrent

chemotherapy remained statistically significant after adjusting for stage,

HPV status, and initial comorbidities (HR 5.00; 95% CI 1.15-21.79) and

after restricting the analysis to the 207 pts that received concurrent

chemotherapy without induction (HR 8.88; 95% CI 1.56-68.23).

Conclusion: Concurrent carboplatin and taxol was associated with worse

OS and cancer-specific outcomes following definitive radiation-based

therapy for SCC of the oropharynx. Cisplatin-based therapy should

continue to be the concurrent therapy of choice for oropharyngeal SCC.

Author Disclosure: D.N. Margalit: None. M. Puzanov: None. B. Rawal:

None. R.I. Haddad: None. J.D. Schoenfeld: None. G. Rabinowits: None.

N. Chau: None. J. Lorch: None. L.A. Goguen: None. D.J. Annino: None.

T. Thomas: None. P.J. Catalano: None. R.B. Tishler: None.

103Increasing Primary Surgical Treatment of T1 and T2 OropharyngealSquamous Cell Carcinoma and the Treatment Selection LearningCurve: National Cancer Data BaseJ. Cracchiolo,1 S.S. Baxi,1 I. Ganly,1 S.G. Patel,1 M.A. Cohen,2

and B.R. Roman1; 1Memorial Sloan Kettering Cancer Center, New York,

NY, 2Weill Cornell Medical College, NY Presbyterian Hospital, New York,

NY

Purpose/Objective(s): There has been new enthusiasm for primary sur-

gical approaches for early T-stage oropharyngeal squamous cell carcinoma

(OPSCC). The objective of this study is to identify national trends in the

choice of primary treatment for T1-T2 OPSCC. We also examined post-

operative pathologic characteristics including margins and evidence of

extracapsular spread (ECS) given the subsequent implications for recom-

mended adjuvant therapy.

Materials/Methods: We identified patients diagnosed with T1 or T2

OPSCC from 2004 to 2012 in the National Cancer Data Base (NCDB) and

categorized them as receiving primary surgical or primary radiation

treatment. We analyzed trends in primary treatment over time using the

Cochran-Armitage test and examined predictors of primary treatment

including tumor, sociodemographic, and hospital factors using a multi-

variable logistic regression. Among primary surgical patients, we

described pathologic factors (ECS and margin status) and clinical to

pathologic upstaging following surgery.

Results: Of 5,193 patients with T1 and T2 OPSCC, 70% received primary

surgical treatment. This number increased from 57% in 2004 to 84% in

2012 (P<.0001). On multivariable analysis, lower clinical N stage and

higher hospital volume were associated with primary surgical treatment.

The highest volume versus lowest volume hospitals treated 83% versus

63% of patients with primary surgery, respectively (odds ratio 2.62, con-

fidence interval 2.09-3.28, PZ.009). Among primary surgical patients,

while clinical T stage was mostly correct (84% of T1 tumors and 79% of

T2 tumors), the clinical lymph node status was less accurate. Patients with

clinical N0 and N1 disease were upstaged 25% and 35% of the time,

respectively. ECS was present in 25% of patients treated with primary

surgical treatment, and positive margins were present in 29%. ECS, T

upstaging, and N upstaging remained stable over time, while positive

margin status decreased from a high of 37% in 2007 to 2008 to 22% in

2012 (P<.0001).

Conclusion: The use of a primary surgical approach for management of

early T-stage OPSCC has increased and is related to both tumor and

nontumor factors. Although margin positivity has decreased over time,

there is an overall nontrivial rate of pathologic adverse features. One

quarter of patients receiving primary surgery had ECS on postoperative

pathologic analysis. Efforts are warranted to improve the primary treat-

ment selection process to avoid triple modality therapy.

Author Disclosure: J. Cracchiolo: None. S.S. Baxi: None. I. Ganly: None.

S.G. Patel: None. M.A. Cohen: None. B.R. Roman: None.

104Improved Survival With 3DCRT Over IMRT in Early-Stage LarynxCancer: A National Cancer Data Base AnalysisM.W. Jackson, A. Amini, B. Jones, and S. Karam; Department of

Radiation Oncology, University of Colorado Denver, Aurora, CO

Purpose/Objective(s): A classic treatment option for early-stage cancer

of the glottic larynx has historically been radiation alone using opposed

lateral photon fields. However, some clinicians have transitioned to more

modern intensity modulated radiation therapy (IMRT) techniques. We used

the National Cancer Data Base (NCDB) to study the impact of treatment

technique on survival.

Materials/Methods: We queried the NCDB for T1-T2 N0 cancer of the

glottic larynx managed with definitive RT alone. Patients with T3/T4,

node-positive, or metastatic disease were excluded, as were patients un-

dergoing surgery or chemotherapy. Retrievable data included age, race,

gender, insurance status, facility type, Charlson-Deyo comorbidity score,

year of diagnosis, stage, and treatment technique recorded as either IMRT

or Conformal or 3-D therapy (3DCRT). Kaplan-Meier overall survival

(OS) estimates were generated, with accompanying univariate (UVA) log-

rank and multivariate (MVA) Cox proportional hazards regression ana-

lyses. Factors demonstrating significance of P<.10 on UVAwere included

in MVA analysis. Propensity score-matched (PSM) analysis was performed

to account for indication bias and mitigate heterogeneity between treat-

ment groups.

Volume 94 � Number 4 � 2016 Posters 873

Results: A total of 1929 patients with complete data were identified with a

median follow-up of 37 months (range 2-128 months). For the entire

cohort, median age was 67 years (range 28-90 years); 88% were male, 84%

were white, 36% had private insurance, 24% were treated at an academic/

research institution, 68% were T1 stage, and 39% were treated with IMRT.

On UVA, OS was significantly decreased with older age, male gender,

nonprivate insurance, increasing comorbidity score, T2 stage, and treat-

ment with IMRT (hazard ratio [HR] 1.32, PZ.002). Five-year OS for the

entire cohort was 69%; it was 72% versus 63% for the 3DCRT and IMRT

groups, respectively. On MVA, older age (HR 1.04, P<.001), male gender

(HR 1.35, PZ.04), having nonprivate insurance, treatment at a community

cancer program (HR 1.41, PZ.04), high comorbidity score (HR 2.78,

P<.001), T2 stage (HR 1.53, P<.001), and IMRT (HR 1.23, confidence

interval [CI]: 1.02-1.49, PZ.029) remained associated with decreased OS.

PSM groups (nZ641 per group) were well-balanced for all factors (X2 p

0.31-0.98) between IMRT versus 3DCRT groups, and Cox proportional

hazard analysis of PSM groups confirmed an OS survival decrement (HR

1.29, CI: 1.07-1.51, PZ.022) associated with IMRT relative to 3DCRT.

Conclusion: This population-based analysis indicates 3DCRT techniques

are associated with improved OS in early-stage cancer of the glottic larynx.

This may be due to marginal miss of the primary tumor or decreased dose

to nearby subclinically involved lymph nodes when highly conformal

techniques are implemented. Physicians should strongly consider a classic

3D approach to early-stage larynx cancer given its documented history of

good outcomes and cost-effectiveness.

Author Disclosure: M.W. Jackson: None. A. Amini: None. B. Jones:

None. S. Karam: None.

105Elective Nodal Irradiation and Patterns of Failure in Head and NeckCancer Following Primary Radiation TherapyJ. Kjems,1 A.B. Gothelf,2 K.E. Hakansson,2 L. Specht,1 C.A. Kristensen,2

and J.T. Friborg1; 1Department of Oncology, Section of Radiotherapy,

Rigshospitalet, University of Copenhagen, Copenhagen, Denmark,2Rigshospitalet, Copenhagen, Denmark

Purpose/Objective(s): The delineation of elective clinical target volumes

in head and neck cancer (HNC) is important, but the extent of lymph node

levels necessary to include is debated. A comprehensive analysis of

recurrence patterns in a large cohort of HNC patients was performed with

emphasis on recurrences in the retropharyngeal region and level IB.

Materials/Methods: Between 2005 and 2012, 942 patients with oro- or

hypopharyngeal, laryngeal, or oral cavity carcinomas were curatively

treated with primary radiation therapy (RT). The median follow-up was 34

months, and 77% of patients received intensity modulated RT. The retro-

pharyngeal region was only routinely included in case of involvement of

the posterior pharyngeal wall, and level IB only in case of involvement of

the oral cavity. In patients with regional recurrence, the anatomical site of

the recurrence was assessed from surgical descriptions or CT scans and

compared to the original RT treatment plan (available from 2007 and

onward). P16 status was available in 282 oropharynx carcinomas

(OPSCC), with 65% p16 positive.

Results: Forty percent of patients (NZ376) had recurrences, 24.2% local

(NZ228), 13.1% regional (NZ123), and 11.6% distant (NZ109). In 700

patients with available treatment plans, retropharyngeal and level IB re-

currences were observed in 2 and 7 patients, respectively. Eight patients

(1.1%) had recurrence in a lymph node level not included in their primary

treatment plan. For OPSCC, the loco-regional control (90% vs 70%), but

not distant control (92% vs 87%), was significantly better in p16-positive

versus p16-negative patients. Although fewer recurrences occurred in the

p16-positive group, patients with recurrence of p16-positive tumors were

more likely to recur in distant sites.

Conclusion: Retropharyngeal or level IB recurrences following primary

HNC RT are rare, and inclusion of these regions in the elective treatment

volumes should be limited to patients with involvement of the posterior

pharyngeal wall and oral cavity, respectively.

Author Disclosure: J. Kjems: None. A.B. Gothelf: None. K.E. Hakans-

son: None. L. Specht: None. C.A. Kristensen: None. J.T. Friborg: None.

106Prognostic Groups for p16-Positive Oropharyngeal Squamous CellCancer (p16-OPSCC) Treated With Chemoradiation Therapy (CRT) inNRG Oncology/RTOG 0129 and 0522C. Fakhry,1 Q. Zang,2 M. Gillison,3 P.F. Nguyen-Tan,4 D.I. Rosenthal,5

R.S. Weber,5 L. Lambert,6 A. Trotti,7 W.L. Barrett,8 W. Thorstad,9

S.S. Yom,10 S. Wong,11 J.A. Ridge,12 S.S.D. Rao,13 W.K. Huh,14

E. Vigneault,15 D. Raben,16 J. Harris,2 and Q.T. Le17; 1Johns Hopkins

University School of Medicine, Baltimore, MD, 2NRG, Philadelphia, PA,3The Ohio State University, Columbus, OH, United States, 4Centre,

Montreal, DE, Canada, 5The University of Texas MD Anderson Cancer

Center, Houston, TX, 6Hopital Notre Dame, Montreal, QC, Canada, 7H.

Lee Moffitt Cancer Center and Research Institute, Tampa, FL, 8University

of Cincinnati, Cincinnati, OH, 9Washington University, St Louis, MO,10University of California San Francisco, San Francisco, CA, 11Medical

College of Wisconsin, Department of Hematology and Oncology,

Milwaukee, WI, 12Fox Chase Cancer Center, Philadelphia, PA,13University of California, Davis, Sacramento, CA, 14University of

Alabama, Birmingham, AL, 15Centre Hospitalier Universitaire de Quebec,

Hotel-Dieu de Quebec, Quebec, QC, Canada, 16Department of Radiation

Oncology, University of Colorado Denver, Aurora, CO, 17Stanford

Radiation Oncology, Stanford, CA

Purpose/Objective(s): Identification of patients with p16-OPSCC with

high probability of cure is essential for therapeutic deintensification with

CRT.

Materials/Methods: Patients with T2-T4, N0-N3 OPSCC enrolled in

Radiation Therapy Oncology Group (RTOG) 0129 or 0522 with available

p16 immunohistochemistry results and tobacco history were eligible for

retrospective analysis of factors associated with overall (OS) and pro-

gression-free survival (PFS). RTOG 0129 compared high-dose cisplatin-

based CRT standard fractionation with accelerated fractionation (AFX),

and RTOG 0522 compared cisplatin-AFX with and without cetuximab.

Recursive partitioning analysis (RPA) identified mutually exclusive patient

groups for OS. OS and PFS were estimated by Kaplan-Meier method and

compared by log-rank. X2 test compared grade 3 toxicities among RPA-

defined risk groups.

Results: Analysis included 402 patients with p16-OPSCC. Median follow-

up for surviving patients was 5.7 years (range, 0.1-9.9). Univariate factors

associated with improved OS included low tobacco exposure (�10 vs >10

pack-years), age (�50 vs >50 years), better performance status (Zubrod

0 vs 1), lower tumor (T2-T3 vs T4), and nodal (N0-N2a vs N2b-N2c vs

N3) stage (for all, hazard ratio<1.0, P<.04). All these factors were inde-

pendently associated with improved OS in multivariable analysis. Three

distinct risk groups for 5-year OS were identified as at low (92%, 95%

confidence interval [CI] 89-96), intermediate (72%, 95% CI 65-80), or

high (53%, 95% CI 39-67) risk of death. Five-year PFS was also higher for

the low-risk group relative to the intermediate- and high-risk groups

(P<.001 for low vs intermediate and PZ.01 for intermediate vs high).

Acute and chronic toxicities were largely similar by risk groups with few

exceptions. In the low-risk group severe acute neutropenia (PZ.007) was

less common and a trend toward less frequent late severe dysphagia

(PZ.08) and mucositis (PZ.08) was observed. Patterns of recurrence

differed by risk groups (PZ.03); the majority of recurrences in the low-

risk group were locoregional (70.6%), while the majority of recurrences

for intermediate- and high-risk groups were distant (54% and 62%,

respectively).

Conclusion: A low-risk group with decreased risk of death and distant

failure was identified. This may inform eligibility for radiation-based

therapeutic deintensification trials.

Acknowledgment(s): National Cancer Institute and Bristol Myers Squibb

grants U10CA21661, U10CA180868, U10CA180822, and U10CA37422.

International Journal of Radiation Oncology � Biology � Physics874

Author Disclosure: C. Fakhry: None. Q. Zang: None. M. Gillison: None.

P. Nguyen-Tan: None. D.I. Rosenthal: None. R.S. Weber: None. L.

Lambert: None. A. Trotti: None. W.L. Barrett: None. W. Thorstad:

None. S.S. Yom: None. S. Wong: None. J.A. Ridge: None. S.S. Rao:

None. W.K. Huh: None. E. Vigneault: None. D. Raben: None. J. Harris:

None. Q. Le: None.

107Validation and Comparison of Prognostic Scoring Systems in aCohort of Human PapillomaviruseAssociated Oropharynx CancersTreated NonoperativelyM.C. Ward,1 J. Scharpf,2 D.J. Adelstein,2 B.B. Burkey,2 E. Lamarre,2

N. Houston,2 and S. Koyfman1; 1Cleveland Clinic Foundation, Cleveland,

OH, 2Cleveland Clinic, Cleveland, OH

Purpose/Objective(s): The recent human papillomavirus (HPV) epidemic

in oropharynx cancer (OPC) presents a demographic separate from the

historical tobacco-associated OPC. This population is comparatively

younger and healthier than previous patients with tobacco-associated

cancers. Comorbidity assessment is a critical component of the pretreat-

ment evaluation, but the optimal system for documentation and research is

unclear in this cohort. The purpose of this study was to validate and

compare 2 common comorbidity scales: the historic Charlson comorbidity

index and the more recent Adult Comorbidity Evaluation-27 (ACE-27).

Materials/Methods: From a retrospective institutional review board-

eapproved tumor registry, patients with HPV-associated OPC treated

nonoperatively from 2001 to 2013 were identified. Both the Charlson score

and the ACE-27 category were recorded. The Kaplan-Meier method was

used to assess overall survival (OS) by comorbidity system with differ-

ences assessed using the log-rank test. Two separate multivariate Cox

proportional hazards models were generated, each with the individual

comorbidity system as well as the predictors of OS statistically significant

on univariate analysis.

Results: Two hundred eighty-eight patients were identified, of which 274

were treated with combined chemoradiation therapy and 14 with radiation

therapy alone. The cervical nodes were involved in 277 patients, and 285

were stage III-IVB. The median oncologic follow-up was 53.5 months.

There were 55 deaths in the group for an estimated 5-year overall survival

(OS) of 83%. Patients with no comorbidities measured by the ACE-27

score uniformly had a Charlson score of zero. However, a distribution of

Charlson scores noted amongst patients with mild, moderate, or severe

ACE-27 categories had a range of Charlson scores: 0-5, 0-4, and 0-8,

respectively, each with a median Charlson score of 1. On univariate

analysis age, Charlson score, ACE-27 category, pack-years of smoking,

and T stage were significantly associated with OS. On multivariate analysis

(MVA), pack-years of smoking, T stage, and each comorbidity score

remained independently associated with OS.

Conclusion: Both the ACE-27 and the Charlson comorbidity index are

validated prognostic scales for the measurement of comorbidity in HPV-

associated head and neck cancer. Both have advantages: although ACE-27

is more sensitive, the Charlson score may retain a degree of granularity

given the range of possible results. Comorbidity should be included in

predictive models for OS in an attempt to account for competing risks in

this relatively healthy subset of head and neck patients.

Abstract 107; Table 1

Variable

MVA with CharlsonMVA withACE-27

HR p HR p

Age 0.995 .763 1.013 .475Charlson 1.629 <.0001 — —ACE-27 1 vs 0 — — 1.446 .359

2 vs 0 — — 2.434 .0223 vs 0 — — 1.882 .192

Pack-years 1.013 .0053 4.773 .024T stage 3-4 vs 1-2 2.3 .0026 2.154 .007

Author Disclosure: M.C. Ward: None. J. Scharpf: None. D.J. Adelstein:

None. B.B. Burkey: None. E. Lamarre: None. N. Houston: None. S.

Koyfman: None.

108The Outcomes of Induction Chemotherapy for Head and NeckCancer PatientsS.Y. Wu,1 F.P. Lee,2 K.C. Lin,3 M.T. Lai,2 C.C. Wu,2 T.M. Chen,2

Y.F. Ding,2 and S.P. Yuan2; 1Department of Radiation Oncology, Wan Fang

Hospital, Taipei Medical University, Taipei, Taiwan, 2Department of

otorhinolaryngology, Wan Fang Hospital, Taipei Medical University,

Taipei, Taiwan, 3Department of Oral and Maxillofacial Surgery, Wan Fang

Hospital, Taipei Medical University, Taipei, Taiwan

Purpose/Objective(s): Until now, the role of induction chemotherapy has

remained a subject of controversy. Our study was to directly compare

survival in patients receiving induction chemotherapy docetaxel or plat-

inum given before concomitant chemoradiation therapy with upfront

chemoradiation therapy alone.

Materials/Methods: The National Health Insurance claims database and

cancer registry databases in Taiwan were linked for the analysis. Head and

neck cancer patients from January 1, 2002 to December 31, 2011 were

included in the study. The inclusion criteria were having a head and neck

cancer (identified according to the International Classification of Diseases,

Ninth Revision, Clinical Modification [ICD-9-CM] codes 140.0-148.9),

being aged >20 years, being classified as American Joint Committee on

Cancer (AJCC) clinical cancer stage III-IV (locally advanced head and

neck cancers without metastasis), and having undergone induction

chemotherapy or concurrent platinum-based chemoradiation therapy

(CCRT). Exclusion criteria were having been diagnosed with cancer before

the head and neck cancer was confirmed, having distant metastasis, AJCC

clinical cancer stage I-II, platinum and docetaxel combined use before RT,

docetaxel use during or after RT, induction chemotherapy beyond 8 weeks

before RT, only 1 course of induction chemotherapy before RT, cetuximab

use, adjuvant chemotherapy within 90 days after completion of RT, less

than 7000 cGy dose of RT, curative head and neck cancers surgery before

RT, nasopharyngeal cancer, carcinoma in situ, sarcoma, head and neck

cancer recurrence, or an unknown gender, and being younger than 20 years

of age. The total number of enrolled head and neck cancer patients was

30,990 persons.

Results: In total, 10,721 stage III-IV head and neck cancer patients

without distant metastasis were included in the study, and the median

follow-up duration was 4.18 (interquartile range, 3.25) years. There were

7968 patients in the CCRT group (arm 1); 503 patients in the induction

chemotherapy with docetaxel group of arm 2, and 2232 patients in the

induction chemotherapy with platinum group of arm 3. We used the CCRT

arm as the control arm to investigate the risk of death after induction

chemotherapy. After adjustments for age, gender, clinical stage, and

comorbidities, the adjusted hazard ratios of overall deaths were 1.37 (95%

confidence interval [CI], 1.22e1.53) in arm 2 and 1.44 (95% CI,

1.36e1.52) in arm 3. In disease-specific survival rate analysis, the adjusted

HRs of head and neck cancers deaths were 1.29 (95% CI, 1.14-1.46) in arm

2 and 1.47 (95% CI, 1.38-1.56) in arm 3.

Conclusion: Our cohort study showed induction chemotherapy with

docetaxel or platinum not only did not improve survival but also resulted in

more all death and head and neck cancer death risk compared with CCRT.

Author Disclosure: S. Wu: None. F. Lee: None. K. Lin: None. M. Lai:

None. C. Wu: None. T. Chen: None. Y. Ding: None. S. Yuan: None.

109Induction Chemotherapy Predicts Cumulative Radiation Dose andFails to Improve Survival in Advanced Head and Neck Cancer, aNational Cancer Data Base AnalysisW. Stokes,1 A. Amini,1 J. McDermott,2 A. Jimeno,2 D. Raben,1

D.W. Bowles,2 and S. Karam1; 1Department of Radiation Oncology,

University of Colorado Denver, Aurora, CO, 2Department of Medical

Oncology, University of Colorado Denver, Aurora, CO

Abstract 110; Table 1

RTOG 0129risk group

TGR < 0.74% (nZ51) TGR > 0.74% (nZ52)

LR failure Distant failure LR failure Distant failure

Low (nZ52) 0/30 1/30 1/22 2/22Intermediate (nZ33) 0/18 1/18 4/15 3/15High (nZ18) 0/3 1/3 10/15 3/15

Volume 94 � Number 4 � 2016 Posters 875

Purpose/Objective(s): The role of induction chemotherapy (IC) in

advanced head and neck squamous cell carcinoma (HNSCC) remains

controversial. In recent randomized trials, the addition of IC to concurrent

chemoradiation (CRT) failed to improve overall survival (OS). This failure

may stem from the studies’ lack of power due to slow accrual and/or from

their inclusion of patients with less advanced nodal disease, prompting the

present analysis of the National Cancer Data Base (NCDB).

Materials/Methods: The NCDB was queried for subjects diagnosed from

2003 to 2011 with T (any) N2b-3 M0 cancers of the oropharynx, hypo-

pharynx, and larynx, who underwent external beam radiation without

surgery. We defined 2 analytic cohorts based on the sequencing of

chemotherapy (CT) and radiation therapy (RT): an IC cohort with start of

CT preceding RT by 43 to 98 days (thus allowing 2-3 cycles of IC as used

in recent trials) and a CRT alone cohort with CT starting within 7 days of

RT start. Logistic regression was used to identify factors associated with

nonguideline-concordant RT dose (ie, <66 Gy), and Cox regression was

used to assess the association of CT sequence on OS.

Results: A total of 6086 CRT and 1917 IC subjects were evaluable. As

compared to the CRT group, the IC cohort tended to be younger and to

have more advanced T and N status and more hypopharynx cancer, were

more likely to receive <66 Gy of RT (20.9% vs 14.9%; P<.01), and

displayed worse OS (median 52.1 vs 64.9 months, P<.01). After adjusting

for age, year, sex, race, location, income, comorbidities, primary site, and

T and N status with multivariate analysis, the IC cohort had increased odds

of receiving <66 Gy (odds ratio 1.42; 95% confidence interval [CI] 1.24-

1.63; P<.01), but their OS did not significantly differ from that of the CRT

cohort (hazard ratio [HR] for mortality 1.07; 95% CI 0.99-1.16; PZ.08).

On subgroup analysis, IC status was not associated with improved OS

among the 2809 subjects with T4 or N3 disease (HR 1.02; 95% CI 0.92-

1.13; PZ.72), the 1107 patients with N3 disease (HR 1.02; 95% CI 0.86-

1.22; PZ.82), or the 351 subjects with T4N3 disease (HR 0.97; 95% CI

0.73-1.28; PZ.81). Among the 5194 patients without T4 or N3 disease, IC

status predicted a slight increase in mortality (HR 1.12; 95% CI 1.00-1.25;

PZ.046).

Conclusion: In this large group of HNSCC patients with advanced nodal

disease from the NCDB, IC subjects were more likely to receive less-than-

definitive doses of RT, and OS was not significantly different from that of

CRT subjects, even on subgroup analyses of increasingly advanced dis-

ease. Failure of prior studies to demonstrate an OS benefit with IC may

have less to do with statistical power or patient selection and more to do

with difficulty completing guideline-concordant care following IC.

Author Disclosure: W. Stokes: Employee; University of Colorado Denver

GME. A. Amini: None. J. McDermott: None. A. Jimeno: None. D.

Raben: None. D.W. Bowles: None. S. Karam: None.

110Patterns of Failure After Definitive Radiation for OropharyngealCancerdShould P16 Status and Tumor Growth Rate Alter theClinical Target Volume?C.T. Murphy,1 M. Dziemainowicz,2 T. Shaikh,1 M. Fareed,1 J.A. Ridge,1

R. Mehra,1 and T.J. Galloway1; 1Fox Chase Cancer Center, Philadelphia,

PA, 2Temple University School of Medicine, Philadelphia, PA

Purpose/Objective(s): To determine patterns of failure after definitive

radiation therapy (RT) for oropharyngeal cancer (OPC) and employ clin-

ical target volume (CTV) expansions based upon p16 status and primary

tumor growth rate (TGR).

Materials/Methods: One hundred three OPC patients with known p16

status, smoking pack-years, and TGR (defined as percent volume growth/

day) managed with chemoradiation therapy (CRT) were analyzed. Diag-

nostic computed tomographic scan images of patients with locoregional

failure (LRF) were imported into a commercially available contouring

system and fused with the radiation treatment planning scan dose distri-

bution using deformable registration. The distance of the LRF from the

edge of radiation prescription volume (RxV) was recorded. Cox propor-

tional hazards model was used to estimate the risk of disease recurrence.

Results: Seventy-nine patients were p16(+) (77%), and 98 were stage III or

IV (95%). Median TGR was 0.74%/day (range 0.01-5.5). Median follow-

up was 30 months (range 0.5-80). Ninety-two percent were treated with

CRT. On multivariate analysis, p16(-) status (hazard ratio [HR] 3.4, 95%

confidence interval [CI] 1.4-8.3) and increasing TGR (HR 4.8, 95% CI 1.6-

14.5) were the strongest predictors of recurrence. Patterns of failure ac-

cording to Radiation Therapy Oncology Group (RTOG) 0129 risk group,

stratified by median TGR are shown in Table 1.

Fourteen of 15 patients with LRF had evaluable postrecurrence scans

and plans. The median time to LRF was 4 months (range 0.75-14 months).

A majority of failures (11/15, 73%) were marginal to high-dose radiation

prescription volume (RxV; ie, the tumor recurred both in and out of the

treated volume). Four recurrences developed completely within the high-

dose RxV. The median distance of the furthest extent of the recurrence

from the field edge was 7 mm; the distance from the RxVedge necessary to

encompass the LRF with adequate margin ranged from 0 to 20 mm. Eighty

percent of failures would have been covered by an expansion of an addi-

tional 12 mm outside of the RxV; 3 patients experienced LRF more than 15

mm outside of the high-dose RxV.

Conclusion: Locoregional failures in low-risk OPC tumors are rare, such

that reductions in CTV margins may be justified in the interest of treatment

deintensification. The dominant pattern of failure in intermediate- and

high-risk tumors appears to be marginal to the high-dose prescription

volume, predominantly among tumors with increased TGR. CTV expan-

sions based upon RTOG 0129 risk group and TGR warrant investigation.

Author Disclosure: C.T. Murphy: None. M. Dziemainowicz: None. T.

Shaikh: None. M. Fareed: None. J.A. Ridge: None. R. Mehra: None.

T.J. Galloway: None.

111Dosimetric Verification of Dental Stent Efficacy in Head and NeckRadiation Therapy Using Modern Radiation Therapy Techniques:Quality of Life (QOL) and Treatment Compliance ImplicationsE. Allan,1 L. Lu,2 H. Hooman,3 A. Chakravarti,2 M. Van Putten,4

and D. Blakaj1; 1The James Cancer Hospital and Solove Research

Institute, Wexner Medical Center at The Ohio State University, Department

of Radiation Oncology, Columbus, OH, 2The Ohio State University Wexner

Medical Center, Columbus, OH, 3Ohio University - Heritage College of

Osteopathic Medicine, Dublin, OH, 4The James Cancer Hospital and

Solove Research Institute, Wexner Medical Center at The Ohio State

University, Department of Maxillofacial Prosthodontics, Columbus, OH

Purpose/Objective(s): Oral mucositis is a commonly observed toxicity

during head and neck radiation therapy. Metallic dental restorations pro-

duce short-range secondary electrons that deposit dose into nearby tissue

causing large, painful ulcers in the adjacent mucosal surfaces. This con-

tributes to patient weight loss during treatment and often leads to breaks in

therapy to replan or to allow for recovery. Various protective dental stents

have been proposed and tested in very simple phantoms and 2-dimensional

(2D) beam arrangements. Our objective was to generate the first quanti-

tative assessment of electron scatter and stent efficacy using an anatomi-

cally realistic phantom and a modern beam configuration and delivery

method in order to better address our patients’ QOL during treatment.

Materials/Methods: We created a tissue-equivalent phantom to simulate a

complete upper and lower jaw with 2 sets of removable gold caps on

opposing molars. We created a set of 4-mm upper and lower ethylene

copolymer dosimetric stents to provide space between the mucosae and

teeth as well as between the upper and lower jaw. The phantom was placed

International Journal of Radiation Oncology � Biology � Physics876

in a cylindrical water bath to simulate head and neck soft tissue. A linear

accelerator was used to deliver 6-MV photons in opposed lateral, intensity

modulated radiation therapy (IMRT), and volumetric modulated arc ther-

apy (VMAT) configurations. We used film for dosimetric measurement in

both the occlusal plane and the vertical plane to simulate mucosal surfaces

of the tongue and the cheek. A single dose of 200 cGy was prescribed to

the base of tongue region. Plans and film measurements were made for the

phantom with and without the stents in place. We then converted the film

reading to a 2D digital dose map using quality assurance software. We

used image processing software to measure the areas on the film that were

enclosed by each isodose line.

Results: Our readout of the 2D dose distributions in the vertical (buccal)

plane clearly demonstrates a reduction in maximum dose of 27% for the

opposed beam and 40% for the IMRT and VMAT with the use of the

protective dental stent. The area of simulated tissue receiving �100 cGy

was 2.5 cm2 higher in the occlusal plane and 5.5 cm2 greater in the buccal

plane when the dental stents were absent. To our knowledge, this is the first

dosimetric analysis of dental stents using an anatomically realistic phan-

tom and modern beam arrangement.

Conclusion: Our observed 40% dose reduction implies that patients’ oral

mucosae adjacent to dental fixtures could receive doses in excess of 100

Gy during a course of definitive IMRT or VMAT radiation therapy to the

head and neck. In this era of increasing IMRT and VMAT use, our results

emphasize the importance of dosimetric stent use to improve QOL and

reduce treatment breaks for our patients undergoing head and neck radi-

ation therapy.

Author Disclosure: E. Allan: None. L. Lu: None. H. Hooman: None. A.

Chakravarti: None. M. Van Putten: None. D. Blakaj: None.

112Contralateral Submandibular Gland Sparing in Head and NeckSquamous Cell Carcinoma Is Safe and Improves Patient-ReportedOutcomes (PROs)Y. Mao,1 S. Samuels,2 E. Sapir,2 and A. Eisbruch2; 1Cancer Center, Sun

Yat-sen University, Guangzhou, China, 2University of Michigan, Ann

Arbor, MI

Purpose/Objective(s): To determine the feasibility and potential benefits

of sparing contralateral submandibular gland (cSMG) during definitive

intensity modulated radiation therapy (IMRT) for head and neck squamous

cell carcinoma (HNSCC).

Materials/Methods: We reviewed treatment plans, mean doses (MDs) to

organs at risk, patient-reported outcomes, and tumor control in 273 patients

with HNSCC treated with definitive IMRT to the bilateral necks at our

institution between 2005 and 2014. Patients were periodically given the

validated xerostomia questionnaire (XQ) and head and neck quality of life

questionnaire (HNQOL) pretreatment and at follow-up.

Results: The majority (73%) of patients had oropharyngeal cancer. The

vast majority of patients, 93.0%, harbored stage III/ IV disease, and 93.0%

of patients received concurrent chemotherapy. Median follow-up was 22.5

months (1e115 months), and 662 surveys were analyzed. The medians of

MDs to oral cavity (OC) was 36 Gy (8-67 Gy), combined parotid glands

(comPG) 31 Gy (6-58 Gy), ipsilateral SMG (iSMG) 66 Gy (28-72 Gy), and

cSMG 37 Gy (8-70 Gy). On univariate analysis, significant predictors of

XQ summary score (SS) included the MDs of OC (rZ0.44, P<.01),

comPG (rZ0.63, P<.01), and cSMG (rZ0.22, PZ.04). On multivariate

analyses, the comPG MD (rZ0.42, PZ.04) and time from treatment (rZ-

0.14, PZ.01) were statistical predictors of the XQSS, and the cSMG MD

(rZ0.21, PZ.09) was a marginally significant predictor for XQSS. At the

6- and 12-month timepoints, the cSMG MD is a significant predictor for

XQSS (rZ0.82, PZ.01 and rZ0.57, PZ.04, respectively), eating domain

(rZ0.90, PZ.01 and rZ0.68, PZ.02, respectively), fasting domain

(rZ0.66, PZ.04 and rZ0.52, PZ.07, respectively) and HNQOL (rZ0.77,

P<.01 and rZ0.49, PZ.02, respectively). Using 39 Gy cSMG MD as a

prespecified threshold based on dose-saliva relationships, regression

modeling showed that patients receiving <39 Gy had significantly favor-

able XQSS in both univariate (rZ-22.61, P<.01 and rZ-15.18, PZ.01,

respectively) and multivariate (rZ-26.50, P<.01 and rZ-13.02, PZ.03,

respectively) analysis at 6 and 12 months. Moreover, patients receiving

<39 Gy had improvement (P<.01) in their XQSS over time after IMRT,

while patients receiving >39 Gy did not (PZ.29) (data will be presented).

Sixty-three percent (172/273) of the patients received <39Gy to the cSMG

to preserve salivary output. In this group, the median of MD to cSMG was

31Gy (8-39 Gy), and the majority (75%) was oropharyngeal cancer. There

were 28 total tumor recurrences with no failures in contralateral level IB,

and only 1 patient failed on the contralateral neck, within the 54 Gy

isodose line and in initially noninvolved level II.

Conclusion: cSMG MD predicts for both patient-reported xerostomia and

QOL after IMRT. cSMG sparing did not compromise disease control. We

recommend keeping the cSMG MD to no more than 39 Gy if clinically

possible.

Author Disclosure: Y. Mao: None. S. Samuels: None. E. Sapir: None. A.

Eisbruch: None.

113Sparing Level IB in Node-Positive, HumanPapillomaviruseAssociated Oropharyngeal Carcinoma: An EarlySafety and Efficacy AnalysisN.P. Joshi,1 A. Juloori,1 M.C. Ward,1 H. Qu,1 J.F. Greskovich, Jr,2

E. Murray,1 J. Potter,1 A. Dorfmeyer,1 P. Xia,1 and S. Koyfman1;1Cleveland Clinic Foundation, Cleveland, OH, 2Cleveland Clinic Florida,

Weston, FL

Purpose/Objective(s): The conformality of modern intensity modulated

radiation therapy (IMRT) allows meaningful avoidance of the subman-

dibular glands (SMG) in select patients, thereby potentially improving late

xerostomia. This study explores the safety and efficacy of this approach in

select node-positive oropharyngeal carcinoma patients, comparing it with

traditional level IB coverage.

Materials/Methods: From a retrospective institutional review board-

eapproved registry, patients with T1-2, node-positive human papilloma-

virus (HPV)-associated oropharyngeal carcinoma treated with definitive

IMRT at a single institution from August 2009 to January 2014 were

identified. While in the earlier portion of this study routine avoidance of

level Ib was uncommon, in recent years, we routinely avoid clinically

uninvolved level Ib, even in the presence of level 2a LN. We reviewed and

verified all treatment plans. Mean doses to the SMG were compared for

instances in which level Ib LN were targeted and level Ib LN were avoi-

ded. We then examined mean doses to the oral cavity as well as predictors

of late grade �2 xerostomia on a per patient basis, for which patients were

divided into 3 groups: bilateral level IB targeted (A), a single level IB

targeted (B), and bilateral IB not targeted (C). We also reviewed every

failure location to identify the rate of level IB regional recurrence. Dif-

ferences in continuous variables were compared using the Wilcoxon rank

sums test and Pearson X2 test was used for categorical variables.

Results: Among 87 patients and 174 level Ib stations, level Ib was targeted

in 95 instances and avoided in 79 instances. Mean SMG doses were

significantly lower when level IB was spared compared to where it was

targeted (44.5% reduction; 37.5 Gy vs 67.5 Gy; P<.0001). We examined

predictors of late grade �2 xerostomia on a per patient basis which

included 17 patients in Group A, 61 in group B, and 9 in group C. Age,

KPS, smoking pack-years, T stage, N stage, and type of chemotherapy did

not differ amongst the 3 groups. Median doses to oral cavity decreased

with increasing level Ib sparing (42.2 Gy [Group A] versus 35.4 Gy [Group

B] versus 30.7 [Group C]; PZ.001). Rates of late grade �2 xerostomia

were numerically but not statistically lower in group C versus A and B

(11% vs 59% and 52%, PZ0.17). With a median follow-up of 29.8

months, no regional failures were identified in levels IB in any patient.

Conclusion: Sparing level IB, either contralaterally or bilaterally, is safe in

T1-2, node-positive HPV+ oropharyngeal cancer. Avoiding level Ib ap-

pears to translate into significantly lower SMG doses as well as oral cavity

doses. Larger studies are needed to validate these early findings and the

impact of this technique on late xerostomia and other functional outcomes.

Volume 94 � Number 4 � 2016 Posters 877

Author Disclosure: N.P. Joshi: None. A. Juloori: None. M.C. Ward:

None. H. Qu: None. J.F. Greskovich: None. E. Murray: None. J. Potter:

None. A. Dorfmeyer: None. P. Xia: Honoraria; Philips. S. Koyfman:

None.

114A Prospective Evaluation of Dysphagia After Transoral RoboticSurgery for Squamous Cell Carcinoma of the OropharynxW.G. Albergotti, R.L. Ferris, U. Duvvuri, S. Kim, T. Wasserman, J. Jordan,

and K. Anthony; University of Pittsburgh Medical Center, Pittsburgh, PA

Purpose/Objective(s): Transoral robotic surgery (TORS) for oropharyn-

geal squamous cell carcinoma (OPSCC) has been associated with

improved long-term dysphagia quality of life as compared to chemo-

radiation. Nevertheless, dysphagia is common in the perioperative period

and has been inadequately characterized. Our primary objective in this

study is to characterize short-term swallowing outcomes after TORS for

OPSCC in a prospective manner in an attempt to improve postoperative

outcomes.

Materials/Methods: Patients undergoing TORS for OPSCC were pro-

spectively enrolled into this study between the dates of June 20, 2014 and

July 31, 2015. Patients were evaluated by a speech-language pathologist

postoperatively for diet recommendations and swallow strengthening ex-

ercises. The Eating Assessment Tool 10 (EAT-10), a 10-item validated

questionnaire measuring swallowing quality of life, was administered on

postoperative day (POD) 1, POD 7, and POD 30. A score >3 is considered

to be indicative of swallowing dysfunction. Medical records were queried

for demographics, clinical history, staging, intraoperative factors, and

postoperative course. Patients were excluded for a history of previous

TORS or radiation to the oropharynx, repeat TORS within 1 month after

enrollment, TORS for nonmalignancy, a procedure on a nonoropharyngeal

aerodigestive subsite, a contraindication to swallowing evaluation, or

incomplete data. Statistical analysis was performed using a paired t test to

compare EAT-10 scores between POD 1 and POD 7 and POD 30.

Results: Fifty-nine patients met initial inclusion criteria. Twenty-four

patients were excluded (8 for nonoropharyngeal procedures, 5 for con-

traindications to swallowing evaluation, 7 for repeat TORS within 1

month, and 4 for incomplete data), leaving 35 patients (26 males, 9 fe-

males) for analysis. The mean age was 58.8 (range 43-74) years. Four of

the 35 patients (11.4%) reported preoperative dysphagia. Twenty of the 35

patients (57.1%) underwent tongue base resection, with the remainder

undergoing radical tonsillectomy. T stages were Tx (3), T1 (18), T2 (13),

T3 (1), all HPV+. All patients were started on an oral diet by POD 1

without instrumental testing. The mean EAT-10 score (0-40) on POD 1 was

21.5 (range 0-37), on POD 7 was 27.7 (range 14-45), and on POD 30 was

11.9 (range 1-33). EAT-10 scores were significantly worse at POD 7

(PZ.003) and significantly better on POD 30 (P<.001) as compared with

initial evaluation. However, at 1 month, only 5 of 34 patients (14.3%) had

normal EAT-10 scores. Mean weights (lbs) decreased significantly over the

month (207.6 vs 198.8, P<.001).

Conclusion:Most patients who undergo TORS experience dysphagia for at

least the first month after surgery. Patients can be counseled that dysphagia

will worsen by postoperative day 7 and then improve, but it likely will not

resolve by 1 month. Swallowing evaluation and therapy should be

considered routine in this cohort of patients.

Author Disclosure: W.G. Albergotti: None. R.L. Ferris: None. U. Duv-

vuri: None. S. Kim: None. T. Wasserman: None. J. Jordan: None. K.

Anthony: None.

115The Influence of Diabetes Mellitus and Metformin on DistantMetastases in Oropharyngeal Cancer: A Multicenter StudyZ.S. Zumsteg,1 B.M. Beadle,2 D.E. Spratt,3 A. Rivera,4 H.D. Skinner,2

J. Osborne,5 A.S. Garden,2 and N. Lee4; 1Cedars-Sinai Medical Center,

Los Angeles, CA, 2The University of Texas MD Anderson Cancer Center,

Houston, TX, 3University of Michigan, Ann Arbor, MI, 4Memorial Sloan

Kettering Cancer Center, New York, NY, 5Memorial Sloan-Kettering

Cancer Center, New York, NY

Purpose/Objective(s): Locoregional control in oropharyngeal cancer

(OPC) following concomitant chemoradiation (CRT) has improved to

unprecedented rates in the modern era. This is largely attributable to the

increasing prevalence of human papillomavirus (HPV)-positive OPC, a

malignancy with increased sensitivity to both radiation therapy (RT) and

chemotherapy when compared to HPV-negative OPC. As a result, distant

metastases are increasing, driving survival outcomes for OPC. Given that

both diabetes mellitus and metformin use have been purported to affect

clinical cancer outcomes, including metastases, we aimed to determine the

impact of these factors in a large population of OPC patients treated in the

modern era.

Materials/Methods: The records of 1745 patients with OPC cancer treated

at 2 large cancer centers with definitive RT from 1998 to 2011 were

retrospectively reviewed. Outcomes assessed included local failure-free

survival (LFFS), regional failure-free survival (RFFS), distant metastasis-

free survival (DMFS), and overall survival (OS).

Results: Median follow-up was 4.3 years. A total of 184 patients had

diabetes at time of diagnosis, including 102 patients taking metformin.

Chemotherapy was administered to 77%, 86%, and 82%, of nondiabetic

patients, diabetic patients taking metformin, and diabetic patients not

taking metformin, respectively (PZ.063). There was no significant dif-

ference between these groups in terms of tumor stage (PZ.18), nodal stage

(PZ.18), or HPV status (PZ.56), although HPV data was available for

only 23% of the cohort. In comparison to patients without diabetes, dia-

betic patients not using metformin had worse 5-year DMFS (89.6% vs

78.7%, PZ.011) and OS (83.0% vs 70.7%, PZ.048). By contrast, diabetic

metformin users had no significant difference in 5-year DMFS (90.1%) or

OS (89.6%) in comparison to nondiabetic patients. On multivariate anal-

ysis with diabetic patients not using metformin as a reference, nondiabetic

patients (hazard ratio [HR] Z 0.54, 95% confidence interval [CI] 0.32-

0.93, PZ.03) and diabetic metformin users (HRZ0.46, 95% CI 0.20-1.04,

PZ.06) had improved DMFS after adjusting for age, stage, smoking, and

chemotherapy use, although this did not reach statistical significance for

diabetic metformin users. Five-year LFFS and RFFS was 93% to 95% in

all groups, with no significant difference with respect to diabetes status or

metformin use.

Conclusion: Diabetic OPC patients not using metformin had significantly

higher rates of distant metastases than nondiabetic patients even after

adjusting for confounding variables, whereas diabetics using metformin

had similar rates of distant metastases to nondiabetics. Further prospective

investigation is warranted to validate metformin’s benefit in OPC.

Author Disclosure: Z.S. Zumsteg: None. B.M. Beadle: None. D.E.

Spratt: None. A. Rivera: None. H.D. Skinner: None. J. Osborne: None.

A.S. Garden: None. N. Lee: None.

116Response-Adapted Volume De-escalation (RAVD) in LocallyAdvanced Head and Neck Cancer: Toxicity and Quality of LifeAnalysesJ.M. Melotek,1 V.M. Villaflor,1 T.G. Karrison,1 R.J. Brisson,1 E.A. Blair,1

L. Portugal,1 K.M. Stenson,2 J.A. de Souza,1 E. Cohen,3 A. Langerman,1

M.T. Spiotto,1 T.Y. Seiwert,1 E.E. Vokes,1 and D.J. Haraf1; 1University of

Chicago, Chicago, IL, 2Rush University, Chicago, IL, 3University of

California San Diego, La Jolla, CA

Purpose/Objective(s): Efforts to reduce treatment toxicity in locally

advanced head and neck squamous cell carcinoma (LA-HNSCC) have

focused on radiation therapy (RT) dose de-escalation in select populations.

We report here the toxicity and quality of life (QOL) outcomes from a

phase 1/randomized phase 2 trial using response to induction chemo-

therapy (IC) as a decision tool to guide the extent of RT volume reduction

in a nonselected population of patients (pts) with LA-HNSCC.

Materials/Methods: Pts with measurable LA-HNSCC received 2 cycles of

IC (cisplatin, paclitaxel, cetuximab � everolimus). Pts with “good”

response (GR), defined as �50% reduction in the sum of gross tumor

International Journal of Radiation Oncology � Biology � Physics878

diameters, received TFHX2 (paclitaxel, fluorouracil, hydroxyurea, and 1.5

Gy twice daily RT every other week) to 75 Gy with the planning target

volume (PTV1) encompassing exclusively gross disease. Pts with <50%

response (NR) were treated with volumes encompassing PTV1 and the

next nodal station at risk (PTV2) to 45 Gy, followed by a sequential boost

to PTV1 to 75 Gy. Physician-assigned acute toxicity (Common Termi-

nology Criteria for Adverse Events version 4.0) and QOL data using the

Functional Assessment of Cancer Therapy-Head and Neck Version 4

(FACT-H&N) instrument and McMaster Radiotherapy Questionnaire were

prospectively collected.

Results: Ninety-four pts were enrolled: median age 57 years (range 27-

76), 84% male, 63% HPV+ oropharynx, 54% �10 pack-year tobacco use,

56% �T3, 88% �N2b. Thirty-seven pts (41.6%) had GR to IC. There were

no significant differences in acute grade 3+ dermatitis (27.0% GR vs

25.5% NR, PZ.87) or mucositis (59.5% GR vs 60.8% NR, PZ.90), but

NR were significantly more likely to have a percutaneous endoscopic

gastrostomy (PEG) tube placed during treatment (50.0% GR vs 72.9% NR,

PZ.031) and be PEG tube dependent at 3-month (22.9% GR vs 56.2%

NR, PZ.002) and 6-month follow-up (5.7% GR vs 31.1% NR, PZ.005).

Mean global FACT-H&N QOL scores were similar at baseline (78.9 GR vs

76.0 NR, PZ.21), but significantly higher in GR at 1 month posttreatment

(73.5 GR vs 67.2 NR, PZ.020) and trended toward significance at 6

months (77.3 GR vs 72.2 NR, PZ.070). Mean subjective mouth dryness

scores on the McMaster Radiotherapy Questionnaire were no different at

baseline (6.1 GR vs 5.6 NR, PZ.14), but significantly higher in GR at 1

month posttreatment (3.9 GR vs 3.0 NR, PZ.034).

Conclusion: RAVD is a novel treatment approach that uses IC response to

determine the extent of RT volume reduction. Pts with GR to IC in whom

elective nodal coverage was omitted experienced less toxicity and

improved QOL compared with NR pts. Further investigation is warranted.

Author Disclosure: J.M. Melotek: None. V.M. Villaflor: Research Grant;

Novartis, Celgene. T.G. Karrison: None. R.J. Brisson: None. E.A. Blair:

None. L. Portugal: None. K.M. Stenson: None. J.A. de Souza: None. E.

Cohen: Consultant; Bayer, Novartis, VentiRx, Merck, Bristol-Myers

Squibb, AstraZeneca, Eisai. Speaker’s Bureau; Biodesix, Bayer. A. Lan-

german: None. M.T. Spiotto: None. T.Y. Seiwert: Research Grant;

Genentech/Roche, Boehringer Ingelheim. Honoraria; Novartis, Bayer/

Onyx, Merck. E.E. Vokes: Consultant; Abbvie, AstraZeneca, Boehringer

Ingelheim, Celgene, Clovis Oncology, GeneCentric, Genentech, Merck,

Synta, VentiRx, Eisai, Lilly, Transgene. D.J. Haraf: None.

117Intrafraction Organ Motion Tracking With Real-Time MRI-GuidedRadiation Therapy for Head and Neck CancerJ.C. Rwigema,1 D.H. Thomas,1 M. Cao,1 T. Yoshizaki,1 and A.M. Chen2;1University of California, Los Angeles, Los Angeles, CA, 2University of

California, Los Angeles- David Geffen School of Medicine, Los Angeles,

CA

Purpose/Objective(s): Although intrafraction organ motion in the head

and neck region can affect the precision of radiation therapy (RT) and

conceivably lead to geographic miss, there are essentially no studies

investigating this phenomenon. Using the first commercially available

magnetic resonance imaging (MRI)-guided radiation therapy system, we

evaluated real-time tongue motion related to deglutition for patients “under

beam” for head and neck cancers in an effort to refine planning target

volume (PTV) margin design.

Materials/Methods: Six patients with head and neck tumors, including

those involving the base of tongue (nZ3) and paranasal sinuses (nZ3)

were available for analysis. All patients were treated to doses ranging from

36 to 70 Gy in 20 to 33 fractions using an MRI-guided RT system with on-

board cine MR imaging. To track swallowing motion, 4 reference points

were mapped on daily sagittal cine MR imaging at these positions: (1)

posterior edge of the mandible alveolar process as a reference point with

minimal or no movement and mobile points at (2) most superior point of

tongue dorsum, (3) mid base of tongue, and (4) inferior-posterior edge

point of base of tongue. A significant displacement in tongue movement

was defined as motion greater than 5 mm from baseline noise or airway-

related movement. Three parameters were measured to quantify tongue

motion: tongue displacement distance, deglutition time as time spent in

motion >5 mm, and number of swallowing events. Patients with base of

tongue tumors were compared with the other patients to determine how the

presence of tumor affected tongue motion.

Results: A total of 160 cine MR sagittal image sets were evaluated, with a

median duration of 798 seconds (range, 552-1068 seconds). For the entire

patient group, the mean number of swallowing events/fraction was

9.7�8.7, resulting in a mean tongue displacement of 7.9�2.7 mm/fraction

and a maximum tongue displacement range of 8.5 to 22.3 mm/fraction.

Overall the mean time spent swallowing was 1.9%�2.8% of the fraction

time with an absolute mean time of 13.3�19.9 seconds per fraction. The

maximum time spent swallowing ranged from 4.3% to 25.1% of the

fraction time with absolute values of 20.5 to 198.8 seconds. When the 3

parameters for tongue motion measurement were compared between pa-

tients with tongue tumors and those with uninvolved tongues, no differ-

ences were observed between the 2 groups (p>.05).

Conclusion: Real-time on-board cine MR imaging allows for quantifica-

tion of in-treatment deglutition-related tongue motion. Our findings

demonstrating significant displacement of internal organs including tumor

suggest that planning target volume (PTV) margins may often be inade-

quate to account for intrafraction motion. These results provide a frame-

work for investigation of individualized adaptive PTV margin optimization

and for the evaluation of potential dosimetric implications.

Author Disclosure: J. Rwigema: None. D.H. Thomas: None. M. Cao:

None. T. Yoshizaki: None. A.M. Chen: None.

118Pathological Factors Predicting the Risk of Distant Metastases forHuman PapillomavirusePositive Oropharyngeal Squamous CellCarcinoma (OPSCC)J.N. Lukens,1 K. Tangsriwong,1 N. Mitra,1 R. Cohen,1 G.S. Weinstein,1

B. O’Malley,1 A. Chalian,1 C. Rassekh,1 K.T. Montone,1 P. Ahn,1

H. Quon,2 and A. Lin1; 1University of Pennsylvania, Philadelphia, PA,2Johns Hopkins University, Baltimore, MD

Purpose/Objective(s): Despite superior locoregional control for HPV+

OPSCC, approximately 10% of patients (pts) will develop distant metas-

tases (DM). Current American Joint Committee on Cancer pathological

staging does not adequately identify pts at high risk of DM. There are

emerging data on risk factors for DM in pts treated with chemoradiation,

but very little information on risk factors for DM in pts treated with initial

surgery. Materials/Methods: We performed a retrospective analysis of 174

HPV+ OPSCC pts with �20 months follow-up, treated with transoral

robotic resection (TORS) and neck dissection (ND) followed by adjuvant

chemoradiation therapy (CRT). Pathologic data include the number and

proportion of involved nodes, lowest level of involvement, and extent of

extracapsular extension (ECE), classified as none, microscopic/focal, or

macroscopic/gross. There were 15 pts with clinical N2c disease, of whom

5 underwent bilateral ND; for these pts we report the sum of positive nodes

from both sides of the neck. The Wilcoxon rank sum test was used for

continuous variables, multivariable logistic regression for predictors of

DM, and Kaplan-Meier for survival probabilities.

Results: With a median follow-up of 38 months, 12 pts (7%) developed

DM. The number of positive nodes was the strongest predictor of the risk

of DM, with the probability increasing in proportion to the number of

positive nodes �4. Risk of DM was 14% for �4 positive nodes (vs 3.4%,

PZ.021), 18% for �5 positive nodes (vs 3%, PZ.002), 22% for �6

positive nodes (vs 3.5%, PZ.002), and 28% for �7 positive nodes (vs

3.4%, P<.001). Median positive nodes for pts with and without subsequent

DM was 7 versus 2, PZ.004. Number of nodes examined was similar

between pts with and without DM (median 41). On univariate analysis,

macroscopic/gross ECE was a predictor of DM, with 20% of these pts

developing DM, vs 4.2% of pts with no or microscopic ECE (PZ.007).

Neither microscopic ECE nor clinical/pathologic stage predicted for sub-

sequent DM. No pts with cN2c disease developed DM. A multivariable

Volume 94 � Number 4 � 2016 Posters 879

model revealed that the number of positive nodes (odds ratio [OR] 1.22,

95% confidence interval [CI] 1.07-1.40, PZ.004) remained a significant

predictor of DM, while macroscopic/gross ECE was of borderline signif-

icance (OR 3.30, 95% CI 0.85-12.8, PZ.084). ROC analysis suggests that

�6 positive nodes is a reasonable cut-point for prediction of the risk of DM

(AUC Z 0.75). Pts with �6 positive nodes had inferior 4-year distant

metastasis-free survival (DMFS) and overall survival (OS) compared to pts

with <6 positive nodes: DMFS 80% versus 95% (PZ.002), and OS 88%

versus 97% (PZ.022).

Conclusion: The number of positive nodes appears to be the strongest

predictor of DM in pts with pathologically staged HPV+ OPSCC under-

going primary surgery followed by adjuvant CRT. Pts with �6 positive

nodes have a 22% risk of developing DM and appear to have inferior OS

than pts with fewer positive nodes. If validated, the number of positive

nodes could be incorporated into a revised pathologic staging system for

HPV+ OPSCC, to select pts for intensified or investigational adjuvant

systemic therapy.

Author Disclosure: J.N. Lukens: None.K. Tangsriwong: None. N. Mitra:

None. R. Cohen: None. G.S. Weinstein: None. B. O’Malley: None. A.

Chalian: None. C. Rassekh: None. K.T. Montone: None. P. Ahn: None.

H. Quon: None. A. Lin: None.

119CYFRA 21-1 as an Instant Prognostic Marker of Tumor Response onRadiation With or Without Chemotherapy in Patients With Larynxand Hypopharynx Squamous Cell CarcinomaJ.J. Mrochem-Kwarciak, T. Rutkowski, A. Wygoda, R. Deja, A. Hajduk,

P. Widlak, and K.A. Skladowski; Maria Sklodowska-Curie Memorial

Cancer Center and Institute of Oncology, Gliwice, Poland

Purpose/Objective(s): Radiation therapy (RT) alone or in combination

with chemotherapy (ChT) remains the standard nonsurgical approach for

head and neck squamous cell carcinoma (HNSCC) patients. But in spite

of a relatively high rate of complete tumor responses, locoregional

relapse is still a major kind of treatment failure for HNSCC. The pres-

ence of survived SCC at the time of treatment completion seems to be

very important to know in aspect of further intensification or salvage

conduct, but current methods of discrimination between residual tumor

and treatment-related outcomes are not suitable for such an instant need.

From this point of view, in the present study, 2 serum markers, CYFRA

21-1 (CYFRA) and SCC Antigen (SCC-Ag), have been monitored over

RT in patients with carcinoma of the larynx (LXC) and hypopharynx

(HPC).

Materials/Methods: Ninety-three consecutive patients with LXC (73%)

and HPC (27%) in stages T1 (8%), T2 (44%), T3 (30%), T4 (18%), and N0

(52%), N1 (6%), N2 (34%), and N3 (8%) were treated over 2 years, be-

tween 2009 and 2011 by RT alone (63%), concurrent RT/ChT (16%), and

induction ChT followed by RT (14%) or concurrent RT/ChT (7%). Both

CYFRA and SCC-Ag were estimated twice, before and at the end of the

treatment.

Results: Both CYFRA and SCC-Ag pretreatment levels correlated directly

with T and N stages. Generally, over the treatment CYFRA and SCC-Ag

levels were decreased and increased, respectively. Median follow-up has

been 36 months. Only for CYFRA levels estimated at the end of the

treatment was a significant correlation with the outcome found: at that

time CYFRA was elevated mainly in patients with partial remission

(persistent tumor presented in primary or/and nodal site) with median 2.33

ng/mL, while patients with complete remission had lowered levels with

median 1.65 ng/mL (PZ.0001). Thus, for patients who at the end of the

treatment had a CYFRA level equal to or above 2 ng/mL, the probability

of 3-year LRC and OS were 42% and 40%, respectively, in comparison

with those with a level below 2 ng/mL (82% and 57%, respectively); in

multivariate analysis, this CYFRA estimation has remained the most

significant prognostic factor for locoregional control (PZ.0003) and OS

(PZ.01), and corresponded with a 4- and 2.5-fold increased risk of

locoregional failure and death, respectively. The measurement of CYFRA

21-1 at the end of the treatment had 84% sensitivity and 67% specificity,

and 57% and 88% of both positive and negative prognostic value,

respectively.

Conclusion: Of the CYFRA and SCC-Ag potential markers, only the first

one has prognostic value. When CYFRA 21-1 is measured at the end of the

treatment (RT or concurrent RT/ChT), it seems to be a powerful prognostic

marker for tumor response. In patients with LXC and HPC who undergo

RT or RT/ChT, with or without induction ChT, persistent, uncured tumor is

very feasible if a CYFRA level equal to or higher than 2 ng/mL is

measured instantly at the end of the treatment.

Author Disclosure: J.J. Mrochem-Kwarciak: None. T. Rutkowski: None.

A. Wygoda: None. R. Deja: None. A. Hajduk: None. P. Widlak: None.

K.A. Skladowski: None.

120Weekly Versus Every-3-Weeks Platinum-Based ChemoradiationRegimens for Head and Neck CancerJ.M. Melotek,1 B.T. Cooper,2 M. Koshy,3 J.S. Silverman,2

and M.T. Spiotto3; 1University of Chicago, Chicago, IL, 2New York

University, New York, NY, 3University of Illinois at Chicago, Chicago, IL

Purpose/Objective(s): The majority of chemoradiation (CRT) trials for

locally advanced head and neck squamous cell carcinoma (HNSCC) have

relied on platinum-based chemotherapy regimens administered every 3

weeks. However, given the increased use of weekly platinum regimens, it

remains unclear how different chemotherapy schedules compare regarding

efficacy and toxicity.

Materials/Methods: We identified 212 patients with HNSCC from a

retrospective database who were treated at a single academic medical

center with concurrent platinum-based CRT given weekly or every 3

weeks. One hundred forty-four patients received chemotherapy every 3

weeks, and 68 patients received chemotherapy weekly. JMP version 10

(SAS Institute) was used to perform statistical analysis using 2-sided tests

and defining significance as P<.05. Discrete variables were compared with

the X2 test, and differences in the medians were assessed using the Wil-

coxon test. Survival curves were plotted using the Kaplan-Meier method,

and significance was assessed using the log-rank test. For univariate

analysis and multivariate analysis, we used Cox proportional hazard or

logistic regression models to compare differences in survival or differences

in categorical variables, respectively.

Results: Patients receiving weekly platinum regimens were more likely to

have increased age (median: 61.4 vs 55.5 years; P<.001), high or very high

Charlson comorbidity index (45.6% vs 27.8%; PZ.01), and more likely to

receive carboplatin-based regimens (6.3% vs 76.5%; P<.001). The 2

regimens had similar locoregional control (hazard ratio 1.10, 95% confi-

dence interval [CI] 0.63-1.88, PZ.72), disease-free survival (PZ.28), and

overall survival (PZ.71). Platinum regimens delivered every 3 weeks were

associated with increased days of hospitalization (median: 3 days vs

0 days; PZ.03) and acute kidney injury (AKI) during radiation therapy

(50.0% vs 22.1%; P<.001). On multivariate analysis, AKI was signifi-

cantly associated with regimens delivered every 3 weeks (odds ratio [OR]:

24.38; 95% CI 3.00-198.03; PZ.003) and high comorbidity scores (OR:

2.74; 95% CI 2.15-5.99; PZ.01).

Conclusion: Our results suggest that platinum-containing CRT regimens

delivered every 3 weeks and weekly have similar disease control, but

weekly platinum regimens are associated with less acute toxicity.

Author Disclosure: J.M. Melotek: None. B.T. Cooper: None. M. Koshy:

None. J.S. Silverman: None. M.T. Spiotto: None.

121All High Risk? Pathologic Extracapsular Extension (ECE) in the Eraof Human PapillomaviruseAssociated (HPV+) Head and NeckCancers (HNC)J.K. Russo,1 A. Snow,2 A. Terrell,3 S.L. Mott,4 M. Laszweski,5

A. Hetland,6 B. Liu,7 C. Fischer,8 C.M. Anderson,4 T.A. Dufan,9

and J.M. Watkins4; 1Bismarck Cancer Center Department of Radiation

International Journal of Radiation Oncology � Biology � Physics880

Oncology, Bismarck, ND, 2University of Iowa Holden Comprehensive

Cancer Center, Dept of Pathology, Iowa City, IA, 3Sanford Health, Dept of

Head and Neck Surgery, Sioux Falls, SD, 4University of Iowa Hospitals &

Clinics, Iowa City, IA, 5St. Alexius CHI Medical Center, Dept of Pathology,

Bismarck, ND, 6Mid-Dakota Clinic and CHI St. Alexius Medical Center,

Dept of Head and Neck Surgery, Bismarck, ND, 7University of South

Dakota Sanford School of Medicine, Sioux Falls, SD, 8Sanford Health,

Dept of Pathology, Bismarck, ND, 9Bismarck Cancer Center, Dept of

Radiation Oncology, Bismarck, ND

Purpose/Objective(s): Nodal ECE is an established risk factor for HNC

recurrence and mortality; however, it is unknown whether ECE impacts

outcomes for HPV+ HNC.

Materials/Methods: This was a retrospective multi-institutional compar-

ative outcomes analysis by HPV/p16 status. Eligible patients had patho-

logic confirmation of ECE for HNC involving the oropharynx (OP), oral

cavity (OC), or unknown primary (UP), and underwent curative-intent

therapy. Patients with metastatic disease at diagnosis, unknown HPV/p16

status, or <3 month follow-up were excluded.

Results: From 2003 to 2014, 76 patients were eligible for this study. The

median age at diagnosis was 60 years (range 29-82), with 46 involving

the OP, 28 the OC, and 2 of UP. Forty-one patients (54%) had HPV+

tumors. All but 5 patients underwent therapeutic neck dissection, and

the primary site was resected in 65 patients. For resected primary cases,

38, 23, and 4 patients underwent adjuvant chemoradiation therapy

(CRT), radiation therapy (RT) alone, and no adjuvant therapy, respec-

tively. For 9 patients who underwent definitive RT, 7 received concur-

rent CRT. Of note, 40% of HPV+ and 35% of HPVe patients did not

receive chemotherapy (PZNS). The median number of nodes excised

and involved were 27 (1-92) and 2 (1-32), respectively. At a median

follow-up of 26.3 months (range, 1.4-104.0; median 34.1 for survivors),

52 patients were alive (48 without recurrence, 4 with salvaged recur-

rence), and 24 patients had died (21 of HNCa). Patterns of failure

included local (nZ6), regional (6), locoregional + distant (6), and

distant only (4). In comparing the HPV+ and HPVe groups, disease-

free and overall survival was superior for the HPV+ group (P<.01; Table

1). HPV+ patients were more likely to be male (93% vs 51%), undergo

definitive RT (30% v 20%), have higher stage disease (73% v 49% stage

IV), have larger nodal size (median 3.6 vs 1.9 cm), and be less likely to

have undergone resection of primary (78% v 94%). There were no

differences in number of lymph nodes sampled or involved or in follow-

up between the groups.

Conclusion: HPV+ HNC with ECE has a favorable prognosis despite

more advanced stage and larger nodal burden. Prognosis in the HPVe

population with ECE remains poor despite multimodality therapy (sur-

gery with CRT). ECE in the HPV+ population should be re-evaluated as

a negative prognostic factor and indicator for therapeutic escalation. It is

noteworthy that approximately half of HPV+ patients in our population

did not receive chemotherarpy; further investigation of this is warranted,

considering ongoing efforts at deintensification of therapy for this

subpopulation.

Author Disclosure: J.K. Russo: Employee; CHI St Alexius. Director of

Clinical Research; Bismarck Cancer Center. A. Snow: None. A. Terrell:

None. S.L. Mott: None. M. Laszweski: None. A. Hetland: Partner;

MidDakota Clinic. Shareholder, Board Member, Chair ENT, Compensa-

tion Committee, Surgery Center Committee; Mid Dakota Clinic. B. Liu:

Student; Sanford Medical School. C. Fischer: None. C.M. Anderson:

Research adviser, uncompensated; Galera Therapeutics. T.A. Dufan:

None. J.M. Watkins: Employee; University of Iowa.

Abstract 121; Table 1

HPV status 3-Year outcome (95% CI)

Overall survival Positive 91% (74-97%)Negative 44% (25-61%)

Disease-free survival Positive 91% (74-97%)Negative 40% (22-57%)

122WITHDRAWN

123Subsite Variation in Oropharyngeal Squamous Cell Carcinomas inthe Era of Human Papillomavirus: Tonsillar Fossa Has ImprovedSurvival Compared to Base of TongueA. Platek,1 V. Jayaprakash,2 M. Platek,2 V. Gupta,1 D. Cohan,1 W. Hicks,1

T. Winslow,3 A. Groman,4 S.S. Dibaj,1 H. Arshad,1 M.A. Kuriakose,1

G.W. Warren,5 and A.K. Singh2; 1Roswell Park Cancer Institute, Buffalo,

NY, 2Roswell Park Cancer Institute, Buffalo, NY, United States, 3Roswell

Park Cancer Insitute, Buffalo, NY, 4Roswell Park Cancer Institue, Buffalo,

NY, 5Medical University of South Carolina, Charleston, SC

Purpose/Objective(s): Previous studies on treatment outcomes of patients

with advanced oropharyngeal squamous cell carcinomas (OPSCC) have

rarely analyzed subsite differences in detail in the era of human papillo-

mavirus (HPV). The purpose of this study was to evaluate subsite-specific

differences in survival between squamous cell carcinomas (SCC) of the

base of tongue (BOT) and the tonsillar fossa (TF) in a cohort likely to have

a high incidence of HPV-associated tumors.

Materials/Methods: Retrospective cohort analysis utilizing data from the

Surveillance, Epidemiology, and End Results (SEER) Program. The SEER

cohort included 8073 primary BOT and TF SCC patients without distant

metastases treated between 2004 and 2011. Primary outcome measures

were subsite-based differences in overall survival (OS) and disease-specific

survival (DSS). Cox proportional hazard ratios were estimated.

Results: Among the 8073 primary BOT and TF SCC patients, 3705 (46%)

were BOT and 4368 (54%) were TF. Median age for BOT and TF patients

was 62 and 58 years, respectively. Other clinical characteristics were

similar between groups, but more TF patients had poorly differentiated

tumors. Overall survival with all stages combined favored TF (P<.01) and

remained superior when stratified by stage. In multivariate analyses

adjusted for age, gender, race, and treatment, the hazard ratio (HR) for OS

was superior for TF tumors in comparison to BOT tumors across all dis-

ease stages (stage I HR 1.28, 95% confidence interval [CI] 1.01-1.64; stage

II HR 1.30, 95% CI 1.08-1.59; stage III HR 1.30, 95% CI 1.14-1.49; stage

IV HR 1.14, 95% CI 1.00-1.30). Similar advantages were noted for DSS

favoring improved outcomes for TF.

Conclusion: In this large, modern cohort, OS and DSS favored outcomes

in TF as compared with BOT. Further study is required to evaluate factors

that influence subsite-based survival differences in TF and BOT patients in

the era of HPV.

Author Disclosure: A. Platek: None. V. Jayaprakash: None. M. Platek:

None. V. Gupta: None. D. Cohan: None. W. Hicks: None. T. Winslow:

None. A. Groman: None. S.S. Dibaj: None. H. Arshad: None. M.A.

Kuriakose: None. G.W. Warren: None. A.K. Singh: None.

124Comparisons of Dysphagia and Quality of Life (QOL) in MatchedPatients with HPV-positive Oropharyngeal Cancer ReceivingChemoirradiation or Cetuximab and IrradiationS. Samuels,1 Y. Tao,1 T. Lyden,1 M. Haxer,1 M.E. Spector,2 K. Malloy,2

B. Carol,1 F.P. Worden,1 M. Schipper,1 and A. Eisbruch1; 1University of

Michigan, Ann Arbor, MI, 2University of Michigan Comprehensive Cancer

Center, Ann Arbor, MI

Purpose/Objective(s): We compared functional outcomes of radiation

therapy (RT) concurrent with cetuximab or with chemotherapy for

matched, good prognosis patients with human papillomavirusepositive

(HPV+) oropharyngeal cancer (OPC).

Materials/Methods: Outcomes of patients with HPV+ OPC with minimal

or no smoking history and non-T4/N3/N2C, treated on a prospective

protocol of RT concurrent with cetuximab (cet-RT), were compared with

similar patients on prospective chemo-RT protocols. Intensity modulated

RT in all patients aimed to spare swallowing organs and salivary glands. In

both groups, videofluoroscopy (VF), Common Terminology Criteria for

Volume 94 � Number 4 � 2016 Posters 881

Adverse Events (CTCAE) observer-rated dysphagia (ORD), and validated

QOL questionnaires (Xerostomia Questionnaire [XQ], Head and Neck

QOL, and University of Washington QOL) were performed before RT and

3 and 12 months following RT. Differences between treatment groups were

assessed using t tests, multiple linear regression, and repeated-measures

analysis.

Results: Twenty-six cet-RT patients were matched to 27 chemo-RT pa-

tients with similar baseline characteristics. In the chemo-RT group (median

follow-up of 52 months), no recurrences occurred. In the cet-RT group

(median follow-up of 20 months), 1 patient had persistent microscopic

disease on salvage neck dissection and 1 distant failure. Both groups had

mild VF-based swallowing dysfunction before treatment, which worsened

at 3 months (P<.02) and persisted at 12 months, and did not differ between

treatment groups for any VF measures (P>.11). For both groups, CTCAE

ORD was low before treatment: almost all were grade 0, worsened at 3

months (39% �grade 1), and then improved at 12 months (16% �grade1).

No differences in ORD were observed between treatment groups (PZ.26).

QOL Summary and domain scores for eating were good before treatment,

worse at 3 months, and then improved (P<.01) to near baseline at 12

months. No differences between the 2 groups were seen in any QOL

domain (P>.10).

Conclusion: Both groups, matched for pretherapy factors, had excellent

clinical outcomes without significant differences in objective or subjective

functions. These data do not support using cetuximab instead of chemo-

therapy for treatment deintensification in HPV+ patients.

Author Disclosure: S. Samuels: None. Y. Tao: None. T. Lyden: None. M.

Haxer: None. M.E. Spector: None. K. Malloy: None. B. Carol: None.

F.P. Worden: None. M. Schipper: None. A. Eisbruch: None.

Abstract 126; Table 1 RTT predicted for survival for definitive and adjuvantradiation.

4-Year actuarial survival

nAccelerated

RTTStandardRTT

ProlongedRTT P

Definitivetreatment

16,841 62.2% 60.5% 51.9% <.001

Adjuvanttreatment

4165 74.5% 65.4% 55.0% <.001

125A Phase 1/2 Study of Nab-Paclitaxel, Cisplatin, and CetuximabWith Concurrent Radiation Therapy for Locally Advanced SquamousCell Cancer of the Head and NeckS.G. Chun,1 R. Hughes,2 B. Sumer,3 L. Myers,2 J. Truelson,2 S. Khan,4

T.W. Ma,4 Y. Xie,4 J.S. Yordy,5 S.A. Chen,6 S. Cooley,7 J. Wu,4 H. Choy,2

and L.A. Nedzi3; 1The University of Texas MD Anderson Cancer Center,

Houston, TX, 2University of Texas Southwestern Medical Center, Dallas,

TX, 3The University of Texas Southwestern School of Medicine, Dallas, TX,4UT Southwestern, Dallas, TX, 5Valley Radiation Therapy Group,

Anchorage, AK, 6Pacific Radiation Oncology, Honolulu, HI, 7University of

Texas at Southwestern Medical Center, Dallas, TX

Purpose/Objective(s): Historic control rates for locally advanced head

and neck cancer (HNC) are suboptimal and Nab-paclitaxel may improve

the efficacy of radiation therapy (RT).

Materials/Methods: Concurrent nab-paclitaxel, cisplatin, cetuximab, and

70-Gy RT was evaluated in patients with HNC. Patients with stage III-IV,

cT2N2-3M0, T3-T4, any NM0 of the oropharynx, hypopharynx, or larynx

were eligible for this phase 1/2 trial. The phase 1 study determined the

maximum tolerated dose (MTD) of nab-paclitaxel. The primary endpoint

of the phase 2 component was 2-year progression-free survival (PFS).

Results: Median follow-up was 24 months for 34 patients enrolled. The

MTD of nab-paclitaxel was 20 mg/m2 when combined with 20 mg/m2

cisplatin and 250 mg/m2 cetuximab. The rate of acute grade 3+ mucositis

was 41%, dermatitis 24%, dysphagia 34%, and hematologic suppression

56%. The 2-year PFS was 60% (95% confidence interval [CI] 0.42-0.78),

local control 71% (95% CI 0.55-0.87), and overall survival (OS) 68%

(95% CI 0.50-0.86). Higher RT dose was associated with improved OS

(hazard ratio [HR] 0.71, PZ.047), and prolonged treatment time was

associated with worse PFS (HR 1.36, PZ.038) on multivariate analysis.

Conclusion: This is the first study evaluating cisplatin, cetuximab, and

nab-paclitaxel in humans. This treatment regimen had similar PFS and

toxicity as historic HNC combined modality trials.

Author Disclosure: S.G. Chun: None. R. Hughes: None. B. Sumer: None.

L. Myers: None. J. Truelson: None. S. Khan: None. T. Ma: None. Y. Xie:

None. J.S. Yordy: None. S.A. Chen: None. S. Cooley: None. J. Wu:

None. H. Choy: None. L.A. Nedzi: None.

126The Impact of Chemoradiation Treatment Time on Survival inPatients With Head and Neck CancerT. Shaikh, E. Handorf, C.T. Murphy, R. Mehra, J.A. Ridge,

and T.J. Galloway; Fox Chase Cancer Center, Philadelphia, PA

Purpose/Objective(s): Lengthening radiation treatment time (RTT) yields

inferior control rates for head and neck squamous cell carcinoma. The

consequences of extended RTT in the setting of chemoradiation are un-

clear. The primary objective was to assess the impact of RTT in head and

neck cancers on overall survival (OS) using a modern dataset.

Materials/Methods: Patients diagnosed with tongue, hypopharynx, larynx,oropharynx, or tonsil cancer were identified using the National Cancer

Data Base. RTT was defined as date of first RT to date of last RT. In the

definitive setting (70 gray [Gy] in 35 fractions), prolonged RTT was

defined as >56 days, accelerated RTT was defined as <49 days, and

standard RTT was defined as 49 to 56 days. In the adjuvant setting (60 Gy

in 30 fractions), prolonged RTTwas defined as >49 days, accelerated RTT

was defined as <42 days, and standard RTT was defined as 42 to 49 days.

X2 tests were used to identify predictors of RTT. The Kaplan-Meier

method was used to compare OS among groups. Cox proportional hazards

model was used for OS analysis in patients with known comorbidity

status.

Results: A total of 21,006 patients met the inclusion criteria, a majority

of which (69%) received concurrent chemotherapy. Of 16,841 patients

receiving primary radiation, 23% had an accelerated RTT, 20% had a

prolonged RTT, and 57% had a standard RTT. Of 4165 patients receiving

adjuvant radiation, 29% had an accelerated RTT, 14% had a prolonged

RTT, and 57% had a standard RTT. Concurrent chemotherapy predicted

for patients with prolonged RTT in the primary (P<.001, 6.7% increase)

and adjuvant radiation groups (P<.001, 3.7% increase). Other predictors

of prolonged RTT in the primary and adjuvant setting included female

gender, African American race, higher Charlson comorbidity index,

uninsured status, community cancer facility, higher stage, lower income,

and lower education. Tumor location (oropharynx or hypopharynx pri-

mary) predicted for prolonged RTT in patients receiving primary radi-

ation but not adjuvant radiation. On multivariable analysis, accelerated

RTT was associated with an improved OS (hazard ratio [HR] 0.878, 95%

confidence interval [CI] 0.781-0.986, PZ.029), while a prolonged RTT

was associated with inferior OS (HR 1.266, 95% CI 1.138-1.410,

P<.001).

Conclusion: Prolonged RTT is associated with worse OS in patients

receiving primary or adjuvant radiation, even in the setting of chemo-

radiation. Expeditious completion of radiation should continue to be a

quality metric for the management of head and neck malignancies.

Author Disclosure: T. Shaikh: None. E. Handorf: None. C.T. Murphy:

None. R. Mehra: None. J.A. Ridge: None. T.J. Galloway: None.

International Journal of Radiation Oncology � Biology � Physics882

127The Effect of an Oral Care Intervention in Decreasing theExpression of Proinflammatory Cytokines in Patients ReceivingChemoradiation for Oral Cancer: A Randomized Clinical TrialN.J. Sanfilippo,1 R. Vasconcelos,2 J. Moya,3 D. Malamud,2 C. Barber,2

B.E. Smith,1 M. DeLacure,4 R. Kerr,2 B. Schmidt,2 D. Myssiorek,4

and P. Corby4; 1NYU Langone Medical Center, New York, NY, 2New York

University College of Dentistry, New York, NY, 3New York University

Langone Medical Center, New York, NY, 4New York University School of

Medicine, New York, NY

Purpose/Objective(s): Oral mucositis (OM) is one of the most debili-

tating adverse effects in patients undergoing radiation therapy (RT),

chemoradiation therapy, or both. Currently, there are no effective ther-

apies or preventive treatments for OM; rather, most suggested treatments

are palliative in nature. Physiologically, chemotherapy (CT) and RT

evoke a profound inflammatory response, resulting in damage to the

vascular endothelium. The release of proinflammatory mediators is

responsible for mucosal injury and compromises the integrity of the

protective epithelial barrier, which can result in an increased suscepti-

bility to infection. The objective of this pilot study was to associate the

effects of a novel oral care protocol on OM severity and to evaluate

salivary proinflammatory cytokines in cancer patients undergoing RT or

CT/RT.

Materials/Methods: A total of 16 subjects undergoing RT or CT/RT were

enrolled prior to starting treatment. All subjects received a baseline stan-

dard of care oral/dental prophylaxis plus fluoride application prior to the

start of RT or CT/RT. Patients were assigned to an oral health interven-

tional group (IG) or control group (CG). Subjects assigned to the CG

followed a biweekly treatment schedule in which they had their teeth

brushed by a dental professional and were asked to follow standard of care

(SOC) oral hygiene instructions at home. Subjects randomized to the IG

received the Oral Mucosal Deterging and Periodontal Debridement

(OMDP) protocol and attended weekly treatment visits at which they had

their teeth brushed by a dental professional, periodontal debridement, tooth

polishing, and flossing. Subsequently, the cleansing and deterging of the

oral mucosal surfaces was performed using a soft-bristled toothbrush and

an antibacterial agent (alcohol-free chlorhexidine mouth rinse). Subjects in

the IG were instructed to continue to follow the OMDP protocol at home.

Stimulated whole saliva samples were collected at baseline (prior to OM

development), on the onset of OM, during cancer treatment, and 2 months

after the end of RT or CT/RT. Changes in the levels of proinflammatory

cytokines were measured.

Results: Salivary inflammatory biomarkers, noted in levels of IL-10, IL-

13, IL-4, and TNF-a had a significant increase in the CG and reduced or

stayed the same under IG.

Conclusion: These results suggest that overall inflammation was consis-

tently higher as compared to baseline with control treatment and lower

than or similar to baseline with the OMDP treatment, providing encour-

agement for the effectiveness of the oral care protocol as a coadjuvant

treatment for this population.

Author Disclosure: N.J. Sanfilippo: None. R. Vasconcelos: None. J.

Moya: None. D. Malamud: None. C. Barber: None. B.E. Smith: None.

M. DeLacure: None. R. Kerr: None. B. Schmidt: None. D. Myssiorek:

None. P. Corby: None.

128Three-Dimensionally Printed Bolus in Head and Neck ElectronRadiation therapyD.A. Elliott,1 J. Walker,2 and J.M. Holland1; 1Oregon Health and Science

University, Portland, OR, 2Oregon Health & Science University, Portland,

OR

Purpose/Objective(s): When treating head and neck cutaneous or other

superficial cancers with electron radiation therapy, custom bolus is often

used to optimize dose distribution to improve treatment volume coverage

and minimize dose to normal tissues. We aim to improve on this custom

bolus technique with patient-specific 3-dimensionally printed bolus.

Materials/Methods: Following computed tomographic simulation, the

patient DICOM imaging data is transferred to the radiation treatment

planning system. Target volumes and organs at risk are contoured and a

plan is created using a standard, noncustom bolus technique. Following

completion of the standard plan, a custom bolus is then created as a

DICOM-RT structure with forward planning in an attempt to decrease

unnecessary dose to organs at risk and increase treatment volume coverage

and dose homogeneity. Custom bolus is tagged with a marker to assist in

orientation and alignment during patient setup. An in-house algorithm is

then used to translate the custom bolus from DICOM-RT to stereo-

lithography file format to transfer to a 3-dimensional printer.

Results: Our institution has successfully modeled custom bolus and con-

verted the bolus to appropriate format for 3-dimensional printing; creating

personalized bolus for individual patients that can be used in electron ra-

diation therapy to the head and neck. Our preclinical model supports

improved treatment volume coverage and decreased dose to normal tissues

with the printed custom bolus. Images of the comparative dosimetric

evaluation of the printed custom bolus to standard bolus electron radiation

therapy will be presented.

Conclusion: We have developed a model to attempt to personalize head

and neck electron radiation therapy with a 3-dimensionally printed custom

bolus. This model is currently under preclinical testing for head and neck

cancer.

Author Disclosure: D.A. Elliott: None. J. Walker: None. J.M. Holland:

None.

129Induction Chemotherapy Followed by Concurrent ChemoradiationTherapy Versus Concurrent Chemoradiation Therapy Upfront inLocally Advanced Oral Cavity Cancer: Systematic Review and Meta-Analysis of Individual DataG.N. Marta, Sr,1 R. Riera,2 R.I. Haddad,3 E. Cohen,4 G. Rabinowits,5

and L.P. Kowalski6; 1Department of Radiation Oncology Hospital Sırio-

Libanes and Instituto do Cancer do Estado de Sao Paulo - Faculdade de

Medicina da Universidade de Sao Paulo, Sao Paulo, Brazil, 2Brazilian

Cochrane Center and Discipline of Emergency Medicine and

EvidenceeBased Medicine / Universidade Federal de Sao Paulo e Escola

Paulista de Medicina (UNIFESPeEPM), Sao Paulo, Brazil, 3Department

of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical

School and Department of Medicine, Brigham and Women’s Hospital,

Boston, MA, 4University of California San Diego, La Jolla, CA, 5Head and

Neck Oncology Program /Harvard Medical School, Dana Farber Cancer

Institute, Boston, MA, 6Department of Head and Neck Surgery and

Otorhinolaryngology AC. Camargo Cancer Center, Sao Paulo, Brazil

Purpose/Objective(s): The initial treatment for locally advanced oral

cavity squamous cell carcinoma (SCC) comprises surgery followed by

radiation therapy with or without chemotherapy. However, nonsurgical

treatment could be an option. This study was performed to evaluate the

effectiveness of induction chemotherapy in nonsurgical protocols for oral

cavity SCC patients.

Materials/Methods: Randomized controlled trials evaluating sequential

therapy (induction chemotherapy followed by chemoradiation therapy

[Chemo group]) versus chemoradiation therapy alone (Control group) in

head and neck cancer were analyzed. Two of the authors independently

evaluated the studies regarding eligibility criteria and risk of bias.

Results: Three studies fulfilled the eligibility criteria. All of them included

different sites of locally advanced head and neck SCC. The individual data

of oral cavity cancer patients were retrieved from 2 studies (Paradigm and

Decide trials). Data from the third study were not retrieved, and this trial

was not considered for analysis. A total of 65 patients were randomly

assigned to the Chemo group (nZ34) and the Control group (nZ31). Both

trials were classified as having low risk of bias. No significant overall

benefit in favor of induction chemotherapy was found regarding mortality

rate (Chemo group, 9/34; Control group, 8/31 [3-year mortality:

Volume 94 � Number 4 � 2016 Posters 883

heterogeneity: X2Z0.00, dfZ1 (PZ.97); I2Z0%; test for overall effect:

ZZ0.04, PZ.97) and progression-free survival (Chemo group 9/34;

Control group 12/31 [3-year progression-free rate: heterogeneity:

X2Z0.07, dfZ1 (PZ.79); I2Z0%; test for overall effect: ZZ1.05,

PZ.29]).

Conclusion: Induction chemotherapy when administered before chemo-

radiation therapy did not improve clinical outcomes in comparison to

upfront chemoradiation therapy in patients with locally advanced oral

cavity cancer.

Author Disclosure: G.N. Marta: None. R. Riera: None. R.I. Haddad:

None. E. Cohen: None. G. Rabinowits: None. L.P. Kowalski: None.

130Risk Factors for Local, Regional, or Distant Recurrence in HumanPapillomavirusePositive Oropharyngeal CancerJ.J. Caudell, T. Strom, A.O. Naghavi, J.M. Frakes, L.B. Harrison,

and A. Trotti; H. Lee Moffitt Cancer Center and Research Institute, Tampa,

FL

Purpose/Objective(s): Human papillomavirus (HPV)-associated squa-

mous cell carcinomas of the oropharynx are increasingly seen as a separate

disease entity from smoking-associated oropharyngeal cancers, with a

much higher likelihood of cure. The current American Joint Commission

on Cancer (AJCC) staging system for HPV-positive oropharyngeal cancer

is not prognostic for outcome. We evaluated a variety of potential prog-

nostic factors in order to propose potential new components for a staging

system.

Materials/Methods: After gaining institutional review board approval, we

queried an institutional database for patients with HPV or p16-positive

nonmetastatic oropharyngeal cancers treated with definitive radiation

therapy (RT), and 245 cases were identified. Patient, tumor, and treatment

factors were abstracted from the charts. In addition, pretreatment imaging

was reviewed, including computed tomography (CT) in 99.6%, positron

emission tomography/CT in 94.3%, and magnetic resonance imaging in

7.3% to obtain precise size, location, number, and extent of primary and

nodes. Outcomes, including local control (LC), regional control (RC),

locoregional control (LRC), and freedom from distant metastases (FFDM)

were calculated from the end of RT and estimated via Kaplan-Meier

method. Comparisons were made via log-rank test.

Results: Median follow-up of patients alive at last contact was 36 months.

All patients were treated with definitive RT alone (nZ38, 15.4%) or

concurrent systemic therapy and RT (nZ209, 84.6%). LC was seen in 239

of 245 patients, for a 3-year LC rate of 97.8%. There were no statistically

significant prognostic factors for local control, including tumor size or

invasion of adjacent structures. RC was achieved in 235 of 245 patients

(95.3% at 3 years). RC was less likely if there were 5 or more nodes (1-4

vs �5, PZ.05), or if a lymph node was present in level 4 (level 3 or above

vs level 4, PZ.005). Distant metastases occurred in 21 patients, for a 3-

year FFDM rate of 91.4%. Lower rates of FFDM were associated with a

lymph node greater than 6 cm (PZ.02), bilateral lymphadenopathy (uni-

lateral vs bilateral, PZ.034), 5 or more nodes (1-4 vs �5, P<.001), or if a

lymph node was present in level 4 (level 3 or above vs level 4, P<.001).

Conclusion: Outcomes for patients with HPV-associated oropharyngeal

cancer treated with definitive RT are excellent. The increasing burden of

adenopathy, either by size, number, or bilateral involvement, or location in

level 4 portended a higher risk of regional failure or metastasis. These

factors may provide a basis for altering staging.

Author Disclosure: J.J. Caudell: None. T. Strom: None. A.O. Naghavi:

None. J.M. Frakes: None. L.B. Harrison: None. A. Trotti: None.

131Low-Level Laser Therapy and Laser Debridement for Managementof Oral Mucositis in Patients With Head and Neck Cancer ReceivingChemotherapy and RadiationE. Allan,1 C. Barney,1 S. Baum,2 T. Kessling,3 V.M. Diavolitsis,1

D. Blakaj,1 J.C. Grecula,1 J.W. Rocco,4 M. Van Putten,3 and A.D. Bhatt1;1The James Cancer Hospital and Solove Research Institute, Wexner

Medical Center at The Ohio State University, Department of Radiation

Oncology, Columbus, OH, 2The Ohio State University Wexner Medical

Center, College of Medicine, Columbus, OH, 3The James Cancer Hospital

and Solove Research Institute, Wexner Medical Center at The Ohio State

University, Department of Maxillofacial Prosthodontics, Columbus, OH,4The James Cancer Hospital and Solove Research Institute, Wexner

Medical Center at The Ohio State University, Department of

Otolaryngology-Head and Neck Surgery, Columbus, OH

Purpose/Objective(s): Oral mucositis (OM) is a very common side effect

of head and neck radiation therapy (RT) and concurrent chemotherapy and

radiation (CRT) leading to severe pain, infection, weight loss, higher rates

of hospitalization, higher financial cost of treatment, and breaks in therapy

resulting in increased morbidity and reduced treatment efficacy. The 2014

guidelines from The Mucositis Study Group of the Multinational Associ-

ation of Supportive Care in Cancer included a suggestion that low-level

laser therapy (LLLT) was useful for the prevention of high-grade OM in

patients receiving head and neck RT alone, but no comment could be made

regarding its efficacy in the therapeutic setting for severe OM during RT or

CRT. Our objective is to describe the technical aspects of the laser regimen

we use in this setting and relate the qualitative experience we have had thus

far.

Materials/Methods: Since 2013, 53 patients were referred to dentistry for

laser therapy for significant and bothersome Radiation Therapy Oncology

Group grade 2-3 OM, either during or after RT/CRT. Our regimen uses 2

lasers; a class IV Er, Cr: YSGG laser (l Z2.780 mm), as well as a class IV

diode laser (l Z940 nm). The first laser is used to debride the entire

surface of the ulcerated areas. Settings are 0.25 to 0.75 watts, 15 to 20 pps,

0 water, and 90 air flow rate. The next laser is then used as biostimulation

for pain relief during CRT and for wound healing after CRT is completed.

Instrument settings are 0.6 watts, 12 joules, continuous wave pulse (CW)

for 20 seconds per site and 0.2 watts, (CW), 4 joules, 20 seconds per site

for each situation, respectively. Treatments were administered once or

twice weekly depending on severity and continued until complete reso-

lution of ulcerations.

Results: Forty-one patients started laser treatments either during or

within 1 month of RT/CRT completion. Twelve patients initiated treat-

ments over 1 month after RT/CRT completion. All patients experienced

full clinical resolution of oral ulcerations. The number of treatments

patients received ranged from 2 to 15 with a median number of 7.

Qualitatively, patients tended to report significant early pain relief,

especially during the first 48 hours following laser treatments. Providers

also felt that patients tended to heal more quickly once treatment was

initiated, though, without a comparison group at this point, no definitive

conclusion can be reached.

Conclusion: This report highlights the technical aspects of LLLT and our

regimen for managing severe mucositis. Our qualitative experience thus far

suggests benefits in pain relief and quicker recovery from severe mucositis.

Additional studies are underway evaluating the feasibility, efficacy, and

quality of life metrics.

Author Disclosure: E. Allan: None. C. Barney: None. S. Baum: None. T.

Kessling: None. V.M. Diavolitsis: None. D. Blakaj: None. J.C. Grecula:

None. J.W. Rocco: None. M. Van Putten: None. A.D. Bhatt: None.

132A Comparative Study of Patient-Reported Quality of Life,Xerostomia, and Dysgeusia in Oropharyngeal Squamous CellCarcinoma (OPSCC) Treated With Volumetric Modulated Arc Therapy(VMAT) or Proton Pencil Beam Scanning (PBS)P. Ahn,1 A. Lin,2 O. Zhou,1 J.N. Lukens,1 and S. Sharma3; 1University of

Pennsylvania, Philadelphia, PA, 2Department of Radiation Oncology,

University of Pennsylvania, Philadelphia, PA, 3University of Pennsylvania,

Philadelphia, PA, United States

Purpose/Objective(s): Treatment for OPSCC can affect a number of

outcomes related to quality of life (QOL). We explore differences in

measures such as xerostomia, dysgeusia, weight loss, pain, and overall

International Journal of Radiation Oncology � Biology � Physics884

functional status in patients treated with either VMAT or PBS with and

without cisplatin for OPSCC.

Materials/Methods: We examined 95 consecutive patients from April

2013 to November 2014 who were treated with definitive radiation

therapy (RT) or adjuvant RT following transoral robotic surgery (TORS).

Individual ipsilateral and contralateral saliva-producing structures

including parotid, submandibular, and sublingual glands, as well as

buccal mucosa, tongue, hard palate, soft palate, and upper and lower lips

were contoured. Patient-reported QOL questionnaires (nZ229),

including the European Organization for Research and Treatment of

Cancer (EORTC)-QLQ30, EORTC-H&N35, Work Status, Dysgeusia,

and GRiX xerostomia survey, were prospectively assessed at consult and

during follow-up visits.

Results: Fifty-seven patients (60%) were treated with VMAT, and 38

patients (40%) were treated with PBS. Fifty-two patients (55%) were

treated after TORS, and 43 (45%) with definitive RT or chemoradiation.

Patients were followed for a median of 6 months. PBS conferred a sig-

nificant 1.4- to 11.3-fold decrease in dose compared to VMAT for

contralateral parotid and submandibular glands, bilateral sublingual

glands, buccal mucosa, tongue, hard palate, soft palate, and lower and

upper lip. Forty-four patients (18 PBS) had 6-month follow-up question-

naires. At 6 months, 65% of patients treated with VMAT had moderate-

severe xerostomia compared to 39% for PBS (PZ.08), with no difference

in mild-severe sticky saliva (54% VMAT vs 61% PBS, PZ.63) or mild-

severe appetite changes (42% VMAT vs 33% PBS, PZ.55). There was a

trend toward improved taste in patients treated with PBS (56% PBS vs

31% VMAT, PZ.1) at 6 months. In the subset of patients who received

cisplatin (nZ24), PBS was associated with significantly higher global

health domain scores (PBS 92 vs VMAT 77, higher number is better,

PZ.04), less pain (PBS 2 vs VMAT 18, PZ.007), and lower painkiller use

(PBS 0 vs VMAT 40, PZ.002).

Conclusion: In our cohort of patients, PBS confers a significant dosimetric

advantage to most saliva-producing oral cavity structures. We find a sug-

gestion that this translates to a decrease in moderate-severe xerostomia as

well as to taste changes for PBS compared to VMAT at 6 months. For

patients receiving concurrent cisplatin, PBS is associated with an improved

global health domain score, lower pain, and less painkiller use. To our

knowledge this is the first study comparing VMAT and proton QOL out-

comes in OPSCC patients. We are continuing to increase patient accrual

and follow-up in this cohort. We aim to expand our current findings and

report additional results in the near future.

Author Disclosure: P. Ahn: Patent/License Fees/Copyright; Pyronia

Medical. A. Lin: None. O. Zhou: None. J.N. Lukens: None. S. Sharma:

None.

133Clinical Outcomes With Mucosal Sparing Radiation Therapy inResected Oropharyngeal CancerM.E. Gamez,1 M.Y. Halyard,1 M.L. Hinni,2 H.E. Richard,2 T.H. Nagel,2

C. Vargas,3 W.W. Wong,2 K. Curtis,1 M.A. Zarka,1 and S.H. Patel1; 1Mayo

Clinic, Scottsdale, AZ, 2Mayo Clinic, Phoenix, AZ, 3Mayo AZ, Phoenix, AZ

Purpose/Objective(s): Oropharyngeal cancer has been treated historicallywith either primary radiation therapy (RT) or open surgery with similar

disease control and survival rates but a greater complication rate in patients

(pts) treated with surgery. However, minimally invasive surgical ap-

proaches such as transoral surgery (TOS) have emerged with comparable

oncologic outcomes and with organ preservation. Currently, there are

ongoing efforts to develop risk-adapted therapies that can potentially

deintensify treatment regimens for favorable groups like human papillo-

mavirus (HPV)-associated oropharyngeal squamous cell carcinomas

(OPSCC). We report the outcomes of HPV-related OPSCC pts that un-

derwent mucosal sparing RT in resected OPSCC.

Materials/Methods: We identified for the present institutional review

boardeapproved analysis a subset of 40 pts treated at Mayo Clinic Ari-

zona with TOS followed by neck only radiation using intensity modulated

radiation therapy (IMRT). These were favorable-risk patients with nega-

tive margins, needing RT secondary to risk factors in the neck: multiple

positive lymph nodes, lymph nodes >3 cm, and/or extracapsular exten-

sion. The median age was 57.8 years. Thirty-eight pts (95%) had transoral

laser microsurgery. Twenty-four pts (60%) had base of tongue tumors, and

16 (40%) had tonsillar tumors. The majority of the pts (70%) were

documented to be HPV positive and had T1 N2b disease. T-stage distri-

bution was as follows: T1, 27 pts (67%); T2, 12 pts (30%); and T3, 1 pt

(3%). N-stage distribution was as follows: N1, 4 pts (10%); N2a, 4 pts

(10%); N2b, 27 pts (67%); and N2c, 5 pts (13%). Therefore, 90% of pts

had N2 disease. Twenty-three pts (57%) received concurrent chemo-

therapy, consisting of cisplatin in 18 (78%) pts and 5 (22%) pts with

cetuximab.

Results: The median follow-up for surviving pts was 51 months (range,

13-155 months). The median RT dose to the neck was 6000 cGy (range,

5400-6400 cGy). There were no local failures and only 1 regional

failure, resulting in 97.5% locoregional control at 4 years. Two pts

developed distant metastatic disease, without evidence of locoregional

recurrence, for a 4-year overall survival of 97%. The 4-year recurrence-

free survival was 94%. RT-related acute toxicities were grade �2

dysphagia in 5 pts (12%). Only 1 patient was percutaneous endoscopic

gastrostomy tube dependent at 1 year, but none were dependent at last

follow-up. Xerostomia grade �2 was noted in 7 pts (17%), oral

mucositis grade �2 in 4 pts (10%), neck stiffness in 2 pts (5%), and

trismus in 1 pt (2.5%).

Conclusion: Our analysis suggests that mucosal sparing RT after TOS in

oropharyngeal cancer pts provides comparable oncologic and improved

functional outcomes in selected pts. Sparing the mucosal surface of the

primary site appears feasible without impacting on survival or locoregional

control. This new treatment paradigm in resected OPSCC is promising and

requires further validation through a prospective trial.

Author Disclosure: M.E. Gamez: None. M.Y. Halyard: None. M.L.

Hinni: None. H.E. Richard: None. T.H. Nagel: None. C. Vargas: None.

W.W. Wong: None. K. Curtis: None. M.A. Zarka: None. S.H. Patel:

None.

134The Use of Predictive Modeling in Adaptive Radiation Therapy forHead and Neck CancerE. Brown,1 R. Owen,2 F. Harden,2 K. Mengersen,2 K. Oestreich,3

W. Houghton,4 M.G. Poulsen,5 S. Harris,6 C. Lin,6 and S. Porceddu7;1Princess Alexandra Hospital Radiation Oncology Department,

Woolloongabba, Australia, 2Queensland University of Technology,

Brisbane, Australia, 3Princess Alexandra Hospital Radiation Oncology

Department, Brisbane, Australia, 4Radiation Oncology Mater Centre,

Brisbane, Australia, 5Mater Radiation Oncology, South Brisbane, QLD,

Australia, 6Royal Brisbane and Women’s Hospital Radiation Oncology

Department, Brisbane, Australia, 7Princess Alexandra Hospital,

Woolloongabba, Australia

Purpose/Objective(s): Adaptive radiation therapy (ART) can account for

the dosimetric impact of anatomic and tumor changes throughout the

course of chemoradiation for locally advanced head and neck cancer

(HNC). However, it is time- and resource-intensive, making identification

of patients most likely to require ART of vital importance. Logistic

regression enables prediction of the risk of events occurring, and classi-

fication and regression tree (CART) analysis is a simple analytic tool that

helps determine the key variables in a population to design an explanatory

model. This makes them attractive tools to use in the setting of HNC ART.

We sought to investigate the utility of logistic regression modeling and

CART analysis in developing predictive models for HNC patients likely to

benefit from ART.

Materials/Methods: Patients with node-positive oropharyngeal squamous

cell carcinoma (OPC) and nasopharyngeal carcinoma (NPC) treated with

curative-intent chemoradiation were enrolled prospectively in the study.

Patients underwent a second planning computed tomographic scan if the

Volume 94 � Number 4 � 2016 Posters 885

change in external contour between the planning scan and daily treatment

scan was >1 cm. The dosimetric impact was assessed and a replan

generated if target volume coverage was inadequate or organs at risk dose

exceeded tolerance. Patient demographics and tumor characteristics were

recorded and compared between patients who were replanned and those

that were not. Univariate and multivariate analyses were performed and

factors found to be significant for replanning included in logistic regression

and CART analysis. To assess the logistic regression and CART analysis

with larger patient numbers, it was repeated on all patients who underwent

a second planning scan, making the assumption that this scenario always

necessitates replanning.

Results: One hundred and ten patients were enrolled between October

2013 and December 2014. The majority were OPC (84.5%) and male

(91.8%) and they were predominantly classified at the T2 (33.6%) and N2

(80%) stage. Twenty-one patients (19.1%) underwent a second planning

scan, and of these, 5 (4.9%) patients underwent a replan. Nodal disease

stage, pretreatment size of the largest node, diagnosis (P<.01), and initial

weight (categorized in 2 groups) (P<.07) were identified as significant for

inclusion in the logistic regression model predicting the need to replan.

When the percentage of patients replanned was increased, nodal disease

stage (PZ.06), pretreatment size of the largest node, diagnosis, initial

weight, and percentage weight change (P<.01) were identified as signifi-

cant for inclusion in the logistic regression model.

Conclusion: Predictive modeling, using logistic regression and CART

analysis, can be utilized in the identification of OPC or NPC patients more

likely to require ART. This could facilitate the efficient implementation of

ART resulting in the appropriate allocation of institutional resources.

Author Disclosure: E. Brown: None. R. Owen: None. F. Harden: None.

K. Mengersen: None. K. Oestreich: None. W. Houghton: None. M.G.

Poulsen: None. S. Harris: None. C. Lin: None. S. Porceddu: None.

135Comprehensive Geriatric Assessment as a Predictor of Tolerance,Quality of Life, and Toxicity in Older Patients Receiving RadiationN.A. VanderWalde,1 A.M. Deal,2 E. Comitz,3 L. Stravers,3 H. Muss,4

B. Reeve,5 E. Basch,6 and B.S. Chera7; 1University of Tennessee Health

Science Center and West Cancer Center, Memphis, TN, 2Lineberger

Comprehensive Cancer Center, Chapel Hill, NC, 3UNC School of

Medicine, Chapel Hill, NC, 4University of North Carolina Hospitals,

Chapel Hill, NC, United States, 5UNC Gillings School of Global Public

Health, Chapel Hill, NC, 6Lineberger Comprehensive Cancer Center and

UNC Chapel Hill, Chapel Hill, NC, 7University of North Carolina

Hospitals, Chapel Hill, NC

Purpose/Objective(s): Comprehensive geriatric assessments (CGA) pre-

dict toxicity from chemotherapy and surgery. This study’s purpose is to

evaluate associations between dysfunction as measured by a CGA and

tolerance to treatment, changes in quality of life (QOL), and patient-re-

ported symptoms (PRS) in patients with lung or head and neck cancer

(HNC) receiving radiation therapy (RT) or chemoradiation therapy (CRT).

Materials/Methods: We conducted a prospective observational cohort

study, evaluating the predictive value of CGA in eligible patients 65 years

and older with HNC or lung cancer undergoing curative-intent RT or CRT.

Pretreatment CGA, QOL (European Organization for Research and

Treatment of Cancer QLQ-C30), and PRS (Patient-Reported Outcomes

version of the Common Terminology Criteria for Adverse Events) were

obtained. Questionnaires were repeated biweekly during RT and at 6 weeks

posttreatment. Dysfunction was defined as scores <14 on the Instrumental

Activities of Daily Living (I-ADL) scale. Poor tolerance to treatment was

defined by hospitalization, >3 day treatment delay, change in RT or CRT

regimen, and death. Associations between I-ADLs and QOL changes be-

tween baseline and end of RT as well as end of RT to 6 weeks postradiation

were evaluated. Associations between I-ADL<14 and tolerance to RT as

well as PRS ratings were also evaluated. Fisher exact test was used.

Results: Of the 50 patients accrued, 46 had evaluable data. Mean age was

72.5 years (range 65-92). Sixty-one percent had HNC. Forty-six percent

received CRT. At baseline, 37% had I-ADL<14. Thirty-five percent

required a gastronomy (G)-tube. Thirty-nine percent had poor tolerance to

RT or CRT. There was no association between I-ADL<14 and tolerance.

HNC patients with I-ADL<14 required a G-tube more often than those

with I-ADLZ14 (86% vs 45%, PZ.09). Complete QOL data through the

end of RTwere available for 77% and through post-RT follow-up for 50%.

Patients with I-ADL<14 had lower baseline QOL scores (Global Health

Status Domain [GHS], P<.01). From baseline to end of RT those with

baseline I-ADL<14 had less of a decline in Role Functioning (RF)

(PZ.01) and GHS (PZ.03) domains. However, from the end of RT to the

6-week follow-up, those with I-ADL<14 were more likely to continue to

drop and less likely to improve in the RF (PZ.02) and Social Functioning

(PZ.03) domains. I-ADL<14 at baseline was also associated with higher

severity of PRS shortness of breath (P<.01), pain (P<.01), loss of taste

(P<.01), cough (PZ.04), dry skin (PZ.02), anxiety (PZ.05), depression

(PZ.01), and concentration interference (PZ.02) during RT.

Conclusion: Pretreatment functional deficits were associated with

continued decline and lack of recovery of QOL in this patient population.

Furthermore, patients with pretreatment functional deficits reported higher

severity of symptoms. Larger studies could further elucidate the CGA’s

predictive value.

Author Disclosure: N.A. VanderWalde: Research Grant; Conquer Cancer

Foundation. A.M. Deal: None. E. Comitz: None. L. Stravers: None. H.

Muss: Consultant; Pfizer. Co-Chair of Committee; Alliance for Clinical

Trials in Oncology. B. Reeve: None. E. Basch: Associate Editor; Journal

of American Medical Association. Chair of Committees; American Soci-

ety of Clinical Oncology. Member of Board; National Cancer Institute.

B.S. Chera: None.

136Skin Cancer of the Head and Neck With Perineural Spread: Patternsof FailureE. Sapir, A.A. Tolpadi, S. Samuels, M. Ibrahim, E. Elalfy, J.B. McHugh,

and A. Eisbruch; University of Michigan, Ann Arbor, MI

Purpose/Objective(s): To analyze patterns of failure of patients with headand neck cutaneous squamous cell carcinoma (HNCSCC) with radiolog-

ical or gross cranial nerve involvement (GCNI), microscopic focal peri-

neural invasion (MFPNI), and microscopic extensive perineural invasion

(MEPNI) managed with or without radiation therapy (RT).

Materials/Methods: After a review of charts, we identified RT plans and

radiologic studies of 106 patients with HNCSCC with PNI or GCNI

who were either observed or treated from 2000 through 2013 with

adjuvant RT. The pathology specimens were prospectively reviewed by

the study’s pathologist (J.M.). Cox proportional hazards models were

used to estimate disease-free survuval (DFS) and recurrence-free sur-

vival (RFS).

Results: Median follow-up for all patients was 19.8 months. In patients

treated with RT, the skin tumor bed was irradiated together with ipsi-

lateral lymph nodes. Median dose to gross disease was 66 Gy and 60

Gy to the course of nerves and structures that were judged at risk.

Chemotherapy was used with RT in 24 cases. Thirty-five patients had

GCNI: the involved nerves were the facial and the branches of the

trigeminal nerves. GCNI distribution was as follows: single nerve

involvement, 11 patients, and >1 nerve GCNI, 14 patients. Clinical

target volume (CTV) included the involved cranial nerves, in 95%

additional high-risk nerves (cranial nerves VII and V1-V3), and in 83%

base of skull ganglions were treated electively. Thirteen of 35 (37%)

patients with GCNI failed in the treated nerves in-field, of whom 2 also

failed in previously untreated neural ganglions, and 2 pts failed along

nerves not electively treated that communicate with involved nerves.

Seventy-eight percent of patients with gross ganglion involvement at

presentation have failed in-field within the ganglions. No relapses

occurred in the electively irradiated neural ganglions and cranial

nerves. Nineteen of thirty patients (63%) with MEPNI in the skin

specimens without evidence of GCNI were treated with RT. CTV in

International Journal of Radiation Oncology � Biology � Physics886

these cases included nerves innervating the involved skin dermatome

but not SBG. Seventy-three percent of observed patients and 31% of

irradiated patients with MEPNI recurred. Two years RFS in nerves

(94% vs 25% respectively; hazard ratio [HR] 0.06, 95% confidence

interval [CI] 0.006-0.5, PZ.01) and DFS (73% vs 40%, HR 0.32, 95%

CI 0.1-0.99, PZ.05) rates were significantly higher in the treated

MEPNI pts compared with the observed. Twenty-seven of 37 patients

(73%) diagnosed with MFPNI were observed; the rest were treated with

RT. Patients with MFPNI had a low rate of neural and overall failure,

and there was no significant benefit to irradiation in these patients.

Conclusion: Our study demonstrates the patterns of failure and the role of

RT for patients with HNCSCC, GCNI, MEPNI. Awareness of these pat-

terns and knowledge of the cranial nerves anatomy should serve as

guidelines for target volume delineation. In patients with MEPNI, RT is

associated with fewer gross perineural recurrences and better DFS,

compared with the observation strategy.

Author Disclosure: E. Sapir: None. A.A. Tolpadi: None. S. Samuels:

None. M. Ibrahim: None. E. Elalfy: None. J.B. McHugh: None. A.

Eisbruch: None.

137Continuity of Care Follow-up Update: Results of a ConsensusMultidisciplinary 8-Week Postradiation Tumor Response EvaluationAlgorithmA.C. Hessel,1 B.M. Beadle,1 R.S. Weber,1 and L.E. Burke2; 1The

University of Texas MD Anderson Cancer Center, Houston, TX, 2MD

Anderson Cancer Center, Houston, TX

Purpose/Objective(s): Patients with human papillomavirus (HPV)-asso-

ciated cancers are often treated with multimodality therapy; they have

excellent overall survival with expectations of long follow-up for both

recurrence and toxicity. In an effort to standardize routine surveillance,

decompress clinic volumes, and reduce resource utilization, the Continuity

of Care (COC) pathway was created. COC is a rotating schedule among all

physicians and imaging obtained at regular intervals to 2 years and then a

transition to survivorship clinic. This study presents an analysis of pathway

compliance for the first posttreatment COC visit and whether that

encounter appropriately triages patients for further assessment of regional

disease persistence and/or lymphadenectomy.

Materials/Methods: The COC project began enrollment in May 2012.

Those patients dispositioned to definitive (chemo)radiation at the

multidisciplinary tumor board were identified as eligible. The majority

of patients were stage III or IV, and COC pathway was designed to

assess the locoregional status at 8 weeks posttreatment by all pro-

viders. At that visit, all providers examined the patient and reviewed

the computed tomographic (CT) scan. At that time, the patient was

dispositioned to surgical intervention, routine COC follow-up in 3

months, or an interim visit in 4 to 6 weeks consisting of an ultrasound

and/or positron emission tomographic scan. If an interim visit was

needed, the decision was made to move to COC follow-up or proceed

with surgery.

Results: In the first year, there were 131 eligible patients with

oropharynx or unknown primary cancers. At the posttreatment 8-week

visit, 2 patients (1.5%) had planned neck dissections, and 129 (98.5%)

had routine imaging. This single evaluation found 66 patients (50.4%) to

have no evidence of disease in both the primary and cervical lym-

phatics; they were transitioned directly to COC follow-up. Ten patients

(7.6%) were taken to surgery, and 53 patients (40.5%) were scheduled

for an interim visit. At the interim visit, 51 were found to have no

evidence of disease and went to routine COC follow-up; only 2 went to

surgery. Of the entire cohort, only 14 patients (10.7%) had surgery; 10

had no tumor, 1 had PTC, 1 had persistent primary disease, and 2 had

persistent neck disease.

Conclusion: The COC pathway was developed to create standardized

cancer surveillance for the head and neck patients. These results show that

an 8-week posttreatment visit with CT imaging is an optimal early

assessment of tumor response, with the majority of patients being able to

transition directly to routine follow-up. For those with ambiguous findings,

there is adequate time for additional imaging or early surgical intervention.

Continued assessment of this standardized process will assess the value of

routine imaging as well as tumor prognosis.

Author Disclosure: A.C. Hessel: None. B.M. Beadle: None. R.S. Weber:

None. L.E. Burke: None.

138Methods for Reducing Normal Tissue Complication Probabilities(NTCP) in Oropharyngeal Cancer (OPC): Dose De-escalation orRemoval of Planning Target VolumesS. Samuels,1 A. Eisbruch,1 K.A. Vineberg,1 C. Lee,1 M.M. Matuszak,1

R.K. Ten Haken,1 and K.K. Brock2; 1University of Michigan, Ann Arbor,

MI, 2Department of Radiation Oncology, University of Michigan, Ann

Arbor, MI

Purpose/Objective(s): Treatment deintensification is being investigated

for patients with human papillomavirus (HPV)-positive OPC. Additionally,

toxicity may be reduced by reducing treatment uncertainty and decreasing

or eliminating the planning target volumes (PTV). The purpose of this

study was to compare the NTCP calculated for swallowing and salivary

structures for standard volumetric modulated arc therapy (VMAT) plans,

dose-reduced plans, and plans eliminating the PTV margin.

Materials/Methods: Treatment plans of 38 patients treated for locally

advanced OPC were obtained. Constrictor and saliva-sparing clinical

VMAT plans to a standard dose of 70 Gy (P70) were created. Using

identical cost functions and planning algorithms, treatment plans were

generated that (1) eliminated the PTVs and treated clinical target volumes

only to standard doses (C70) and (2) maintained the PTV but dose de-

escalated to 60 Gy (P60). NTCP mean dose models for the pharyngeal

constrictors (PC), glottis/supraglottic larynx (GSL), parotid glands (PG),

and submandibular glands (SMG) were analyzed. To account for uncer-

tainty in NTCP models, mean change in NTCP >5% was considered the

minimal clinically important difference. Percent volume PTV overlap

(vPTV) was generated to identify geometric parameters for organs at risk

that predict for NTCP improvement. Overlap percentages with high-dose

PTVs (66-70 Gy, vPTVhigh) and low-dose PTVs (56-59 Gy, vPTVlow) were

evaluated separately. Paired t tests and analyses of variance (ANOVAs)

were performed to compare treatments.

Results: Compared to the P70 plans, the dose-reduced P60 plans had a

small decrease in NTCP in all tissues except the ipsilateral SMG (P<.01);

however, only the ipsilateral PG (iPG) exceeded a 5% difference (23.9%

vs 16.2%). Compared to the P70 plans, the PTV-reduced C70 plans also

had a small decrease in NTCP (P<.01), and similarly, only the iPG (23.9%

vs 17.5%) and contralateral SMG (cSMG) (NTCP 32.1% vs 22.9%)

exceeded a 5% difference. For the iPG vPTVlow, only patients in the third

and fourth quartiles (13%-18%, >18%) demonstrated a significant

improvement in the NTCP greater than 5% for the P60 (mean D from

P70Z7.9 and 15.0% for the third and fourth quartiles, respectively,

ANOVA P<.001) and the C70 plans (mean D from P70Z7.7 and 11.5%

for the third and fourth quartiles, respectively, ANOVA PZ.002). For the

cSMG, only patients in the top vPTVlow quartile (>22%) showed a sig-

nificant improvement in NTCP with either the dose reduction (mean D

from P70Z14.5%, P<.01) or PTV elimination (mean D from

P70Z21.1%, P<.01). Additionally, analysis of both SMGs revealed a

vPTVlow upper limit of 35%, beyond which no benefit of using either

strategy was observed.

Conclusion: PTV elimination and dose-reduction strategies resulted in

lowering the NTCP of the iPG, and only PTVelimination was successful in

lowering the NTCP of the cSMG. vPTVlow analysis of the iPG and cSMG

revealed a benefit for patients with >13% or 22% overlap, respectively.

Clinical validation of these strategies is required.

Volume 94 � Number 4 � 2016 Posters 887

Author Disclosure: S. Samuels: None. A. Eisbruch: None. K.A. Vine-

berg: None. C. Lee: None. M.M. Matuszak: None. R.K. Ten Haken:

None. K.K. Brock: None.

139WITHDRAWN

140Selectively Sparing the Submandibular Gland When Level IB LymphNodes Are Included in the Radiation Target Volume: A Safety andToxicity Analysis in Cancers of the Oropharynx and Oral CavityJ.F. Greskovich, Jr,1 N.P. Joshi,2 A. Juloori,3 M.C. Ward,2 H. Qu,2

E. Murray,2 J. Potter,2 A. Dorfmeyer,2 P. Xia,2 and S. Koyfman2;1Cleveland Clinic Florida, Weston, FL, 2Cleveland Clinic Foundation,

Cleveland, OH, 3Cleveland Clinic, cleveland, OH

Purpose/Objective(s): Submandibular gland metastases are extremely

rare in head and neck cancer, even in the presence of level Ib lymph node

(LN) involvement. In recent years, we have contoured the submandibular

gland (SMG) and selectively attempted to limit its dose exposure even in

patients in whom the level Ib LN station is targeted. This study reports our

preliminary feasibility and safety experience with selective submandibular

gland sparing and its dosimetric impact.

Materials/Methods: We identified 174 patients with squamous cell cancer

(SCC) of the oral cavity or oropharynx, with T1-2, N0-3, M0 disease in

whom at least a single-level Ib lymph node region was included in the

target volume. All patients were treated from 2009 to 2014 with definitive

or postoperative intensity modulated radiation therapy with or without

chemotherapy. Patients with recurrent disease, or who were treated with

reirradiation or a split course technique were excluded. Patient, tumor, and

treatment-related factors were abstracted from the medical record. The

treatment plans for each patient were reviewed and verified for level Ib

targeting (unilateral vs bilateral) as well as if the submandibular gland was

excluded from the target volume and sparing was attempted during plan-

ning. Mean doses were calculated for each submandibular gland and the

oral cavity.

Results: A total of 174 patients met criteria for inclusion. Patients had a

median age of 59 years and median KPS of 90 at diagnosis. One hundred

and forty seven patients had SCC of the oropharynx, and 27 patients had

SCC of the oral cavity. One hundred and thirty-four patients were treated

definitively while 40 were treated postoperatively. Of the 174 included

patients, 142 were treated with concurrent chemotherapy. Among the

190 level Ib LN stations that were deliberately targeted in the clinical

treatment volume, 32 submandibular glands were contoured, excluded

from the target volume and avoided during treatment planning. Mean

doses to the spared submandibular glands were able to be reduced by

12% (66.9 Gy vs 58.9 Gy). In a subgroup analysis of 26 patients who had

bilateral level Ib LN targeted, we compared the dosimetric outcomes of 4

patients in whom bilateral submandibular glands were avoided with

those of 22 patients in whom bilateral submandibular glands were

included in the target volumes. In addition to a 12% reduction in mean

submandibular dose (66.3 Gy vs 58.2 Gy), mean oral cavity dose was

reduced by 14.5% (43.4 Gy vs 37 Gy). None of these patients experi-

enced any level Ib LN failures.

Conclusion: Selective sparing of the submandibular gland when targeting

the level Ib nodes in oral cavity and oropharynx cancer is feasible, reduces

the mean doses to submandibular glands and oral cavity, and does not

result in increased level Ib nodal failure rates. Future studies with larger

numbers are needed to validate this preliminary finding and examine the

impact of this technique on functional outcomes.

Author Disclosure: J.F. Greskovich: None. N.P. Joshi: None. A. Juloori:

None. M.C. Ward: None. H. Qu: None. E. Murray: None. J. Potter:

None. A. Dorfmeyer: None. P. Xia: None. S. Koyfman: None.

141Assessment of Morbidity and Arm Function Following RadialForearm Free Flap in Head and Neck Cancer PatientsY. Patil1 and A. Mincara2; 1University of Cincinnati Medical Center/

University of Cincinnati College of Medicine Cincinnati, OH, 2University

of Cincinnati, Cincinnati, OH

Purpose/Objective(s): The objective of this study is to evaluate the

presence of postoperative arm (donor site) dysfunction by assessing

disability following surgical reconstruction using radial forearm free flap

(RFFF) in head and neck cancer patients. Three clinically validated

questionnaires including a general, disease-specific, and site-specific sur-

vey were administered: Short Form 36 Health Survey (SF-36), Short

Musculoskeletal Function Assessment Questionnaire (SMFA), and Dis-

abilities of the Arm, Shoulder and Hand (DASH) Questionnaire, respec-

tively. There are currently no retrospective or prospective studies assessing

the presence of postoperative disability in RFFF patients. No studies are

available on the correlation of the general, disease-specific, and site-spe-

cific disability surveys. This study will help to guide the head and neck

cancer surgeon in microvascular free flap selection as well as guide

postoperative therapy.

Materials/Methods: Patients who met study enrollment criteria (including

no previous neck surgery, arm surgery, shoulder surgery, neuromotor dis-

orders, congenital anatomical abnormalities, or systemic disorders)

completed the questionnaires, which measured various aspects of function

including function index, bothersome index, physical component score,

mental component score, and other important quality of life outcomes.

Postoperative patient scores for each survey were compared to normative

data in the general U.S. population.

Results: The cohort consisted of 23 individuals with an average age of

65�9.5 years; 56.5% were male and 43.5% were female. One-sample z

tests were performed to compare the means of the cohort to the general

population. The cohort observed a significantly higher mean score,

compared to the general U.S. population, for the DASH (13.24 vs 10.1,

PZ.04) and Bothersome Index of the SMFA (21.47 vs 13.77, PZ.05),

indicating that the cohort observed a greater disability compared to the

general population. In addition, the cohort observed a significantly lower

mean score, compared to the general U.S. population, for the PCS of the

SF-36 (39.38 vs 50, P<.01), indicating greater disability.

Conclusion: Since the validated SF-36 questionnaire detected a significant

difference between patients and age- and sex-controlled population norms,

RFFF does cause disability of the donor site in head and neck cancer

patients. SF-36, DASH, and SMFA all showed physical disability

following a RFFF. Shorter and more concise questionnaires, such as the

SF-36 physical component score subset, are equally valid in evaluating

donor site morbidity. This can serve as a concise and efficient clinical tool

in evaluating the present of donor site disability and the need for post-

operative physical therapy.

Author Disclosure: Y. Patil: None. A. Mincara: None.

142Initial Clinical Outcomes From a Prospective Phase 1 Trial ofHypofractionated Stereotactic Body Radiation Therapy for Early-Stage Glottic Larynx CancerD.L. Schwartz,1 S.G. Chun,2 C. Ding,1 A. Sosa,1 L.A. Nedzi,1 J.S. Yordy,3

S.A. Chen,4 R.D. Timmerman,1 and B. Sumer1; 1The University of Texas

Southwestern School of Medicine, Dallas, TX, 2The University of Texas

MD Anderson Cancer Center, Houston, TX, 3Valley Radiation Therapy

Group, Anchorage, AK, 4Pacific Radiation Oncology, Honolulu, HI

Purpose/Objective(s): The primary objective of this trial is to confirm

safety and feasibility of hypofractionated stereotactic body radiation

therapy (SBRT) for early-stage glottic laryngeal cancer and to determine

the most rapid fractionation scheme tolerated without dose-limiting

toxicity (DLT).

Materials/Methods: Seventeen consecutive patients with a diagnosis of

carcinoma in situ, or cT1a-T2N0M0 carcinoma of the glottic larynx

International Journal of Radiation Oncology � Biology � Physics888

cancer were enrolled in this institutional review boardeapproved single-

institution prospective phase 1 clinical trial. Patients required biopsy-

proven squamous cell carcinoma histology or squamous cell variants

(sarcomatoid, verrucous, basaloid, and papillary subtypes) involving the

true vocal cord. Absence of DLT permitted enrollment in incrementally

shorter bioequivalent fractionation schedules starting with 50 Gy in 15

fractions (fx), followed by 45 Gy in 10 fx and 42.5 Gy in 5 fx.

Maximum combined clinical target volume (CTV) and planning target

volume (PTV) expansion was limited to 5 mm beyond involved laryn-

geal structures. Patients were treated daily in the 50 Gy in 15 fx arm,

and every other day, up to 3 times per week in subsequent arms, with

robotic radiosurgery.

Results: Median follow-up interval was 8 months (range: 1.9-18.6

months), with 5 patients followed for at least 1 year. The study cohort

enrolled 16 males and 1 female with a median age of 61 years (range:

46-93 years). There were 11 former smokers, 2 never smokers, and 4

who continued smoking during and following treatment. Stage distri-

bution was 12 cT1 (71%) and 5 cT2 (29%). Mean gross tumor volume,

CTV, and PTV were 1.94�1.88, 3.93�2.23, and 8.63�5.69 cm3,

respectively. The first 4 patients received 50 Gy in 15 fx, and the next 13

patients received 45 Gy in 10 fx, all without DLT. At median follow-up,

local control was 92%; however, 2 patients with cT2 subglottic disease

developed recurrence in the high-dose region requiring salvage laryn-

gectomy. Two additional patients with cT1 disease required laser

stripping for persistent in situ or invasive disease, again in the high-dose

region. No marginal misses have occurred. Maximum treatment-asso-

ciated toxicity was grade 3 dysphagia in 1 patient treated with 45 Gy in

10 fx who continued to smoke >1 pack per day after treatment and

required protective tracheostomy for grade 4 laryngeal edema. Overall

survival is 100%.

Conclusion: Hypofractionated SBRT for early-stage glottic laryngeal

cancer appears safe and feasible at our first 2 fractionation levels. Disease

control appears comparable to standard treatment, although as expected,

unfavorable cT2 tumors with subglottic extension remain at higher risk for

local failure at these fraction sizes. Patients who continue to smoke heavily

may be at higher risk for SBRT-associated toxicity. We are currently

enrolling patients at a fractionation schedule of 42.5 Gy in 5 fx, with

analysis ongoing.

Author Disclosure: D.L. Schwartz: None. S.G. Chun: None. C. Ding:

None. A. Sosa: None. L.A. Nedzi: None. J.S. Yordy: None. S.A. Chen:

None. R.D. Timmerman: None. B. Sumer: None.

143Survival Is Independent of Radiation Therapy Treatment CenterExperience After Surgical TherapyD. Martin,1 B. Kumar,2 N. Brown,3 A. Agrawal,2 E. Ozer,2 S. Kang,2

J.W. Rocco,2 R. Carrau,2 D. Schuller,2 J.C. Grecula,4 A.D. Bhatt,4

D. Blakaj,4 T. Teknos,2 V.M. Diavolitsis,4 and M. Old2; 1The Ohio State

University College of Medicine - Department of Otolaryngology,

Columbus, OH, 2The James Cancer Hospital and Solove Research

Institute, Wexner Medical Center at The Ohio State University, Department

of Otolaryngology-Head and Neck Surgery, Columbus, OH, 3The Ohio

State University College of Medicine Department of Biomedical

Informatics Center for Biostatistics, Columbus, OH, 4The James Cancer

Hospital and Solove Research Institute, Wexner Medical Center at The

Ohio State University, Department of Radiation Oncology, Columbus, OH

Purpose/Objective(s): Current National Comprehensive Cancer

Network guidelines recommend that optimal treatment for patients with

head and neck cancer (HNC) requires the full range of support services

and specialists with expertise. Recent data further supported this

recommendation by way of a study comparing overall survival in stage

III/IV HNC patients receiving primary radiation or chemoradiation at

historically low-accrual centers (HLAC) versus historically high-accrual

centers (HHAC). Patients at HHAC had greater overall survival. This

study examined the impact of radiation therapy center treatment expe-

rience on overall survival in a shared surgical cohort of patients with

advanced oral cavity and oropharyngeal squamous cell carcinoma

(OCSCC/OPSCC).

Materials/Methods: A retrospective longitudinal study was conducted of

333 patients with primary OCSCC (120) or OPSCC (213) treated surgi-

cally at a National Cancer Instituteedesignated cancer center from 2000 to

2012 who received adjuvant radiation therapy. Radiation therapy location

was either at our institution (HVC) or a local radiation center (OSH).

Survival curves were plotted using the Kaplan-Meier method. Cox pro-

portional hazards models were used to assess univariate associations be-

tween potential predictors for death. Unadjusted hazard ratios (HR) and

95% confidence intervals (CI) are reported. Comparisons between loca-

tions were assessed via 2-sample t tests or Mann-Whitney tests, depending

on distribution of the data for quantitative outcomes or X2 or Fisher exact

tests for categorical outcomes.

Results: One hundred thirty-nine patients received adjuvant radiation at

our institution, whereas 194 received it locally. Median follow-up times

were 4.5 years (HVC) and 4.2 years (OSH). The demographic and clinical

characteristics were similar except for significant differences in age (HVC,

56.5 years; OSH, 59.9 years) and T stage (T3/T4: HVC, 25%; OSH, 43%).

There was no significant difference between N classification, American

Joint Committee on Cancer stage, or smoking status. Significant predictors

of overall survival in univariate model were age, extracapsular spread,

marital status, T classification, smoking status, and pack years. After

adjusting for these variables, location of adjuvant therapy was not a sig-

nificant predictor of survival (HR 0.735, PZ.0728).

Conclusion: While recent landmark data demonstrated improved overall

survival in HNC patients treated with primary radiation or chemoradiation

at high-volume centers, this study did not show a statistically significant

difference in survival in surgically treated patients who received post-

operative therapy at a high-volume center versus their local center.

Although this study is limited due to its retrospective nature, it highlights

the important role that our community radiation oncologists have in the

treatment of HNC patients.

Author Disclosure: D. Martin: None. B. Kumar: None. N. Brown: None.

A. Agrawal: None. E. Ozer: None. S. Kang: None. J.W. Rocco: None. R.

Carrau: None. D. Schuller: None. J.C. Grecula: None. A.D. Bhatt:

None. D. Blakaj: None. T. Teknos: None. V.M. Diavolitsis: None. M.

Old: None.

144Impact of Post-Chemoradiation Therapy Selective Neck Dissectionon Patient-Reported Quality of LifeK. Wang,1 R.J. Amdur,2 W.M. Mendenhall,3 R. Green,4 S. Aumer,4

T. Hackman,4 A. Zanation,4 J.P. Zevallos,4 S. Patel,4 M. Weissler,4

and B.S. Chera4; 1UNC Lineberger Comprehensive Cancer Center,

University of North Carolina School of Medicine, Chapel Hill, NC,2University of Florida Hospitals, Gainesville, FL, 3University of Florida

Health Proton Therapy Institute, Jacksonville, FL, 4University of North

Carolina Hospitals, Chapel Hill, NC

Purpose/Objective(s): Neck dissections are often performed after

definitive chemoradiation therapy (CRT) and may add morbidity. There

is a lack of patient-reported quality of life (QOL) data for post-CRT

neck dissections. We herein report prospectively collected QOL out-

comes in patients with human papillomavirus (HPV)-associated

oropharyngeal squamous cell carcinoma enrolled on a multi-institu-

tional phase 2 clinical trial who underwent CRT followed by selective

neck dissection (ND).

Materials/Methods: Inclusion criteria were (1) T0-3, N0-2c, M0; (2)

HPV+ or p16+; and (3) minimal smoking history. Patients received 60-

Gy intensity modulated radiation therapy (IMRT) with concurrent

weekly cisplatin, followed by biopsy of the primary site and selective

ND of (at least) pretreatment-positive lymph node levels. Patients

receiving more than a biopsy at the primary site were excluded from

this analysis. Patient-reported QOL outcomes were obtained pre-CRT,

6 weeks post-CRT (pre-ND), and at acute and late post-ND timepoints.

Abstract 145; Table 1 Percentage of patients with extracapsular extension bysize and number of involved nodes. Raw numbers in parentheses.

No. of involvednodes

Lymph node size

�2.0 cm 2.1-3.0 cm 3.1-4.0 cm >4.0 cm

1 20.0 (1/5) 9.1 (1/11) 18.8 (3/16) 0 (0/10)�2 40.0 (2/5) 54.5 (6/11) 36.4 (4/11) 66.7 (6/9)

Volume 94 � Number 4 � 2016 Posters 889

We used the following questionnaires: European Organization for

Research and Treatment of Cancer (EORTC) QLQ-C30 (general

QOL), EORTC H&N-35 (head and neckespecific QOL), EAT-10

(swallowing), and NDII (neck dissection impairment index). We

compared post-CRT and post-ND composite and itemized scores using

2-tailed t tests.

Results: Thirty-eight enrolled patients had a neck dissection, performed at

a median of 9 weeks post-CRT. Pretreatment N stage was 10 N1/N2a, 23

N2b, and 5 N2c. Median number of levels dissected was 2 (range, 1-6) and

median number of nodes dissected was 12 (range, 1-46). Five patients had

bilateral ND. Median time from ND to postoperative assessment was 2

months (range, 1-7 months) for acute and 18 months (range, 8-37 months)

for late timepoints. There was a significant worsening of the neck

dissection impairment index (NDII) score from post-CRT to acute (87.4 vs

80.2, PZ.023) but not late (PZ.558) post-ND timepoints. The acute

decrease in the NDII was greater in patients with >12 nodes dissected

(PZ.007) and was overall correlated with the total number of nodes

dissected (Spearman PZ.027). Overall EORTC-QLQ-C30, H&N-35, and

EAT-10 scores did not worsen post-ND.

Conclusion: The use of post-CRT selective neck dissection did not worsen

general (EORTC QLQ-C30), head and neckespecific (EORTC QLQ-H&N

35), or swallowing (EAT-10) QOL scores. An instrument designed to

assess the specific impact of neck dissection (NDII) on QOL did show a

significant decrease in acute, but not late NDII QOL score after neck

dissection with a greater acute decrement associated with a higher number

of dissected nodes.

Author Disclosure: K. Wang: None. R.J. Amdur: None. W.M. Men-

denhall: None. R. Green: None. S. Aumer: None. T. Hackman: None. A.

Zanation: None. J.P. Zevallos: None. S. Patel: None. M. Weissler: None.

B.S. Chera: None.

145Identifying Predictors of Extracapsular Extension (ECE) in PatientsConsidering Primary Surgery for Human PapillomaviruseAssociatedOropharyngeal Squamous Cell Carcinoma (OPSCC)I.H. Lee,1 J. Darbinian,2 D. Gurushanthaiah,3 and K.H. Wang3; 1Kaiser

Permanente Northern California, Santa Rosa, CA, 2Kaiser Permanente

Northern California - Division of Research, Oakland, CA, 3Kaiser

Permanente Northern California, Oakland, CA

Purpose/Objective(s): Patients with OPSCC often have a choice between

primary resection (with risk-adapted adjuvant therapy) versus primary

chemoradiation (CRT). Human papillomavirus (HPV)-associated OPSCC

tends to present with more advanced nodal disease with relatively small

primaries, so presence or absence of ECE is often a key determinant of

whether adjuvant chemotherapy is recommended in addition to radiation.

Our goal was to determine whether nodal size is a significant risk factor for

ECE in this population.

Materials/Methods: This is a retrospective analysis of patients with

node-positive OPSCC that underwent neck dissection in an integrated

health care system from January 2009 to June 2014. Cases were iden-

tified from the institutional tumor registry, which is part of the Sur-

veillance, Epidemiology, and End Results program. Data were collected

from both institutional electronic databases and medical chart review.

Patients were excluded if they had radiation or excisional biopsy prior to

neck dissection. Pathology reports were reviewed for HPV status (based

on IHC for p16 or HPV), presence or absence of ECE, size of largest

involved node, and number of involved nodes. Bivariate analyses were

conducted to compare clinical characteristics by whether or not ECE

was present using X2 or Fisher exact tests for categorical variables and t

tests, analyses of variance, and nonparametric tests for continuous

variables.

Results: We identified 95 patients with node-positive OPSCC who un-

derwent neck dissection prior to any radiation or chemotherapy. Of these,

11 were excluded due to missing data on HPV status. Of the remaining 84,

78 (93%) had HPV-associated disease, and among these, 23 (29.5%) were

found to have ECE on final pathology. There was no significant association

between nodal size and probability of ECE. In particular, there were 10

patients with a single involved node >4 cm in size, and none was found to

have ECE. However, patients with multiple involved nodes appeared more

likely to have ECE than those with a single involved node (50 vs 12%,

P<.001).

Conclusion: Among patients with HPV-associated OPSCC, large nodal

size does not appear to be associated with an increased risk of ECE. On

the other hand, having multiple positive nodes is associated with an

increased risk of ECE. Patients with a single large node (ie, N2a dis-

ease) that opt for primary surgery may have a relatively low risk of ECE

and thus may have a good chance of avoiding adjuvant concurrent

chemotherapy.

Author Disclosure: I.H. Lee: Partner; The Permanente Medical Group.

Research Grant; Kaiser Permanente. J. Darbinian: None. D. Guru-

shanthaiah: Partner; The Permanente Medical Group. Running depart-

ment; The Permanente Medical Group. K.H. Wang: None.

146WITHDRAWN

147Evaluation of a Knowledge-Based Planning Model for Head andNeck Cancer Patients Treated in the Setting of a Clinical TrialJ.A. James, D. Melancon, R. Carter, B. Wang, and N.E. Dunlap; University

of Louisville, Louisville, KY

Purpose/Objective(s): Automated knowledge-based treatment planning

techniques are designed to minimize plan variability, improve quality, and

increase planning efficiency. Knowledge-based planning can be particu-

larly useful for head and neck cancer due to the difficulty in optimizing

multiple target volumes and organs at risk. In the setting of a clinical trial,

strict compliance criteria are used which is often time consuming and

difficult to achieve. We sought to investigate the use of automated planning

in the setting of a head and neck clinical trial in order to improve planning

efficiency while maintaining compliance with the recommended dosi-

metric criteria.

Materials/Methods: Volumetric modulated arc therapy (VMAT) and he-

lical tomotherapy plans from 54 previously treated bilateral head and neck

patients (36 definitive, 18 postoperative) were used to train a knowledge-

based treatment planning model. The model was used to replan 12 VMAT

patients enrolled in head and neck clinical trials. The 12 patients include 6

patients treated on the Radiation Therapy Oncology Group (RTOG) 0920

protocol and 6 patients treated on the RTOG 3501 foundation study. Ra-

diation therapy quality management software was used to analyze both the

automated plans and the original plans that were created using manual

optimization. The plans were compared using relevant dosimetric data and

optimization times.

Results: The model-based plans provided an average decrease in mean

dose for the lips, oral cavity, parotids, pharynx, and esophagus of 1.10 Gy

(PZ.159), 3.25 Gy (PZ.007), 3.33 Gy (PZ.096), 8.07 Gy (PZ.048), and

6.65 Gy (PZ.0002), respectively. The maximum dose to 0.03 mL of the

PRV spinal cord increased from an average of 41.93 Gy using manual

optimization to 43.81 Gy using the model (PZ.678). Dose uniformity to

the high-risk planning target volume (PTV) decreased significantly for the

model based plans. The average decrease in minimum dose and average

International Journal of Radiation Oncology � Biology � Physics890

increase in maximum dose to 1 mL of the high-risk PTV was 3.47 Gy

(PZ.0008) and 2.66 Gy (PZ.034), respectively. Nine of the 12 model-

based plans failed compliance criteria for minimum dose due to stringent

dose coverage criteria specified by clinical trial protocols; however, all

plans were deemed clinically acceptable. Optimization times were reduced

from 75 minutes for manual optimization to 10 minutes for automated

optimization.

Conclusion: Complex head and neck plans created using a knowledge-

based treatment planning model provided overall increased sparing of

organs at risk with clinically acceptable dose coverage and uniformity. The

automated optimization significantly increased planning efficiency over

the more traditional manual optimization process. However, a generalized

head and neck model may be insufficient to meet minimum dose criteria

specified by many clinical trial protocols and may require manual inter-

vention to meet compliance.

Author Disclosure: J.A. James: None. D. Melancon: None. R. Carter:

None. B. Wang: organize physics activities; RSS. Organize meeting;

AAPM. N.E. Dunlap: None.

148Nodal Yield Threshold for Early-Stage Clinically Node-Negative OralCavity CancerJ.L. Shah,1 J.J. Chen,1 M. Kaplan,2 R. von Eyben,3 Q.T. Le,1

and W. Hara1; 1Stanford Radiation Oncology, Stanford, CA, 2Stanford

Otolaryngology, Stanford, CA, 3Stanford University, Stanford, CA

Purpose/Objective(s): A recently published randomized study

demonstrated the superiority of an elective neck dissection over a

therapeutic (salvage) neck dissection for patients with early-stage oral

cavity cancer. Further data have suggested that resection of at least 18

lymph nodes improves outcomes for patients with head and neck

cancer. We hypothesized that for early-stage oral cavity cancer, a

lower threshold of resected nodes would be adequate to confer survival

benefit.

Materials/Methods: We retrospectively reviewed 551 consecutive pa-

tients treated at our institution from 1998 to 2013 for oral cavity

squamous cell carcinoma with surgical resection followed by adjuvant

therapy if indicated. Patients treated for recurrent disease and metastatic

disease were not included. For this study, we included patients with T1

or T2 tumors who were clinically node-negative and underwent a neck

dissection as part of their primary surgical treatment. We used iterative

Kaplan-Meier modeling to determine the number of resected nodes that

would confer a survival benefit in this cohort. We then used a Cox

regression model to control for pathologic and demographic factors

known to affect prognosis.

Results: A total of 81 patients met the inclusion criteria for this cohort

study with the following breakdown by subsite: 68% oral tongue, 19%

floor of mouth, 6% buccal mucosa, 5% alveolar ridge, 1% retromolar

trigone, 1% lip, and 0% hard palate. Iterative modeling revealed a

survival benefit in patients who had at least 10 lymph nodes resected

(PZ.006). Only 13 patients had fewer than 10 lymph nodes resected.

Sixty-one percent of these patients were pathologically node-negative

compared to 71% in those with more than 10 lymph nodes resected

(PZNS). Locoregional failure at 2 years was 26% versus 40% in pa-

tients with 10 or more versus fewer than 10 lymph nodes resected,

respectively (PZNS). When stratified by perineural invasion (PNI),

patients with the best overall survival were those without PNI who had

at least 10 nodes resected (2-year overall survival [OS] 91%). Patients

with PNI who had at least 10 nodes resected (2-year OS 79%) did no

better than those with fewer than 10 nodes resected regardless of PNI

status (PZ.01). After controlling for PNI, depth of invasion, poor dif-

ferentiation, lymphovascular invasion, and active smoking status at

diagnosis, the effect of fewer than 10 lymph nodes resected was still

significant (hazard ratio 2.9, PZ.05).

Conclusion: Even for early-stage clinically node-negative oral cavity

cancer, there may be a nodal yield threshold that defines an adequate neck

dissection. Our data suggest that this threshold is 10 lymph nodes.

Author Disclosure: J.L. Shah: None. J.J. Chen: None. M. Kaplan: None.

R. von Eyben: None. Q. Le: None. W. Hara: None.

149The Role of Brachytherapy in Treatment of Oral Tongue CancerJ. Chadha,1 K.S. Hu,2 A. Jacobson,3 M. Persky,4 S. Schantz,5 T. Tran,4

M. Urken,6 Z. Li,2 B. Culliney,7 and L.B. Harrison8; 1Mount Sinai St.

Luke’s & Roosevelt, New York, NY, 2NYU Langone Medical Center, New

York, NY, 3NYU Langone Cancer Center, New York, NY, 4NYU Langone,

New York City, NY, 5New York Eye and Ear, New York, NY, 6Mount Sinai

and Beth Israel, New York City, NY, 7Mount Sinai Beth Israel Medical

Center, New York, NY, 8H. Lee Moffitt Cancer Center and Research

Institute, Tampa, FL

Purpose/Objective(s): Brachytherapy (BT) is a useful modality in both

the definitive and adjuvant treatment of oral tongue cancer either as stand-

alone treatment or in combination with external beam radiation therapy

(EBRT), surgery, and/or chemotherapy whether in the primary or recurrent

setting. It can be used as a boost treatment to deintensify the dose of EBRT,

as a definitive treatment option alternative to surgical management, and as

a stand-alone adjuvant therapy in patients with positive surgical margins.

We review our 10-year experience.

Materials/Methods: Between January 2004 and December 2014, 39 pa-

tients (pts) with oral tongue cancer received Ir-192 LDR BT as part of their

treatment course. Pt characteristics were as follows: median age of 53

years (range 20-88); 27 males, 12 females; T stage: 20 T1, 15 T2, 1 T3, 2

T4, 1 multifocal; N stage: 26 N0, 4 N1, 4 N2, 5 Nx; 31 newly diagnosed

(18 T1, 10 T2, 1 T3, and 2 T4); and 8 recurrent (3 T1, 4 T2, and 1

multifocal). The combination of EBRT and BT was prescribed for 20 pts,

and 19 pts received BT alone. Eight pts were treated definitively with

EBRT+BT (nZ6; T2N0, multifocal primary, T2N2b, T3N0, T4N1,

T4N2c) or with BT and neck dissection (nZ2, both T2N0). Thirty-one

were treated adjuvantly with EBRT+BT (nZ14) or with BT alone (nZ17).

Indications for EBRT were elective nodal RT (nZ3), LN+ (nZ3), PNI

(nZ7), and recurrence (nZ1). In the pts receiving adjuvant BT only, in-

dications for BT were close/positive margin (nZ11), PNI (nZ2), or both

close/positive margin+PNI (nZ4). Four of 17 adjuvant BT only pts did not

have elective node dissection (LND), while 13 did have LND. The median

EBRT dose was 54 Gy (30.6-70 Gy), and 1 pt received protons. The

median BT dose in the EBRT+BT group was 20 Gy (10-27 Gy). The

median BT dose in the BT alone group was 45 Gy (30-60 Gy). The median

number of catheters used was 4 (2-12). Tracheostomy was performed in

the majority of pts, and 8 pts also had neck dissection at the time of BT

catheter placement. Nine pts in the EBRT+BT group received systemic

therapy.

Results: For the entire cohort, the median follow-up is 40.5 months (7-118

months). The 3-year local control (LC), regional control (RC), and overall

survival (OS) is 88%, 75.4%, and 83%, respectively. Among the defini-

tively treated pts, 3-year LC, RC, and OS is 100%, 85.7%, and 60%,

respectively. Among the 14 treated with adjuvant EBRT+BT, the 3-year

LC, RC, and OS is 91.7%, 78.8%, and 100%, respectively. Among the 17

pts receiving adjuvant BT alone, 3-year LC, RC, and OS is 82%, 67.3%,

and 84.6%, respectively. Among this subset, LND impacted on regional

control (3-year RC 84% vs 25%, PZ.017)

Conclusion: Oral tongue BT offers the potential for highly individualized

treatment in multiple clinical scenarios. It offers highly conformal therapy

as definitive or adjuvant therapy either as a boost or stand-alone therapy.

High local control rates can be achieved in pts with positive margins with

BT alone. LND is important to maintain high levels of regional control

particularly in those receiving BT alone.

Author Disclosure: J. Chadha: None. K.S. Hu: None. A. Jacobson: None.

M. Persky: None. S. Schantz: None. T. Tran: None. M. Urken: None. Z.

Li: None. B. Culliney: None. L.B. Harrison: None.

1

Volume 94 � Number 4 � 2016 Posters 89

150No Difference in Outcome Comparing 5-Fluorouracil, Hydroxyurea,and Twice Daily Radiation Therapy (FHX) to High-Dose Cisplatinand Radiation Therapy (CRT) in Advanced T-Stage Head and NeckSquamous Cell Carcinoma (ATSHNSCC): A Retrospective StudyO.R. Adeniran,1 H. Kim,2 M.T. Spiotto,3 L. Feldman,4 P. Hu,2 L. Chen,5

A. Nallari,2 J. Qu,6 B. Haverkos,7 S. Rastogi,2 and A. Chilkulwar2;1UIC Cancer Center, Chicago, IL, 2University of Illinois - Chicago,

Chicago, IL, 3University of Chicago, Chicago, IL, 4University of Illinois at

Chicago, Chicago, IL, 5University of Illinois, Chicago, IL,6Mayo Clinic, Rochester, MN, 7Ohio State University, Columbus, OH

Purpose/Objective(s): In ATSHNSCC, direct comparisons between

definitive treatment regimens are lacking. From 1998 to 2003, the che-

moradiation therapy regimen employed was FHX, while from 2004 to

2009 it was CRT. To investigate whether one of these regimens was

superior, we performed a review of patient outcomes that included

overall survival (OS), disease-free survival (DFS), and long-term toxicity

(LTT).

Materials/Methods: The institutional review board approved this study,

and the institution’s tumor registry database was queried for all patients

(pts) with (T3 or T4) ATSHNSCC. Medical records were examined for

variables that included age, race, gender, and Eastern Cooperative

Oncology Group performance status (PS). Tumor-specific data included

subsite, stage, grade, treatment received, date of last follow-up, recurrence,

and/or death. The following LTT’s were assessed: difficulty speaking (DS),

difficulty eating (DE), feeding tube placement (FT), tracheostomy (TO),

mandibular osteoradionecrosis (MORN), mucocutaneous fistula (MCF),

dry mouth (DM), and soft-tissue fibrosis (STF).

Results: Three hundred and seventeen pts met eligibility and are included

in this analysis. Median age is 57 years. Gender (%): 75 males and 25

females. Race (%): 35 Caucasian (C), 47 African American (AA), 12

Hispanic (H), 3 Asian (A), and 3 other (O). PS (%): 0Z35.6, 1Z54.5,

2Z9.6, and 3Z0.4. A 1o subsite (%): oral cavity (OC) 26.1, oropharynx

(OP) 37.1, larynx 18.2, nasopharynx (NP) 5.4, and hypopharynx (HP)

10.4. Tumor grade (%): high 19.2, moderate 65.1, and low 15.7. T stage

(%): T1Z5.6, T2Z10.4, T3Z34.4, and T4Z50.0. The ratio of pts

receiving FHX to CRTwas 3:1. There was a higher percentage of advanced

PS (2 and 3) in FHX- versus CRT-treated pts. A higher percentage of FHX

treated pts had OC 1o (31 vs 14) and CRT treated pts had a higher per-

centage of OP 1o (34 vs 48) tumors. There was no significant difference in

median survival (MS) between treatment groups (FHX 6.7 vs CRT 6.4

years). Kaplan-Meier survival analysis showed no difference in (%) OS

(43 vs 50) or DFS (74 vs 60) between FHX- and CRT-treated pts. Dif-

ferences were seen in the rate of LTT. DE was more frequent in CRT-

versus FHX-treated pts (83.3 vs 66.4 %). There was a higher frequency of

DM and FT in CRT- versus FHX-treated pts (41.1% vs 33.3% and 66.7%

vs 58.5%, respectively). MORN was more frequent with FHX- (12.9%)

versus CRT-treated pts (4.2%). There was no difference in outcome by race

except DMwas less frequent (%) in AA (24.6) versus C (42.4), H (48.2), or

A/O (50).

Conclusion: In ATSHNSCC, no survival outcome differences were seen

between FHX- and CRT-treated patients. Some differences in LTT were

seen.

Author Disclosure: O.R. Adeniran: None. H. Kim: None. M.T. Spiotto:

None. L. Feldman: None. P. Hu: None. L. Chen: None. A. Nallari: None.

J. Qu: None. B. Haverkos: None. S. Rastogi: None. A. Chilkulwar:

None.

151Radiation Dose Escalation Is Associated With Improved Survival inCervical Esophageal CancerN. Jegadeesh,1 Y. Liu,2 B. Oyelade,1 T. Gillespie,1 R.J. Cassidy, III,1

F. Fernandez,2 N. Saba,1 J.J. Beitler,2 and J.C. Landry2; 1Winship Cancer

Institute of Emory University, Atlanta, GA, 2Emory University, Atlanta, GA

Purpose/Objective(s): The management of cervical esophageal cancer

(CEC) remains controversial. Organ preservation is a treatment goal and

thus chemoradiation therapy (CRT) has become the de facto standard of

care in this site. Given the paucity of cases, randomized evidence does not

exist to guide an optimal radiation dosing paradigm. It is unclear whether

the lower radiation doses used for thoracic esophageal cancer or the higher

doses used for more anatomically similar hypopharyngeal cancer are

preferable. We hypothesize that higher radiation dose is associated with

improved survival in CEC.

Materials/Methods: The National Cancer Data Base was queried for stage

I-III squamous cervical esophageal cancers that completed radiation

therapy (RT) or CRT with curative intent. We dichotomized patients to

dose cohorts of 30 to 54 Gy and >54 to 80 Gy to represent thoracic

esophageal and head and neck dosing schema, respectively. Kaplan-Meier,

log-rank test, and multivariable Cox proportional hazards regression were

performed with overall survival (OS) as the primary outcome variable.

Martingale residual analysis and log-rank statistic were used to identify an

optimal dose.

Results: Five hundred sixty patients treated between 1998 and 2011

were eligible for analysis; 299 (53.4%) received 30 to 54 Gy and 261

(46.6%) received >54 Gy. Of the total patient cohort, 86.2% received

chemotherapy. The high radiation dose cohort was associated with

more advanced stage, more recent year of diagnosis, lower mean tumor

size, and usage of intensity modulated RT (all P<.05). There was no

significant difference in chemotherapy use or Charlson comorbidity

index between groups. Median survival was 16.5 versus 18.7 months in

low- versus high-dose groups (PZ.101). On multivariable analysis,

there was a trend toward worse OS with the lower dose cohort (hazard

ratio [HR] 1.25, 95% confidence interval [CI] 1.00-1.57, PZ.054). In a

separate analysis, increasing radiation dose between 30 and 80 Gy, as a

continuous variable, was associated with improved OS (HR 0.98, 95%

CI 0.97-0.99, PZ.004). A linear relationship for improved OS was

identified on Martingale residual plot from 30 to 80 Gy. A dose of

53.2 Gy was identified as an optimal cut point to dichotomize

survival.

Conclusion: CEC treated with increasing radiation dose was associated

with improved OS in this large population-based analysis. This series

represents the largest study describing radiation dose in this population. In

the absence of prospective evidence, these results may help guide therapy

in this uncommon disease.

Author Disclosure: N. Jegadeesh: None. Y. Liu: None. B. Oyelade: None.

T. Gillespie: None. R.J. Cassidy: None. F. Fernandez: None. N. Saba:

None. J.J. Beitler: None. J.C. Landry: None.

152Dosimetric Assessment of the Contralateral Parotid 6 Months AfterRadiation Therapy for Head and Neck TumorsB.S. Chera,1 P. Mavroidis,2 M. Kostich,3 R.J. Amdur,4 W.M. Mendenhall,5

N.C. Sheets,6 R. Green,1 A. Price,3 S.K. Das,2 and L.B. Marks7;1University of North Carolina Hospitals, Chapel Hill, NC, 2University of

North Carolina, Chapel Hill, NC, 3University of North Carolina at Chapel

Hill, Chapel Hill, NC, 4University of Florida Hospitals, Gainesville, FL,5University of Florida Health Proton Therapy Institute, Jacksonville, FL,6Rex/UNC, Raliegh, NC, 7UNC School of Medicine, Chapel Hill, NC

Purpose/Objective(s): To estimate the correlation between different

dosimetric indices of the contralateral parotid gland to the severity of

patient-reported dry mouth 6 months following deintensified chemo-

radiation therapy (CRT).

Materials/Methods: Forty-four patients were treated on a prospective

multi-institutional phase 2 study assessing the efficacy of deintensified

CRT in patients with favorable-risk, human papillomavirus (HPV)-

associated oropharyngeal squamous cell carcinoma. All patients received

60 Gy intensity modulated radiation therapy (IMRT) with concurrent

weekly intravenous cisplatinum (30 mg/m2). Radiation dose to the

contralateral parotid was standardly reduced with the dosimetric goal

being mean dose <26 Gy. All patients reported symptoms using the novel

Abstract 152; Table 1

Parotid dosimetricindices

None/mild/moderatedry mouth (nZ36)

Severe/very severedry mouth (nZ7)

Dmean 24.1�7.0 Gy 29.0�8.7 GyV30 27.5%�20.2%. 41.7%�21.3%WARPF 34.9%�18.4% 49.0%�18.9%

International Journal of Radiation Oncology � Biology � Physics892

patient-reported ouctome version of the Common Terminology Criteria for

Adverse Events (PRO-CTCAE). Patients reported the severity of the dry

mouth (none, mild, moderate, severe, or very severe). We correlated in-

dividual patient dosimetric data from the contralateral parotid gland to

their 6-month posttreatment dry mouth PRO-CTCAE responses. Severe/

very severe responses were considered clinically signicant and were used

as the cut point for our analyses. We specifically assessed parotid: Dmean,

V30, and the weighted average reduction of the regional parotid function

(WARPF). WARPF is a quantity that is calculated by weighting each dose

of the parotid dose-volume histogram (DVH) using a regional parotid

function reduction curve. Performance of these dosimetric indices was

assessed through the area under to receiver operating characteristic curve

(ROC).

Results: Average parotid Dmean, V30, and WARPF for patients reporting

severe/very severe versus none/mild/moderate dry mouth are shown in

Table 1. The corresponding values of the area under the ROC curve for

Dmean, V30, and WARPF were 63.9%, 70.2%, and 73.8%, respectively.

Conclusion: Patients who report severe/very severe dry mouth 6 months

post deintensified CRT had on average higher contralateral Dmean, V30,

and WARPF values. WARPF performs better as descriptor for patient-

reported dry mouth, and further investigation could verify its suitability as

a dosimetric constraint in treatment planning.

Author Disclosure: B.S. Chera: None. P. Mavroidis: None. M. Kostich:

None. R.J. Amdur: None. W.M. Mendenhall: None. N.C. Sheets: None.

R. Green: None. A. Price: None. S.K. Das: None. L.B. Marks: None.

153Postoperative Radiation Therapy for HumanPapillomaviruseAssociated Oropharyngeal CarcinomaS.S.D. Rao,1 D. Jackson,1 A.M. Chen,2 Q. Luu,1 A. Bewley,1

D.G. Farwell,1 and M.E. Daly1; 1University of California, Davis,

Sacramento, CA, 2University of California, Los Angeles- David Geffen

School of Medicine, Los Angeles, CA

Purpose/Objective(s): Human papillomavirus (HPV)-associated

oropharyngeal squamous cell carcinoma (OPSCC) is associated with

improved oncologic outcomes compared to nonvirally mediated

oropharyngeal cancers, and treatment protocols are under development to

de-escalate therapy while maintaining excellent tumor control. Limited

single-institution studies suggest that traditional pathologic prognostic

features may not adequately risk-stratify HPV-associated OPSCC. We

reviewed our institutional experience using postoperative radiation ther-

apy (RT) for p16-positive OPSCC and assess predictors of recurrence and

survival.

Materials/Methods: The medical records of 53 consecutive patients with

p16-positive OPSCC treated with surgery and postoperative RT between

October 2003 and February 2014 were retrospectively reviewed. Fifty

(94%) underwent resection of the primary tumor and ipsilateral (nZ44)

or bilateral (nZ6) neck dissection while the remaining 3 (6%) under-

went laser resection of the primary alone for clinically node-negative

tumors. Thirty (57%) received postoperative concurrent chemotherapy.

All patients underwent postoperative intensity modulated RT to a me-

dian dose of 66 Gy (range: 38-70 Gy). Clinical and pathologic features

including American Joint Committee on Cancer (AJCC) tumor (T) and

nodal (N) stage, surgery-to-RT interval, number of involved nodes,

nodal ratio, extracapsular extension (ECE), lymphovascular space in-

vasion, (LVSI), and perineural invasion (PNI) were recorded, and their

influence on locoregional control (LRC), distant control (DC), and

overall survival (OS) was assessed using the log-rank method and Cox

regression.

Results: At a median follow-up of 38.1 months, the 3-year estimates of

LRC, DC, and OS among all patients were 95%, 85%, and 89%, respec-

tively. ECE and LVSI were identified in 26 (49%) and 10 (19%) patients,

respectively. The median interval from surgery to RT was 7.1 weeks

(range: 4.6-18.3 weeks). Increasing number of involved nodes predicted

for decreased LRC (PZ.02), DC (PZ.003), and OS (PZ.002). Similarly,

increasing nodal ratio was correlated with worse LRC (PZ.05) and OS

(PZ.02) with a trend toward worse DC (PZ.07). The presence of ECE,

PNI, and LVSI were not predictive of LRC, DC, or OS (P>.05 for all). By

contrast, AJCC stage did not predict for LRC or DC (P>.05). An

increasing interval from surgery to RTwas associated with decreased LRC

(PZ.02).

Conclusion: Our data confirm other studies suggesting that HPV-associ-

ated OPSCC has distinct prognostic factors for recurrence. In our series

nodal disease burden, as measured by number and ratio of involved nodes,

was the primary determinant of LRF, DF, and OS. As with other head and

neck cancers, interval from surgery to RT influences LRC. Other patho-

logic factors historically predictive in head and neck cancer did not

strongly impact local or distant failure.

Author Disclosure: S.S. Rao: None. D. Jackson: None. A.M. Chen: None.

Q. Luu: None. A. Bewley: None. D.G. Farwell: None. M.E. Daly: None.

154Overall Survival After Endoscopic Surgery Versus Radiation as theInitial Treatment in Oropharyngeal Squamous Cell CarcinomaJ.K. Molitoris,1 S.M. Bentzen,2 S.E. Strome,3 D.P. Zandberg,2

K.J. Cullen,2 N. Hanna,4 and M.D. Chuong2; 1University of Maryland

Medical Center, Baltimore, MD, 2University of Maryland School of

Medicine, Baltimore, MD, 3University of Maryland, Baltimore, MD,

United States, 4Department of Surgical Oncology, University of Maryland

Medical Systems, Baltimore, MD

Purpose/Objective(s): Chemoradiation therapy (CRT) and radiation

therapy (RT) offer the potential for organ preservation over surgery for

patients with squamous cell carcinoma of the oropharynx (OPC). However,

CRT can result in significant morbidity, which has led to increasing interest

in minimally invasive endoscopic head and neck surgery (eHNS) tech-

niques such as transoral laser microsurgery and transoral robotic surgery.

While there are increasing data supporting the use of eHNS for HPV+

OPC, which has a favorable prognosis, the use of eHNS is not well

described for human papillomavirus-negative (HPV-) OPC patients, who

have a poorer prognosis. We used the National Cancer Data Base (NCDB)

to evaluate factors associated with the use of eHNS as well as overall

survival (OS) in HPV- OPC patients compared to definitive CRT/RT.

Materials/Methods: This study included T1-3N0-2bM0 HPV- OPC

patients diagnosed between 2010 and 2012 who received either definitive

CRT/RT or eHNS as initial treatment. Binary logistic regression (LR) was

performed to determine factors associated with receiving eHNS. Cox

regression (CoxR), adjusted for patient and tumor factors, was used to

compare OS hazard ratios (HR). Propensity score matching separated

patients into cohorts with low, moderate, and high likelihood of receiving

eHNS for Kaplan-Meier survival comparison.

Results: We identified 1562 HPV- OPC patients with a median age of 59

years. Initial therapy was CRT/RT (84.2%) or eHNS (15.8%). Among

eHNS patients, 9.8% were cT3, 61.8% were cN0-1, 32.8% had positive

margins, and 64.8% received adjuvant therapy (14.8% RT, 50.0% CRT).

Among RT patients, 29.1% were cT3, 45% were cN0-1, 84.0% received

concurrent chemotherapy, and 25.3% had neck dissections after RT.

Factors associated with an increased likelihood of eHNS on LR were

tonsillar primary, treatment at an academic center, and lower clinical stage.

In CoxR, eHNS was not associated with improved OS (HR 0.631; confi-

dence interval [CI] 0.421-1.134; PZ.144). In propensity score matching,

2-year OS was improved for eHNS versus CRT/RT patients in the high

likelihood cohort (94.5%; CI 91.4-97.6 vs 81.0%; CI 78.1-83.9; PZ.011),

Volume 94 � Number 4 � 2016 Posters 893

but not the low likelihood cohort (80.1%; CI 71.9-88.6 vs 75.8% CI

73.0-78.6, PZ.782).

Conclusion:While HPV- OPC patients who were selected to receive eHNS

in the NCDB generally had more favorable clinical tumor stage, nearly

two-thirds also received adjuvant CRT/RT. While recognizing the potential

influence of patient selection that is not captured by covariates available in

the NCDB, we did not find a significant OS difference among patients who

received initial CRT/RT or eHNS. Prospective studies are needed to clarify

whether a subset of HPV- OPC patients derive benefit from eHNS and

whether the quality of life is impacted in eHNS patients by the high

incidence of adjuvant therapy.

Author Disclosure: J.K. Molitoris: None. S.M. Bentzen: None. S.E.

Strome: None. D.P. Zandberg: None. K.J. Cullen: None. N. Hanna:

None. M.D. Chuong: None.

155Predictors of Unplanned Hospital Admissions and/or Feeding TubePlacement in Patients With Oropharyngeal Cancer Treated WithCisplatin-Based ChemoradiationB. Harr,1 J. Bodmann,1 N.P. Joshi,2 M.C. Ward,2 D. Ives,1 M. Rahe,1

C.A. Reddy,1 J. Hamker,1 J.F. Greskovich, Jr,3 T. Nwizu,1 D.J. Adelstein,1

and S. Koyfman2; 1Cleveland Clinic, Cleveland, OH, 2Cleveland Clinic

Foundation, Cleveland, OH, 3Cleveland Clinic Florida, Weston, FL

Purpose/Objective(s): We sought to identify factors that were associated

with increased risks of unplanned emergency room and hospital admis-

sions and/or feeding tube placement (a complicated treatment course) in

patients with oropharyngeal cancer treated with cisplatin-based chemo-

radiation (CRT).

Materials/Methods: We identified patients with oropharynx cancer who

were treated with CRT and intensity modulated radiation therapy (IMRT)

between 2009 and 2014 from our institutional review boardeapproveddatabase. We examined patient-, tumor-, and treatment-related variables

and their association with an adverse event (AE), defined as an unplanned

emergency room or hospital admissions and/or feeding tube placement

during or within 90 days of completing CRT. Univariate and Multivariate

analysis (MVA) was performed using logistic regression analysis to

identify factors associated with AE.

Results: Of the 96 patients included in this study, 90% were men, 69%

were married, 92% were HPV/p16 positive, 53% had >10 pack-years

smoking history, and 63% were working at diagnosis with a mean

Charlson comorbidity index of 0.7. High-risk disease features were

uncommon, as 12% had T4 disease and 22% had N2c/3 disease. Ninety

percent were treated with bilateral IMRT, and the remainder (10%) were

treated unilaterally. In addition, 62.5% were treated with 2 doses of

bolus (100 mg/m2) cisplatin, with the remaining treated with either 3

bolus doses (32.3%) or weekly (40 mg/m2, 52%) cisplatin. Median

follow-up was 22.6 months, and the 2-year recurrence-free survival was

83%. A total of 53 patients (55%) had an AE, of which 25 had an

admission only, 5 had a feeding tube only, and 23 had both. On MVA,

higher N stage (N2c/3) was associated with significantly increased risk

of AE (odds ratio [OR] 5.6; PZ.01), while being unmarried carried a

similar trend (OR 2.5; PZ.06). Patient age, performance status, gender,

T stage, HPV status, and comorbidities were not associated with higher

risks of AE.

Conclusion: Unplanned hospital or ER admissions and feeding tube

placement is common among oropharyngeal cancer patients treated with

cisplatin-based CRT. Patients with higher N stage disease (and larger

high-dose treatment volumes) are at significantly higher risk of AE,

and unmarried patients demonstrate a similar trend. Using these results,

we can better identify at-risk patients, and efforts can be made to

reduce AE.

Author Disclosure: B. Harr: Stock; Johnson & Johnson. J. Bodmann:

None. N.P. Joshi: None. M.C. Ward: None. D. Ives: None. M. Rahe:

None. C.A. Reddy: None. J. Hamker: None. J.F. Greskovich: None. T.

Nwizu: None. D.J. Adelstein: None. S. Koyfman: None.

156Transoral Robotic-Assisted Resection Approach for IdentifyingUnknown Primaries of the Head and NeckN. Khan,2 J. Kass,1 M. Teng,2 B. Miles,2 and E. Genden2;1Boston University, Boston, MA, 2Icahn School of Medicine at Mount

Sinai, New York, NY

Purpose/Objective(s): In traditional squamous cell carcinoma (SCCA),

unknown primaries are rare (2%-3%), and a site can be identified in 67%

of cases. Recent case reports have suggested that transoral robotic-assisted

resection (TORS) of the base of tongue (BOT) can improve upon tradi-

tional methods with identification of a primary tumor in up to 90% of

patients. We reviewed our use of TORS in identifying unknown primaries

as well as the usefulness of preoperative positron emission tomographic

(PET) scans. In addition, we reviewed the postoperative management and

oncologic outcomes.

Materials/Methods: This is an institutional review boardeapproved

retrospective chart review of 280 consecutive cases between 2011 and

2015. Unknown primary was defined as follows: regional squamous cell

carcinoma in the neck, confirmed by biopsy, with a negative clinical exam

(including flexible fiberoptic nasolaryngoscopy) and negative preoperative

computed tomographic (CT) scan. In nearly all cases (nZ20 of 21), a

preoperative PET scan was also performed.

Results: Twenty-one cases were identified. Mean age at presentation was

56 years (range 42-72 years). All but 1 patient was male, and all were

human papillomavirus (HPV) positive. Neck disease was most commonly

unilateral and multimodal: N1 (nZ3), N2a (nZ5), N2b (nZ11), N2c

(nZ1), N3 (nZ1). The largest node at presentation averaged 3.18 cm

(range 1.7-6 cm). A primary tumor was found in 76% of cases (nZ16 of

21). The average size of the primary tumor was 0.78 cm (range 0.1-1.8

cm). Preoperative PET scans were unable to identify a primary tumor in

80% of cases (nZ14 of 20). In the 6 cases in which a primary site was

anticipated, it was 67% accurate (nZ4 of 6). A primary tumor could not be

found in the remaining 2. Operative plans were either unilateral (nZ14) or

bilateral (nZ7) resection of the tonsil and/or BOT. Mean hospital stay was

1.7 days (range 1-3 days). Postoperative treatments varied widely: 4 were

observed (including 2 cases in which the primary was not identified). The

remaining were radiated with either 60 or 66 Gy to the primary site and

ipsilateral or bilateral necks. Decisions on postoperative radiation and

fields were based on a multidisciplinary review of the pathologic report

including nodal disease burden, extracapsular spread, and/or perineural

innervation in the primary tumor. Postoperative CRT with cisplatin was

used in 2 cases. All patients are currently disease free with a mean follow-

up of 12 months (range 0.4-32 months).

Conclusion: TORS-assisted approaches to the unknown primary improve

upon the traditional approach; however, there exists opportunity for

improvement. PET-CT is helpful when definitively positive, but the scans

often added little information. Oncologic outcomes are favorable in

HPV-associated disease despite wide variation on postoperative manage-

ment. In particular, a prospective clinical trial addressing observation

versus XRT would be helpful in determining treatment algorithms in the

subset of patients in which the primary site remains unknown.

Author Disclosure: J. Kass: None. N. Khan: None. M. Teng: None. B.

Miles: None. E. Genden: None.

157Retrospective Analysis of Prophylactic Gabapentin on Pain andWeight Loss in Patients Undergoing Radiation Therapy forOropharyngeal CancerT. Dong,1 A. Reed,2 G.C. Jones,3 D. Scoble,4 J. Deeken,4 and G.K. Bajaj4;1Howard University College of Medicine, Washington, DC, 2Walter Reed

National Military Medical Center, Bethesda, MD, 3Tri-Cities Cancer

Center, Kennewick, WA, 4Inova Comprehensive Cancer and Research

Institute, Falls Church, VA

Purpose/Objective(s): Pain and weight loss are common side effects of

radiation therapy (RT) for head and neck cancer. Narcotic pain medication

International Journal of Radiation Oncology � Biology � Physics894

(NPM) is commonly used but often results in incomplete pain control and

is associated with its own side effects. Single-institution retrospective

reports have provided evidence that prophylactic gabapentin (GP) may

reduce the need for NPM in patients undergoing RT, however the ideal

dosing schedule is unknown. The purpose of this study was to evaluate

whether the use of 300 mg 3 times daily (TID) of GP in a community

oncology center will result in reduced need for NPM or reduced weight

loss (WL) in patients undergoing RT with or without chemotherapy for

oropharyngeal cancer.

Materials/Methods: We performed a retrospective chart review of all

patients treated with RT for oropharyngeal cancer in our clinic over the last

33 months. From these data, we stratified patients by the use of GP and

compared the amount of NPM, the time to initiation of NPM, and the

amount of WL between patients who did and did not receive GP.

Two-tailed t tests were used to determine significance with a for signifi-

cance set at P�.05.Results: From November 2012 through July 2015, 64 patients with

oropharyngeal cancer completed their prescribed RT course and at least 1

month of post-RT follow-up at our institution. Thirty-one patients received

GP (at least 300 mg TID) within the first 2 weeks of RT. Patients who

received GP had less unintentional WL (9.03 vs 15.82 lbs, PZ.004) and

initiated NPM later in their RT course (34.6 vs 22.3 days, P<.001). On

subset analysis, patients who underwent upfront surgery and patients with

p16-positive disease had less unintentional WL with the use of GP.

Oropharyngeal subsite did not influence the effect of GP. No adverse

effects were attributed to GP.

Conclusion: Patients who initiated at least 300 mg TID of GP within the

first 2 weeks of RT for oropharyngeal cancer experienced less WL and a

longer time to initiation of NPM compared to those that did not initiate GP.

These differences were independent of oropharyngeal subsite. No adverse

effects were attributed to GP. Patients who underwent upfront surgery and

patients with p16-positive disease derived the greatest benefit from the use

of GP. These data support continued exploration of GP use in this patient

population. Our institution intends to continue to explore the effect of

higher GP dosing in future patients to elucidate potential dose-response

relationships.

Author Disclosure: T. Dong: None. A. Reed: None. G.C. Jones: None. D.

Scoble: None. J. Deeken: None. G.K. Bajaj: None.

158Squamous Cell Cancer of an Unknown Primary Head and Neck Site:Is Upfront Neck Dissection Still Relevant in the Era ofChemoradiation?M.J. Amsbaugh,1 C.A. Perez,1 J. Gaskins,1 C.L. Silverman,1 J. Bumpous,1

K. Potts,1 A.A. Rajeurs,2 R. Redman,1 and N.E. Dunlap1; 1University of

Louisville, Louisville, KY, 2University of Louisville School of Medicine,

Louisville, KY

Purpose/Objective(s): We sought to identify prognostic factors for

survival and the role of upfront neck dissection in the era of definitive

chemoradiation for patients with squamous cell carcinoma (SCC) of an

unknown primary head and neck site.

Materials/Methods: All patients with SCC of an unknown primary site of

the head or neck were reviewed from a prospective database. Patient de-

mographic, treatment, and toxicity data were collected. Toxicity was coded

according to Common Terminology Criteria for Adverse Events version

4.03. The Kaplan-Meier method was used to estimate survival. The

log-rank test and proportional hazards regression were used to analyze

factors influencing outcomes using a least absolute shrinkage and selection

operator (LASSO) to select relevant variables.

Results: Of 2258 patients seen in our multidisciplinary clinic with a new

diagnosis of head and neck cancer between 2003 and 2013, no primary site

was identified in 66 patients. Fifteen patients were treated with definitive

chemoradiation (CRT), 14 patients were treated with radiation alone, and

37 patients received an upfront neck dissection followed by adjuvant

radiation or CRT. With a median follow-up of 22.4 months for surviving

patients, actuarial emergency of a primary, neck failure, progression-free

survival (PFS), and overall survival (OS) at 5 years were 11.2%, 21.4%,

42.3%, and 65.9%, respectively. On multivariate analysis, OS was

significantly correlated with level II and level IV neck involvement.

Patients with level II involved at presentation had better OS (hazard ratio

[HR] 0.236, 95% confidence interval [CI] 0.084-0.662), and patients with

level IV had worse OS (HR 3.53, 95% CI 1.27-9.83). Patients who were

current smokers (PZ.054) or who had N2b or greater stage disease

(PZ.088) trended toward worse survival. Patients who underwent an

upfront neck dissection followed by appropriate adjuvant therapy had a

lower rate of local failure (PZ.003), locoregional failure (PZ.068), and

emergence of the primary (PZ.001) with no difference in PFS (PZ.189)

or OS (PZ.641) compared to patients who underwent definitive chemo-

radiation. Patients who received definitive chemoradiation had more

advanced neck disease at presentation (P<.001). Grade 3 or higher toxicity

occurred in 25 of 66 patients (37.9%) following radiation therapy. Upfront

neck dissection did not predict for increased toxicity.

Conclusion: Survival for patients with SCC of an unknown primary site of

the head and neck is likely determined by occult primary site and subse-

quent presenting cervical lymph node metastases. Upfront neck dissection

is well tolerated and reduces failures in the head and neck compared to

definitive chemoradiation but does not improve overall survival.

Author Disclosure: M.J. Amsbaugh: Employee; University of Louiville.

C.A. Perez: Employee; University of Louiville. J. Gaskins: None. C.L.

Silverman: None. J. Bumpous: None. K. Potts: None. A.A. Rajeurs:

None. R. Redman: None. N.E. Dunlap: Honoraria; Osler.

159A Comparison of Split-Field and Whole-Field Intensity ModulatedRadiation Therapy and Volumetric Modulated Arc Therapy forLaryngeal Sparing in Oropharynx CancerV. Takiar,1 S. Quinlan-Davidson,2 S. Tung,2 A.S. Garden,2

W.H. Morrison,2 D.I. Rosenthal,2 G.B. Gunn,2 C.D. Fuller,2 B.M. Beadle,2

H. Wang,2 C. Wang,2 and J. Phan2; 1University of Cincinnati, Cincinnati,

OH, 2The University of Texas MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): We assessed the optimal intensity modulated

radiation therapy (IMRT) technique for treatment of nonlaryngeal head

and neck cancers with regard to clinical target volume (CTV) coverage and

sparing of organs at risk (OARs), particularly the larynx. The split-field

approach (SF-IMRT) has been advocated to avoid unnecessary dose to the

nondiseased larynx, whereas concerns for match-line failures have led to

the use of a whole-field approach (WF-IMRT). Volumetric arc-based

IMRT (VMAT) is the “next generation” IMRT offering faster treatment

time and potentially less radiation exposure. Here we compare SF-IMRT,

WF-IMRT, and VMAT in a large cohort of oropharyngeal carcinoma

patients.

Materials/Methods: Seventy oropharynx carcinoma patients definitively

treated with SF-IMRT (40 treated unilaterally and 30 treated to the bilat-

eral neck; T4 or N3 excluded) were replanned with WF-IMRT (7-11

noncoplanar beams) and VMAT (2-arcs) per institutional planning goals.

OARs were delineated per Radiation Therapy Oncology Group 1016. All

contours underwent institutional quality assurance peer review prior to

planning. Plans were compared using conformity (CI) and heterogeneity

(HI) indices of the high-dose CTV and Dmeanof OAR. Differences were

analyzed by paired t tests.

Results: All plans achieved institutional planning objectives and were

clinically acceptable. VMAT resulted in 4 to 5 times faster treatment

delivery, required 15% to 40% fewer monitor units, improved target

coverage (V100), demonstrated superior CI and HI, and had fewer hot

spots (V105/V107 of 1.6%/0.02%) compared to WF-IMRT (2.2%/0.4%) or

SF-IMRT (6.5%/3.3%) (P<.05 for all). For bilateral plans, the larynx dose

was similar for SF-IMRT (25.8 Gy), WF-IMRT (24.4 Gy), and VMAT

(24.7 Gy). SF-IMRT had 10% to 15% higher pharyngeal constrictor dose

(P<.05) but lower arytenoid dose (11.9 Gy) compared to WF-IMRT (15.0

Gy) and VMAT (16.0 Gy) (P<.05). Arytenoid dose correlated with

Volume 94 � Number 4 � 2016 Posters 895

SF-IMRT isocenter location, with each 0.25-cm distance superiorly from

organ (up to 2 cm) corresponding to a 1.3-Gy drop in mean arytenoid dose

(P<.01). For unilateral plans, the larynx dose was lower for VMAT (14.4

Gy) than WF-IMRT (15.1 Gy) and SF-IMRT (16.1 Gy) (P<.05). The

larynx dose was lower by 10.1 Gy for unilateral compared to bilateral

plans. The larynx dose was higher by 4.1 Gy for plans with level 3

adenopathy compared to plans with adenopathy above the hyoid or that

were node-negative (P<.05).

Conclusion: In the treatment of oropharyngeal cancer, <30 Gy mean

laryngeal dose is achievable with SF-IMRT, WF-IMRT, or VMAT tech-

niques. VMAT should be considered in appropriate candidates when

shorter delivery times, fewer monitoring units, and increased dose homo-

geneity are warranted.

Author Disclosure: V. Takiar: None. S. Quinlan-Davidson: None. S.

Tung: None. A.S. Garden: None. W.H. Morrison: None. D.I. Rosenthal:

None. G.B. Gunn: None. C.D. Fuller: None. B.M. Beadle: None. H.

Wang: None. C. Wang: None. J. Phan: None.

160Nurse-Led Multiprofessional Team to Improve Adherence toTreatment for Patients With Head and Neck Cancer ReceivingRadiation Therapy in a Community HospitalC. Johnson,1 K. Ottaviano,1 J.M. Smudde,2 and N. Mohideen2;1Northwest Community Hospital, Arlington Heights, IL,2Northwest Community Hospital, Arlington Heights, IL, United States

Purpose/Objective(s): The incidence of head and neck cancer associ-

ated with human papillomavirus (HPV) has increased in the Unite States.

This community hospital has also seen an increase in this population.

Our baseline data for the first half of 2014 revealed 31% of 13 patients

treated with concurrent chemoradiation experienced treatment interrup-

tion, and 38% had weight loss >5 kg. The purpose of this project was to

identify a patient-specific protocol using a nurse-led multiprofessional

team that would be used for all patients with head and neck cancer to

reduce the toxicity, reduce treatment interruptions, improve nutrition,

and thereby improve patients’ quality of life, adherence to treatment, and

outcomes.

Materials/Methods: A checklist was created for the patient and the health

care team along with an education packet for the patient. Checklist and

packet items included the following: team member identification; dental

care recommendations for the dentist with a return document for care

received (new process); feeding tube placement education/care of tube;

weekly dietician consultation; prehabilitation speech therapy to teach

swallow exercises (new process); tobacco cessation specialist consultation;

oral protectant prescription to start with treatment; social work screening;

nurse twice weekly assessments and evaluation of oral cavity; and radia-

tion, medical and surgical oncologists’ oversight/management of the

combined modality treatment. Adherence was assessed by no interruptions

in treatment, defined as unplanned treatment-related interruptions greater

than 5 days, and the patient’s weight maintained. Interruptions greater than

5 days are statistically significant in changing the outcome of treatment.

Results: The head and neck protocol checklist was implemented July

2014. Initial data are promising and suggest effectiveness. Thirty-two

patients were treated with head and neck cancers from July 2014 through

June 2015. Of 32 patients, 17 had feeding tubes. Twenty-two percent or 7

patients experienced weight loss greater than 5 kg, with a range from 5.2 to

11.8 kg and a mean of 7.8 kg. Only 1 patient (3%) experienced a treatment

break greater than 5 days in length due to treatment-related symptoms.

That patient had an 8-day break mainly due to fatigue. One patient had a

4-day break due to mucositis, 1 had a 3-day break for mucositis, and

another had a 2-day break due to gastrointestinal issues. One patient had a

tooth abscess early in treatment and missed 5 days.

Conclusion: This study confirms the importance of a nurse-led multi-

professional team to improve the quality of care for patients undergoing

head and neck cancer treatment to impact adherence to treatment.

Author Disclosure: C. Johnson: None. K. Ottaviano: None. J.M.

Smudde: None. N. Mohideen: None.

161High-Dose Versus Weekly Cisplatin Definitive ChemoradiationTherapy for Human PapillomaviruseRelated OropharyngealSquamous Cell Carcinoma of the Head and NeckC.A. Perez,1 M.J. Amsbaugh,1 W. Claudino,2 M. Yusuf,1 X. Wu,2

S.N. Rai,2 T. Roberts,2 L. Wilson,1 L. Hall Volz,1 S. Khanal,1 A.B. Jenson,1

E. Cash,1 J. Bumpous,1 C.L. Silverman,1 P. Tennant,1 N.E. Dunlap,1

and R. Redman1; 1University of Louisville, Louisville, KY,2University of Louisville School of Medicine, Louisville, KY

Purpose/Objective(s): To compare the outcomes and toxicity of high-

dose cisplatin (HDC) versus weekly (WC) definitive concurrent chemo-

radiation (CRT) for patients with human papillomavirus (HPV)-related

squamous cell carcinoma (SCC) of the oropharynx.

Materials/Methods: Patients with p16-positive SCC of the oropharynx

treated using CRT with cisplatin at a single institution between 2010 and

2014 were reviewed from a prospective database. Patient demographic,

treatment, toxicity, and outcome data were collected. Toxicity was coded

according to Common Terminology Criteria for Adverse Events version

4.03. The Kaplan-Meier method was used to estimate overall survival (OS)

and event-free survival (EFS).

Results:We identified 56 patients during the study period. One patient was

excluded because of the presence of 2 synchronic primaries at presenta-

tion. Of the 55 patients included, 85% were male and 57% had a history of

smoking >10 pack-years. Median age at time of diagnosis was 55.4 years

(range 40.3 to 80.0 years). ATotal of 22 patients were treated with HDC at

doses of 100 mg/m2 on days 1 and 22 of the CRT, and 33 were treated with

WC at doses of 40 mg/m2 every week. Groups were well balanced in

respect to sex (PZ.454) and smoking history (PZ.799). The median total

dose of cisplatin for both cohorts was 200 mg/m2, with a mean of 195 mg/

m2 for the HDC group and 189 mg/m2 for the WC group. After a median

follow-up of 21 months, there was 1 local failure and no distant failures in

the HDC cohort. In the WC group, there were 6 total failures (2 local and 4

distant). One patient in the WC died without a failure 2 months after

therapy due to infection. At last follow-up, EFS was improved for patients

treated with HDC compared to WC (95% vs 79%; PZ.043), despite a

longer follow-up in patients with HDC (26 months vs 16 months). There

was no significant difference in OS (95% vs 91%; PZ.421). In terms of

toxicity, there was a nonsignificant trend toward higher rates of weight loss

in the HDC arm compared with the WC arm, both in all weight loss grades

(87% vs 100%, PZ.128) and weight loss more than grade 1 (60% vs 77%;

PZ.190). However, gastric tube dependence at 6 months following ther-

apy was similar between groups (13% vs 12%, PZ1.000). Also, acute

renal injury of any grade (55% vs 36%, PZ.183) and grade 3 or 4 he-

matological toxicity (20% vs 23%, PZ.761) were similar between groups.

Conclusion: In patients with HPV-positive oropharyngeal SCC, definitive

CRTwith HDC andWC have a similar toxicity profile in terms of weight loss,

gastric tube dependence 6 months after therapy, hematologic and renal injury

rate. HDC has better EFS when compared with WC, and this seems to be

driven by increased distant failure rates, although theOSwas similar. A longer

follow-up will be needed to better assess whether the OS remains similar.

Author Disclosure: C.A. Perez: None. M.J. Amsbaugh: None. W.

Claudino: None. M. Yusuf: None. X. Wu: None. S.N. Rai: None. T.

Roberts: None. L. Wilson: None. L. Hall Volz: None. S. Khanal: None.

A.B. Jenson: None. E. Cash: None. J. Bumpous: None. C.L. Silverman:

None. P. Tennant: None. N.E. Dunlap: None. R. Redman: None.

162Successful Development, Implementation, and Assessment of aClinical Care Path (CP) Guide for the Treatment of HumanPapillomavirus (HPV)-Initiated Oropharynx CancerD.J. Adelstein,1 S. Koyfman,2 B.B. Burkey,1 N. Houston,1 K. Tullio,1

E. Lamarre,1 J. Scharpf,1 R. Lorenz,1 M. Khan,1 T. Nwizu,1

J.F. Greskovich, Jr,3 D. Ives,1 M. Rahe,1 J. Bodmann,1 B. Harr,1

C. Cummings,1 and B. Bolwell1; 1Cleveland Clinic, Cleveland, OH,2Cleveland Clinic Foundation, Cleveland, OH, 3Cleveland Clinic Florida,

Weston, FL

International Journal of Radiation Oncology � Biology � Physics896

Purpose/Objective(s): Sufficient evidence now exists to allow develop-

ment of CP guides for most cancers. The goals of these CP guides are to

standardize treatment, optimize outcomes, minimize toxicities, and control

costs. We propose that meaningful metrics can be identified and used to

assess both compliance with these CP guidelines and the quality of clinical

care.

Materials/Methods: In a multidisciplinary effort initiated in 2013, a

clinical CP guide was developed for the use of definitive chemoradiation

therapy (CRT) in patients with newly diagnosed stage III-IVB HPV-initi-

ated squamous cell carcinoma of the oropharynx, the most common head

and neck cancer diagnosis at our institution. The guide was widely vetted

among the head and neck oncology team and approved by consensus at

both our main institution and regional satellites. It included diagnostic and

staging criteria for CP eligibility, pretreatment evaluation, treatment spe-

cifics, and supportive care guidelines. Concurrent radiation therapy with

either cisplatin or cetuximab was considered an acceptable treatment

choice. CP-related metrics were defined and retrospectively analyzed.

Results: All head and neck cancer patients treated at our institution

between January 2014 and June 2015 were reviewed. There were 82

CP-eligible patients identified, 60 of whom were treated according to the

CRT CP. The 22 patients not treated on the CP included 15 who underwent

primary surgery and 7 treated with radiation alone (4 with early-stage III

tumors and 3 deemed inappropriate for chemotherapy). All 60 of the CRT

CP-treated patients had their diagnosis of HPV-initiated squamous cell

oropharynx cancer confirmed; all had multidisciplinary involvement

including otolaryngology, radiation, and medical oncology; and all were

given a formal clinical stage. All 60 received a radiation dose of at least 70

Gy using intensity modulated radiation therapy, concurrently with either

high-dose cisplatin every 3 weeks (41 patients), weekly cisplatin (14 pa-

tients), or a standard cetuximab regimen (5 patients). Among the cisplatin-

treated patients, a total cisplatin dose >200 mg/m2 was targeted and was

achieved in 96% of them. A total radiation treatment duration of <7weeks

was targeted and was achieved in 88% of patients. A requirement for

disease restaging after treatment within 3 to 4 months was defined and was

accomplished in all patients at a median of 94 (max 123) days.

Conclusion: CP development for locoregionally advanced HPV-initiated

oropharynx cancer can successfully standardize patient management.

Meaningful metrics can be defined and used to assess CP compliance.

Author Disclosure: D.J. Adelstein: None. S. Koyfman: None. B.B. Bur-

key: None. N. Houston: None. K. Tullio: None. E. Lamarre: None. J.

Scharpf: None. R. Lorenz: None. M. Khan: None. T. Nwizu: None. J.F.

Greskovich: None. D. Ives: None. M. Rahe: None. J. Bodmann: None.

B. Harr: None. C. Cummings: None. B. Bolwell: None.

163Identification of Anatomic Correlates of Failure in Patients WithT4a Larynx CancerA.S.R. Mohamed,1 B.H. Pham,2 J.A. Messer,3 W.H. Morrison,1

M. Zafereo,1 A.C. Hessel,1 S. Lai,1 M.S. Kies,1 R. Ferrarotto,1

A.S. Garden,1 R.S. Weber,1 D.I. Rosenthal,1 and C.D. Fuller1;1The University of Texas MD Anderson Cancer Center, Houston, TX,2The Chicago Medical School at Rosalind Franklin University of Medicine

and Science, North Chicago, IL, 3The University of Texas Medical School,

Houston, TX

Purpose/Objective(s): To identify whether anatomic patterns of primary

tumor invasion in T4a laryngeal squamous cell carcinoma predict local

control (LC) in patients treated with nonsurgical larynx preservation (LP)

approaches versus patients treated with total laryngectomy (followed by

postoperative radiation therapy (TL-PORT)).

Materials/Methods: We retrospectively reviewed the medical records of

patients with T4a (according to the seventh edition of the cancer staging

manual of the American Joint Committee on Cancer) squamous cell

laryngeal cancer treated between 1983 and 2011 at a single institution.

Anatomic invasion by primary tumor of any combination of cartilage

(cricoid or thyroid), muscles, or soft tissue (thyroid gland or esophagus)

was recorded. Areas of invasion were then categorized by number of

structures and multiplicity, such as single cartilage, muscle, or soft tissue

versus multiple (eg, cartilage and muscle/soft tissue), then grouped into a

binary of single (SIP) or multiple invasion pattern (MIP). Kaplan-Meier

and log-rank tests were used to evaluate LC. Cox regression multivariate

analysis was performed.

Results: Two hundred twenty patients with a median follow-up of 68

months were analyzed. Sixty (27%) received LP with RT (85% of whom

also received chemotherapy), and 160 patients (73%) received TL-PORT.

Eighty percent of patients in the LP group had SIP, and 20% had MIP. In

comparison, 62% of TL-PORT patients had SIP, while 38% had MIP. The

5-year LC was significantly better in LP patients with SIP compared to

MIP (88% vs 0%, P<.0001). However, the 5-year LC was not statistically

different in patients with SIP compared to MIP in the TL-PORT group

(86% vs 97%, PZ.4). On multivariate analysis, MIP in the LP cohort was

the only independent predictor of LC (hazard ratio 19.4; 95% confidence

interval 5.5-80.3, P<.0001), but in the TL-PORT group, none of the

studied variables (age, sex, ethnicity, larynx subsite of origin, performance

status, nodal stage, radiation dose, and chemotherapy use) was predictive

of LC.

Conclusion: LC for T4a laryngeal cancer patients with SIP and MIP is

favorable in the TL-PORT group, while LP should be eschewed for even

minimal multianatomic site involvement but remains an option for selected

cases with SIP.

Author Disclosure: A.S. Mohamed: None. B.H. Pham: None. J.A.

Messer: None. W.H. Morrison: None. M. Zafereo: None. A.C. Hessel:

None. S. Lai: None.M.S. Kies: None. R. Ferrarotto: None. A.S. Garden:

None. R.S. Weber: None. D.I. Rosenthal: None. C.D. Fuller: Research

Grant; National Institutes of Health/National Cancer Institute, SWOG/

Hope Foundation, General Electric Healthcare/MD Anderson Center,

Elekta AB/MD Anderson, MD Anderson Cancer Center.

164Human Papillomavirus Status and Long-Term Outcomes for StageIII-IV Squamous Cell Carcinoma of the Oral Cavity, Oropharynx, andHypopharynx Treated With a Multimodal Intensification RegimenP.M. Wald,1 D. Blakaj,2 B. Swanson,1 D. Schuller,3 H. Hamedani,1

A. Agrawal,1 E. Ozer,1 T. Teknos,3 M. Old,1 J.W. Rocco,3 A.D. Bhatt,1

V.M. Diavolitsis,2 L. Wei,1 and J.C. Grecula2; 1The Ohio State University

Wexner Medical Center, Columbus, OH, 2The James Cancer Hospital and

Solove Research Institute, Wexner Medical Center at The Ohio State

University, Department of Radiation Oncology, Columbus, OH,3The James Cancer Hospital and Solove Research Institute, Wexner

Medical Center at The Ohio State University, Department of

Otolaryngology-Head and Neck Surgery, Columbus, OH

Purpose/Objective(s): From February 1993 until December 2000, a

series of 3 consecutive prospective phase 2 trials were carried out at a

single institution to evaluate an intensified treatment regimen. Previously

reported outcomes showed excellent rates of protocol compliance, disease

control, and overall survival. Our objectives are to (1) analyze tumor

human papillomavirus (HPV)/p16 status for these patients, (2) provide

long-term updates on disease control and overall survival, and (3) deter-

mine whether favorable outcomes were achieved for our HPV-negative

cohort. We hypothesize that our intensification protocols provide favorable

disease control for HPV-negative patients.

Materials/Methods: One hundred twenty-seven adult patients with clinicalstage III-IV squamous cell carcinoma of the oral cavity, oropharynx, or

hypophyarynx were previously treated according to 1 of 3 “intensification

protocols” consisting of preoperative chemoradiation (910 cGy in 7 bid

fractions with concurrent cisplatin), surgical resection with intraoperative

radiation therapy (750 cGy), and adjuvant chemoradiation (4500 cGy in 20

fractions with concurrent cisplatin or cisplatin/paclitaxel). We retrospec-

tively reviewed patient charts to update long-term disease control and

overall survival outcomes. For those who were lost to follow-up, the social

security death index was used to gather survival data. The date and

location of locoregional and distant disease failures were recorded. Long-

term locoregional and systemic disease control rates were calculated at

Volume 94 � Number 4 � 2016 Posters 897

1-, 3-, 5-, 10-, and 15-year follow-up intervals. Previously obtained tissue

samples are in the process of HPV/p16 testing, results pending.

Results: Median follow-up was 54 months (range 0-260 months). Median

overall survival was 81 months. Survival rates at 1, 5, 10, and 15 years

were 77%, 54%, 42%, and 32%, respectively. Twenty-seven percent of

patients (34 of 127) are still alive without disease an average of 208

months since starting treatment. Fifteen patients had locoregional failures.

Twenty-one patients developed distant metastatic disease. The most

common first sites of metastatic disease were lung (15), liver (4), bone (1),

and intracranial extension (1). Twelve patients had grade 5 toxicities,

including neutropenic sepsis (4), cardiorespiratory arrest (3), myocardial

infarction (1), stroke (1), and unknown (3).

Conclusion: These intensification regimens result in excellent disease

control and long-term survival with acceptable levels of protocol

compliance and toxicity. If HPV analysis shows favorable results in our

HPV-negative cohort, this would suggest a need for a phase 2 multi-

institutional intensification protocol for HPV-negative patients.

Author Disclosure: P.M. Wald: None. D. Blakaj: None. B. Swanson:

None. D. Schuller: None. H. Hamedani: None. A. Agrawal: None. E.

Ozer: None. T. Teknos: None. M. Old: None. J.W. Rocco: None. A.D.

Bhatt: None. V.M. Diavolitsis: None. L. Wei: None. J.C. Grecula: None.

165Chondroradiation Necrosis and Late Radiation-Related TissueChanges in the Larynx: Twenty-FoureYear University of WisconsinExperienceT. Gessert, C. Britt, A.P. Wojcieszynski, A. Wieland, P.M. Harari,

and G.K. Hartig; University of Wisconsin, Madison, WI

Purpose/Objective(s): To review our series of patients with chondror-

adiation necrosis (CRN) of the larynx and review the current literature to

help define the challenges with both the diagnosis and treatment of this

uncommon complication of therapy.

Materials/Methods: Medical records were reviewed for 612 patients who

underwent primary or salvage radiation for laryngeal cancer from 1991 to

2015. Two hundred ninety-three patients were treated initially at the

University of Wisconsin Hospitals and Clinics. Patients receiving treat-

ment at an outside institution were excluded. Records were reviewed to

identify patients with a diagnosis of CRN and the characteristics, treat-

ments, and outcomes related to those patients. Patients were not included if

they were known to have persistence or recurrence of disease.

Results: Of the 293 patients, 7 cases (2.4%) of probable CRN were

identified. Of these, 4 had supraglottic tumors and 3 had glottic tumors. All

7 patients received an initial radiation dose of 70 Gy or greater. All 7 cases

of laryngeal radiation reaction were classified as Chandler grade III or IV.

Six of 7 patients exhibited dysphagia and/or edema, and 4 of 7 had

laryngeal obstruction. Computed tomography scans were performed on all

patients and positron emission tomographic (PET) imaging was performed

on 6 of 7 patients. Of the 6 patients undergoing PET imaging, 2 were read

as likely recurrence. Five patients underwent at least 1 biopsy, and 3 pa-

tients ultimately underwent total laryngectomy, all of which were negative

for disease. Most of our patients received supportive treatment. Six patients

underwent tracheotomy, and 3 underwent gastronomy tube (G-tube)

placement, of which 1 remained tracheotomy dependent and one remained

G-tube dependent. The health care burden consisted of a total of 10

emergency room visits and 8 hospital admissions in total for the 7 patients.

Conclusion: CRN is an uncommon and challenging diagnosis. Early

diagnosis aids in laryngeal preservation, but failure rates of supportive

treatment for advanced disease remain high.

Author Disclosure: T. Gessert: None. C. Britt: None. A.P. Wojcieszynski:

None. A. Wieland: None. P.M. Harari: None. G.K. Hartig: None.

166Correlating Patient-Reported Outcomes and Provider-DocumentedAdverse Events During Radiation Therapy for Head and Neck CancerJ.R. Niska,1 M.Y. Halyard,1 A.D. Tan,2 P.J. Atherton,2 S.H. Patel,1

and J.A. Sloan2; 1Mayo Clinic, Phoenix, AZ, 2Mayo Clinic, Rochester, MN

Purpose/Objective(s): To correlate changes in real-time electronic pa-

tient-reported outcomes (ePROs) with provider-documented adverse

events (AEs) at multiple time points during the course of radiation therapy

(RT) for head and neck (H&N) cancer.

Materials/Methods: Sixty-five H&N RT patients completed an electronic

real-time 12-item Linear Analogue Self-Assessment (LASA) at baseline,

before biweekly appointments, and at last week of RT. Changes in LASA

item scores between time points were calculated. Clinical data were

collected from the institutional medical record. AEs were recorded at

baseline, before biweekly appointments, and at the last week of RT. We

graded AEs using the Common Terminology Criteria for Adverse Events

version 4.0. LASA item scores were categorized according to maximum

grade AE experienced by each patient at any given time point and analyzed

using Wilcoxon methodology.

Results: Over the course of H&N RT, a majority of patients reported a

clinically significant decrease in most ePRO quality of life (QOL)

domains. QOL domains for which patients most commonly reported a

clinically significant decrease were fatigue (77.8% of patients), social

activity (75.4%), and overall QOL (74.2%). At the last week of RT,

patients reported average LASA scores worse than baseline for all QOL

domains except level of support and financial concerns. During treatment,

all patients experienced at least grade 2 AE with 35.4% experiencing at

least grade 3 AE. At the end of treatment, every patient had either a grade

2 or grade 3 AE. At baseline, patients experiencing a maximum grade 2

AE reported worse physical WB (48.6 for grade 2, 70.5 for grade 1,

PZ.0190). At week 1, the following QOL domains were statistically

different according to maximum grade AE: overall QOL (77.4 for grade 1,

64.5 for grade 2, and 10.0 for grade 3, PZ.0222), mental WB (81.3 for

grade 1, 66.0 for grade 2, and 30.0 for grade 3, PZ.0073), physical WB

(73.9 for grade 1, 56.0 for grade 2, and 10.0 for grade 3, PZ.0065), and

fatigue (60.9 for grade 1, 42.5 for grade 2, and 20.0 for grade 3,

PZ.0336).

Conclusion: H&N RT is associated with diminished QOL and significant

AEs. Diminished QOL at early time points appears to correlate with

maximum grade AE for multiple ePRO QOL domains. Patients adapting to

their circumstances or providers addressing AEs may account for the lack

of correlation between ePRO QOL domains and maximum grade AE

beyond week 1 of RT.

Author Disclosure: J.R. Niska: None. M. Halyard: Medical School Dean;

Mayo Medical School. A.D. Tan: None. P.J. Atherton: None. S.H. Patel:

None. J.A. Sloan: None.

167Predictors of Dysphagia After Treatment With SubmandibularGland-Sparing Radiation Therapy for Advanced-StageOropharyngeal Squamous CancerU. Parvathaneni,1 M. Nyflot,2 J.J. Liao,1 G.E. Laramore,1

and M.F. Gensheimer1; 1University of Washington, Seattle, WA,2Department of Radiation Oncology, University of Washington,

Seattle, WA

Purpose/Objective(s): Dysphagia is common after head and neck radia-

tion therapy and has a detrimental effect on quality of life. Sparing the

submandibular gland contralateral to the primary tumor (cSMG) from high

radiation dose has recently been shown to reduce xerostomia. Xerostomia

is known to worsen posttreatment dysphagia. SMG saliva is rich in mucins,

which act as an essential lubricant during the swallowing process. The goal

of this study was to determine whether cSMG sparing treatment is asso-

ciated with reduced late post-treatment dysphagia in advanced-stage

oropharynx cancer patients.

Materials/Methods: Patients treated with definitive bilateral neck intensitymodulated radiation therapy for stage III/IV oropharyngeal squamous

cancer were eligible for this retrospective study. Those with disease

recurrence were excluded. Salivary glands and swallowing-related organs

at risk, including pharyngeal constrictor muscles, were contoured in a

uniform manner. Primary (objective) endpoint was time from end of

radiation treatment to freedom from gastrostomy tube dependence.

International Journal of Radiation Oncology � Biology � Physics898

Secondary (subjective) endpoint was presence or absence of grade

2+ observer-rated late dysphagia. Cox proportional hazards regression and

logistic regression were used to assess influence of normal tissues doses on

dysphagia endpoints.

Results: Sixty-nine patients were included. Ninety-seven percent received

concurrent systemic therapy. Fifty-seven percent had cSMG mean dose

<50 Gy, a level shown to predict for xerostomia. Eighty-four percent of

patients had a gastrostomy tube placed. On univariate analysis, the

strongest predictor of time to freedom from gastrostomy tube dependence

was cSMG dose (hazard ratio 0.97 per Gy [95% confidence interval 0.95-

0.98], P<.0001). This relationship persisted on multivariate analysis

(PZ.04). Patients with cSMG dose less than median (42 Gy, nZ34) had

significantly shorter time to freedom from gastrostomy tube dependence:

median 1.9 versus 3.5 months, P<.0001. No patient with cSMG dose less

than median was feeding tube dependent 1 year posttreatment. Higher

cSMG dose was associated with worse observer-rated dysphagia on uni-

variate but not multivariate analysis.

Conclusion: cSMG sparing radiation therapy with the goal of improving

posttreatment xerostomia may improve dysphagia as well.

Author Disclosure: U. Parvathaneni: None. M. Nyflot: None. J.J. Liao:

None. G.E. Laramore: None. M.F. Gensheimer: None.

168A Phase 1 Trial of Ketogenic Diet With Concurrent Chemoradiation(CRT) in Head and Neck Squamous Cell Carcinoma (HNSCC)C.M. Anderson, E. Loth, E. Opat, K. Bodeker, L. Ahmann, J. Parkhurst,

W. Sun, M. Furqan, D. Laux, H. Brown, S. Vollstedt, K. Follmer, D. Ma,

D. Spitz, M. Fath, J. Buatti, and B.G. Allen; University of Iowa Hospitals

& Clinics, Iowa City, IA

Purpose/Objective(s): Ketogenic diet (KD) combined with CRT reduced

tumor growth and improved survival in preclinical models. We hypothe-

sized stage III-IVb HNSCC patients would be able to remain compliant

with KD because of percutaneous endoscopic gastrostomy (PEG) tube

requirement during CRT.

Materials/Methods: This institutional review boardeapproved phase 1

clinical trial enrolled stage III-IVb definitive and postoperative HNSCC

patients receiving concurrent platinum-based CRT. PEG placement was

required up front, but subjects were encouraged to continue KD by mouth

due to benefits in continued swallowing activity on ultimate function. KD

recipe ideas and shakes were provided to subjects for daily consumption

for 5 weeks concurrent with CRT. Shakes could also be delivered by way

of PEG tube for caloric requirements. Fingerstick ketones (FK) were

checked Monday through Friday, and serum beta-hydroxybutyrate (BHB),

glucose, and uric acid were checked weekly. Lipid panel was checked at

week 3. In addition, serum oxidative stress parameters were assessed prior

to, during, and after completing KD.

Results: Median follow-up for all enrolled subjects (nZ9) from comple-

tion of RT was 7.1 months (range: 0-12.9 months). Three of 9 subjects

successfully completed 5 weeks of KD as prescribed. Successful subjects

used scheduled antiemetics, consumed shakes by way of PEG tube as

opposed to orally, and had strong social support. Median days on KD for

those who discontinued was 6 (range: 0-8). Reasons for discontinuation

included additional stress of diet compliance (1 of 6 patients), Common

Terminology Criteria for Adverse Events version 4.0 grade 2 nausea (2 of 6

patients), grade 3 nausea (1 of 6 patients), grade 3 fatigue (1 of 6 patients),

and grade 4 hyperuricemia (1 of 6 patients). SAEs (nZ6) included paro-

tiditis, nausea, vomiting, neutropenic fever, hyperuricemia, and pancrea-

titis, of which hyperuricemia and pancreatitis were possibly attributed to

KD and considered DLTs. The trial was temporarily suspended after the

grade 4 hyperuricemia (12.7 nd/dL; nl ref 2.4-7.0); the protocol was

amended to initiate allopurinol and address treatment of diet-related

hyperuricemia. In those who completed KD, the median days FK were

elevated and weeks the BHB levels were above baseline were 24 days

(range: 19-25) and 5 weeks (range: 4-5), respectively. FK levels were

impacted by dexamethasone given for chemotherapy and by carbohydrates

in liquid medications. Median uric acid levels were 5.3 nd/dL (range:

4.4-5.4). Lipids remained normal. Serum oxidative stress markers,

as assessed by protein carbonyls, increased linearly with increasing days

on KD.

Conclusion: While challenging despite PEG availability, KD compliance

is possible when combined with concurrent CRT for HNSCC. Enrollment

continues and updated results will be presented.

Acknowledgment(s): Research supported by NIH U54TR001356 and

KetoCal 4:1 provided by Nutricia Pharmaceuticals.

Author Disclosure: C.M. Anderson: Employee; University of Iowa Hos-

pitals & Clinics. Uncompensated; research is unrelated to this project;

Galera Therapeutics, Inc.. E. Loth: Student/Student Employee; University

of Iowa Hospitals & Clinics. E. Opat: None. K. Bodeker: None. L.

Ahmann: None. J. Parkhurst: None. W. Sun: None. M. Furqan: None.

D. Laux: None. H. Brown: None. S. Vollstedt: None. K. Follmer: None.

D. Ma: Student and Research Assistant; University of Iowa. D. Spitz:

Faculty member; University of Iowa. M. Fath: Assistant Research Sci-

entist; University of Iowa. J. Buatti: None. B.G. Allen: Physician scien-

tist; University of Iowa Hospitals & Clinics. Research Grant; ASTRO.

169Prognostic Significance of Extracapsular Spread and Perineural andLymphovascular Invasion in Patients with HPV- and Non-HPVeRelated Oropharyngeal Squamous Cell Carcinoma.R. Borucki,1 S.A. Nguyen,1 E.A. Nicolli,1 S. Sridharan,2 T. Day,1

and D. Neskey1; 1Medical University of South Carolina, Charleston, SC,2Georgetown University, Georgetown, DC

Purpose/Objective(s): Approximately 15,000 cases of oropharyngeal

squamous cell carcinoma (OPSCC) are diagnosed each year in the United

States. Previously these cancers had a strong association with chronic

alcohol and tobacco use. Over the past decade the increasing incidence of

the human papillomavirus (HPV) has been identified as a major etiologic

agent of oropharyngeal cancer and currently is associated in over 70% of

cases in the United States. Although the prognostic significance of HPV is

clear, the impact of traditional pathologic parameters remains unclear.

Therefore, the goal of this study is to compare surgical pathologic vari-

ables in non-HPVe and HPV-associated OPSCC. We hypothesize the

presence of lymphovascular invasion will portend a poor prognosis in

HPV-related OPSCC, whereas perineural invasion and extracapsular

spread will be associated with decreased survival in non-HPVerelateddisease.

Materials/Methods: We conducted a retrospective chart review of 240

patients treated for oropharyngeal SCC at a tertiary care cancer center. The

primary outcome measures were overall and disease-free survival. The

endpoints assessed were clinical and pathologic T and N stage and pres-

ence of extracapsular spread, perineural invasion, or lympohvascular

invasion.

Results: Of the 240 patients with OPSCC reviewed, 116 patients under-

went surgery as part of their primary treatment and had pathologic vari-

ables and p16 status available for review. Of this cohort, 70 patients had

HPV-related disease, while 46 patients had non-HPVerelated disease. The

presence of perineural invasion was associated with decreased disease-free

survival in both HPV and non-HPVerelated OPSCC (PZ.003 and

PZ.003, respectively). Lymphovascular invasion portended a worse

overall and disease-free survival in HPV-related disease (PZ.050 and

PZ.006, respectively) but not non-HPVerelated (PZ.717 and PZ.942,

respectively). In contrast, extracapsular spread did not correlate with

decreased overall or disease-free survival in either HPV-related OPSCC

(PZ.899 and PZ.575, respectively) or non-HPVerelated OPSCC

(PZ.772 and PZ.843, respectively).

Conclusion: In the current series, the presence of perineural invasion

portended a worse prognosis in patients with OPSCC independent of HPV,

whereas there was not an associated prognostic significance of extrac-

apsular spread regardless of HPV status. In contrast, lymphovascular in-

vasion appears to correlate with decreased survival in patients with

Volume 94 � Number 4 � 2016 Posters 899

HPV-related OPSCC. Based on these findings the indications for adjuvant

therapy following surgery for OPSCC may be different for HPV- and non-

HPVerelated disease.

Author Disclosure: R. Borucki: None. S.A. Nguyen: None. E.A. Nicolli:

None. S. Sridharan: None. T. Day: None. D. Neskey: None.

170Initial Experience Using Transoral Robotic Surgery for Advanced-Stage (T3) Tumors of the Head and NeckJ. Kass,1 C. Pool,2 M. Teng,2 B. Miles,2 and E. Genden2; 1Boston

University, Boston, MA, 2Icahn School of Medicine at Mount Sinai, New

York, NY

Purpose/Objective(s): Currently transoral robotic surgery (TORS) using

the DaVinci system is FDA approved for use in benign and malignant

early-stage T1 and T2 tumors. At institutions with a robust TORS program,

there is an increased comfort level with both the technology and transoral

anatomy as such larger tumors are being excised. We sought to identify

how many stage T3 oropharyngeal squamous cell carcinomas (OPSCC)

have been attempted using robotic surgery at a high-volume institution

(>50 cases per year). In particular, we evaluated the following outcomes:

initial clinical T stage, margin status, hospital stay, tracheostomy, and

percutaneous endoscopic gastrostomy (PEG) tube dependence and use of

adjuvant therapy.

Materials/Methods: We conducted an institutional review boardeapproved

retrospective chart review of 280 consecutive cases between 2011 and 2015.

All pathology reports were reviewed, and those with final pathology

demonstrating stage T3 and T4 (pT3-4) OPSCC were included. One patient

with incomplete records was excluded.

Results: Fourteen patients were identified as pT3, and 1 patient was

excluded for incomplete preoperative staging. Mean age was 65 years

(range 50-85 years) and 64% (nZ9 of 14) were human papillomavirus

positive. All but 1 of the patients was male, and 57% (nZ8 of 14) had a

history of tobacco use. Preoperative clinical staging for the cohort was cT1

(nZ2), cT2 (nZ9), cT3 (nZ2), and cT4 (nZ1). Tumors were patholog-

ically staged as either pT3 (nZ13) or pT4a (nZ1). Overall length of stay

was 5 days (range 2-8 days). Negative margins were achieved in 86% of

patients (nZ12 of 14). One patient required a tracheostomy, and 1 patient

remained PEG tube dependent. All patients had postoperative XRT, and

the 2 patients with positive margins had chemotherapy (either docetaxel

and cetuximab or cisplatin).

Conclusion: TORS may be used to resect advanced T-stage malignancies;

however, there are risks of positive margins, tracheostomy, and PEG tube

dependence, as well as hospital stays that exceed the typical 1 to 2 days.

The potential benefit of avoiding 3-modality treatment (surgery, radiation,

and chemotherapy) should only be considered in tumors in which negative

margins can be achieved.

Author Disclosure: J. Kass: None. C. Pool: None. M. Teng: None. B.

Miles: None. E. Genden: None.

171Predictive Factors Associated With Late Pulmonary Toxicity AfterDefinitive Intensity Modulated Radiation Therapy With InductionChemotherapy for Head and Neck CancerH. Inokuchi, M. Yoshimura, and M. Hiraoka; Kyoto University Graduate

School of Medicine, Kyoto 606-8507, Japan

Purpose/Objective(s): Pulmonary toxicity after radiation therapy in pa-

tients with head and neck cancer (HNC) represents an underreported

complication, which is suggested from the long-term update of Intergroup

Radiation Therapy Oncology Group 91-11. Recently, intensity modulated

radiation therapy (IMRT) has been introduced in the treatment of HNC that

may reduce the incidence of late toxicity. We evaluate the incidence and

predictors of pulmonary toxicity in our clinical cohort of patients with

HNC who received IMRT.

Materials/Methods: We reviewed a database of 181 consecutive patients

with squamous cell carcinoma of the head and neck who received IMRT

for whole neck irradiation field during 2002 to 2013 at our hospital. Of

these patients, the data of 80 patients (diagnosed with oropharyngeal,

hypopharyngeal, or laryngeal cancer) who underwent at least 1 cycle of

induction chemotherapy and were scheduled for definitive chemoradiation

therapy (CRT) were queried in this retrospective review. Patients were

ineligible if they received less than 60 Gy or had a follow-up of less than

12 months with critical event. Time to pneumonia was calculated based on

treatment start date to the date of last follow-up note or the first date that

pneumonia was noted by clinical follow-up imaging. Several clinical and

dosimetric parameters correlated with the swallowing structures and sali-

vary glands were assessed.

Results: The average follow-up after treatment was 26 months (range,

12-72 months). Of 80 patients with HNC, 22 patients (27.5%) were

recognized as the episodes of pneumonia at a median of 20 months after

initiating treatment. The 1-year and 3-year cumulative incidence of aspi-

ration pneumonia was 6.0% and 25.4%, respectively. Many of the patients

(15%) had silent aspiration, not associated with clinical symptoms. Four

patients (5%) developed fatal aspiration pneumonia at a median of 12

months (range, 0.9 to 62 months) from the start of IMRT. Among the

patients with survivors, log-rank test identified risk factors (P<.05) for

aspiration pneumonia, including the planning target volume receiving

greater than 63 Gy, severe body weight loss during CRT, a serum albumin

change from baseline, the volume of the supraglottis receiving greater than

60 Gy, and the mean dose of the major salivary glands. On multivariate

analysis using logistic regression, a serum albumin lower than 2.5 g/dL

(PZ.02) and the mean major salivary glands dose greater than 48 Gy

(PZ.03) were significantly associated with the progression of symptom-

atic pneumonia.

Conclusion: Various factors contributed to the risk for aspiration

pneumonia, which is an important late toxicity of noncancer-related

mortality among our advanced HNC cohorts who received IMRT with

induction chemotherapy. Further studies with a longer follow-up are

warranted to identify patients at risk for severe aspiration pneumonia

after CRT.

Author Disclosure: H. Inokuchi: None. M. Yoshimura: None. M.

Hiraoka: None.

172Contralateral Neck-Sparing Radiation Therapy in Select PatientsWith Locally Advanced Oropharyngeal Cancer After Primary SurgeryWith Neck DissectionW. Smith,1 R.L. Bakst,1 E. Genden,1 M. Yao,1 E. Demicco,1

K. Misiukiewicz,1 M. Posner,1 and V. Gupta2; 1Icahn School of Medicine

at Mount Sinai, New York, NY, 2Roswell Park Cancer Institute, Buffalo, NY

Purpose/Objective(s): Current guidelines recommend postoperative ra-

diation therapy (RT) in locally advanced oropharyngeal cancers to include

bilateral neck lymph nodes (LNs). In select cases ipsilateral neck RT alone

may be sufficient. We evaluated the efficacy of sparing the contralateral

LNs in select postoperative patients with base of tongue (BOT) or locally

advanced tonsillar cancers.

Materials/Methods: We queried our institutional review boardeapproved

database in the Department of Radiation Oncology for patients with BOT

or locally advanced (T3-4, N2a-b, or T1-4N1-2b with extracapsular

extension [ECE]) tonsillar squamous cell carcinoma who underwent

resection with LN dissection and received adjuvant contralateral neck-

sparing RT at our institution. On preoperative work up, none of the patients

had suspicious LNs in the contralateral neck by physical exam or by

positron emission tomographic/computed tomographic (PET/CT) scans.

Most patients with BOT cancers had bilateral LN dissections with path-

ologically negative contralateral LNs (mean number of contralateral LNs

dissected 20.3; range, 10-39), whereas those with tonsillar cancers only

had ipsilateral LN dissections. Patients were followed posttherapy with

physical exam every 2 to 3 months and with PET/CT imaging every 3 to 6

months. We determined the number of locoregional failures (LRF),

contralateral LN failures, distant metastases, and deaths.

Abstract 172; Table 1 Number of patients per TNM stage

Base of tongue Tonsil

N0 N1 N2a N2b N0 N1 N2a N2b

T1 0 1 2 2 T1 0 2* 3 2T2 3 0 0 2 T2 0 0 1 4T3 0 0 0 0 T3 0 1 0 0T4 0 0 0 0 T4 0 0 0 0

*Inclusion criteria met by ECE+

International Journal of Radiation Oncology � Biology � Physics900

Results: Twenty-three patients met eligibility criteria. Staging is listed in

Table 1. Of those tested, most had human papillomavirus-positive (HPV+)

(19 of 21Z90%) and p16-positive (21 of 22Z95%) tumors by PCR and

IHC staining, respectively. Indications for RT were LVI (nZ6; 26%), PNI

(nZ2; 9%), close (nZ6; 26%) or positive (nZ3; 13%) margins, N2

(nZ16; 70%), and/or ECE (nZ6; 26%). The radiation fields included the

primary site only (nZ4; 17%), ipsilateral neck only (nZ6; 26%), or pri-

mary and ipsilateral neck (nZ13; 57%) to a mean radiation dose of 62.2

Gy (range, 60-66 Gy). Three patients (13%) received concurrent cisplatin.

Only 1 patient (4%) required percutaneous endoscopic gastrostomy tube

for nutrition. After a mean follow-up time of 22 months (range, 6-47

months), there were no LRFs, contralateral LN failures, distant metastases,

or deaths.

Conclusion: The absence of locoregional and contralateral LN failures or

metastases provides promising evidence as to the safety of contralateral

neck-sparing RT in select, mostly HPV+, postoperative SCC patients with

BOT or locally advanced tonsillar cancers. Longer follow-up and more

patients are needed to validate these findings.

Author Disclosure: W. Smith: None. R.L. Bakst: None. E. Genden:

None. M. Yao: None. E. Demicco: None. K. Misiukiewicz: None. M.

Posner: None. V. Gupta: None.

173Simultaneously Integrated Boost in the Chemoradiation of PatientsWith Locally Advanced Squamous Cell Carcinoma of the Head andNeckdOwn ExperienceA.V. Mikhailov, N.A. Vorobyov, V.P. Sokurenko, E.V. Smirnova,

G.I. Andreev, D.D. Puchkov, A.V. Smirnova, and N.A. Plakhotina; DTC

IIBS named after S.Berezin, Saint-Petersburg, Russia

Purpose/Objective(s): The high recurrence rate and short disease-free

survival after curative treatment of squamous cell carcinoma of the head

and neck (HNSCC) are still actual problems of modern radiation oncology.

Irradiation with higher total doses extends the treatment course and leads

to the tumor repopulation phenomenon, which decreases effectiveness of

treatment. It is possible to overcome the phenomenon of repopulation by

reducing the duration of treatment by increasing single doses delivered to

the tumor.

Materials/Methods: One hundred two patients with histologically

confirmed HNSCC of II-III clinical stages received platinum-based che-

moradiation between 2013 and 2015 in a radiation therapy department at

our clinic. The treatment volumes were formed as follows: the primary

lesion (gross tumor volume [GTV] + 0.5-1.0 cm clinical target volume

[CTV] + planning target volume [PTV] 0.3 cm) was treated to the total

dose equivalent to 70 Gy of conventional fractionation, the upper neck

(levels I-III + PTV 0.5 cm) to 60 Gy, the lower neck (levels IV-V + PTV

0.5 cm)eequivalent to 50 Gy. Single doses to these volumes were 2.21 Gy,

2.0 Gy, and 1.8 Gy, respectively. GTV delineation was performed using

computed tomography (CT) and magnetic resonance imaging, and some

patients had positron emission tomographic/CTscans with 18F-FDG. The

irradiation regimen was once a day, 5 days a week, and the course duration

was 6 weeks (30 fractions). A treatment planning system and medical

linear accelerator were used. The volume, treated to the equivalent of 70

Gy in conventional fractionation, ranged from 164 to 370 mL. Intensity

modulated radiation therapy with simultaneously integrated boost was

used. Tolerances of the eye, lens, optic nerves and chiasm, brain stem,

spinal cord, parotid gland, intact mucosa of the mouth, and pharynx were

not exceeded. Patient positioning accuracy was controlled by kV-imaging

daily and cone beam CT weekly.

Results: Eighty-six of 102 patients (84.3%) received full chemoradiation

course without a break. Radiation toxicity manifested with grade 2 oral and

pharyngeal mucositis and grade 2 radiation epidermitis. Sixteen patients

(15.7%) took a break of 5 to 7 days after the 23 to 25 fractions due to the

development of grade 3 mucositis (PTV70Gy in these cases exceeded 330

mL). After 1 month, we saw almost complete relief of radiation mucositis

and dermatitis. During further observation, patients noted the satisfactory

salivation and there was no incidence of xerostomia. Progressive disease in

the first 6 months after irradiation registered in 4 patients with highly

differentiated HNSCC.

Conclusion: Using the technique of simultaneously integrated boost with

intensity modulated radiation therapy during chemoradiation therapy of

HNSCC can increase effectiveness of treatment by reducing the duration

of radiation therapy course while maintaining satisfactory tolerability.

Author Disclosure: A.V. Mikhailov: None. N.A. Vorobyov: None. V.P.

Sokurenko: None. E.V. Smirnova: None. G.I. Andreev: None. D.D.

Puchkov: None. A.V. Smirnova: None. N.A. Plakhotina: None.

174Use of Supportive Care in Patients With Metastatic Squamous CellCarcinoma of the Head and Neck (MSCCHN)E. Nash Smyth,1 J. Kubisiak,2 L. Bowman,1 L. Li,1 M. Chace,2

E. Kubisiak,2 D. Gilden,2 A.B. Lin,1 and D.M. Gilden2; 1Eli Lilly and

Company, Indianapolis, IN, 2JEN Associates, Inc., Cambridge, MA

Purpose/Objective(s): Supportive care strategies are critical to mini-

mizing treatment and disease-related complications while maximizing

quality of life for MSCCHN patients (pts). The study objective was to

evaluate the frequency of supportive care use in U.S. pts with MSCCHN.

Materials/Methods: Patients registered in Surveillance, Epidemiology and

End Results (SEER) as having SCCHN between 2005 and 2009 and linked

to Medicare claims from 2002 to 2010 were evaluated. Metastases were

identified by the presence of stage IVC as denoted in SEER records or

secondary/distant cancer diagnoses in Medicare claims. Supportive care

use as per National Comprehensive Cancer Network (NCCN) guidelines

was evaluated by therapy status. All analyses were descriptive.

Results: A total of 4616 pts with 93,095 pt-months were eligible. First-line

therapy for metastatic disease was received by 1902 pts. A greater pro-

portion of pt-months on systemic therapy were spent receiving supportive

care versus time not spent on systemic therapy (Table 1). The percentage

of pt-months receiving supportive care for pts who permanently dis-

continued systemic therapy (ie, postsystemic) was comparable to the

percentage of pt-months prior to initiating therapy and/or not receiving

systemic treatment for metastatic disease. During >1-month gaps in sys-

temic therapy (ie, treatment interruptions), there were larger absolute de-

creases in the proportion of pt-months on symptom management

(�41.4%), nutritional support (�21.5%), and infection treatment (�6.2%)

versus time on systemic treatment.

Conclusion: These results help to characterize supportive care treatment

considerations for U.S. patients with MSCCHN. A greater proportion of pt

time on systemic treatment received supportive care across NCCN cate-

gories versus pts not receiving systemic treatment. The majority of sup-

portive care use dropped during gaps in treatment, perhaps in association

with the decrease in systemic therapy side effects.

Author Disclosure: E. Nash Smyth: Stock; Eli Lilly. J. Kubisiak: contract

research; Eli Lilly, Janssen Pharmaceuticals. L. Bowman: Stock; Eli Lilly.

L. Li: Stock; Eli Lilly. M. Chace: contract research; Eli Lilly, Janssen

Pharmaceuticals. E. Kubisiak: contract research; Eli Lilly, Janssen Phar-

maceuticals. D. Gilden: contract research; Eli Lilly. A.B. Lin: Stock; Eli

Lilly. D.M. Gilden: Independent Contractor; Eli Lilly and Company,

Janssen Pharmaceuticals.

Abstract 174; Table 1

Supportive care categories Untreated/ 1st-line 2nd-line 3rd-line �4th-line Treatment interrupion Postsystemic

Total cohort, pt-months 56,032 4707 1438 587 298 2838 27,195Symptom management 42.2% 93.8% 95.9% 96.4% 95.6% 52.4% 41.7%Infection treatment 19.3% 40.9% 34.6% 32.7% 36.2% 26.5% 18.1%Nutritional support 28.0% 69.4% 72.1% 73.1% 74.2% 47.9% 32.3%Speech/swallowing therapy 19.7% 34.6% 33.7% 37.0% 26.2% 35.4% 22.2%Durable medical equipment 18.1% 33.7% 33.3% 32.7% 27.2% 31.5% 18.8%Tracheostomy care 8.8% 13.8% 14.0% 11.8% 16.8% 14.4% 7.7%Wound management 1.9% 6.3% 4.8% 2.4% 2.3% 3.4% 1.5%Dental care/xerostomia 3.6% 5.2% 4.7% 6.3% 4.7% 3.7% 3.4%Depression management 12.2% 22.6% 18.8% 17.5% 15.8% 14.0% 10.9%Pain management 18.5% 35.2% 31.0% 31.5% 34.2% 26.0% 19.9%Social work 1.4% 2.4% 1.9% 1.5% 1.3% 1.5% 1.5%Audiology care 3.3% 3.3% 2.5% 2.4% 3.0% 4.9% 3.3%

Volume 94 � Number 4 � 2016 Posters 901

175Prognostic Value of Midtreatment Nodal Response toChemoradiation in Oropharyngeal Squamous Cell Carcinomas:Implications for Treatment ModificationK.S. Hu,1 R. Stewart,2 A. Jacobson,3 M. Persky,4 S. Schantz,5 T. Tran,6

M. Urken,7 B. Culliney,2 Z. Li,1 and L.B. Harrison8; 1NYU Langone

Medical Center, New York, NY, 2Mount Sinai Beth Israel Medical Center,

New York, NY, 3NYU Langone Cancer Center, New York, NY,4NYU Langone, New York City, NY, 5New York Eye and Ear, New York, NY,6NYU Langone Medical Center, New York City, NY, 7Mount Sinai and Beth

Israel, New York City, NY, 8H. Lee Moffitt Cancer Center and Research

Institute, Tampa, FL

Purpose/Objective(s): P16 status and smoking history have emerged as

prognostic factors in oropharyngeal squamous cell carcinoma (OPX).

Daily cone beam computed tomography (CBCT) used for image guidance

can provide real-time volumetric assessment of disease response. It is not

known if early tumor response to radiation therapy (RT) can have an

impact on disease outcomes. This study examines the prognostic impli-

cations of midcourse nodal response to RT.

Materials/Methods: Forty-four patients (pts) with node-positive OPX

underwent definitive concurrent chemoradiation therapy (CCRT) with

daily CBCT between April 2012 and July 2014. Pts had a mean age of 60

years (46-74), 95.5% were male, and 4.5% were female. All received 70

Gy to gross disease, and 98% received concurrent chemotherapy. In all,

70.5% were p16-positive, 20.5% were p16-negative, and 9.1% had

unknown p16 status. A total of 56.8% had <10 pack-year history (PYH),

and 43.2% had >10 PYR. Nodal status was as follows: N1, 11% (nZ4);

N2a, 5% (nZ2); N2b, 55% (nZ24); N2c, 25% (nZ11); and N3, 5%

(nZ2). Nodal volumes were measured on CBCTs from treatment days 1,

10, 20, and 35. Nodal decrease (ND) was determined based on percentage

of shrinkage from day 1 volume.

Results: At a median follow-up of 17 months (2-31), the 2-year actuarial

estimates of disease-free survival (DFS), local control (LC), regional

control (RC), distant metastasis (DM), and overall survival (OS) were

87%, 94%, 92%, 89%, and 92%, respectively. ND at days 10 or 35 did not

correlate with outcomes; however, ND at day 20 (D20) and smoking status

did. Pts with ND above the median decrease at D20 (40%) had improved 2-

year RC (100% vs 78.4%, PZ.03) and trended toward improved 2-year

DFS (95.5% vs 72.7%, PZ.06), LC (100% vs 85%, PZ.08), and OS

(100% vs 81.0%, PZ.11), compared to those with D20 ND <40%.

Similarly, smoking status <10 PYH correlated with improved 2-year RC

(100 vs 77%, PZ.02) and OS (100% vs 66%, PZ.003) with a trend to

improved 2-year DFS (100% vs 79%, PZ.08) and LC (100% vs 84%,

PZ.07) compared to those with >10 PYH. Among smokers, a D20ND

>40% versus <40% had 2-year RC of 100% versus 49.4% (PZ.04). Two-

year DM rates were higher in >10 PYH smokers (30% vs 0%, PZ.01) and

p16-negative pts (29% vs 4%, PZ.01) compared to <10 PYH and p16-

positive pts, respectively. Among p16-positive pts, D20ND >40% versus

<40% had 2-year LC and RC of 100% versus 77.9% (PZ.05). On Cox

MVAwith forward analysis evaluating D20ND and smoking, D20ND was

independently prognostic for LRC (PZ.039) and trended toward signifi-

cance for DFS (PZ.075).

Conclusion: ND of >40% by treatment day 20 is associated with improved

LRC, with a trend to improved DFS. Smoking history was strongly

associated with poorer RC and OS. Pts who were p16-negative and

smokers had higher rates of DM. The prognostic value for LRC of mid-

treatment ND can potentially be used to select pts for locoregional treat-

ment intensification strategies (eg, pts >10 PYH with D20ND <40%) and

those for treatment de-escalation (<10 PYH pts with D20ND >40%).

Author Disclosure: K.S. Hu: leadership; ASTRO. R. Stewart: None. A.

Jacobson: None. M. Persky: None. S. Schantz: None. T. Tran: None. M.

Urken: None. B. Culliney: None. Z. Li: None. L.B. Harrison: None.

176The Fate of Head and Neck Cancer Patients After (Chemo)radiationTherapydRisk Group Stratified Outcome Analysis and TreatmentEffect Modeling in the Presence of Competing RisksK. Hakansson,1 J.H. Rasmussen,2 L. Specht,1 S.M. Bentzen,3

and I. Vogelius1; 1Department of Oncology, Section of Radiotherapy,

Rigshospitalet, University of Copenhagen, Copenhagen, Denmark,2Department of Otorhinolaryngology, Head & Neck Surgery and

Audiology, Rigshospitalet, University of Copenhagen, Copenhagen,

Denmark, 3University of Maryland School of Medicine, Baltimore, MD

Purpose/Objective(s): Tumor control after (chemo)radiation therapy for

head and neck squamous cell carcinoma (HNSCC) is known to be affected

by a number of risk factors. However, the literature on the risk of

competing endpoints, potentially impacting optimal management, is

scarce. In this study, the treatment outcome has been analyzed while taking

competing events into account. The purpose was to assess the potential

benefit from intensified local treatment.

Materials/Methods: A cohort of 286 patients treated for HNSCC was

stratified by low, intermediate, and high risk, using a previously published

model. Persistent disease, locoregional relapse, distant metastasis (DM),

and death from other causes at 24 months after treatment were analyzed

using a competing risk model. The effect of an intensified radiation therapy

strategy (FDG-PET-based dose painting) was estimated while accounting

for competing risks, expanding a previously published TCP model. It was

assumed that the intensified treatment did not impact the competing events

of DM and death and that patients achieving locoregional control (LRC) as

a result of intensified treatment suffer the same risks of DM and death as

other patients of their risk group.

Results: Two years after treatment, 10% (confidence interval [CI]: 4.8-21),

43% (CI: 34-54), and 78% (CI: 71-85) of the high-, intermediate-, and low-

risk patients, respectively, were alive with no sign of tumor. The modes of

failure from the competing risk model are shown in Table 1 (confidence

interval data not shown); note that percentages add up to 100% in contrast to

Kaplan-Meier analysis. The estimated effect of treatment intensification

differed depending on risk group, both in magnitude and in the disease-free

survival (DFS)/LRC ratio,whichwas lowest for the high-risk group (Table 1).

Abstract 176; Table 1 Two-year outcome, analyzed using a competing riskmodel, and expected benefit from local treatment intensification.

Lowrisk

Intermediaterisk

Highrisk

Persistent disease 4.6% 10% 35%Locoregional relapse 7.6% 18% 27%Distant metastasis 3.4% 13% 11%Death from other causes 6.1% 16% 17%Alive (disease-free survival) 78% 43% 10.0%Expected benefit from local

treatment intensificationLRC (%-points) 7.5 15 23DFS (%-points) 6.8 10 6.7

International Journal of Radiation Oncology � Biology � Physics902

Conclusion: The outcome of high-risk patients is discouraging; in the

<40% of patients achieving LRC, >70% develop DM or die from other

causes within 24 months. The link between improvement in LRC and DFS

depends heavily on risk group, as the benefit is negatively impacted by the

risk of DM and death from other causes. Estimates of LRC without

competing risks can be misleading in assessing benefits of novel treatments

and when counseling patients.

Author Disclosure: K. Hakansson: None. J.H. Rasmussen: None. L.

Specht: Research Grant; Varian Medical Systems. Honoraria for teaching

courses held (to department); Varian Medical Systems. S.M. Bentzen:

None. I. Vogelius: Research Grant; Varian Medical Systems. Honoraria for

teaching courses held (to department); Varian Medical Systems.

177Clinical Predictors of Locoregional Failure in Advanced LaryngealCancer Treated With Definitive Chemotherapy and RadiationA.M. Khan,1 M.C. Ward,2 D.J. Adelstein,1 S. Koyfman,2 C.A. Reddy,1

P. Bhateja,1 P. Funchain,1 E. Lamarre,1 B.B. Burkey,1 M. Khan,1

J. Scharpf,1 B. Prendes,1 J.F. Greskovich, Jr,3 R. Lorenz,1 N.P. Joshi,2

M. Rahe,1 D. Ives,1 B. Harr,1 J. Bodmann,1 and T. Nwizu1;1Cleveland Clinic, Cleveland, OH, 2Cleveland Clinic Foundation,

Cleveland, OH, 3Cleveland Clinic Florida, Weston, FL

Purpose/Objective(s): Definitive chemoradiation (CRT) has become an

established organ preservation treatment for patients with locoregionally

advanced laryngeal cancer. Although most patients will experience long-

term disease control after CRT, locoregional failure remains a concern for

a subset of patients. This study is aimed to identify predictors of locore-

gional failure after larynx-preservation therapy in patients with stage III-IV

laryngeal cancer.

Materials/Methods: From an institutional review boardeapproved tumor

registry, all patients with American Joint Committee on Cancer stage III-

IVB squamous cell carcinoma of the larynx treated with definitive con-

current chemoradiation therapy between 1993 and 2014 were identified.

Patient, tumor, and treatment characteristics were collected. The Cox

proportional hazards method was used to identify predictors of locore-

gional failure. Statistical significance was inferred at the 0.05 level for all

hypothesis testing.

Results: A total of 117 patients were included in the study with a median

follow-up of 31 months (range: 1.8e 185). The median age at diagnosis

was 59 years (range: 43- 77), 74% were male, 57% were current smokers,

and 21% had a history of alcohol abuse. Fourteen percent of patients

presented with T2 tumors, 70% with T3 tumors, and 16% had T4 disease.

Sixty-two percent were node-positive; 23% with N1 disease, 4% with N2a,

14% with N2b, 21% with N2c, and 1% with N3 disease. Twenty-two

percent were glottis, and 67% were supraglottic. Radiation therapy (RT)

was administered once (68%) or twice daily (31%) to a total dose of 70 to

74.4 Gy, with either a 3-field approach (3D-RT) in 61% or, more recently,

using intensity modulated radiation therapy (IMRT) in 37%. Chemo-

therapy consisted of either cisplatin and 5-fluorouracil (60%) or single-

agent cisplatin (33%) at standard dosing. Twenty (17%) locoregional

failures were observed. Univariate analysis found a significantly higher

locoregional failure rate in those patients with both significant alcohol and

�10 pack-year smoking history who were currently smoking at the time of

diagnosis (hazard ratioZ3.66, 95% confidence interval 1.39 - 9.71,

PZ.009). T stage, N stage, chemotherapy regimen, or radiation therapy

dose or fractionation were not associated with locoregional failure.

Furthermore, 85% of the locoregional failures occurred within 2 years,

with a median time to locoregional failure of 16 months.

Conclusion: Overall, these data suggest that patients treated with definitiveCRT for stage III-IVB laryngeal cancer who fail locoregionally do so

within 16 months. Heavy alcohol history combined with significant

smoking history and continued smoking at time of treatment was a sig-

nificant predictor of failure. Emphasis on smoking cessation and alcohol

reform is warranted.

Author Disclosure: A.M. Khan: None.M.C. Ward: None. D.J. Adelstein:

None. S. Koyfman: None. C.A. Reddy: None. P. Bhateja: None. P.

Funchain: None. E. Lamarre: None. B.B. Burkey: None. M. Khan:

None. J. Scharpf: None. B. Prendes: None. J.F. Greskovich: None. R.

Lorenz: None. N.P. Joshi: None.M. Rahe: None. D. Ives: None. B. Harr:

None. J. Bodmann: None. T. Nwizu: None.

178Tumor Board Checklists Affect Pretreatment Clinic ReferralPatternsW. Swegal and S. Chang; Henry Ford Health System, Detroit, MI

Purpose/Objective(s): Checklists have been utilized within the surgical

community in recent years as a way to ensure that certain crucial steps or

processes are completed. They are a process-related effort to ensure quality

patient care. Patients with newly diagnosed head and neck cancer require

multidisciplinary preoperative evaluation and counselling prior to their

treatment, and postsurgical patients require proper review for referrals for

adjuvant treatments. A checklist system was implemented at our tumor

board during 2013 in order to improve quality of care and adherence to

National Comprehensive Cancer Network guidelines.

Materials/Methods: We conducted a retrospective analysis of newly

diagnosed head and neck cancer patients presented at our institution’s

multidisciplinary tumor board between the years of 2010 and 2015. The

year 2013 was considered the point when we started to use checklists

during the tumor board discussion. We compared 100 newly diagnosed

patients before the checklist was implemented to 100 patients afterward.

Compliance with tumor board recommendations and pretreatment evalu-

ation within 1 month were compared between groups. Pretreatment eval-

uation included appointments with a clinical psychologist, speech

language pathologist, and nutritionist. Analysis was also performed on

referral for audiogram and dental evaluation, as well as whether patients

were reviewed for clinical trials.

Results: Preliminary analysis suggests that appropriate and completed

referrals to medical oncology were similar between the 2 groups (P�.99).This was similar to radiation oncology, where 93% patients in the pre-

checklist era and 100% patients from the postchecklist group were

correctly referred or not referred (PZ0.3). Difference in referral patterns

to pretreatment evaluation and counseling trended toward significant.

Referral patterns for speech language pathology (PZ.19), clinical psy-

chology (PZ.25), and nutrition (PZ.06) all trended toward a significant

difference with overall more appropriate referrals occurring after the

checklist was implemented. This was a similar trend with optimization

clinic (PZ.23) and audiology (PZ.06). Appropriate referrals to dental

(PZ.006) were significantly increased with the implementation of the

checklist, and a great proportion of patients were reviewed for clinical

trials (PZ.002).

Conclusion: The use of the checklist during the tumor board presentation

helped to increase follow through and adherence to recommendations for

pretreatment evaluation. We hope to further highlight the benefits of the

Volume 94 � Number 4 � 2016 Posters 903

tumor board checklist with review of all newly diagnosed head and neck

patients prior to and after implementation.

Author Disclosure: W. Swegal: None. S. Chang: None.

179Dramatic Reduction in the Need for Feeding Tube Use in HumanPappilomavirusePositive Oropharyngeal Cancer in the IntensityModulated Radiation Therapy EraS. Koyfman,1 M.C. Ward,2 N. Houston,2 N.P. Joshi,1 B. Harr,2 T. Nwizu,2

D.J. Adelstein,2 P. Xia,1 and J.F. Greskovich, Jr3; 1Cleveland Clinic

Foundation, Cleveland, OH, 2Cleveland Clinic, Cleveland, OH,3Cleveland Clinic Florida, Weston, FL

Purpose/Objective(s): Human papillomavirus (HPV)-positive oropha-

ryngeal cancer patients have improved cancer-related outcomes and appear

to tolerate treatment with less morbidity. We hypothesize that feeding tube

rates are decreasing selectively in this population.

Materials/Methods: Patients treated with definitive chemoradiation

(CRT; >66 Gy) for head and neck squamous cell cancer between 2007

and 2014 were included in this institutional review obardeapproved

study. Patients were treated with 3-dimensional CRT in the earlier years

of the study and intensity modulated radiation therapy (IMRT) since

2009. All patients were treated with concurrent chemotherapy, typically

cisplatin based (90%), or with cetuximab (10%). At our institution,

prophylactic feeding tubes (FTs) are typically avoided, and reactive FTs

are used if patients lose excessive amounts of weight or experience

dehydration requiring admission. Analysis of Variance trend test was

performed to identify the significance of temporal trends in the use of

FTs over time.

Results: Of 421 patients included in this study, 194 (46%) had an FT

placed. Of these, corpaks were used in 84%, while percutaneous endo-

scopic gastrostomy (PEG) tubes were used in the remainder. HPV-

positive oropharyngeal cancers comprised 280 pts (66%), while the

remainder included pts with cancers of the larynx (15%), HPV-negative

oropharynx (11%), hypopharynx (4.5%), and nasopharynx (2.5%).

Median age was 58 years, 84% were male, and 76% had a smoking

history (median pack-years, 27). Median dose of RT was 72 Gy (68-78

Gy). Rates of FT use steadily declined for patients with HPV-positive

oropharynx cancer, with a highly significant decrease from 80% in 2007

to 24% in 2014 (R2Z0.931; PZ.0001). This pattern was not seen for all

other (HPV-negative) patients (see Table 1). Median duration of FT use

was 60 days and did not differ by HPV status. Overall, only 40 pts (10%)

had an FT beyond 90 days following IMRT, and only 15 patients (3.6%)

had an FT at last follow-up.

Conclusion: FT use has become the exception rather than the rule in pa-

tients with HPV-positive oropharyngeal cancer undergoing definitive CRT.

Routine FT use should be avoided in this population.

Author Disclosure: S. Koyfman: None. M.C. Ward: None. N. Houston:

None. N.P. Joshi: None. B. Harr: None. T. Nwizu: None. D.J. Adelstein:

None. P. Xia: None. J.F. Greskovich: None.

Abstract 179; Table 1

Year HPV+ oropharynx feeding tube All other feeding tube

2007 80% 34.37%2008 59.25% 24.24%2009 53.33% 32.43%2010 54% 52.94%2011 29.26% 61.53%2012 41.86% 46.66%2013 25% 38.46%2014 23.91% 21.87%

180Proliferation Saturation Index Predicts Oropharyngeal SquamousCell Cancer Gross Tumor Volume Reduction to ProspectivelyIdentify Patients for Adaptive Radiation TherapyT. Lewin,1 J. Kim,2 K. Latifi,3 J. Poleszczuk,4 J. Bull,1 H. Byrne,1

J.F. Torres-Roca,3 E.G. Moros,3 R. Gatenby,2 L.B. Harrison,3

J. Heukelom,5 A.S.R. Mohamed,6 D.I. Rosenthal,6 C.D. Fuller,6

J.J. Caudell,3 and H. Enderling4; 1University of Oxford, Oxford, United

Kingdom, 2H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL,3H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL,4H. Lee Moffitt Cancer Center, Tampa, FL, 5Netherlands Cancer Institute,

Amsterdam, Netherlands, 6The University of Texas MD Anderson Cancer

Center, Houston, TX

Purpose/Objective(s): Radiation therapy (RT) for human papillomavirus

(HPV)-positive oropharyngeal squamous cell cancer (OPX) provides 3-

year locoregional control (LRC) rates of 75% to 95%. Given the excellent

cancer control outcomes for OPX following RT, there is significant interest

in reducing therapy intensity, such as adaptive radiation therapy (ART), for

rapidly shrinking tumor volumes to improve organ at risk sparing. How-

ever, we have limited understanding of which patients will be suitable for

ART. Proliferation saturation index (PSI) is defined as a ratio of tumor

volume to host tissue carrying capacity. Tissue carrying capacity (K) is an

integral measure of the maximum tumor volume that can be supported by

the current tumor environment including oxygen and nutrient availability,

immune surveillance, and acidity. The PSI can be measured from 2

pretreatment computed tomographic (CT) scans separated in time. We

hypothesized that PSI may be able to define a subgroup of patients with

rapidly shrinking tumors during RT who may gain the most benefit

from ART.

Materials/Methods: From an MDACC institutional database, 9 patients

with OPX treated with RT and concurrent chemotherapy were identified.

Patient demographics, treatment, and outcomes were extracted. PSI was

estimated from routine pre-RT and RT planning simulation computer to-

mography scans and the Gompertz tumor growth model. Gross tumor

volume (GTV) was contoured on weekly CT scans obtained during

treatment to estimate tumor shrinkage. GTV reductions during fractionated

RT were calculated as a function of PSI.

Results: Gompertz tumor growth with patient-specific PSI values fit

retrospective data with high confidence (R2>0.93). Median PSI was 0.1

(range 0.01-0.3). Median GTV was 23.6 cm3 (range 3.5-62). Of the 9

patients, 3 had GTV reductions of >20% within the first 3 weeks. This was

correlated with PSI values less than 0.07.

Conclusion: PSI derived from standard of care CT images may be able to

prospectively identify candidates for ART prior to treatment.

Author Disclosure: T. Lewin: None. J. Kim: None. K. Latifi: None. J.

Poleszczuk: None. J. Bull: None. H. Byrne: None. J.F. Torres-Roca:

None. E.G. Moros: None. R. Gatenby: None. L.B. Harrison: None. J.

Heukelom: None. A.S. Mohamed: None. D.I. Rosenthal: None. C.D.

Fuller: None. J.J. Caudell: None. H. Enderling: None.

181The Effect of Smoking and Human Papillomavirus Status onSurvival in Oropharyngeal Squamous Cell Carcinoma PatientsUndergoing Concurrent ChemoradiationA. Platek,1 V. Jakaprakash,1 M. Platek,2 V. Gupta,1 D. Cohan,1 W. Hicks,1

T. Winslow,3 A. Groman,4 S.S. Dibaj,1 H. Arshad,1 M.A. Kuriakose,1

G.W. Warren,5 and A.K. Singh2; 1Roswell Park Cancer Institute, Buffalo,

NY, 2Roswell Park Cancer Institute, Buffalo, NY, United States, 3Roswell

Park Cancer Insitute, Buffalo, NY, 4Roswell Park Cancer Institue, Buffalo,

NY, 5Medical University of South Carolina, Charleston, SC

Purpose/Objective(s): The effect of smoking and human papillomavirus

(HPV) status on the survival of oropharyngeal squamous cell carcinoma

International Journal of Radiation Oncology � Biology � Physics904

(OPSCC) undergoing definitive concurrent chemoradiation (CCRT)

remains unclear. The purpose of this review was to examine these effects

on survival outcomes among a single institute population.

Materials/Methods: This retrospective review of OPSCC patients treated

with CCRT between 2008 and 2015 was conducted. All tumors were

examined for HPV 16/18 status (+/). Smoking status and other clinical

characteristics were abstracted from the electronic medical record. Former

smokers are patients who quit within a month of diagnosis or treatment.

Descriptive summaries, overall survival (OS), and multivariate cox pro-

portional hazard ratios (HR) were completed.

Results: Out of 134 patients, 94 patients (71%) had HPV-positive (HPV+)

tumors. Age, gender, and overall stage were not statistically different be-

tween patients with HPV+ or HPV-negative (HPV-) tumors. HPV+ patients

had higher tumor grade (P<.01). Patients with HPV+ tumors had a higher

percentage of never smokers than patients with HPV- tumors (10.3% vs

26.6%). Median pack-years were 18 for the HPV+ group versus 30 for the

HPV- group. OS did not differ for HPV+ versus HPV- patients. Within

HPV+ patients, current and former smokers had significantly worse OS

than never smokers (P<.01). The same was true for HPV- patients, but the

log-rank test for this group did not reach statistical significance (PZ.06).

The 3-year survival rate for former smokers in each group was similar

(HPV+: 0.77 (0.60, 0.88) and HPV-: 0.75 (0.52, 0.88). Hazard ratios for OS

among current smokers compared to never/former smokers in each patient

group were statistically significantly higher.

Conclusion: Current smoking is associated with poor prognosis, inde-

pendent of HPV status in OPSCC patients treated with CCRT. Former

smokers have similar outcomes irrespective of HPV tumor status. The

amount of tobacco a patient is exposed to before diagnosis cannot be

altered, but every effort should be made to get patients to quit smoking as

soon as possible before CCRT regardless of HPV status.

Author Disclosure: A. Platek: None. V. Jakaprakash: None. M. Platek:

None. V. Gupta: None. D. Cohan: None. W. Hicks: None. T. Winslow:

None. A. Groman: None. S.S. Dibaj: None. H. Arshad: None. M.A.

Kuriakose: None. G.W. Warren: None. A.K. Singh: None.

182Positive Effect of Surgery Regardless of Stage on OropharynxSubsites Base of Tongue and Tonsillar-Fossa: A SEER AnalysisA. Platek,1 V. Jayaprakash,2 M. Platek,2 V. Gupta,1 D. Cohan,1 W. Hicks,1

T. Winslow,3 A. Groman,4 S.S. Dibaj,1 H. Arshad,1 M.A. Kuriakose,1

G.W. Warren,5 and A.K. Singh2; 1Roswell Park Cancer Institute, Buffalo,

NY, 2Roswell Park Cancer Institute, Buffalo, NY, United States,3Roswell Park Cancer Insitute, Buffalo, NY, 4Roswell Park Cancer Institue,

Buffalo, NY, 5Medical University of South Carolina, Charleston, SC

Purpose/Objective(s): There is increasing interest in the use of robotic

surgery in carcinoma of the oropharynx. The purpose of this study was to

examine survival outcomes among 2 oropharyngeal subsites (tonsillar-

fossa [TF] and base of tongue [BOT]).

Materials/Methods: We conducted a retrospective cohort analysis utiliz-

ing data from the Surveillance, Epidemiology, and End Results (SEER)

Program. The SEER cohort included 8073 primary BOT and TF SCC

patients without distant metastases treated between 2004 and 2011. Pri-

mary outcome measures were subsite-based differences in overall survival

(OS) and disease-specific survival (DSS) for TF and BOT patients stratified

by overall stage and comparing treatment method for each subsite. Cox

proportional hazard ratios were estimated for each group.

Results: For all stages combined, both BOT and TF patients who received

surgery with radiation had superior OS (P<.01). The same was true when

analyses were stratified by stage within each subsite. Multivariate hazard

ratios adjusted for age, gender, race, and tumor grade for OS were sta-

tistically significantly higher for both BOT and TF patients who did not

receive surgery compared to those who did receive surgery for each stage.

Conclusion: In this SEER cohort, OS was superior in both BOT and TF

patients who received surgery with adjuvant radiation. OPSCC survival

may be improved by treating more BOT and TF patients with surgery and

adjuvant radiation. As modern, less invasive surgical techniques such as

transoral robotic surgery gain wider acceptance, approaches that combine

surgery and radiation (with or without chemotherapy) while minimizing

morbidity and lack of function should be attempted.

Author Disclosure: A. Platek: None. V. Jayaprakash: None. M. Platek:

None. V. Gupta: None. D. Cohan: None. W. Hicks: None. T. Winslow:

None. A. Groman: None. S.S. Dibaj: None. H. Arshad: None. M.A.

Kuriakose: None. G.W. Warren: None. A.K. Singh: None.

183Organ Preservation in Patients With Orbit-Invasive SinonasalCancerM.B. Yusuf,1 M.J. Amsbaugh,2 C.L. Silverman,2 J. Bumpous,2

C.A. Perez,2 K. Potts,2 P. Tennant,2 R. Redman,2 M.H. Bertke,2

and N.E. Dunlap2; 1University of Louisville School of Medicine, Louisville,

KY, 2University of Louisville, Louisville, KY

Purpose/Objective(s): The optimal treatment strategy for patients with

orbital encroachment of sinonasal cancer is not known. We sought to

determine whether organ preservation with neoadjuvant therapy was

possible in patients who were determined to require an exenteration for

locally advanced sinonasal cancer.

Materials/Methods: All patients presenting to our institutional head and

neck cancer clinic from 2005 to 2014 were reviewed from a prospective

database. Patient characteristics, disease, treatment, and toxicity data were

collected. Kaplan-Meier methods were used to estimate survival. Log-rank

test and cox proportional hazards modeling were used to examine pre-

dictive factors.

Results: Twenty patients were determined by our multidisciplinary team to

require an orbital exenteration as part of definitive treatment of their

sinonasal cancer. Fourteen patients underwent orbital preservation (OP).

Neoadjuvant treatment consisted of concurrent chemoradiation (CRT) in

all patients aside from 2 who refused chemotherapy and underwent radi-

ation therapy (RT) alone. Six patients received primary surgery followed

by adjuvant CRT. Treatment groups were well balanced with respect to

age, sex, site, stage, and histology. Five patients (36%) receiving neo-

adjuvant therapy had complete disease response at time of surgery. With a

median time of follow-up of 18.8 months, actuarial exenteration-free

survival was 62% at 2 years for patients undergoing OP. No patients

receiving neoadjuvant therapy required an immediate posttreatment orbital

exenteration; however, 1 patient required salvage orbital exenteration when

she experienced local failure 35 months after treatment. At 2 years, there

were no significant differences in local-regional control (75% vs 60%,

PZ.997), progression-free survival (40% vs 36%, PZ.493), or overall

survival (40% vs 58%, PZ.815) between patients receiving OP or upfront

surgery (PZ.815). On multivariate analysis, tumor site predicted for local-

regional control (maxillary: 86%, nasal cavity: 80%, ethmoid sinus: 40%,

PZ.050). Patients with nasal cavity tumors demonstrated a trend toward

improved OS with a hazard ratio of 0.240 (95% confidence interval .053 -

1.08). Treatment was well tolerated in both groups. One patient treated

with OP experienced grade 3 radiation-related toxicity, developing

dysphagia which improved with medical management. However, 1 patient

developed renal failure following his first dose of cisplatin and subse-

quently died before completion of treatment.

Conclusion: For patients with orbital encroachment of sinonasal cancer,

neoadjuvant CRT can preserve the orbit while offering similar survival and

disease control compared to upfront orbital exenteration.

Author Disclosure: M.B. Yusuf: None. M.J. Amsbaugh: None. C.L.

Silverman: None. J. Bumpous: None. C.A. Perez: None. K. Potts: None.

P. Tennant: None. R. Redman: None. M.H. Bertke: None. N.E. Dunlap:

None.

Volume 94 � Number 4 � 2016 Posters 905

184Magnitude and Timing of Gross Tumor Volume Response toNeoadjuvant Chemotherapy and Concurrent Chemoradiation in theTreatment of Locally Advanced Nasopharyngeal CarcinomaJ.A. Giambattista,1 N. McVicar,1 M. Martin,1 C. Ho,1 B. Maas,1 J. Hay,1

J. Wu,1 M. Keyes,2 and E. Berthelet1; 1BC Cancer Agency, Vancouver, BC,

Canada, 2University of British Columbia, Vancouver, BC, Canada

Purpose/Objective(s): Curative radiation therapy (RT) for locally

advanced nasopharyngeal carcinoma (NPC) is based on the gross tumor

volume (GTV), but the magnitude and timing of GTV changes during

combined modality therapy remain unclear. This study analyzes GTV

changes at phases of induction chemotherapy and sequential concurrent

chemoradiation therapy (CRT) in patients with locally advanced NPC.

Materials/Methods: Subjects included 13 patients with newly diagnosed

stage III-IV NPC who underwent treatment between 2011 and 2014.

Criteria for eligibility included 2 cycles of neoadjuvant chemotherapy, at

least 5 cycles of concurrent chemotherapy and 3 magnetic resonance im-

aging (MRI) scans at specific phases of treatment (T0: before treatment,

T1: postinduction, and T3: 3 months after CRT). The induction phase

consisted of 2 cycles of gemcitabine and cisplatin. The CRT phase con-

sisted of weekly cisplatin and RT delivered using volumetric modulated

arc therapy (VMAT). The total dose was 70 Gy over 35 daily fractions

administered 5 days/week. A subset of 3 patients received an additional

MRI 4 to 5 weeks into CRT (T2). Primary gross tumor volume (GTVp)

was defined as the GTV and adjacent involved retropharyngeal lymph

nodes. Tumor volumes were delineated on gadolinium-enhanced fat-

saturation T1 weighted MRIs by 2 observers. Mean values are reported +/-

one standard deviation.

Results: Preliminary analysis included 6 (out of 13) subjects. The mean

initial GTVp was 62.7 +/- 32.8 mL. The mean GTVp response after in-

duction phase was 21.4% +/- 12.3% with a mean rate of volume change of

0.31 +/- 0.19 mL/day which corresponded to a 0.56% +/- 0.35% daily

reduction in tumor volume. The total mean GTVp response after

completion of treatment (T3) was 77.6% � 21.6%. Subgroup analysis of

subjects who underwent an additional MRI showed a mean GTVp

reduction of 42.5% � 22.6% and a mean rate of volume change of 0.87 �0.08 mL/day which corresponded to a 1.7% � 0.93% daily reduction in

tumor volume (from T1 to T2).

Conclusion: Preliminary results suggest that the GTVp progressively

diminishes following both induction chemotherapy and CRT. The mean

GTVp response after 4 to 5 weeks of CRT exceeded the response observed

after induction chemotherapy by a factor of 2. The mean rate of volume

change at 4 to 5 weeks of CRTwas threefold the rate seen during induction

chemotherapy. These observations may support the optimal timing of

imaging for replanning in the context of adaptive field RT. Analysis of the

full NPC patient dataset is ongoing and will be reported.

Author Disclosure: J.A. Giambattista: None. N. McVicar: None. M.

Martin: None. C. Ho: None. B. Maas: None. J. Hay: None. J. Wu: None.

M. Keyes: None. E. Berthelet: None.

185C-Reactive Protein as a Biomarker of Radiation Therapy andChemotherapy Toxicity Monitoring in Patients With Head and NeckCancerA. Wygoda, J.J. Mrochem-Kwarciak, M. Kentnowski, T. Rutkowski,

B. Pilecki, A. Heyda, A. Hajduk, U. Dworzecka, I.J. Gawron,

K. Grabinska, L.P. Michalecki, P. Polanowski, and K.A. Skladowski;

Maria Sklodowska-Curie Memorial Cancer Center and Institute of

Oncology, Gliwice, Poland

Purpose/Objective(s): C-reactive protein (CRP) is a serum protein

elevated in a variety of illnesses, including cancer. Also some data exist

that CRP level is correlated with acute mucositis in patients (pts) with head

and neck cancer (HNC) treated with radiation therapy (RT) and

significantly elevated when dose/fraction number is increased. The purpose

of this study is to show how CRP behaves before and after HNC RTwith or

without chemotherapy.

Materials/Methods: From 2010 to 2015, 401 pts (318 men and 83 women)

with a mean age of 58 years (range 20-81) with oral cavity (5),

nasopharynx (30), oropharynx (133), hypopharynx (47), larynx (180), or

unknown primary (6) cancer were under treatment. Tumor stage was

evaluated in all pts: T0-6, T1-60, T2-147, T3-119, T4-69, and N0-174, N1-

52, N2-143, and N3-32. Four schedules of treatment have been used: RT

alone (RTA, 178 pts), concurrent chemoradiation therapy (CRC, 93 pts),

and induction chemotherapy (IC) followed by RTA (46 pts) or CRC (84

pts). CRP (normal level, <2.87mg/L) was measured as follows: before IC

(if applied) or before RTA/CRC, every week during RTA/CRC, and,

finally, 1 month after treatment. Total number of CRP measurements was

3622.

Results: Initial CRP level was normal in 56% of pts; in 44% it was

elevated (median 6.05 mg/L) but not correlated with tumor advancement.

After IC, normalization of CRP was observed in 73% of pts (median CRP

decreased from 3.12 to 1.24 mg/L). During RT progressive, week by week

up to week 6, elevation of CRP was observed for RTA and CRC with no

difference between both treatments. However, evident impacts of initial

CRP level and IC were noted; in patients who had abnormal initial CRP

and/or IC, significantly higher CRP levels were observed over RTA and

CRC. There was no correlation observed between CRP and total dose or

overall volume of radiated tissue.

Conclusion: CRP may be a useful laboratory parameter in monitoring of

toxicity in HNC patients undergoing RTA and CRC. IC leads to significant

reduction/normalization of CRP in the majority of patients, which may

reflect an anti-inflammatory effect in tumor burden. Similar CRP levels in

patients treated with RTA and CRC may suggest that tumor and normal

tissue radiation sensitivity is an essential proinflammatory factor.

Author Disclosure: A. Wygoda: None. J.J. Mrochem-Kwarciak: None.

M. Kentnowski: None. T. Rutkowski: None. B. Pilecki: None. A. Heyda:

None. A. Hajduk: None. U. Dworzecka: None. I. Gawron: None. K.

Grabinska: None. L.P. Michalecki: None. P. Polanowski: None. K.A.

Skladowski: None.

186Adaptive Radiation Therapy in Adults With NasopharyngealCarcinoma: A Retrospective Study at 1 InstitutionN. Alsafadi,1 M. Khan,2 M. Abrar,2 F. Saeedi,2 Y. Albarakati,2

A. Alshaikh,2 M. Algarni,2 B. AL Ahmad,2 C. Chantel,2 S. Jaber,2

and E. Noor2; 1Radiation Oncology. King Abdulaziz Medical City PNOC,

Jeddah, Saudi Arabia, 2PNOC-KAMC, JEDDAH, Saudi Arabia

Purpose/Objective(s): To determine whether nasopharyngeal carcinoma

(NPC) patients benefit from adaptive radiation therapy (ART) in terms of

tumor coverage and/or organ at risk (OAR) sparing. We also aim at

determining patients with higher risk for plan deterioration according to

potential predictive factors.

Materials/Methods: According to our institution protocol, all NPC pa-

tients treated radically with or without concurrent chemotherapy receive

adaptive radiation therapy (ART). We reviewed the planning parameters of

40 such patients treated from January 2013 to December 2014. Since our

protocol calls for induction chemotherapy in locally advanced disease,

none of our patients had bulky neck nodes (more than 6 cm) at the start of

radiation therapy. All patients were treated using volumetric modulated arc

therapy (VMAT) with simultaneous integrated boost with 2 dose levels, 70

Gy and 60 Gy, all in 35 fractions. Patients were replanned systematically

twice during radiation therapy on days 15 and 29. The resulting ART plans

were compared to the non-ART (NART) plans generated retrospectively

for the purpose of this study only. This review focuses on target coverage

and OARs doses. Three potential predictive factors for plan deterioration

were analyzed. These include the maximum positional shift (MPS) on day

15, neck separation reduction (NSR) on day 15, and percentage weight loss

(PWL) on day 21.

International Journal of Radiation Oncology � Biology � Physics906

Results: For the entire series, ART plans had significantly better planning

target volume (PTV) coverage. The mean percentage dose covering 95%

of PTV (D95) was 96.44% for ART versus 95.58% for NART (PZ.0003)

for PTV70, and 94.63% for ART versus 93.94% for NART (PZ.0009) for

PTV60. Using our treatment setup, only MPS was able to define 2 risk

groups for plan deterioration. The high-risk group (HRG) comprises 27

patients with MPS>1 mm and the low-risk group (LRG) comprises 13

patients with MPS�1 mm. The mean PTV70 (D95) for ART versus N-ART

for the HRG and LRG were 96.44% versus 95.36% (PZ.0001), and

96.15% versus 96.02% (PZ.686). The same was true for PTV60 with

94.93% versus 93.36% (PZ.0033), and 94.62% versus 94.00% (PZ.104).

When compared to ART, 6 out of 40 NART plans had a PTV70 and/or

PTV60 (D95) drop exceeding 3% (PZ.0255), 5 of them in the HRG. The

ART plans achieved a significantly lower spinal cord dose maximum

(SCMD) of 42.64 Gy versus 44.47 Gy (PZ.0001). Only 1 patient in this

group had SCMD exceeding 46 Gy as compared to 13 in NART plans

(PZ.0012). The improvement in the mean parotid dose with ART in this

series did not reach statistical significance (33.35 Gy versus 34.20 Gy,

PZ.063).

Conclusion: ART plans resulted in improved tumor coverage and better

spinal cord sparing compared to the corresponding NART plans in the

present NPC patient series. Further evaluation of the value of the predictive

factors will be carried out in a larger sample of patients.

Author Disclosure: N. Alsafadi: None. M. Khan: None. M. Abrar: None.

F. Saeedi: None. Y. Albarakati: None. A. Alshaikh: None. M. Algarni:

None. B. AL Ahmad: None. C. Chantel: None. S. Jaber: None. E. Noor:

None.

187Trends in Head and Neck Squamous Cell Carcinoma (HNSCC) ClinicalResearch as Reported in the American Society of Clinical Oncology(ASCO) Proceedings From 1996 to 2015J. Moreira,1 M. Agulnik,1 and A. Rademaker2; 1Northwestern University

Feinberg School of Medicine, Division of Hematology/Oncology, Chicago,

IL, 2Northwestern University Feinberg School of Medicine, Division of

Preventive Medicine-Biostatistics, Chicago, IL

Purpose/Objective(s): The direction of clinical research can be ascer-

tained by reviewing ASCO Proceedings abstracts. We aimed to identify

trends in HNSCC as reported in the ASCO annual meetings from 1996 to

2015 to determine whether clinical research findings lead to an improve-

ment in patient survival.

Materials/Methods: All abstracts in the head and neck section of the

ASCO Proceedings from 1996 to 2015 were reviewed. Abstracts on locally

advanced, recurrent, or metastatic HNSCC were further explored.

Descriptive summary information was recorded regarding number of au-

thors, randomization, trial phase, presence of novel drug or combination

therapies, timing of chemotherapy, and disease-free and overall survival.

Results: From 1996 to 2015, there were a total of 2294 head and neck

cancer abstracts. From 1996 to 2006, 207 abstracts were presented at the

ASCO annual meeting. From 2007 to 2015, 158 abstracts were presented,

of which 59 were presented at oral abstract or poster discussion sessions.

The average number of authors from 1996 to 2007 was 8.82, and 10.77

from 2007 to 2015, with an average number of authors from 1996 to 2015

of 9.66. Sixty percent of studies from 1996 to 2006 focused on locally

advanced disease, compared to 62% from 2007 to 2015. From 1996 to

2006, 27% were randomized studies versus 41% from 2007 to 2015. From

1996 to 2006, 57% of abstracts explored concurrent chemoradiation

therapy, versus 55% of abstracts from 2006 to 2015. Fifty percent of ab-

stracts from 1996 to 2006 investigated a novel drug or novel drug com-

bination, compared to 95% of abstracts from 2007 to 2015. Six percent of

abstracts from 1996 to 2006 demonstrated a statistically significant

improvement in survival compared to 7% of abstracts from 2007 to 2015.

Amongst those abstracts from 1996 to 2006, all were of patients with

locally advanced disease, with the majority (7 of 12) showing a survival

benefit of radiation administered with chemotherapy or cetuximab.

Amongst those from 2007 to 2015, 9 or 11 were in patients with locally

advanced disease. Three studies demonstrated the survival benefit of in-

duction chemotherapy, 2 studies demonstrated the benefit of TPF versus

PF, and 3 studies demonstrated the survival benefit of novel monoclonal

antibodies (ie, nimotuzumab).

Conclusion: Since 1996, the number of clinical trials of HNSCC presented

at the ASCO annual meeting has increased. There has been a marked

increase in the exploration of either novel drugs or novel drug combina-

tions. This has not translated into statistically significant improvements in

overall survival because most trials were nonrandomized, phase 2 studies.

More randomized clinical trials are needed to build on the successes seen

in the treatment of HNSCC. Novel drugs and/or drug combinations will

likely lead to paradigm shifts in the treatment of human papillomavir-

usepositive HNSCC.

Author Disclosure: J. Moreira: None.M. Agulnik: None. A. Rademaker:

None.

188The Effect of Body Mass Index on Outcomes of CervicalLymphadenectomyP. DeVries, C. Nocon, A. Wieland, G.K. Hartig, and T.M. McCulloch;

University of Wisconsin, Madison, WI

Purpose/Objective(s): To determine which factors, including body mass

index (BMI), affect the outcomes of neck dissections as they relate to

lymph node yield and postoperative complications.

Materials/Methods: A retrospective chart review was performed on 354

subjects who had undergone a neck dissection between 1992 and 2010 at

an academic medical center. BMI, type of neck dissection, lymph node

yield, demographic information, comorbidities, cancer histology, and

history of radiation were identified. Univariate analyses of variances and

independent t tests were performed to determine the effects of BMI on the

lymph node yield and postoperative complications of neck dissections.

Results: There was an association between increased BMI and increased

lymph node yields after controlling for history of radiation (PZ.001).

Patients who were categorized as obese (BMI�30) had a significantly

higher lymph node yield compared to those with a BMI less than 30

(PZ.005). Obese and nonobese patients averaged 8.44 and 6.95 nodes per

neck level dissected, respectively. There were no associations between

postoperative complications and an increased BMI. The average number of

lymph nodes per neck level dissected for patients with N1, N2a, N2b, N2c,

and N3 staged cancer were 8.52, 7.81, 7.21, 6.56, and 5.34, respectively.

Conclusion: These results show an increase in lymph node yield in patients

with elevated BMI, no history of radiation, and an earlier N-staged cancer.

There is no consensus on what constitutes an adequate lymph node yield in

neck dissections. This study demonstrates that patient- and disease-specific

features may impact the lymph node yield.

Author Disclosure: P. DeVries: None. C. Nocon: None. A. Wieland:

None. G.K. Hartig: None. T.M. McCulloch: None.

189Nodal Volume as a Prognostic Factor in Locally Advanced Head andNeck Cancer Treated With Concurrent Chemoradiaton With IGRT: AnUpdated Analysis From an Indian InstitutionB. Dua,1 K.S. Chufal,2 and G. Jadhav1; 1APOLLO HOSPITAL, New delhi,

India, 2Batra Hospital and Medical Research Centre, delhi, India

Purpose/Objective(s): Nodal volume as a prognostic factor has been

extensively evaluated in head and neck cancer; however, there is still no

consensus. While the older studies have used 3-dimensional techniques,

outdated volume calculation methods are retrospective, and not all patients

have consistently received chemotherapy, the newer studies done with

intensity modulated radiation therapy (IMRT) have mostly subjected pa-

tients to an elective neck dissection following chemoradiation (CRT),

confounding the association between nodal volume and prognosis. We

attempted to analyze nodal volume as a prognostic factor in head and neck

Volume 94 � Number 4 � 2016 Posters 907

cancer treated with CRTwithout an elective neck dissection and to the best

of our knowledge, no similar study with IMRT exists in the published

literature.

Materials/Methods: We enrolled 87 patients of stage III-IV cancer of the

oropharynx (57), and hypopharynx (30), who subsequently received

definitive concurrent CRT with IMRT. A lymph node was considered

pathological as per the the Van der Berkel criteria, and the total nodal

volume (TNV) was the sum of all lymph node volumes calculated by

volume algorithm from the planning computed tomographic scan. The

impact of TNV on locoregional relapse-free survival (LRFS), response

(RR), overall survival (OS), local control (LC), and regional control (RC)

was assesed. Survival analysis was by Kaplan-Meier method with log-rank

testing for assessing significance between groups. Univariate analysis was

done by Mann-Whitney/ X2 test, multivariate analysis was done by logistic

regression forward stepwise method, and a model to predict OS and RC

was generated. An ROC curve analysis was done for estimation of cutoffs.

A classification tree for OS was generated using CART analysis (CHAID

method).

Results: The 2-year OS, LRFS, RR, LC, and RC were 64%, 56%, 65%,

63%, and 83%, respectively. The N-stage distribution was N0 (11), N1

(28), N2A (10), N2B (17), N2C (17), and N3 (4). The mean TNV was 19

mL (0-139 mL) and mean volumes per N stage were N1: 8 mL and N2:

20.7 mL with N2A: 15.6 mL, N2B: 18.8 mL, N2C: 27.3 mL, and N3: 118

mL. The mean TNV in survivors and nonsurvivors was 12.7 mL and 29.5

mL and in regionally controlled and uncontrolled patients was 9.3 mL and

68.7 mL. On multivariate analysis, the TNV was a significant prognostic

factor for OS and RC. ROC curve analysis found an optimal volumetric

cutoff point of 15 mL for OS and RC. The 2-year OS and RC for the <15

mL and >15 mL groups were 78% and 30% (PZ.001) and 100% and 52%

(PZ.001), respectively. Similar results were obtained on subset analysis of

our oropharyngeal patients with 2-year OS of 75% and 24% for the <15

mL and >15 mL groups (PZ.001). The likelihood of mortality increased

by 2% for 1-mL increase in nodal volume for the entire cohort and 2.8%

for our oropharyngeal subgroup.The likelihood of regional failure

increased by 8% for 1-mL increase in nodal volume.

Conclusion: TNV is an independent prognostic factor for OS and RC in

head and neck cancer. TNV can identify patients for consideration of

elective neck dissection following concurrent CRT, that is, for patients

with TNV>15 mL.

Author Disclosure: B. Dua: None. K.S. Chufal: None. G. Jadhav: None.

190HPV Status May Have Limited Value as a Prognostic Factor inPostoperative Squamous Cell Carcinoma of the Head and NeckCompared to Extranodal Extension and Lymphovascular SpaceInvasionM.H. Bertke, J.N. Shaughnessy, M.K. Forsthoefel, E. Cash,

C.L. Silverman, M.J. Amsbaugh, J. Bumpous, K. Potts, R. Redman,

C.A. Perez, and N.E. Dunlap; University of Louisville, Louisville, KY

Purpose/Objective(s): The significance of human papillomavirus (HPV)

in the postoperative setting in squamous cell carcinoma of the head and

neck (HNSCC) is not well defined. Our study aims to determine the impact

of HPV after primary surgery for HNSCC in relation to traditional risk

factors.

Materials/Methods: We retrospectively evaluated a cohort of 927 patients

treated at our institution between 2009 and 2014 for HNSCC. Of these, 128

were nonmetastatic, treated with definitive surgery with or without adju-

vant radiation therapy (RT) or chemoradiation therapy (CRT). Oral cavity

(63.3%, nZ81), laryngeal (21.1%, nZ27), oropharyngeal (10.1%, nZ13),

and unknown primary or other subsite (5.5%, nZ7) cases were included.

Patients were treated with surgery alone (nZ19, 14.8%) or surgery plus

adjuvant treatment (RT 38.3%, nZ49; CRT 46.9%, nZ60). Patient, dis-

ease, and treatment factors were analyzed as potential prognostic factors in

proportional hazard regression models adjusted for age and site of disease.

Results:Median follow-up for survivors was 18.4 months (range 0.6-66.9).

Most patients had locally advanced disease with 63% having pathologi-

cally positive nodal disease. The majority (78%) of the 110 patients with

known HPV status were negative. Of the 81 node-positive patients un-

dergoing a neck dissection, 62% demonstrated extranodal extension

(ECE), and 42% demonstrated lymphovascular space invasion (LVSI). At

2-year follow-up, 18 (14.1%) experienced locoregional treatment failure

(LRF), and 12 (9.4%) experienced DF without LRF. Positive margins

related to poorer disease-free survival (DFS) and overall survival (OS).

ECE, perineural invasion, and LVSI positivity predicted poorer LRF, DFS,

and OS. HPV status alone did not predict LRF, DFS or OS, but an inter-

action analysis revealed that, compared to HPV-positive/ECE-negative

patients, both HPV-positive and HPV-negative patients with ECE experi-

enced significantly poorer OS (85.7% vs 60% and 47%, respectively;

Ps<.018). When all risk factors were considered simultaneously, LVSI

emerged as a significant predictor of OS (hazard ratio [HR]Z4.06, 95%

confidence interval [CI], 1.72-9.57, PZ.001) with ECE showing marginal

significance for prediction of OS (HRZ2.184, 95% CI, .975-4.891,

PZ.058).

Conclusion: Although HPV status plays a major role in the affecting

outcomes of nonoperative HNSCC, its influence on patients in the post-

operative setting appears limited compared to traditional postoperative risk

factors such as ECE and LVSI. Further prospective analysis with a longer

follow-up period is warranted.

Author Disclosure: M.H. Bertke: None. J.N. Shaughnessy: None. M.K.

Forsthoefel: None. E. Cash: None. C.L. Silverman: None. M.J. Ams-

baugh: None. J. Bumpous: None. K. Potts: None. R. Redman: None.

C.A. Perez: None. N.E. Dunlap: None.

191Long-Term Outcomes for Head and Neck Cancer Patients With N3Neck DiseaseA. Wieland, M.E. Witek, A.P. Wojcieszynski, T. Kennedy, G.K. Hartig,

and P.M. Harari; University of Wisconsin, Madison, WI

Purpose/Objective(s): The aim of the present study was to evaluate

clinical outcomes for patients with squamous cell carcinomas of the head

and neck presenting with N3 nodal disease.

Materials/Methods: Following identification of eligible cases from our

institutional review boardeapproved head and neck cancer database, we

performed a retrospective analysis of patients with N3 nodal disease be-

tween 1989 and 2014. We limited this review to patients who received

curative intent therapy at 1 institution and had squamous cell carcinoma

histology. Primary sites included oropharynx (57%), unknown primary

(20%), hypopharynx (10%), larynx (7%), oral cavity (3%), and naso-

pharynx (3%). Treatment of the primary tumor included chemoradiation

therapy (43%), radiation therapy (29%), and surgery (15%). Neck

dissection was performed in 90% of patients. Overall survival, locore-

gional control, and rates of distant metastases were defined.

Results:Mean age was 59 years (range 41-85). Median follow-up was 23.4

months (range 1.2-226.8). Overall survival at 1, 3, and 5 years was 73%,

44%, and 35%, respectively. Locoregional control was 84%, 67%, and

63%, respectively. Distant metastases-free survival at 1, 3, and 5 years was

78%, 61%, and 58%, respectively.

Conclusion: Head and neck cancer patients with N3 nodal disease his-

torically have a poor long-term prognosis. However, approximately 30% of

patients are long-term survivors, particularly when effective locoregional

control is achieved. Factors that correlate with more favorable long-term

outcome are reviewed from this study and from the published N3 head and

neck literature. It appears worthy to approach most N3 patients with

curative intent in the absence of distant metastases at presentation.

Author Disclosure: A. Wieland: None. M.E. Witek: None. A.P.

Wojcieszynski: None. T. Kennedy: None. G.K. Hartig: None. P.M.

Harari: None.

International Journal of Radiation Oncology � Biology � Physics908

192Retrospective Analysis of Cisplatin Nephrotoxicity in Patients WithHead and Neck Cancer Receiving Outpatient Treatment WithConcurrent High-Dose Cisplatin and Radiation TherapyM. Porosnicu,1 J. Faig,2 R.C. Taylor,1 K.A. Rodriguez Porosnicu,3

M. Murea,1 M. Bonomi,1 and K.M. Greven1; 1Wake Forest University

School of Medicine, Winston Salem, NC, 2Icahn School of Medicine at

Mount Sinai, New York, NY, 3Columbia University in the City of New York,

New York, NY

Purpose/Objective(s): Cisplatin remains the pivotal chemotherapy in

squamous cell carcinoma of the head and neck (SCCHN), with neph-

rotoxicity considered the dose-limiting toxicity. The purpose of our

study was to propose an outpatient high-dose cisplatin (op-HD-cis)

protocol aimed at preventing nephrotoxicity and to analyze the results of

its use in patients with SCCHN treated with concurrent radiation

therapy (RT).

Materials/Methods: We retrospectively evaluated 82 patients with

SCCHN treated with op-HD-cis concurrent with RT at our institution.

Acute kidney injury (AKI) and chronic kidney disease (CKD) were

defined by Kidney Disease Improving Global Outcomes criteria. Asso-

ciated factors were identified using analysis of covariance models for

categorical variables and adjusted Pearson correlations for continuous

variables.

Results: The incidence of AKI during treatment was 34.2%. With a me-

dian follow-up of 25.7 months, the average decrease in eGFR was 12.57

mL/min/1.73m2 (SDZ18.58). At 1-year and at last follow-up, 5.4% and

4.4% of patients had eGFR <60 mL/min/1.73m2, respectively. Predictors

associated with AKI and CKD were lower baseline weight and creatinine,

higher baseline creatinine clearance, smoking, female gender, African

American race, hypertension, and increased hydration and magnesium

replacement requirements.

Conclusion: We encountered limited early and late nephrotoxicity.

Importantly, nephrotoxicity was not the main dose-limiting toxicity.

Apparent low baseline serum creatinine in patients with low body weight

should not decrease the protocol intensity. Our results emphasize the

importance of close monitoring and additional replacement of water and

electrolytes as needed. A consistent method of measuring and reporting

chemotherapy-induced nephrotoxicity would be a valuable contribution to

the literature.

Author Disclosure: M. Porosnicu: None. J. Faig: None. R.C. Taylor:

None. K.A. Rodriguez Porosnicu: None. M. Murea: None. M. Bonomi:

None. K.M. Greven: None.

193Presence of Preradiation Therapy Feeding Tube Associated WithPoor Prognostic Subset of Postoperative p16-PositiveOropharyngeal CarcinomaV. Verma,1 J. Liu,2 D. Adkins,2 J.F. Piccirillo,2 B. Nussenbaum,2

W.L. Thorstad,2 and H.A. Gay2; 1Department of Radiation Oncology,

University of Nebraska Medical Center, Omaha, NE, 2Washington

University School of Medicine, St. Louis, MO

Purpose/Objective(s): Timing of feeding tube insertion relative to radi-

ation therapy (RT) could be associated with prognosis in patients with head

and neck cancers. This study assesses feeding tube insertion time as well as

several other clinical variables that are associated with prognoses in

postoperative p16-positive oropharyngeal squamous cell carcinoma that

undergo adjuvant RT or chemoradiation therapy.

Materials/Methods: Three hundred seventy-six consecutive patients with

oropharyngeal cancer were identified from a large prospectively main-

tained academic institutional database from 1997 to 2009. Two hundred

twenty of these had adjuvant RT, and 97 were p16 positive. Of these

patients, 42 never had a feeding tube placed (NO-FT), 23 had one placed

before RT (B-FT), and 42 had one placed during or after RT (DA-FT).

Feeding tubes were not placed prophylactically. Analysis was conducted

between these 3 groups for differential tumor, patient, treatment, and

feeding tube characteristics, as well as differences in overall survival

(OS), disease-free survival (DFS), and distant metastasis-free survival

(DMFS).

Results: Five-year OS for the NO-FT, DA-FT, and B-FT groups was 90%,

86%, and 50%, respectively. Five-year DFS for each group was 88%, 84%,

and 43%, respectively. Kaplan-Meier survival curve analysis and log-rank

statistical analysis showed statistically inferior OS and DFS for the B-FT

group (P<.001). On multivariate analysis, timing of feeding tube place-

ment as well as smoking history was associated with OS and DFS.

Feeding tube insertion prior to RT was associated with higher tumor size

and depth, T (but not N) and overall stage, comorbidities, receipt of

chemotherapy, and less use of transoral laser microsurgery/transoral bovie

surgeries. Additionally, the time from surgery to RT completion was also

statistically longer in the B-FT group. The feeding tube was permanent in

52% of patients in the B-FT group versus 16% in the DA-FT group

(PZ.0075).

Conclusion: The presence of a feeding tube at the time of RT consultation,

due to associated poor prognostic variables of early feeding tube insertion,

can serve as a surrogate marker to identify a subset of p16-positive

oropharyngeal cancer patients with a poor prognosis.

Author Disclosure: V. Verma: None. J. Liu: None. D. Adkins: None. J.F.

Piccirillo: None. B. Nussenbaum: None. W.L. Thorstad: None. H.A.

Gay: None.

194Prognostic Indications of p16 and Smoking Status in Predicting theNeed for Posttreatment Neck Dissection After ChemoradiationTherapy in Head and Neck Squamous Cell CarcinomaZ.J. Cappello,1 M. Eid,2 E. Cash,3 L. Wilson,4 P. Tennant,3 J. Bumpous,5

and K. Potts3; 1University of Louisville School of Medicine, Department of

Otolaryngology, Louisville, KY, 2University, Louisville, KY, 3University of

Louisville, Louisville, KY, 4University of Louisville, James Graham Brown

Cancer Center, Louisville, KY, 5University of Louisville School of

Medicine; James Graham Brown Cancer Center, Louisville, KY

Purpose/Objective(s): This study aimed to determine the effect of p16

and smoking status on response to chemoradiation therapy and the need for

posttreatment neck dissection in patients with head and neck squamous

cell carcinoma (SCC).

Materials/Methods: This study was a retrospective medical records

review of patients with head and neck SCC treated with chemotherapy

and radiation therapy. Based on ICD-9 and CPT codes, patients with

SCC who received definitive chemoradiation therapy were cross-refer-

enced with those also undergoing neck dissection. Preliminary correla-

tions determined age at diagnosis and N stage at presentation should be

adjusted in hypothesis tests. Logistic and Cox proportional hazard

regression models tested risk group differences in the likelihood of

undergoing neck dissection, disease-free survival (DFS), and overall

survival (OS).

Results: A total of 152 patients with oral cavity (nZ25, 16%), oropha-

ryngeal (nZ106, 67.9%), laryngeal (nZ20, 12.8%), and unknown primary

diagnoses (nZ5, 3.2%) were included. Patients were stratified by risk

group: 65 (41.7%) were p16 negative (high risk), 46 (29.5%) were p16

positive with a greater than 10 pack-year smoking history (intermediate

risk), while 41 patients (26.3%) were p16 positive with a fewer than 10

pack-year history (low risk). Of these, 45 patients (29.6%) underwent a

neck dissection after definitive chemotherapy and radiation. After models

were adjusted for age, N stage, and total radiation dose, the low (PZ.005)

and intermediate (P<.001) risk groups were significantly less likely to

undergo neck dissection than were high-risk patients. No significant risk

group differences in DFS were observed. High-risk patients demonstrated

significantly poorer OS (PZ.026). However, model adjustment suggested

Volume 94 � Number 4 � 2016 Posters 909

this result was better explained by age (PZ.004) and total radiation

received (PZ.007).

Conclusion: In this data series, high-risk patients were significantly more

likely to undergo a neck dissection after definitive chemoradiation treat-

ment. However, risk group stratification appeared to have no prognostic

value when examining DFS or OS among this group of patients with head

and neck SCC.

Author Disclosure: Z.J. Cappello: None. M. Eid: None. E. Cash: None.

L. Wilson: None. P. Tennant: None. J. Bumpous: None. K. Potts: None.

195Can Aggressive Intravenous Hydration Prevent Cisplatin-InducedRenal Dysfunction?J. Bodmann,1 L. Rybicki,1 B. Harr,1 D. Ives,1 M. Rahe,1 T. Nwizu,1

S. Koyfman,2 and D.J. Adelstein1; 1Cleveland Clinic, Cleveland, OH,2Cleveland Clinic Foundation, Cleveland, OH

Purpose/Objective(s): Renal dysfunction is a well-recognized toxicity

from cisplatin chemotherapy, and preemptive intravenous (IV) hydration

during cisplatin administration is universally recommended. Subsequent

IV hydration is often useful for patients with treatment-induced nausea,

vomiting, or dehydration and in those who develop acute cisplatin-

induced renal failure, but its role in preventing chronic renal dysfunction

is unclear. We retrospectively reviewed a homogeneous cohort of head

and neck cancer patients treated with definitive chemoradiation therapy

(CRT) using high-dose cisplatin to identify predictors of chronic renal

dysfunction and address the preventative value of aggressive IV

hydration.

Materials/Methods: Between 2013 and 2015, the records of all patients

with p16-positive/human papillomavirusepositive oropharynx cancer

treated with definitive CRT using high-dose intermittent bolus cisplatin

(100 mg/m2 every 3 weeks) at our institution were reviewed. Data

recorded and analyzed included age; race; gender; performance status;

smoking history; alcohol use; tumor stage; total cisplatin dose; baseline,

peak, and 3-month posttreatment serum creatinine; along with number of

occasions that IV hydration was given during the course of CRT. Renal

dysfunction was defined either as an increase in creatinine of at least

25% over baseline or as any creatinine increase above our institutional

norms.

Results: We identified 78 patients for review. Their median age was 58

years (range 40-77). Most were male (86%), Caucasian (90%), and had

stage IVa disease (96%). There were 42 (54%) former smokers, 18 (23%)

active smokers, and 14 (18%) with a history of heavy alcohol con-

sumption. All patients began treatment with normal renal function. Two

doses of cisplatin were given to 68 patients (87%); the other 10 patients

(13%) received 3 doses. During treatment, 39 patients (50%) experienced

a creatinine elevation above normal. Although the 3-month posttreatment

creatinine remained greater than 25% above baseline in 13 patients

(17%), it was greater than our institutional norms in only 2 (3%). There

were 16 patients who did not receive any extra IV hydration. Additional

hydration was given on 1 to 18 occasions to the other 62 patients. No

pretreatment factors, including cisplatin dose, were identified that could

predict for posttreatment renal dysfunction. There was no significant

correlation identified between the number of hydration occasions and the

percentage change between baseline and 3-month posttreatment creati-

nine (PZ.54).

Conclusion: Although acute renal failure is common after high-dose

cisplatin administration, chronic renal dysfunction is rare. No clinical

parameters proved predictive, and aggressive IV hydration after treatment

was not helpful.

Author Disclosure: J. Bodmann: None. L. Rybicki: None. B. Harr:

Stock; Johnson & Johnson. D. Ives: None. M. Rahe: None. T. Nwizu:

None. S. Koyfman: None. D.J. Adelstein: None.

196T3 Squamous Cell Carcinoma of the Glottic Larynx: PrimaryRadiation Treatment OutcomesA.P. Wojcieszynski,1 R. Toya,2 G.K. Hartig,1 T.M. McCulloch,1 C. Britt,1

T. Gessert,1 G.D. Avey,1 and P.M. Harari1; 1University of Wisconsin,

Madison, WI, 2Kumamoto University Hospital, Kumamoto, Japan

Purpose/Objective(s): Patients with T3 squamous cell carcinoma of the

glottic larynx are commonly treated with organ preservation approaches

using radiation or chemoradiation. In this report, we review the University

of Wisconsin experience with T3 glottic cancer treated with nonsurgical

therapy over the last 23 years.

Materials/Methods: Twenty patients with T3 squamous cell carcinoma of

the glottic larynx treated with primary radiation between 1992 and 2014

were identified from our institutional review boardeapproved head and

neck cancer database. Median age at diagnosis was 67 years (range:

24-86). TNM staging included 17 patients with T3N0, 2 with T3N1, and 1

with T3N2 disease. All but 1 patient had a history of tobacco use with a

median 24 pack-year smoking history. Median tumor volume on pre-

treatment computed tomographic scan was 3.4 mL (range: 1.1-9.9).

Twelve patients were treated with radiation alone while 8 patients received

concurrent chemotherapy with cisplatin (nZ5), lapatinib (nZ2), or

cetuximab (nZ1). Median radiation dose was 70 Gy with 10 patients

receiving QD radiation and 10 receiving BID or QD/BID hybrid frac-

tionation regimens. Median overall treatment time was 44 days (range:

31 e 49).

Results: With a mean follow-up time of 6.2 years, local control rates at 5

and 10 years were 79% and 66%, respectively. Five- and 10-year larynx

preservation rates were 74% and 59%. Five patients experienced local

failure, with 1 patient failing both locally and regionally, and 1 patient

experienced distant only failure in the lung and liver. Median time to

locoregional failure was 13.4 months. Ultimately, larynx preservation was

achieved in 15 patients (75%). Six patients (30%) had gastrostomy tubes

placed during treatment for a median duration of 5 months. Median weight

loss during treatment was 6 pounds. Three patients (15%) required hos-

pitalization either during treatment or within the first month after treat-

ment. Of 4 patients (20%) who required a temporary tracheostomy during

treatment, only 1 patient maintained a tracheostomy beyond 6 months.

Median voice quality at the time of last follow-up per Voice Handicap

Index-10 scoring was 15.5 on a 0 to 40 scale (0Zbest voice, 40Zworst

voice) with range 6 to 27 in our treatment cohort.

Conclusion: Patients with T3 glottic cancer treated with primary radiation

approaches experienced high rates of local control (79% and 66% at 5 and

10 years) and larynx preservation (74% and 59% at 5 and 10 years). Patient

numbers were too small to identify a significant impact of radiation

fractionation, overall treatment time, tumor volume, or chemotherapy on

outcome. Treatment tolerance and voice quality outcome were favorable

for the majority of our T3 glottic cancer patients.

Author Disclosure: A.P. Wojcieszynski: None. R. Toya: None. G.K.

Hartig: None. T.M. McCulloch: None. C. Britt: None. T. Gessert: None.

G.D. Avey: None. P.M. Harari: None.

197Case Report: Dosimetric Comparison of Oral Cavity Dose WithDifferent Tongue Positions in 5 Patients Treated With IntensityModulated Radiation Therapy for Oropharyngeal CancerW.J. Kil, C.M. Kulasekere, V. Harwalkar, J.R. Bugno, L.D. Jefferson,

J.Z. Baskin, C.J. Nock, R.D. Derrwaldt, R. Raju, and C.L. Hatch; Louis

Stokes Cleveland VA Medical Center, Cleveland, OH

Purpose/Objective(s): To compare radiation dose to oral cavity (OC)

with different tongue positions in patients treated with intensity modulated

radiation therapy (IMRT) for oropharyngeal cancer (OPC).

International Journal of Radiation Oncology � Biology � Physics910

Materials/Methods: Five patients with OPC who were unable to use bite

block or tongue blade underwent computed tomographic simulations with

neutral tongue position and stuck-out tongue for planning IMRT (IMRT-N

and IMRT-S, respectively). Planning objectives were to deliver 70 Gy, 63

Gy, and 56 Gy in 35 fractions to 95% of PTVs using simultaneously

integrated boost technique, with mean doses (Dmean) of <26 Gy to the

parotid gland, <30 Gy to the submandibular gland, and <36 Gy to the oral

cavity (OC). For other organs at risk (OARs), Radiation Therapy

Oncology Group (RTOG) recommended dose constraints were applied.

Planning was optimized to minimize doses to OARs without compro-

mising coverage of PTVs. Radiation plans with IMRT-N and IMRT-S were

compared.

Results: IMRT-N and IMRT-S showed equivalent radiation target cover-

ages with sparing OARs, except OC. Dmean of OC was 34.8�1.6 Gy and

30.7�3.0 Gy with IMRT-N and IMRT-S, respectively (PZ.006). OC

volume receiving �36 Gy (V36) was 34.1�19.0% using IMRT-N

compared to 28.0�15.9% with IMRT-S (PZ.03). Dmean of oral tongue was

39.6�2.7 Gy and 31.7�3.5 Gy in IMRT-N and IMRT-S, respectively

(PZ.01). The distance from palate to surface of tongue tend to increase in

IMRT-S with 0.9�0.3 cm compared to IMRT-N with 0.1�0.2 cm

(PZ.003).

Conclusion: In 5 patients with OPC, an IMRT-S significantly reduced

radiation dose to oral cavity, specifically oral tongue, compared to

IMRT-N without compromising radiation target coverage or other

OARs. For the patients with OPC who were unable to use bite block or

tongue blade, stuck-out tongue during radiation therapy can provide less

radiation dose to OC, which could lead to fewer radiation-related side

effects in OC.

Author Disclosure: W. Kil: None. C.M. Kulasekere: None. V. Harwal-

kar: None. J.R. Bugno: None. L.D. Jefferson: None. J.Z. Baskin: None.

C.J. Nock: None. R.D. Derrwaldt: None. R. Raju: None. C.L. Hatch:

None.

198Concurrent Chemoradiation in Oropharyngeal Cancer: DoesOmission of 1 or More Cycles of Chemotherapy Diminish Survival orTumor Control?T. Dorius,1 N.R. Bennion,1 L. Smith,1 B. Vanbriggle,1 W. Zhen,1

and A.K. Ganti2; 1University of Nebraska Medical Center, Omaha, NE,2VA Nebraska-Western Iowa Health Care System, Omaha, NE

Purpose/Objective(s): Concurrent chemotherapy and radiation improves

outcomes in patients with locoregionally advanced oropharyngeal cancer

(LA-OPC). The standard regimen of high-dose cisplatin and 70 Gy

radiation is associated with significant toxicities, often resulting in a

decreased relative dose intensity of radiation and chemotherapy. The

effect of such dose reduction on outcomes is unclear. We review our

institution’s results to determine factors that predict for improved

outcomes in LA-OPC.

Materials/Methods: All patients were treated with concurrent chemo-

radiation for LA-OPC at an academic medical center from 2007 to 2014.

Patient, tumor, and treatment variables were abstracted from the patient

records. Overall survival and data regarding relapse were collected.

Overall survival and disease-free survival were compared using Kaplan-

Meier curves. The log-rank test was used to determine statistical

significance.

Results: The 47 patients analyzed had a median age at diagnosis of 56.7

years. The majority were male (89.4%) and had a history of tobacco use

(74.5%). Among the 35 patients whose tumors were tested, 30 stained

positive for p16. All but 1 patient reviewed had at least stage III disease.

Thirty-six (77%) patients were alive at last follow-up. The median follow-

up time for the surviving patients is 2.9 years (range: 0.6-8.2 years).

Among those who received cisplatin, there was no significant difference in

overall survival or disease-free survival based on receipt of all 3 planned

doses versus 1 to 2 doses (PZ.93). Only 9 patients received cetuximab;

there was no significant difference in overall survival when compared to

cisplatin (PZ.60). Six patients had a feeding tube placed prior to starting

chemoradiation; an additional 20 (43%) required placement of a feeding

tube during their chemoradiation course. Patients who lost >10 kg were

noted to have a significantly improved overall and disease-free survival

(PZ.035).

Conclusion: Patients who lost >10 kg weight during therapy had

improved outcomes. One possible explanation includes that the develop-

ment of toxicity correlates with improved response. Omission of 1 or more

cycles of cisplatin did not affect outcomes. These findings need to be

further elucidated in the setting of larger retrospective series and pro-

spective analyses and, if proven to be true, may have treatment

implications.

Author Disclosure: T. Dorius: None. N.R. Bennion: None. L. Smith:

None. B. Vanbriggle: None. W. Zhen: None. A. Ganti: None.

199Prognostic Significance of p16 in Squamous Cell Carcinoma of theLarynx and HypopharynxJ.J. Meshman,1 J. Wang,2 R.K. Chin,1 E. Abemayor,1 M. St. John,1

S. Bhuta,1 and A.M. Chen2; 1University of California, Los Angeles, Los

Angeles, CA, 2University of California, Los Angeles- David Geffen School

of Medicine, Los Angeles, CA

Purpose/Objective(s): To evaluate the prognostic significance of p16

expression in patients with squamous cell carcinoma of the larynx (LSCC)

and hypopharynx (HSCC).

Materials/Methods: The records of all patients with locally advanced,

nonmetastatic LSCC/HSCC at a single institution were reviewed.

P16INK4A (p16) protein expression was evaluated by immunohistochem-

istry (IHC). The Kaplan-Meier method was used to estimate overall sur-

vival (OS) and locoregional control (LRC). In select cases, HPV status was

evaluated for high-risk and low-risk human papillomavirus (HPV) geno-

types by in situ hybridization (ISH).

Results: Thirty-one patients (23 LSCC; 8 HSCC) were identified.

Seventeen patients (55%) were p16 negative; 14 (45%) were p16 positive.

The primary treatment modality was radiation therapy for 22 patients

(71%) and surgery for 9 (29%). Nineteen patients (61%) were evaluated

for high-risk HPV and low-risk HPV genotypes by IHC, of whom 2 pa-

tients (11%) were positive for high-risk HPV and 1 (5%) was positive for

low-risk HPV. For high-risk HPV, the positive predictive value (PPV),

sensitivity, and specificity of p16 were 20%, 100%, and 53%, respectively.

There was no significant difference in the 2-year actuarial rates of OS

(91% vs 64%, PZ.34) or LRC (51% vs 46%, PZ.69) between the p16-

positive and p16-negative patients.

Conclusion: In this small cohort of 31 LSCC and HSCC patients, p16 was

not a significant predictor of either LRC or OS. Furthermore, p16 was

poorly associated with HPV as identified by ISH.

Author Disclosure: J.J. Meshman: None. J. Wang: None. R.K. Chin:

None. E. Abemayor: None. M. St. John: None. S. Bhuta: None. A.M.

Chen: None.

Volume 94 � Number 4 � 2016 Posters 911

200Impact of Weight Loss on Grip Strength in Head and Neck CancerPatients Receiving Radiation TherapyA. Hilbert, J. McGuire, S. Rentz, D.A. Elliott, and J.M. Holland; Oregon

Health and Science University, Portland, OR

Purpose/Objective(s): Malnutrition is a common complication for head

and neck cancer (HNC) patients receiving radiation therapy (RT) or che-

moradiation therapy (CRT). The Academy of Nutrition and Dietetics

(AND) and the American Society of Parenteral and Enteral Nutrition

(ASPEN) define moderate and severe malnutrition as the presence of 2 or

more of the following: inadequate energy intake, weight loss, loss of body

fat, loss of muscle mass, fluid accumulation, or reduced grip strength. In

malnutrition, decreased muscle function may respond more quickly to

nutritional deprivation than changes in weight. Research suggests that grip

strength is a feasible tool for the bedside assessment of muscle function.

The purpose of this study was to implement grip strength assessment as

part of the nutrition assessment for HNC patients receiving RT or CRT and

to investigate the relationship between grip strength and weight loss before

and after 7 weeks of RT.

Materials/Methods: Grip strength, assessed by dynamometry, and weight

were recorded the first and last week of RT. Each hand was measured 3

times and the average was calculated. Hand dominance was noted. Sub-

jects were classified as malnourished if they met the AND/ASPEN

malnutrition criteria for weight loss and reduced grip strength. Reduced

grip strength was defined as a grip strength 2 standard deviations (SD)

below the normative standards provided with the dynamometer. Descrip-

tive statistics were used to describe the subject population, changes in grip

strength, and changes in weight. Correlation between change in grip

strength and change in weight was performed using the Pearson

correlation.

Results: Eleven subjects, with a median age of 59 years (53-74), received

RT (nZ4) or CRT (nZ7). Mean change in grip strength for the left hand

was -1.0�3.8 kg (SD) and for the right hand was -0.7�5.6 kg (SD). The

mean weight loss was 6.6�6.4 kg (SD). Decreased grip strength correlated

weakly with weight loss for both the left (rZ0.248, 90% confidence in-

terval [CI] -0.205-0.501, PZ.462), and right hand (rZ0.180, 90% CI

-0.396-0.686, PZ.596). Two subjects (18%) at the initial treatment visit

and 1 subject (9%) at the last treatment visit qualified as malnourished.

Conclusion: Although this study found no statistical correlation between

weight loss and grip strength in the small series of HNC patients receiving

RT or CRT, collecting data on grip strength using the dynamometer proved

feasible. To more accurately diagnose malnutrition using the criteria of

reduced grip strength, further research is needed to establish grip strength

reference ranges.

Author Disclosure: A. Hilbert: None. J. McGuire: None. S. Rentz: None.

D.A. Elliott: None. J.M. Holland: None.

201Parapharyngeal Space Tumor: Submandibular Approach WithoutMandibulotomyK. Luna-Ortiz,1 O. Villa-Zepeda,2 J. Carrillo,3 E. Molina-Frias,4

and A. Gomez-Pedraza5; 1Instituto Nacional de Cancerologia, Mexico

City, Mexico, 2Instituto Nacional de Cancerologia, Mexico City, Mexico,3Instituto Nacional de Cancerologia, Mexico City, Mexico,4Instituto Nacional de Cancerologia, Mexico City, Mexico,5Instituto Nacional de Cancerologia, Mexico City, Mexico

Purpose/Objective(s): From all head and neck tumors, only 0.5% occur

in the parapharyngeal space, and 80% are benign tumors. Surgery is the

cornerstone of treatment. The deep location of this space and of sur-

rounding structures have given origin to a large number of surgical

approaches for resection of parapharyngeal tumors. Our experience using

the transcervical approach on 14 patients who had surgery upfront for

treatment of these neoplasms is described.

Materials/Methods: A retrospective descriptive case series analysis of

patients with parapharyngeal space tumors treated from January 2009 to

July 2015 was performed.

Results: Fourteen patients were included: 11 females (76.9%) and 3 males

(23.1%) with a mean age of 46.9 years (range 20-75 years). The most

common symptom reported at presentation to our clinic was a foreign body

sensation in the oropharynx and pain. Navigation-guided surgery was used

in 2 cases. Mean size of tumors was 4.7 cm. Complete resection of lesions

was performed in all cases, and the only major complication was major

bleeding in 1 case (7.1%).

Conclusion: According to the present analysis, the submandibular

transcervical approach is an effective and safe technique that allows

resection of large parapharyngeal tumors, even those close to the skull

base. It has minimal morbidity, prevents morbidity associated man-

dibulotomy, and allows extension to a transparotid, transmandibular, and

even to an infratemporal fossa approach. Navigation is indicated in tu-

mors �2 cm. The submandibular transcervical approach should be

considered upfront for tumors �6 cm, preferentially �0.5 cm distant

from skull base.

Author Disclosure: K. Luna-Ortiz: None. O. Villa-Zepeda: None. J.

Carrillo: None. E. Molina-Frias: None. A. Gomez-Pedraza: None.

202WITHDRAWN

203Prognostic Value of Lymph Node Status Is Greater Than LymphNode Ratio and AJCC N Staging for Head and Neck Squamous CellCarcinomasT. Roberts,1 A.D. Colevas,1 W. Hara,1 F.C. Holsinger,1 I. Oakley-Girvan,2

and V. Divi1; 1Stanford University, Stanford, CA, 2Cancer Prevention

Institute of California, Fremont, CA

Purpose/Objective(s): Changes in the epidemiology of head and neck

cancers have created a need for new lymph node prognostics. The lymph

node ratio (LNR) has been proposed as an alternative staging system, but

this metric has limitations that may attenuate its prognostic value. This

study sought to test the prognostic value of the lymph node status (pN) and

compare it to the LNR and American Joint Committee on Cancer (AJCC)

N staging.

Materials/Methods: The Survival, Epidemiology, and End Results

(SEER) database was used to identify surgical cases from 2004 to 2012.

The study sample was grouped based on AJCC N stage, LNR, and pN and

analyzed using the Kaplan-Meier method and multivariate Cox propor-

tional hazard models. Models were compared using the Akaiki Informa-

tion Criterion (AIC). The sample was also analyzed by site of primary

tumor.

Results: We identified 12,437 patients in the SEER database for analysis.

Distribution of nodal staging was 5282 N0 patients, 2483 N1 patients,

4454 N2 patients, and 218 N3 patients. Twenty-four percent of patients

had an oropharyngeal primary. Kaplan-Meier survival curves showed

improved prognostic ability for pN and LNR staging relative to the AJCC

system. Tumors with >5 positive nodes were associated with the worst

overall survival (5-year survival rate Z 16%). Oropharyngeal tumors had

better outcomes for all groupings in all staging systems as compared to

those in nonoropharyngeal sites. Using the pN staging system, >5 positive

nodes in oropharyngeal tumors was strongly associated with decreased

survival (5-year survival rate Z 53%), while patients with 0 to 5 positive

nodes had similar 5-year survival.Multivariate regression outputs demon-

strated more prognostic hazard ratios and a lower AIC for the pN model

compared to the AJCC N stage and LNR models (Table 1). Hazard ratios

were 1.78 (95% confidence interval [CI], 1.62-1.95) for 1 positive node,

2.53 (95% CI, 2.32-2.75) for 2-5 positive nodes, and 4.64 (95% CI, 4.18-

5.14) for >5 nodes.

Conclusion: The pN models demonstrated superior prognostic value

compared to the LNR and AJCC N staging in the overall study sample

Abstract 203; Table 1 Multivariate regression outputs showing hazard ratiosfor AJCC N stage, pN, and LNR groups with 95% confidence intervals.

AJCC Nstage HR (95% CI) pN HR (95% CI) LNR HR (95% CI)

N1 1.98 (1.81-2.16) 1 1.78 (1.62-1.95) 0-6% 1.81 (1.66-1.98)N2a 1.95 (1.72-2.42) 2-5 2.53 (2.32- 2.75) 6%-12.5% 2.52 (2.29 -2.77)N2b 2.85 (2.66-3.16) >5 4.64 (4.18 -5.14) �12.5% 3.69 (3.37 -4.05)N2c 3.78 (3.30-4.15)N3 2.52 (2.05-3.28)

International Journal of Radiation Oncology � Biology � Physics912

and site-specific analyses. Future modifications of the nodal staging

system should be based on the lymph node status, with consideration

given to a separate system for oropharyngeal cancers. Patients with

more than 5 positive nodes have significantly worse survival in all

subsites and should be considered for alternative treatment

regimens.

Author Disclosure: T. Roberts: None. A.D. Colevas: None. W. Hara:

None. F. Holsinger: None. I. Oakley-Girvan: None. V. Divi: None.

204Dosimetric Predictors of Hypothyroidism After Radical IntensityModulated Radiation Therapy for Nasopharyngeal CarcinomaV.H. Lee,1 A.S. Chan,2 C.W. Choi,3 D.L. Kwong,4 K.O. Lam,4 C.C. Tong,5

H.C. Sze,4 S.C. Ng,5 and T.W. Leung5; 1Department of Clinical Oncology,

Queen Mary Hospital, Li Ka Shing Faculty of Medicine, The University of

Hong Kong, Hong Kong, Hong Kong, 2Department of Clinical Oncology,

Li Ka Shing Faculty of Medicine, Queen Mary Hospital, The University of

Hong Kong, Hong Kong, Hong Kong, 3Department of Systems Engineering

and Engineering Management, City University of Hong Kong, Hong Kong,

Hong Kong, 4The University of Hong Kong, Hong Kong, Hong Kong,5Department of Clinical Oncology, Li Ka Shing Faculty of Medicine,

The University of Hong Kong, Hong Kong, Hong Kong

Purpose/Objective(s): We investigated for dosimetric predictors of

hypothyroidism after radical intensity modulated radiation therapy (IMRT)

for nonmetastatic nasopharyngeal carcinoma (NPC).

Materials/Methods: Patients with nonmetastatic NPC treated with radical

IMRT from 2008 to 2013 were reviewed. They were regularly monitored

clinically with serum thyroid function tests before and after IMRT.

Univariable and multivariable analyses were performed for demographic

and dosimetric predictors of biochemical and clinical hypothyroidism.

Results: A total of 149 patients fulfilled eligibility criteria. After a median

follow-up duration of 3.1 years (range, 1.3-8.2 years), 33 patients (22.1%)

and 21 patients (14.1%) developed biochemical and clinical hypothyroid-

ism, respectively. Eight patients (24.2%) who had biochemical hypothy-

roidism developed clinical hypothyroidism later. Univariable and

multivariable analyses revealed that volume of the thyroid (PZ.002,

multivariable), VS60 (the absolute thyroid volume spared from 60 Gy or

less) (P<.001, multivariable), and VS45 (P<.001, multivariable) of the

thyroid were significant predictive factors of biochemical hypothyroidism.

The freedom from biochemical hypothyroidism was longer for those whose

VS60�10 mL (mean 90.9 months vs 62.6 months; P<.001) and VS45�5

mL (mean 91.9 months vs 65.2 months; PZ.001). Similarly, multivariable

analyses revealed that VS60 (PZ.001) and VS45 (PZ.003) were significant

predictors of clinical hypothyroidism. The freedom from clinical hypothy-

roidism was longer for those whose VS60�10 mL (91.5 months vs 73.3

months, PZ.002) and VS45�5 mL (91.5 months vs 75.9 months, PZ.007).

Receiver-operating characteristics (ROC) analyses revealed that the area-

under-the-curve (AUC) value for VS60was significantly higher than that for

VS45 for both biochemical (0.751 vs 0.660; PZ.009) and clinical hypo-

thyroidism (0.718 vs 0.607; PZ.007), respectively.

Conclusion: VS60 and VS45 of the thyroid significantly predicted post-

IMRT hypothyroidism and should be considered important priorities of

dose constraints during IMRT optimisation for NPC.

Author Disclosure: V.H. Lee: None. A.S. Chan: None. C. Choi: None.

D.L. Kwong: None. K. Lam: None. C. Tong: None. H.C. Sze: None. S.C.

Ng: None. T. Leung: None.

205Clinical Presentation of Oropharyngeal Squamous Cell Carcinoma inthe Modern Era: Does Risk Stratification Using HumanPapillomavirus and Smoking Status Matter?M.J. Amsbaugh,1 M.B. Yusuf,2 E. Cash,1 C.L. Silverman,1 J. Bumpous,1

C.A. Perez,1 R. Bert,1 R. Redman,1 and N.E. Dunlap1; 1University of

Louisville, Louisville, KY, 2University of Louisville School of Medicine,

Louisville, KY

Purpose/Objective(s): Patterns of cervical lymph node metastases

(LNMs) in patients with oropharyngeal squamous cell carcinoma

(OPSCC) have historically been described. Here we compare the distri-

bution of cervical LNMs in the human papillomavirus (HPV) era to his-

torical controls. We also evaluated the influence of HPV and smoking

status on the clinical presentation of OPSCC.

Materials/Methods: All patients with OPSCC presenting to our institutional

multidisciplinary clinic from January 2010 to June 2015were reviewed froma

prospective database. Subsite, stage, smoking status, and involved anatomical

nodal levels were tabulated. Variance in clinical presentation was examined

using X2, Kruskal-Wallis, Mann-Whitney, and logistic regression analyses.

Results: Of 291 patients, HPV/p16 status was available for 243. Preliminary

analyses confirmed that patients could be grouped based on anatomical site

(Group 1: tonsil, base of tongue [BOT], vallecula, nZ217; Group 2:

pharyngeal wall, palate, and other nZ26) and risk level (low risk: HPV/p16

positive and �10 pack-year history of smoking, nZ78; intermediate risk:

HPV/p16 positive and >10 pack-year history of smoking, nZ82; and high

risk: HPV/p16 negative, nZ80). Overall, 188 patients (77.3%) had at least 1

cervical LNM at time of presentation (99.4% ipsilateral, 16.5% contralat-

eral). Locally advanced tumor stages were uncommon at presentation

(22.2% T1, 37.0% T2, 26.7% T3, 10.7% T4a, and 2.5% T4b); however,

advanced nodal disease was more common (22.6% N0, 14.8% N1, 7.0%

N2a, 39.1% N2b, 11.9% N2c, and 4.5% N3). Patients who were current

smokers were more likely to present with metastases to level IIa than were

nonsmokers (odds ratio [OR]Z4.56, PZ.032). Rates of HPV/p16 positivity

(P<.001), never having smoked (PZ.016), and cervical LNM (PZ.023)

were significantly higher for patients in Group 1. Low-risk Group 1 patients

presented with nodal stage N2a at a much higher than expected frequency

(PZ.007), and high-risk patients presented with tumor stage T4 at a much

higher than expected frequency (PZ.003). Previous smokers with a BOT

subsite were more likely to have clinically positive ipsilateral necks than

nonsmokers (ORZ1.8, PZ.038). There were no significant differences in T

stage or N stage based on HPV/p16 positivity or smoking (P>.05).

Furthermore, distribution of cervical LNMs was not associated with HPV/

p16 positivity, risk group, or anatomical group (P>.05). When data were

compared to historical series, no significant differences were seen in the

patterns of cervical lymph node metastases for patients with OPSCC.

Conclusion: HPV/p16 positivity, smoking status, risk group, and

anatomical group influenced the clinical presentation of patients with

OPSCC. Historical series describing the patterns of cervical LNMs in

patients with OPSCC remain clinically relevant.

Author Disclosure: M.J. Amsbaugh: Employee; University of Louiville.

M.B. Yusuf: None. E. Cash: None. C.L. Silverman: None. J. Bumpous:

None. C.A. Perez: Employee; University of Louiville. R. Bert: None. R.

Redman: None. N.E. Dunlap: Honoraria; Osler.

206Demographics, Disparities, and Survival in Young Patients WithOral Cavity Squamous Cell Carcinoma: A Population-Level Analysisof 3828 CasesK. Zhan, E.A. Nicolli, and T. Day; Medical University of South Carolina,

Charleston, SC

Purpose/Objective(s): To characterize and identify prognosticators in oralcavity squamous cell carcinomas (OCSCC) in young patients. To deter-

mine whether type of hospital and insurance status correlates with survival.

Materials/Methods: We conducted a retrospective review of the National

Cancer Data Base from 1998 to 2012 of OCSCC in patients younger than

Volume 94 � Number 4 � 2016 Posters 913

45 years of age. Relevant demographic, tumor, and survival variables were

extracted for analysis. Hospitals were divided into community cancer

programs (100-500 annual cancer cases) and comprehensive community or

academic/research programs (�500 annual cancer cases). Cox regression

was used to identify predictors of survival.

Results: We identified 54,565 OCSCC patients, 7.6% of whom are younger

than 45 years of age (nZ3828). Of these patients, 80%were between 35 and

44 years of age. More males were affected (65.7%) than females. Cauca-

sians represented 86.3% of cases, followed by African Americans (9.5%)

and patients of “other” races (4.2%). Private insurance (65.6%) was most

common, with Medicaid (17.6%), uninsured (11.7%), and Medicare (5.1%)

comprising the rest. Overall survival at 2 and 5 years was 76% and 66%,

respectively. The oral tongue subsite was most common (55.4%), followed

by floor of mouth (FOM; 28.5%), gingiva/retromolar trigone (15.4%), and

buccal mucosa (0.7%). An increasing incidence of oral tongue cancers was

seen, while FOM cancers showed a decreasing trend over the study period.

A minority of cases was treated at low-volume community cancer centers,

which saw more stage I-II disease. Uninsured and Medicaid patients had

more advanced stage III-IV disease (P<.001), while those with private in-

surance had more early-stage disease. Further analysis including treatment,

insurance, demographics, and survival was performed. cStage I-II patients

without private insurance were more likely to receive some form of

chemotherapy. Ethnicity, insurance status, income, age group, pathologic

stage, and positive surgical margins are significant prognosticators on

univariate analysis. In multivariate analysis, high pathologic stage, non-

private insurance, treatment at a low-volume community center, and posi-

tive margins remained predictors of worse survival.

Conclusion: In young patients with oral cavity cancers, differences in

treatment, presentation, and survival were seen in those with health

disparities. In addition to staging and surgical margins, treatment at

low-volume community cancer centers and nonprivate insurance status

predicted worse survival.

Author Disclosure: K. Zhan: None. E.A. Nicolli: None. T. Day: None.

207Geographic Variation in the Costs and Outcomes of Treatment forHead and Neck CancerV. Divi,1 L. Tao,2 A. Whittemore,1 and I. Oakley-Girvan2;1Stanford University, Stanford, CA, 2Cancer Prevention Institute of

California, Fremont, CA

Purpose/Objective(s): Advanced head and neck cancer (HNC) is a

complex group of diseases that requires the input and coordination of

multiple providers. While there are general guidelines for treatment, there

is also considerable variation in how patients are treated and how long they

survive after treatment. It is unclear how the treatment variations relate to

treatment costs and survival.

Materials/Methods: We identified Medicare patients with advanced HNC

treated in 12 states between 2004 and 2009 using the linked database

containing Medicare and Surveillance Epidemiology and End Results

(SEER) data. Average cost per patient during the year following diagnosis

was calculated for each state. Costs included inpatient hospital, outpatient,

physician, and durable medical equipment charges. Three-year overall

survival was also calculated for each of the states.

Results: A total of 3678 patients were included in the final study. There was

significant cost variation among the states. Utah incurred the lowest total

costs within 1 year of HNC diagnosis ($51,857 per patient; standard devi-

ation [SD], $4,797), whereas New Jersey had the highest costs ($81,071 per

patient; SD, $2,347). Three-year overall survival also varied among the

states, ranging from 45 months in Kentucky to 58 months in Washington.

There was no correlation between expenditures and length of survival.

Conclusion: Despite significant variation in both expenditures and survival

among the states, we found no correlation between costs and mean survival

time, suggesting that more costly care did not lead to improved outcomes.

Author Disclosure: V. Divi: None. L. Tao: None. A. Whittemore: None. I.

Oakley-Girvan: None.

208Patient Immunosuppression and the Association With Cancer-Specific Outcomes After Treatment of Squamous Cell Carcinoma ofthe OropharynxD.N. Margalit,1 J.D. Schoenfeld,1 B. Rawal,2 M. Puzanov,3 R.I. Haddad,4

G. Rabinowits,2 N. Chau,2 J. Lorch,5 A. Lavigne,3 L.A. Goguen,1

D.J. Annino,1 T. Thomas,1 P.J. Catalano,2 and R.B. Tishler1; 1Dana-Farber

Cancer Institute / Brigham and Women’s Hospital, Boston, MA, 2Dana-

Farber Cancer Institute, Boston, MA, 3Brigham&Women’s Hospital, Boston,

MA, 4Department of Medical Oncology, Dana-Farber Cancer Institute,

Harvard Medical School and Department of Medicine, Brigham and

Women’s Hospital, Boston, MA, 5Dana Farber Cancer Institute, Boston, MA

Purpose/Objective(s): Patients (pts) with a history of immunosuppression

such as hematologic malignancy or transplant have previously been shown

to have an increased risk of a second malignancy and worse cancer-specific

outcomes particularly with respect to oral cavity cancers. We aimed to

determine the impact of immunosuppression on recurrence and survival

after treatment for squamous cell carcinoma of the oropharynx (OPSCC).

Materials/Methods: We analyzed 336 pts with OPSCC treated with

definitive intensity modulated radiation therapy (IMRT) to a median dose

of 70 Gy (median 2 Gy/fraction) at a single institution. We identified all pt

comorbidities at the initial consultation prior to therapy. Pts were

considered to be immunosuppressed if they had a history of hematologic

malignancy, stem-cell or solid-organ transplant, or AIDS. The association

of immunosuppression with overall survival (OS) and recurrence was

assessed with single and multivariable Cox proportional hazard regression

models. Predictors that were significant in the univariate model with a

P value of �.05 were included in the multivariable model. The Fine &

Gray method was used to adjust for competing risks.

Results: The primary cohort comprised 336 pts with OPSCC of whom

85% were male, the median age was 57.3 years, and primary sites were

tonsil (49.4%), base of tongue (46.7%), and other (3.9%). Most pts were

HPV positive (72.3%) with stage IVA being the most common American

Joint Committee on Cancer stage (71.5%) followed by IVB (9.2%) and

other. Pts received upfront concurrent chemoradiation (58%) or sequential

therapy (38.4%); 3.6% received radiation alone. With a median follow-up

of 3.8 years, pts with a history of immunosuppression had a 3-year OS of

71.4% with a cumulative incidence of local-regional failure (LRF) and

distant failure of 32.1%. They had worse OS (adjusted hazard ratio [HR]

3.95; 95% confidence interval [CI] 1.13-13.80), worse recurrence-free

survival (adjusted HR 3.72; 95% CI 1.18-11.75), and increased risk of LRF

(unadjusted HR 3.80; 95% CI 1.31-11.09; adjusted HR 3.27; 95% CI 0.97-

11.01) and increased risk of distant failure (adjusted HR 3.72; 95% CI

1.18-11.73) compared with the primary cohort of pts.

Conclusion: Patients with a diagnosis associated with immunosuppression

prior to definitive treatment for OPSCC have worse OS and recurrence-free

survival and increased rates of local and distant failure after treatment. This

clinical finding lends support to exploring the role of optimal immune function

for cancer clearance and survival after treatment for oropharyngeal cancer.

Author Disclosure: D.N. Margalit: None. J.D. Schoenfeld: None. B.

Rawal: None. M. Puzanov: None. R.I. Haddad: None. G. Rabinowits:

None. N. Chau: None. J. Lorch: None. A. Lavigne: None. L.A. Goguen:

None. D.J. Annino: None. T. Thomas: None. P.J. Catalano: None. R.B.

Tishler: None.

209WITHDRAWN

210Compliance to Radiation Therapy for Head and Neck Cancer in aSafety-Net Health SystemB.P. O’Donnell,1 M.S. Ludwig,2 H.J. Manley,3 J.A. Asper,2

and M. Bonnen2; 1The University of Texas Health Science Center at

Houston, Houston, TX, 2Baylor College of Medicine, Houston, TX,3Southern Methodist University, Dallas, TX

International Journal of Radiation Oncology � Biology � Physics914

Purpose/Objective(s): To identify variables associated with compliance

to radiation therapy (RT) for treatment of head and neck cancer in a safety-

net health system.

Materials/Methods: We systematically screened electronic health record

data from a large safety-net health system serving over 300,000 patients

annually between October 2012 and August 2015 and identified 181 pa-

tients who were treated with radiation therapy (RT) in a curative approach

for American Joint Committee on Cancer stage I-IV head and neck cancer.

The number of days of missed treatments and reasons given for them by

the patients were identified. Demographics, access to care, toxicity, drug

use, comorbidities, psychiatric diagnosis, and treatment variables were

analyzed for impact on compliance. Disease-free (DFS) and overall sur-

vival (OS) of patients with more than or fewer than 5 days of missed

treatments were compared using Kaplan-Meier analyses.

Results: A total of 181 patients (mean age 54.8 years) were identified who

completed RT with curative intent, and 141 (77.9%) of these patients

received chemotherapy (CT). Of these patients, 45 (31.9%) patients

received induction CT, and 136 (96.4%) received concurrent CT. There

were 166 patients (90.0%) who completed all fractions of RT, and 43

patients (23.8%) completed their RTwith no missed days. There were 9240

total treatment days (mean of 51.1 treatment days), of which 844 (9.1%)

were missed or delayed treatment days. Patient-offered reasons for the

delay were classified as: failure to coordinate care (nZ34, 4.0%), trans-

portation issues (nZ54, 6.4%), acute toxicity (nZ229, 25.9%), inclement

weather (nZ20, 2.4%), comorbidities (nZ107, 12.7%), delays related to

PFG feeding tubes (nZ18, 2.1%), tolerance to set-up (nZ32, 3.8%), pa-

tient initiated (nZ209, 24.8%), and unknown (nZ150, 17.8%). Not

owning a vehicle was associated with missing more days due to trans-

portation issues (PZ.04). Male gender (PZ.017) and weight loss during

radiation (PZ.029) were associated with more total missed days of radi-

ation. Variables that were associated with 5 or more missed treatment days

were the lowest Karnofsky Performance Status Score achieved during RT

(PZ.002), maximum pain score during RT (PZ.042), skin sensation

toxicity (PZ.008), history of methamphetamine abuse (PZ.009) (but not

other substances), and multiagent induction CT (PZ.027). In patients

whose RT treatment was completed with fewer than or exactly 5 days of

delay, DFS was improved by 9.4 months (22.5 vs 13.1, PZ.012), and OS

was improved by 7.0 months (28.2 vs 15.3, PZ.0001).

Conclusion: This study concurs with past studies that missing treatments is

detrimental to overall survival. Therefore, it is prudent in a safety-net

hospital to identify specific barriers to treatment compliance in order to

design optimal interventions, such as a questionnaire that identifies high-

risk patients so that we can intervene by allocating resources appropriately

in order to improve outcomes.

Author Disclosure: B.P. O’Donnell: None. M.S. Ludwig: None. H.J.

Manley: None. J.A. Asper: None. M. Bonnen: Selecting new residents

and training current residents; Baylor College of Medicine. Manages af-

fairs of the Radiation Oncology Department; Baylor College of Medicine.

211Free Open Source REDCap Software to Track EORTC QLQ-30 andH&N-35 Quality of Life Scores and Allow Real-Time ClinicalManagement of Individual PatientsA. Platek,1 A. Ostrowski,1 M. Platek,2 A. Beattie,3 S. Fung-Kee-Fung,2

V. Gupta,1 D. Cohan,1 W. Hicks,1 H. Arshad,1 M.A. Kuriakose,1

and A.K. Singh2; 1Roswell Park Cancer Institute, Buffalo, NY, 2Roswell

Park Cancer Institute, Buffalo, NY, United States, 3New York University

Medical Center, New York, NY

Abstract 211; Table 1

Date 8/19/13 10/14/13 *11/13/13

Global Health Status (QOL) 50 41.66 16.66Physical Functioning 93.33 93.33 73.33Emotional Functioning 66.66 75 25Cognitive Functioning 100 50 50

*Parkinson disease diagnosis suspected and referral made. **Parkinson dise

Purpose/Objective(s): Treatment for head and neck cancer involves a

combination of therapies that are highly toxic impacting both quality of

life (QOL) and overall recovery period. The purpose of this project is to

prospectively monitor QOL in a cohort of head and neck cancer patients

while integrating QOL findings into patient care.

Materials/Methods: QOL is currently being tracked in a head and neck

radiation medicine clinic using the European Organization for Research

and Treatment of Cancer (EORTC) Quality of life (QLQ)-30 and EORTC

head and neck module (H&N)-35 questionnaires which was translated into

a digital platform using a Research Electronic Data Capture (REDCap)

survey format. Patients complete the survey at the beginning of treatment,

end of treatment, and at each follow-up appointment (3 months, 6 months,

1 year). The REDCap survey program build enables automatic computa-

tion of scores upon patient completion. Results are therefore available for

immediate review by the clinician.

Results: A total of 561 QOL surveys have been completed by 200 patients

to date. Patients have completed surveys up to a period of 20 months

following treatment.

In this example the physician was able to diagnose a worsening of the

patient’s Parkinson disease based on the results of the QOL, physical

functioning, emotional functioning, and cognitive functioning scores. The

physician was then able to refer her for immediate and appropriate

therapy.

Conclusion: Prospectively tracking QOL before, during, and after treat-

ment provides clinicians with a more comprehensive understanding of

factors related to changes in their patients’ QOL scores, allowing them to

provide immediate treatment or appropriate referrals. Using a digital

platform (REDCap Surveys) is a novel method for tracking and managing

QOL factors in real time in a head and neck cancer population.

Author Disclosure: A. Platek: None. A. Ostrowski: None. M. Platek:

None. A. Beattie: None. S. Fung-Kee-Fung: None. V. Gupta: None. D.

Cohan: None. W. Hicks: None. H. Arshad: None. M.A. Kuriakose:

None. A.K. Singh: None.

212Basal Cell Adenocarcinoma of the Major Salivary Glands: APopulation-Level Study of 509 CasesK. Zhan and E.J. Lentsch; Medical University of South Carolina,

Charleston, SC

Purpose/Objective(s): We sought to better characterize the demographic,

tumor, and long-term survival characteristics of basal cell adenocarcinoma

(BCAC) of the major salivary glands with the National Cancer Data Base

(NCDB), the world’s largest cancer database.

Materials/Methods: We conducted a retrospective review of the National

Cancer Data Base (NCDB) from 1998 to 2012 for all cases of major

salivary gland basal cell adenocarcinoma (BCAC) with histologic code

8147/3. Relevant demographic, tumor, and survival variables were

extracted for analysis. Cox univariate and multivariate regression analysis

was used to identify predictors of survival.

Results: Out of 36,224 major salivary gland cancers in the NCDB, we

found 509 cases of BCAC (1.4%), 88% of which were in the parotid

glands, 11.2% submandibular, and 0.8% in the sublingual glands. Age at

diagnosis ranged from 18 to 92 years (average 64). No gender preference

was found (50.7% male). Most tumors were 2 to 4 cm in size (47.3%).

Regional (11.9%) and distant metastases (1.8%) were uncommon. Occult

nodal disease was rare (5.7%). Of available grade information, 22.4% were

labeled high grade (9% of all cases). Survival between low- and high-grade

1/29/14 4/23/14 **7/22/14 10/15/14 4/15/15

100 83.33 66.66 91.66 50100 100 73.33 93.33 86.66100 66.66 50 91.66 50100 100 50 100 66.66

ase medication adjustment needed and referral made.

Volume 94 � Number 4 � 2016 Posters 915

BCAC was not significantly different (PZ.336), although they were

significantly different in various tumor and staging features. Overall, early-

stage disease was more common than advanced-staged disease (67.8% vs

32.2%), with stage I disease (43.4%) being most common. Overall survival

at 5 and 10 years was 79% and 62%. While numerous variables were found

to significantly impact survival on univariate regression analysis, only age

�65 years (hazard ratio [HR] 2.55, 95% confidence interval [CI]

1.56-4.19), P<.001) and high T stage (HR 1.85, 95% CI 1.16-2.95,

PZ.010) remained significant prognosticators in our multivariate model.

For high T stage disease, surgery with radiation had significantly better

survival than surgery alone.

Conclusion: Basal cell adenocarcinoma is a rare salivary malignancy with

a good prognosis, traditionally understood as a low-risk malignancy. The

presence of “high-grade” tumors may suggest a more aggressive variant,

although survival was not significantly different. Overall, regional and

distant metastases were uncommon. Radiation with surgery may help for

higher T-stage disease. Old age and high T stage were significant pre-

dictors of worse survival.

Author Disclosure: K. Zhan: None. E.J. Lentsch: None.

213Oncocytic Carcinoma of the Major Salivary Glands: A PopulationStudy of 278 CasesK. Zhan and E.J. Lentsch; Medical University of South Carolina,

Charleston, SC

Purpose/Objective(s): To better characterize disease features and

describe the long-term survival of patients with major salivary gland

oncocytic carcinoma, a poorly understood salivary malignancy with fewer

than 100 cases ever reported.

Materials/Methods: We conducted a retrospective review of the National

Cancer Data Base (NCDB) from 1998 to 2012 for all cases of major

salivary gland oncocytic carcinoma (OC) with histologic code 8290/3.

Relevant demographic, tumor, and survival variables were extracted for

analysis. Cox univariate and multivariate regression was used to identify

predictors of survival.

Results: Out of 36,224 cases of major salivary gland cancer in the NCDB,

we found 278 cases of major salivary OC. Median age was 67 years, with

patients ranging from 16 to 90 years old. Most patients were male (61.9%),

Caucasian (84.8%), and presented with parotid disease (88.5%). Sub-

mandibular gland OC represented 10.5% of cases and none in the sub-

lingual glands. All cases were unilateral. Regional metastasis was found in

36.4% of parotid OC and 50% of submandibular OC. High-grade cancers

had significantly more nodal disease than low (66.7% vs 17.8%, P�.001).Occult nodal disease was found in 10.1% of overall cases. Distant

metastasis was rare (4.7% overall). Overall survival at 5 and 10 years was

64% and 39%. In high-grade lesions, surgery with radiation had signifi-

cantly better survival than surgery alone. In multivariate analysis, distant

metastasis (hazard ratio [HR] 13.87 [2.98-64.52], PZ.001) and regional

metastasis (HR 3.01 [1.20-7.55], PZ.019) are significant negative prog-

nosticators while positive margins approached significance (HR 3.01

[1.20-7.55], PZ.074).

Conclusion: Oncocytic carcinoma has a poor long-term prognosis and

lymph node metastases are common. Distant and regional metastasis are

significant predictors of decreased survival.

Author Disclosure: K. Zhan: None. E.J. Lentsch: None.

214Quality of Life Assessment Through OHIP-14 in Head and NeckCancer Patients Undergoing Anticancer Therapy and DentalTreatmentA.C. Scaraficci, P.S.S. Santos, M.P.L. Battisti, and R. Bastos; Bauru

School of Dentistry - University of Sao Paulo, Bauru, Brazil

Purpose/Objective(s): This study aimed to evaluate, by applying the Oral

Health Impact Profile questionnaire (OHIP-14), the impact of dental

treatment prior to head and neck radiation therapy (HNR), and the impact

of dental treatment during HNR on the quality of life of patients with head

and neck cancer. This study also aimed to evaluate the degree of oral

mucositis and pain symptoms in these patients.

Materials/Methods: The OHIP-14 questionnaire was applied in 4 phases.

The first phase took place before initial dental preparation and prior to the

HNR. The second phase took place after dental preparation and before

starting HNR sessions. The third phase was the period during radiation

therapy treatment. The fourth phase was after completion of both dental

and radiation therapy treatments. The degree of oral mucositis presented

was also rated, according to the WHO scale, and also the pain reported by

the patient, through the Visual Analogue Scale (VAS) during the third

phase.

Results: There was no negative impact on any of the patients when

comparing to the results of the questionnaires applied in the first and

second phases, indicating that the dental treatment contributed to main-

taining or improving the quality of life for them. But when comparing the

results of the questionnaires applied in the third phase with those in the

fourth phase, only 5.55% of patients had a worsening dental condition. By

correlating the data obtained in the third step of applying OHIP-14 and the

degree of oral mucositis found, there was a negative impact with respect to

psychological discomfort and physical disability. The correlation between

the VAS and oral mucositis was 0.63, indicating that the level of pain was

directly related to the degree of oral mucositis.

Conclusion: There was an improvement in quality of life in relation to the

oral health of cancer patients who received dental treatment prior to HNR.

The higher the degree of oral mucositis presented, the greater the negative

impact on the patient’s quality of life and the greater the degree of pain

reported by the patient.

Author Disclosure: A.C. Scaraficci: None. P.S. Santos: None. M.P.

Battisti: None. R. Bastos: None.

215Risk of Second Oral Cavity Cancer and Survival After a Primary Solidor Hematological MalignancyR. Shameem,1 M.S. Hamid,2 J. Bauman,1 T.J. Galloway,1 M. Lango,1

J.A. Ridge,1 and R. Mehra1; 1Fox Chase Cancer Center, Philadelphia, PA,2Wayne State University, Detroit, MI

Purpose/Objective(s): Treatment and outcomes of second primary oral

cavity (sOC) cancer in patients with a prior diagnosis of a nonhead and

neck malignancy are not well understood. We sought to determine dif-

ferences by analyzing data from a nationwide cancer registry.

Materials/Methods: The Surveillance Epidemiology and End Results

database (1973-2012) was used to identify adult cases of malignancies in

the oral cavity (lip, tongue, gum, floor of the mouth, palate, and mouth).

Cases were categorized as localized, regional, or distant using the sum-

mary staging variable. The variable “First Malignant Primary Indicator”

was used to differentiate between de novo oral cavity (dnOC) and sOC

cases. For inclusion, an sOC case was required to be diagnosed after a

latency period of �60 months from a prior diagnosis of hematologic or

solid tumor malignancy (excluding head and neck and nonmelanoma skin

cancer). Overall survival (OS) was calculated using the Cox regression

model to determine the impact of sOC cancer on disease-specific survival,

adjusting for age, stage, race/ethnicity, primary cancer type, latency period,

radiation, and surgery. A X2 test was used to detect for any significant

difference between the 2 groups. Odds ratio (OR) association was adjusted

using multiple logistic regression method.

Results: Overall 2546 (4.1%) sOC and 59,206 (95.9%) dnOC cases were

included. The median interval for developing sOC cancer after a primary

solid or hematological malignancy was 112 months (61-448). All primary

cancer subtypes except for female genital tract (149 months [63-405])

endometrial (142 months [61-432]), chronic lymphocytic leukemia (134

months [65-315]), and melanoma (126 months [61-391]) had a median

latency period of less than 10 years. The most frequent stage at initial

diagnosis for dnOC cancer was regional (46.7%) compared to localized

(53.0%) in sOC cancer, P<.01. Compared to dnOC cases, sOC cases were

International Journal of Radiation Oncology � Biology � Physics916

more likely to occur in the �75 years of age group (45.6% vs 19.2%,

P<.01). Odds of being treated with radiation were less in sOC compared to

dnOC cancer (OR: 0.79, confidence interval [CI]: 0.66-0.96, P<.05). sOC

cancer patients were less likely to receive radiation with a prior diagnosis

of breast cancer (OR: 0.60, 95% CI: 0.39-0.95, P<.05). No differences in

surgical intervention were seen between sOC and dnOC cases. Median OS

was longer in dnOC in comparison to sOC cases (61 vs 43 mo) (hazard

ratio [HR]: 1.19, 95% CI: 1.13-1.27, P<.01). Oral cancer-related survival

was worse in sOC compared to dnOC cases (HR: 1.16, 95% CI: 1.1-1.20).

Conclusion: Elderly age is more common in sOC versus dnOC cancer, and

patients with sOC cancer receive radiation less frequently than do those

with dnOC cancer. Although sOC tumors were more likely to be limited to

local disease, survival was worse in sOC compared to dnOC cancer. Future

research should seek to understand whether treatment decisions for the

sOC are driving this difference in survival.

Author Disclosure: R. Shameem: None. M.S. Hamid: None. J. Bauman:

None. T.J. Galloway: None. M. Lango: None. J.A. Ridge: None. R.

Mehra: None.

216Demographics of Individuals Presenting to Community OralScreenings as Compared to Head and Neck Cancer Patients Seen inan Academic Multidisciplinary ClinicB. Cavanaugh,1 C.A. Perez,1 N.E. Dunlap,1 C.L. Silverman,1 Z. Khan,1

L. Wilson,1 K. Potts,1 P. Tennant,1 J. Bumpous,2 X. Wu,3 S.N. Rai,3

and R. Redman1; 1University of Louisville, Louisville, KY, 2University of

Louisville School of Medicine; James Graham Brown Cancer Center,

Louisville, KY, 3University of Louisville School of Medicine, Louisville, KY

Purpose/Objective(s): Community oral screening for head and neck

cancer is widely promoted, although the utility of such screening is largely

unknown. We sought to characterize the demographics of individuals

presenting for free community oral cancer screenings as compared to the

demographics of head and neck cancer patients seen in an academic

multidisciplinary clinic.

Materials/Methods: Demographics of the screened population were

collected retrospectively from information provided voluntarily by subjects

at the time of screening. All subjects older than 18 years of age screened in

calendar years 2012 and 2013 were included. Patient demographics and

cancer type were collected retrospectively from a prospectively maintained

database of patients seen in an academic head and neck cancer multidis-

ciplinary clinic during the same time period. The discrete characteristics

are compared using a X2 test and continuous characteristics are compared

using a t test for normally distributed data and Wilcoxon rank sum test for

nonnormally distributed data.

Results: A total of 519 screened subjects and 491 cancer patients met

inclusion criteria. The gender distribution of the screened population

differed significantly from that of the cancer patient population, with males

representing 74.08% of cancer patients but only 32.11% of the screened

population (P<.001). Screened subjects were also significantly younger

than the cancer patients (median age 52 years and 61 years respectively,

P<.05). The cancer patients were 3 times more likely than screened in-

dividuals to reside in a zip code classified as rural by the Office of Man-

agement and Budget (18.13% vs 5.03%). Referral rate of screened

individuals (including referral to dentistry, dermatology, or otolaryn-

gology) did not differ significantly by gender, age, or screening location

(hospital, state fair, community health fair, or baseball game). In this

cohort of patients, male cancer patients were more likely than female

patients to present to the multidisciplinary clinic with a diagnosis of

squamous cell carcinoma. Other diagnoses including cutaneous malig-

nancies, thyroid cancer, and salivary gland cancers comprised 40.2% of

presenting diagnoses in female cancer patients.

Conclusion: Individuals presenting voluntarily for community oral cancer

screenings are demographically distinct from patients diagnosed with head

and neck cancer and evaluated at an academic multidisciplinary clinic.

Author Disclosure: B. Cavanaugh: None. C.A. Perez: None. N.E.

Dunlap: None. C.L. Silverman: None. Z. Khan: None. L. Wilson: None.

K. Potts: None. P. Tennant: None. J. Bumpous: None. X. Wu: None.

S.N. Rai: None. R. Redman: None.

217Squamous Cell Carcinoma of the Buccal Mucosa: Clinical Outcomesof Patients Treated at a Tertiary Care Hospital in PakistanY.A. Rashid,1 A.A. Rasheed,1 A.A. Jabbar,1 S. Najeeb,2 S. Awan,1

K.A. Khan,1 B.M. Qureshi,1 and S. Akhtar1; 1Aga Khan University

Hospital, Karachi, Pakistan, 2Aga Khan Medical College, Karachi,

Pakistan

Purpose/Objective(s): Squamous cell carcinoma (SCC) of the buccal

mucosa is one of the commonest oral cavity cancers in South East Asia. It

is usually associated with the use of tobacco, gutka, and betel quid. There

is a dearth of data on behavior and outcomes of these tumors, not only

from this region but also worldwide. The aim of our study is to share the

demographics, prognostic factors, and clinical outcomes of these patients.

Materials/Methods: Retrospective chart review was conducted after approval

from the ethical review committee.All patients with localized/locally advanced

buccal mucosa SCC presenting and receiving treatment at our institute between

January 2006 and December 2013 were included. Patients with lesions origi-

nating from other oral structures extending into buccal mucosa and who did not

complete their treatment at the institute were excluded. Demographic data,

patterns of clinical presentations, risk factors, surgical pathology, and details of

treatment were reviewed. SPSS v19 was used to perform statistical analysis.

Data was analyzed by Kaplan-Meier and log-rank test.

Results: A total of 220 patients were included. One hundred seventy-six

(80.4%) were males and 43 (19.6%) were females. Mean age at presentation

was 48 years (range: 24-83). One hundred eighty-five (84%) patients were

users of tobacco and betel quid. Nonhealing oral ulcer was the main pre-

senting symptom in 161 patients (73.5%) followed by cheek swelling in 41

patients (19.3%). All patients underwent surgery of which 157 (71.7%)

received adjuvant treatment. Pathological stage for patients who underwent

surgery was as follows: stage IZ14.6%, IIZ16.4%, IIIZ23.3%, and

IVAZ45.7%. For patients with stages I, II, III, and IVA disease, median

relapse-free survival rates were 15, 22, 12, and 18 months, respectively

(PZ.006). Median overall survival for stages I, II, III, and IVAwere 57, 67,

61, and 12 months, respectively (PZ.004). Pathological staging/margin

status were found to be independent prognostic factors affecting survival.

One hundred eleven (50.4%) of these patients relapsed. Eighty (72.1%) of

those had loco-regional recurrence while 31 (27.9%) developed distant

metastasis. The main sites of metastasis were lungs (51%) and bones (16%).

Conclusion: Squamous cell carcinoma of the buccal mucosa is a common

malignancy in our region which is strongly correlated with tobacco/betel

quid abuse. Despite optimal treatment more than half of the patients in our

study relapsed. Further studies are required to characterize and optimize

management of this malignancy.

Author Disclosure: Y.A. Rashid: None. A.A. Rasheed: None. A.A. Jab-

bar: None. S. Najeeb: None. S. Awan: None. K.A. Khan: None. B.M.

Qureshi: None. S. Akhtar: None.

218Accuracy of 3-Tesla Magnetic Resonance Imaging for the InitialEvaluation of Tongue CarcinomaY. Patil; University of Cincinnati Medical Center/University of Cincinnati

College of Medicine, Cincinnati, OH

Purpose/Objective(s): With a higher signal-to-noise ratio than 1.5 T, in

theory, our study aims to further define 3T MRI’s role in the initial eval-

uation of oral tongue cancer.

Materials/Methods: In this prospective study, 29 patientswith histologicallyproven squamous cell carcinoma of the tongue underwent preoperative 3T

MRI (2009-2012); 25 patients were included and 4 patients were excluded.

Using 3TMRI scans of the head and neck, tumor thickness was measured or

reconstructed, and cervical lymph node metastases were evaluated.

Results: Mean tumor thickness did not significantly differ between

measured (18.2�7.3 mm) or reconstructed (17.9�7.2 mm) images.

Volume 94 � Number 4 � 2016 Posters 917

Compared with histology findings, mean measured thickness was 8.3 mm

higher by MRI while mean reconstructed thickness was 5.51 mm greater

(P<.001). Correlation between MRI depth of invasion and cervical me-

tastases was confirmed: of 21 patients with 3T MRI showing cervical nodal

metastases, 20 had undergone neck dissection. While no patients with a

3T-MRI depth of invasion <5 mm had cervical metastases, 11 of 21 pa-

tients with thickness >5 mm did at surgery. 3T MRI had 83% sensitivity

and 82% specificity for detecting positive cervical lymph node metastases.

Conclusion: We identified 3T MRI’s effectiveness in evaluating the extent

of squamous cell tongue carcinoma. With accuracy provided by direct

measurement of tumor thickness, reconstructions are unnecessary. 3T MRI

has a higher sensitivity, specificity, and negative predictive value than 1.5T

MRI when predicting nodal stage. 3T MRI was useful for detection of

malignant adenopathy with extracapsular spread.

Author Disclosure: Y. Patil: None.

Abstract 220; Table 1 Five-year LC and OS stratified by TGV cut point of0.2 mL/day

n 5-Year LC 5-Year OS

TGV �0.2 mL/day 23 68% 48%TGV <0.2 mL/day 70 97% 80%

219Predicting Outcomes Using Pre- and Posttreatment PET/CT inLocoregionally Advanced Squamous Cell Carcinoma of the Head andNeckA. Orton,1 J. Zang,2 J. Frandsen,3 M. Dziemainowicz,4 S. Lloyd,1

D.C. Shrieve,3 and Y.J. Hitchcock5; 1University of Utah, Salt Lake City,

UT, 2Xijing Hopsital, Xian 710032, China, 3University of Utah Huntsman

Cancer Institute, Salt Lake City, UT, 4Temple University, Philadelphia, PA,5Huntsman Cancer Hospital, University of Utah, Salt Lake City, UT

Purpose/Objective(s): Definitive chemoradiation therapy (CRT) is the

most common treatment for locoregionally advanced squamous cell car-

cinoma of the head and neck (LASCCHN). Contrast-enhancing (CE)

positron emission tomography/computed tomography (PET/CT) imaging

provides both anatomic and metabolic information that can be used in the

upfront setting to stage patients, as well as after completion of treatment to

assess tumor response. We collected prospective CE PET/CT data to

determine the prognostic utility of the metabolic and anatomic information

provided by these scans.

Materials/Methods: One hundred thirty-seven patients with LASCCHN

underwent pretreatment CE PET/CT with repeat imaging obtained 3

months after therapy. Both the primary tumor and dominant lymph node

were assessed for size in greatest dimension, volume in cm3, and max

SUV. Analysis of the predictive value of each parameter on both the pre-

and posttreatment scan for locoregional recurrence (LRR), distant metas-

tasis (DM), and overall survival (OS) was performed using univariate Cox

regression. Factors found to be significant on univariate analysis were

entered into a multivariate model.

Results: Pre- and posttreatment CE PET/CT data were available for 137

patients from 2005 to 2013. The most common primary sites were

oropharynx (66%), larynx (15%), hypopharynx (7%), and oral cavity (6%).

One hundred twenty-six patients received definitive CRT, 11 patients

received radiation only. Themost common treatment regimenwas 6750 cGy

(simultaneous infield boost) delivered in 225 cGy fractions with concurrent

weekly cisplatin (40 mg/m2). Median follow-up time was 27 months. Over

the study period, 9 patients experienced isolated LRR, 12 patients failed at

distant sites only, 3 patients failed simultaneously at local and distant sites,

and 21 patients died. Persistently positive PET after definitive CRT was

predictive of lower OS (hazard ratio 8.2, P<.001). The overall LRC rate was

91%. On univariate analysis, imaging factors that predicted for LRR were

limited to findings on posttreatment CE PET/CT. These included greatest

dimension of primary tumor measured by CTand max PET SUVof primary

tumor and dominant lymph node. On multivariate analysis, only the post-

treatment size of the primary tumor remained independently predictive of

LRR. The posttreatment size of the primary tumor and max SUV of the

primary tumor and dominant node were found to be significantly associated

with DM on univariate analysis. Only the posttreatment max SUV of the

primary tumor remained significant on multivariate analysis.

Conclusion: Persistently elevated metabolic tumor activity on post treat-

ment PET is predictive of DM and OS, while persistent anatomic CT

abnormality is associated with LRR. Pretreatment imaging was useful for

staging but was not predictive of outcome of therapy.

Author Disclosure: A. Orton: None. J. Zang: None. J. Frandsen: None.

M. Dziemainowicz: None. S. Lloyd: None. D.C. Shrieve: None. Y.J.

Hitchcock: None.

220Impact of Pretreatment Volumetric Tumor Growth Velocity onOncologic Outcomes in Oropharyngeal Squamous Cell CancerS. Perni,1 A.S.R. Mohamed,2 A.S. Garden,2 G.B. Gunn,2 D.I. Rosenthal,2

J.G. Scott,3 and C.D. Fuller2; 1Columbia University College of Physicians

and Surgeons, New York, NY, 2The University of Texas MD Anderson

Cancer Center, Houston, TX, 3H. Lee Moffitt Cancer Center, Tampa, FL

Purpose/Objective(s): Volumetric tumor growth velocity (TGV) reflects

in vitro tumor aggressiveness, but its prognostic value has not been

empirically investigated in vivo. We hypothesized that higher pretreatment

TGV would have negative prognostic effects on oncologic outcomes in

patients with oropharyngeal squamous cell cancer.

Materials/Methods: We performed a retrospective analysis of 3-dimen-

sional TGV in patients with oropharyngeal cancer treated with radiation at

a single institution between 2004 and 2008. Ninety-three patients met

inclusion criteria of squamous histology, 2 pretreatment CTs with radio-

graphically visible tumors, and scan time gap of more than 2 weeks.

Volumetric data were collected using commercial software to segment

primary tumor targets. Linear growth rates were calculated from the serial

scans. Demographic, treatment, and outcome information was obtained

from patient charts. Recursive partitioning analysis was used to identify

cut points associated with outcomes. Kaplan-Meier calculations were used

to evaluate disease control and overall survival (OS). Comparisons be-

tween groups were made using Wilcoxon tests. Cox regression, univariate,

and multivariate analyses were also performed.

Results: Median follow-up was 59 months (range 7-118). The majority of

patients (60%) underwent concurrent chemoradiation, 18% underwent

definitive radiation, and 22%underwent induction chemotherapy followed by

radiation or chemoradiation. Mean pretreatment TGV was 0.16�0.23 mL/

day.Recursive partitioning analysis identifiedvelocity cut pointof 0.2mL/day

associated with local recurrence. Patients with TGV�0.2 mL/day had

significantly worse 5-year local control (LC; P<.0001) and OS (P<.0001)

than patients with lower TGV, as shown in Table 1 below.

Univariate analysis also showed that higher TGV predicts worse LC

(hazard ratio [HR] 13.0; 95% confidence interval [CI] 3.1-87.5, PZ.0003)

and OS (HR 3.8; 95% CI 1.7-8.5, PZ.001). Multivariate analysis of age,

sex, ethnicity, T stage, overall American Joint Committee on Cancer stage,

and human papillomavirus status suggested that higher TGV is an inde-

pendent predictor of LC (HR 11.6; 95% CI 2.2 e 95.0, PZ.003) and OS

(HR 3.0; 95% CI 1.2-7.6, PZ.02).

Conclusion: Oropharyngeal squamous TGV�0.2 mL/day is a substantive

negative prognostic indicator for LC and OS. This novel quantitative

CT-based volumetric assessment of tumor growth velocity suggests a

simple methodology for stratification of oropharyngeal squamous cell

cancer patients into a distinct risk group for which treatment and screening

strategies could be optimized.

Author Disclosure: S. Perni: None. A.S. Mohamed: None. A.S. Garden:

None. G.B. Gunn: None. D.I. Rosenthal: None. J.G. Scott: None. C.D.

Fuller: Research Grant; NIH/NCI, SWOG/Hope Foundation, Elekta AB/

MD Anderson Cancer Center, MD Anderson Cancer Center. In-kind

International Journal of Radiation Oncology � Biology � Physics918

Donation; General Electric Healthcare/MD Anderson Cancer Center.

Clinical Scientist Loan Repayment Program; NIH/NCI.

221WITHDRAWN

222The Value of Positron Emission TomographyeComputedTomography in Predicting Occult Nodal Metastasis in RecurrentLaryngeal CancerA. Rosko,1 A. Birkeland,1 A. Shuman,2 M. Prince,2 C. Bradford,2

G.T. Wolf,2 F.P. Worden,1 A. Eisbruch,1 A. Srinivasan,1 and M.E. Spector2;1University of Michigan, Ann Arbor, MI, 2University of Michigan

Comprehensive Cancer Center, Ann Arbor, MI

Purpose/Objective(s): Squamous cell carcinoma (SCC) of the larynx is

routinely treated with radiation (XRT) alone or concurrently with

chemotherapy (CRT) with the goals of disease cure and organ preservation.

Many patients, however, develop local recurrence after XRT or CRT and

require salvage laryngectomy. There is controversy regarding the role of

neck dissection in salvage laryngectomy as the frequency of occult nodal

disease in clinically N0 patients has been reported to be from 3% to 17%,

with a higher rate in T4 (34%) and supraglottic tumors (28%). Positron

emission tomographyecomputed tomography (PET-CT) is increasingly

used in laryngeal SCC to evaluate treatment response, detect recurrence,

and aid in staging. The objective of this study was to evaluate the pre-

dictive value of PET-CT in identifying occult nodal metastasis in patients

undergoing salvage laryngectomy.

Materials/Methods: A retrospective review of 46 clinically N0 patients

with no nodal disease on physical exam or other imaging modalities (CT or

magnetic resonance) who underwent salvage laryngectomy with neck

dissection from January 1, 2002 to December 31, 2014 was performed.

Patients were included if a PET-CT was performed prior to surgery.

PET-CT positivity was determined by the interpretation of the reading

radiologist, and radiologic interpretations of “suspicious” or “positive”

were considered positive. Of the patients, 41.3% (19) were initially treated

with XRT, 54.3% (25) were treated with CRT, and 4.3% (2) were treated

with excision followed by adjuvant XRT or CRT. Of the recurrent tumors,

52.2% (24) were located in the supraglottis, 45.6% (21) in the glottis, and

2.2% (1) in the subglottis. The recurrent tumor was staged as a T4 in

41.3% (19) of patients, T3 in 32.6% (15) of patients, T2 in 21.7% (10) of

patients, and T1 in 4.3% (2) patients.

Results: There were 12 (26.1%) patients with nodal metastasis on final

pathology. Of these patients, 3 (25%) were detected by PET-CT whereas 9

(75%) of patients had a false negative scan. There were 34 (73.9%)

patients without nodal disease, 1 (2.9%) of whom had a false positive

PET-CT while 33 (97.1%) had a true negative PET-CT. The sensitivity of

PET-CT in detecting occult nodal metastasis when compared to final pa-

thology of neck dissection specimens was 25.0% (95% confidence interval

[CI], 5.5%-57.2%) with a specificity of 97.1% (95% CI, 84.7%-99.9%),

positive predictive value (PPV) of 75% (95% CI, 19.4%-99.4%), and

negative predictive value (NPV) of 78.6% (95% CI, 63.2%-89.7%).

Conclusion: PET-CT has reasonable specificity but poor sensitivity in

detecting occult nodal disease. The NPV of 78.6% makes PET-CT an

imperfect predictor of nodal disease in the setting of recurrent laryngeal

cancer.

Author Disclosure: A. Rosko: None. A. Birkeland: Employee; University

of Michigan. A. Shuman: Employee; University of Michigan. M. Prince:

Abstract 222; Table 1 PET-CT test results compared to final pathology.

Positive nodes Negative nodes

Positive PET-CT 3 1Negative PET-CT 9 33

None. C. Bradford: Guided the organization and its endeavors; American

Head and Neck Society. Attend meetings and help mold health initiatives;

National Institute of Health. G.T. Wolf: None. F.P. Worden: None. A.

Eisbruch: None. A. Srinivasan: None. M.E. Spector: None.

223The Long-Term Predictive Value of Posttreatment PositronEmission TomographyeComputed Tomography Imaging in Head andNeck Squamous Cell CarcinomaS. Suppiah,1 M.A. Michel,2 M.E. Stadler,3 B.L. Massey,3 B.H. Campbell,3

S. Wong,4 D. Wang,5 C.J. Schultz,1 and J.R. Robbins1; 1Medical College of

Wisconsin, Department of Radiation Oncology, Milwaukee, WI, 2Medical

College of Wisconsin, Department of Radiology, Milwaukee, WI, 3Medical

College of Wisconsin, Department of Otolaryngology and Communication

Sciences, Milwaukee, WI, 4Medical College of Wisconsin, Department of

Hematology and Oncology, Milwaukee, WI, 5Rush University Medical

Center, Department of Radiation Oncology, Chicago, IL

Purpose/Objective(s): 18F-fluorodeoxyglucoseepositron emission to-

mography (PET) scans can be a helpful tool in managing head and neck

squamous cell carcinoma. The purpose of this study is to evaluate the

predicative significance of a negative posttreatment PET scan performed

on long-term outcomes.

Materials/Methods: One hundred twenty-nine patients who received

radiation therapy with or without chemotherapy at our institution from

2002 to 2011 and had posttreatment PET/CT imaging were retrospec-

tively reviewed under an institutional review boardeapproved protocol.

Median age at diagnosis was 58 years (range: 25 - 81 years), and there

were 97 males and 32 females. Primary site was oral cavity (3%),

oropharynx (68%), hypopharynx (8%), larynx (18%), and unknown

(3%), and the following disease stages were present: stage I (7%), stage

II (8%), stage III (27%), and stage IV (58%). The median dose deliev-

ered was 70 Gy (range 66-72 Gy). Ninety were treated with concurrent

chemotherapy. The median time from the end of therapy to PET/CT was

4.1 months. Posttreatment PET/CT scans were simply categorized as

negative (no residual FDG avidity), positive (residual avidity or

increased FDG avidity at initial or new sites), or equivocal based on

initial radiological report.

Results: Median follow-up was 52 months (range: 4 - 142 months). By

response on posttreatment PET, 71 patients were classified as negative, 45

as positive, and 13 equivocal. Overall disease recurrence (local, locore-

gional, or distant) was identified in 5 patients (7%) in the PET-negative

(PN) group, 24 patients (53%) in the PET-positive (PP) group, and 5 pa-

tients (38%) in the PET-equivocal (PE) group. For disease recurrence, a

negative posttreatment PET scan had a negative predictive value of (NPV)

of 95%. Most recurrences occurred within 2 years of treatment (87.5%).

The median time for recurrence was 20.5 months, 6.6 months, and 12.2

months, respectively for the PN, PP, and PE groups. The 2-year recurrence-

free survival was highest for PN patients; 95.6% compared to 47.7%, and

59.2% for PP and PE groups (P<001). Likewise the 5-year disease-specific

survival (DSS) and overall survival (OS) were also significantly higher for

PN patients; 92.6% and 88.1% for PN, 49.9% and 41.2% for PN, and both

58.7% for PE (P<.001 for both DSS and OS).

Conclusion: For patients treated with radiation therapy for head and neck

squamous cell carcinoma, a negative surveillance PET/CT completed

around 4 months after treatment portends good long-term outcomes with

excellent disease control and survival with limited recurrences. A negative

posttreatment PET has an NPV of 95% for recurrence after long-term

follow-up. While not all patients with a positive or an equivocal post-

treatment PET developed tumor recurrence, close follow-up and early

intervention should be considered for this population.

Author Disclosure: S. Suppiah: None. M.A. Michel: None. M.E. Stadler:

None. B.L. Massey: None. B.H. Campbell: None. S. Wong: None. D.

Wang: None. C.J. Schultz: None. J.R. Robbins: None.

Abstract 225; Table 1

Imaging Recurrence control

DW-MRI positive 24 8 32DW-MRI negative 1 27 38

35 35 70FDG-PET/CT positive 34 19 53FDG-PET/CT negative 1 16 17

35 35 70

Volume 94 � Number 4 � 2016 Posters 919

224Longitudinal Characterization of MRI Kinetics in IrradiatedDysphagia-Related Structures for Nasopharyngeal CarcinomaPatients Receiving IMRTJ.A. Messer,1 A.S.R. Mohamed,1 K.A. Hutcheson,1 Y. Ding,2 J. Wang,2

H. Eichelberger,3 C. French,3 D.I. Rosenthal,1 G.B. Gunn,2

and C.D. Fuller2; 1The University of Texas MD Anderson Cancer Center,

Houston, TX, 2MD Anderson Cancer Center, Houston, TX,3The University of Texas Medical School, Houston, TX

Purpose/Objective(s): Radiation therapy (RT) for nasopharyngeal carci-

noma (NPC) is often effective at curing disease but can injure the sur-

rounding organs at risk, including the muscles responsible for swallowing,

with possible resultant short-term or long-term dysphagia. However, no

study has tracked serial (ie, acute and late) quantitative dose-response

magnetic resonance imaging (MRI) parameter kinetics in a uniform NPC

dataset. We aim to characterize serial MRI signal intensity (SI) changes in

dysphagia-associated volumes of interest (VOIs) as a function of the

radiathin therapy dose.

Materials/Methods: In this retrospective study, we extracted data on 77

patients with stage III-IV NPC who had been treated with curative in-

tensity modulated RT (IMRT). The mean T1- and T2-weighted MRI SIs

were recorded for the superior pharyngeal constrictor (SPC) and soft palate

(SP) at baseline, early-after IMRT, and last follow-up, with normalization

to reference structures receiving <5 Gy. RT dose grids were restored for

dose response analysis. Statistical methods included a nonparametric

analysis test and recursive partitioning analysis (RPA).

Results: The median time to early post-RT follow-up was 4 months, and the

median time to late post-RT follow-up was 41 months. The mean dose to the

SPCwas 62.4Gy (standard deviation [SD], 8.7Gy), and themean dose to the

SP was 66.8 Gy (SD, 7.3 Gy). All structures had a significant increase in T2

SIs early after treatment compared to baseline, irrespective of the mean dose

given (SPC and SP, 0.47�0.12 and 0.56�0.12 at baseline vs 0.73�0.18 and

0.82�0.17, respectively,P<.0001 for both). At last follow-up, the increase in

T2 SI subsided completely for SPC and partially for SP. The T1 SI did not

change significantly in early follow-up images of both structures; on late

follow-up, patients with mean doses >62.25 Gy had significant decrease in

the corresponding T1 SI for SPC (1.6� 0.4 vs 1.3� 0.4,PZ.007) compared

to baseline but decreased nonsignificantly for SP (1.7�0.5 vs 1.6�0.5,

PZ.09). No significant changes in T1 SI were noted with doses below 62.25

Gy for both structures. Continuous RPA showed a cutoff value of magnitude

0.57 for alterations in T1 SI, with a Dmean of 63.8 Gy (95% confidence in-

terval [CI], 61.6-66.0) for those with decrease >0.57 compared to 56.7 Gy

(95% CI, 52.2-61.1) for those not achieving threshold. A sigmoidal fit was

used to create a normal tissue complication probability curve for T1 alter-

ations as a function of dose (observed R2Z0.928).

Conclusion: Serial MRI acquisitions enable the identification of both early

and late radiation-induced changes in swallowing structures after definitive

IMRT for NPC. Decreased SI on late T1 images may indicate muscle

fibrosis and is associated with higher RT doses to the SPC, while increased

SI on early T2 images is associated with acute edema that subsides after

therapy.

Author Disclosure: J.A. Messer: None. A.S. Mohamed: None. K.A.

Hutcheson: None. Y. Ding: None. J. Wang: None. H. Eichelberger:

None. C. French: None. D.I. Rosenthal: None. G.B. Gunn: None. C.D.

Fuller: None.

225Prospective Comparative Study of Diffusion-Weighted MRI VersusFDG-PET for Detection of Recurrence After (Chemo)radiation forHead and Neck Squamous Cell CarcinomaJ. Driessen,1 C. Terhaard,2 M. Philippens,2 and W. Grolman3; 1UMCU,

Utrecht, Netherlands, 2UMC Utrecht, Utrecht, Netherlands, 3UMCU,

Utrecht, Netherlands

Purpose/Objective(s): High-dose (chemo)radiation for head and neck

squamous cell carcinoma (HNSCC) may result in late edema and necrosis,

resembling recurrent disease. FDG-positron emission tomography/

computed tomography (PET/CT) has a high negative predictive value for

recurrent disease; however, it is limited by the positive predictive value.

The diagnostic accuracy of conventional magnetic resonance imaging

(MRI) with diffusion-weighted (DW) MRI to detect a local recurrence has

been compared with the standard FDG-PET/CT.

Materials/Methods: Seventy-four patients clinically suspected of local

recurrence after (chemo)radiation for laryngeal, hypopharyngeal, or

oropharyngeal cancers were prospectively included in this study and un-

derwent an MRI including diffusion-weighted imaging (DW-MRI) and an

FDG-PET/CT. Qualitative assessment of DW-MRI and FDG-PET/CT was

performed by an experienced radiologist resp. nuclear physician blinded

for the other modality. Reference standard was the absence of a biopsy-

proven local recurrence within 6 months following imaging.

Results: Four patients were excluded due to contraindications or disrup-

tion of the MRI (eg, claustrophobia or patient stature). Seventy-three

percent (51 of 70) of the FDG-PET/CTs were positive compared to only

46% (32 of 70) of the DW-MRI. FDG-PET/CT had a diagnostic accuracy

of 72% compared to 73% for MR-DWI. The negative predictive value of

FDG-PET/CT was 94% compared to 71% for MR-DWI. The positive

predictive value of FDG-PET/CTwas 64% compared to 75% for DW-MRI.

See Table 1 for the complete results.

Conclusion: In this study, DW-MRI showed superior positive predictive

value but inferior negative predictive value compared to FDG-PET/CT.

False negative results will cause delay in the detection of recurrence and

therefore will potentially influence the chance of successful salvage sur-

gery. Therefore, based on these results, we consider FDG-PET/CT to be

superior to MR-DWI in the follow-up of HNSCC after (chemo)radiation

therapy. Future improvements in DWI techniques which will decrease

artefacts and enhance contrast, may enable DWI to resemble FDG-PET/CT

negative predictive value without compromising the higher positive pre-

dictive value of DWI.

Author Disclosure: J. Driessen: None. C. Terhaard: None. M.

Philippens: None. W. Grolman: None.

226Assessment of Laryngeal Motion Dynamics Using 4D-ComputedTomography and Dynamic Magnetic Resonance ImagingH. Bahig, P.F. Nguyen-Tan, E.J. Filion, D. Blais, D. Roberge, J. de Guise,

and L. Lambert; Centre Hospitalier de l’Universite de Montreal, Montreal,

QC, Canada

Purpose/Objective(s): With increasing interest in reduced-volume in-

tensity modulated radiation therapy (IMRT) and intrafraction image

guidance for early glottic cancer, we sought to better understand laryngeal

motion. The aim of our study was to determine the dynamics of laryngeal

motion during the course of a radiation treatment using combined 4-

dimensional computed tomography (4D-CT) and cine- magnetic resonance

imaging (MRI) information.

Materials/Methods: This prospective study included patients with T1-2N0glottic cancer treated with radical radiation therapy. Dynamic sagittal MRI

was obtained pretreatment and midtreatment to assess for inadvertent

swallowing frequency and respiratory motion. Pre- and midtreatment 4D-

CT allowed for assessment of larynx excursion during swallowing and

breathing as well as evaluation of gross tumor volume, clinical target

volume (CTV), and planning target volume (PTV) coverage during mo-

tion. In addition, bone registration of simulation CT with pre- and

International Journal of Radiation Oncology � Biology � Physics920

midtreatment 4D-CT allowed for assessment of larynx resting position

shift compared to simulation CT. Student t and Fisher tests were used for

statistical analysis.

Results: Twenty patients were included. Pretreatment median swallowing

frequency over 2 minutes was 1 time (0-5) without instruction versus 0 (0-

1) with instruction not to swallow, PZ.03. Midtreatment median fre-

quency was 0, with or without instruction. On 4D-CT, median amplitude of

deglutition was 22 mm (14-30), 0 mm (0-3), 6 mm (3-9), and 0 mm (0-3)

in the superior, inferior, anterior, and posterior directions, respectively;

these correlated well with amplitudes seen on MRI. Median swallowing

duration was 2.4 (1.3-3.1) seconds. Assuming 1 deglutition during a 2-

minute treatment and a 1-cm craniocaudal PTV, the CTV would spend a

median of 1.7 seconds (0-2) outside of the 95% isodose volume. Median

amplitude of respiratory laryngeal motion was 4 mm (2-6) and 2 mm (1-2)

in the SI and anteroposterior (AP) directions. There were no statistically

significant differences between pre- and midtreatment amplitudes. Mid-

treatment 4D-CT identified a shift in larynx resting position compared to

simulation CT in 40% of cases (median 4 [4-8] mm in SI and 3 [2-4] mm

in AP).

Conclusion: Our study supports minimal occurrence and minimal dosi-

metric impact of swallowing during treatment; PTV volumes need not

account for the entire swallowing motion. Respiratory motion up to 6 mm

was identified and should be taken into account when considering reduced

volume IMRT planning. A larynx shift, potentially related to an anatomy

change, occurs during treatment in a significant proportion of patients and

warrants for the use of laryngeal cone beam CT match rather than bone

match. Our data will help define PTV margins for a future phase 2 study of

single vocal cord IMRT.

Author Disclosure: H. Bahig: None. P. Nguyen-Tan: None. E.J. Filion:

None. D. Blais: None. D. Roberge: None. J. de Guise: None. L.

Lambert: None.

227Correlating Clinical Outcomes With Changes in Tumor Volume and18F-FDG PET Characteristics During Radiation Therapy for Head andNeck Squamous Cell Carcinoma (HNSCC)Y.M. Mowery,1 I. Vergalasova,1 D.S. Yoo,1 S.K. Das,2 W. Hara,3

and D.M. Brizel1; 1Duke University Medical Center, Durham, NC,2University of North Carolina at Chapel Hill, Chapel Hill, NC,3Stanford Radiation Oncology, Stanford, CA

Purpose/Objective(s): Early indication of treatment failure may guide

therapeutic intensification. We evaluated changes in tumor volume and

standard uptake value (SUV) parameters during (chemo)radiation therapy

for HNSCC. We hypothesized that greater decreases in tumor size and

SUV during treatment would correlate with more favorable clinical

outcome.

Materials/Methods: HNSCC patients scheduled for definitive (chemo)

radiation therapy were enrolled in an institutional review board-

eapproved study. Positron emission tomographic/computed tomographic

(PET/CT) scans were performed at initial simulation and after 20 Gy.

Gross tumor volumes for primary tumor (GTVP) and lymph nodes

(GTVN) were delineated on pre- and intratreatment CT scans coregis-

tered to PET images. Primary and nodal tumor volumes, as well as

SUVmax and SUVmean of GTVP and GTVN were measured. SUV

Abstract 227; Table 1

All (nZ72)

Incompleteresponse(nZ10)

Completeresponse,(nZ62) P

D GTVP -46 (-55, -40) -46 (-81, 13) -46 (-55, -40) .9D GTVN -36 (-44, -28) -21 (-40, 11) -43 (-49, -29) .01D SUVmax of GTVP -39 (-45, -33) -31 (-46, -13) -39 (-45, -33) .6D SUVmax of GTVN -35 (-45, -28) -28 (-63, 12) -37 (-47, -31) .6D SUVmean of GTVP -29 (-38, -23) -28 (-46, 7) -28 (-38, -20) .9D SUVmean of GTVN -30 (-35, -23) -28 (-50, 12) -31 (-36, -20) .8

parameters were normalized to blood pool uptake. Treatment response

and recurrence were assessed by diagnostic PET/CT at 12 weeks post-

treatment, serial history/physical exam, and biopsy if indicated. Signif-

icance of relative changes from baseline was assessed by Wilcoxon

signed rank test. Distribution of measurements between groups was

compared by Mann-Whitney test.

Results: Seventy-two patients were evaluated: 51 oropharynx (44 p16

positive, 2 p16 negative, 5 unknown), 4 oral cavity, 7 larynx, 4 hypo-

pharynx, and 6 nasopharynx. Sixty-five received concurrent chemotherapy.

Stage distribution was 1 stage I, 3 stage II, 10 stage III, 53 stage IVA, and 5

stage IVB. Median follow-up was 18 months (interquartile range 9.6-25).

Table 1 shows median percent relative change (95% confidence interval)

on intra- versus pretreatment scans. Primary and nodal tumor volumes,

SUVmax, and SUVmean decreased significantly from baseline (P<.0001).

Except for GTVN, relative changes were not associated with complete

response posttreatment. Twelve recurrences have occurred (3 locoregional,

5 locoregional and distant, 4 distant). Pretreatment tumor volumes and

SUV parameters were not associated with outcome. Intratreatment

SUVmax of GTVP (median 6.2 vs 8, PZ.05), SUVmean of GTVP (2.5 vs 4,

PZ.02), SUVmax of GTVN (4.4 vs 7.2, PZ.02), and GTVN (7.2 vs 21.6 cc,

PZ.004) were significantly lower for patients with complete treatment

response and no subsequent recurrence.

Conclusion: Tumor size and SUV decreased significantly after 20 Gy.

Although this study is limited by sample size, the results suggest that lower

intratreatment SUVmax and SUVmean of GTVP, SUVmax of GTVN, and

nodal tumor volume are associated with improved outcome. Analyses are

ongoing to assess whether metabolic tumor volume and other SUV

parameters predict for clinical outcomes.

Author Disclosure: Y.M. Mowery: None. I. Vergalasova: None. D.S. Yoo:

None. S.K. Das: None. W. Hara: None. D.M. Brizel: None.

228Metabolic Tumor Volume Changes Measured by 18F-FDG-PET/CTAfter 1 Cycle of Induction Chemotherapy Is an Early Predictor ofRadical Chemoradiation Therapy Outcome in Head and NeckSquamous Cell CarcinomaK.H. Wong,1 L. Welsh,2 D. Mcquaid,2 A. Dunlop,2 I. Murray,2 Y. Du,2

S. Chua,2 R. Panek,2 A. Riddell,2 D.M. Koh,2 S. Bhide,1 C. Nutting,2

K. Harrington,1 and K. Newbold2; 1The Institute of Cancer Research,

Sutton and London, United Kingdom, 2The Royal Marsden NHS

Foundation Trust, Sutton and London, United Kingdom

Purpose/Objective(s): To assess the predictive value of 18F-FDG positron

emission tomography/computed tomography (PET/CT) parameter changes

during induction chemotherapy (IC) for response to radical chemoradiation

therapy (CRT) in head and neck squamous cell carcinoma (HNSCC).

Materials/Methods: This is an ongoing prospective, single-institution

study in which patients underwent 18F-FDG-PET/CT with thermoplastic

shell immobilization before and 2 weeks following each cycle of IC (first

cycle, IC1; second cycle, IC2). Following IC, patients received radical

CRT (65 Gy in 30 fractions) over 6 weeks with concomitant chemotherapy

on days 1 and 29. Treatment response was assessed at 3 months from

completion of CRT with clinical examination, magnetic resonance imag-

ing, and 18F-FDG-PET/CT. Patients with evidence of residual or pro-

gressive disease were classed as nonresponders. Reductions in tumor

SUVmax (maximum standard uptake value) and MTV (metabolic tumor

volume with a pre-determined SUV threshold of 3.5) following IC were

compared between responders versus non-responders with Mann-Whitney

U test. The significance threshold was set at P<.05.

Results: Twenty patients with stage III-IVA HNSCC were included in this

preliminary analysis. The median age was 63 years (47-69). All patients

underwent 2 cycles of IC except 3 patients who stopped after 1 cycle due

to poor tolerance. One patient did not receive concomitant chemotherapy

due to persistent myelosuppression. In 17 evaluable patients, there was no

significant difference in the changes from baseline between IC1 and IC2

for MTV (PZ.80) and SUVmax (PZ.10), indicating that most metabolic

response to IC was observed early with little changes with subsequent

Volume 94 � Number 4 � 2016 Posters 921

cycle. There were 15 responders and 5 nonresponders (all had local failure

with 2 also having distant metastases). Pretreatment MTV differed

between the 2 groups (21.9�16.5 cm3 vs 56.7�24.6 cm3, PZ.01) but not

SUVmax (PZ.17). Responders showed a significantly greater mean MTV

reduction following IC1: 84.3�19.4% vs 29.6�55.8% than did non-

responders (PZ.03). The SUVmax reduction was not discriminative be-

tween the 2 groups following IC1 (PZ.12) but did reach statistical

significance following IC2 (PZ.01). A combination of pretreatment

MTV<50 cm3 and MTV reduction >60% following IC1 gives a sensi-

tivity, specificity, and positive and negative predictive value of 87%, 80%,

93%, and 67%, respectively in predicting complete response to CRT.

Conclusion: Our preliminary data suggested that ascertaining MTV

response (or lack of it) as early as IC1 could help to identify patients at risk

of treatment failure in a timely way, which allows more scope to consider

alternative treatment strategies such as radiation therapy dose escalation or

surgery. On the other hand, SUVmax response to IC was more gradual and

unsuitable for early assessment, that is, prior to IC2. This study will be

further reported in the future with more patients and 2 years of disease-free

survival.

Author Disclosure: K.H. Wong: None. L. Welsh: None. D. Mcquaid:

None. A. Dunlop: None. I. Murray: None. Y. Du: None. S. Chua: None.

R. Panek: None. A. Riddell: None. D. Koh: None. S. Bhide: None. C.

Nutting: None. K. Harrington: None. K. Newbold: None.

229Prognostic Value of 18-Fluorodeoxyglucose in IndependentTraining and Validation Sets of Patients With HNSCC LargelyExplained by Association With Tumor VolumeJ.H. Rasmussen,1 I. Vogelius,2 J. Kjems,2 B.M. Fischer,3 S.M. Bentzen,4

and L. Specht2; 1Department of Otorhinolaryngology, Head & Neck

Surgery and Audiology, Rigshospitalet, University of Copenhagen,

Copenhagen, Denmark, 2Department of Oncology, Section of

Radiotherapy, Rigshospitalet, University of Copenhagen, Copenhagen,

Denmark, 3Department of Clinical Physics, Nuclear Medicine & PET, PET

& Cyclotron Unit, Rigshospitalet, University of Copenhagen, Copenhagen,

Denmark, 4Division of Biostatistics and Bioinformatics, University of

Maryland Greenebaum Cancer Center, and Department of Epidemiology

and Public Health, University of Maryland School of Medicine,

Baltimore, WA

Purpose/Objective(s): A prognostic model including 18-fluorodeox-

yglucose positron emission tomography (FDG-PET) performed better in a

training set of 287 patients than a model omitting FDG uptake. Here, we

validate the prognostic models in an independent validation cohort of 327

patients.

Materials/Methods: The previously published training set included 287

patients treated from 2005 to 2009. The validation cohort included 327

patients treated from (later) 2009 to 2012. All patients had squamous cell

carcinoma of the head and neck (HNSCC) and were treated with

intended radical (chemo)radiation therapy. Univariate and multivariate

analyses were performed using the Cox model with any failure as an

event. Maximum standard uptake value (SUVmax) and gross tumor

volume (GTV) were entered as continuous variables, with hazard ratio

(HR) specified per interquartile range. Smoking status and prescription

of concomitant cisplatin were also entered. Association between vari-

ables was tested using Spearman rank correlation. The trained prognostic

model was used to generate 4 risk groups based on the predicted risk of

Abstract 229; Table 1

Univariatetraining

set HR (CI)

Multivariatetraining

set HR (CI)

Univariatetest setHR (CI)

Multivariatetest setHR (CI)

GTV 1.7 (1.4-2.1) 1.6 (1.3-1.9) 1.7 (1.4-2.1) 1.9 (1.5-2.4)SUVmax 2.6 (1.7-4.1) 1.7 (1.0-2.9) 1.8 (1.3-2.5) 1.1 (0.7-1.5)Tobacco 1.9 (1.4-2.6) 1.6 (1.2-2.2) 1.8 (1.4-2.5) 1.7 (1.2-2.3)Cisplatin 0.4 (0.3-0.5) 0.4 (0.3-0.6) 0.6 (0.4-0.9) 0.5 (0.3-0.7)

any failure after 2 years (0%-10%, 10%-30%, 30%-60%, and >60%).

The Kaplan-Meier method was used to compare outcomes in the 4 risk

groups.

Results: The median follow-up was 32 and 29 months in the training set

and in the test set, respectively, and 120 of 121 patients relapsed

(training/test). In a univariate analysis, SUVmax was highly significant

in both the training and the test set (P<.001). The magnitude of

HRSUVmax decreased markedly when moving from univariate to multi-

variate analysis (see Table 1) with a test for difference on the pooled

data yielding PZ.01. This is explained by the significant correlation

between tumor volume, GTV, and SUVmax in both sets (rZ0.373 and

rZ0.339 in the training and test set, respectively; P<.0001 in both sets).

SUVmax did not show independent prognostic value in the test set after

adjusting for tumor volume, but there was no significant difference

between the HRs for high SUV uptake in the test set and training set

(PZ.19). The 4 risk groups developed from the training set were

validated in the test set with survival curves from the Kaplan-Meier

analysis. The observed 2-year freedom from failure in the 4 risk groups

generated by the model in the test set were 90%, 81%, 61%, and 30%.

A simpler model, ignoring FDG uptake, provided equivalent

prognostication.

Conclusion: The prognostic value of the original model was validated in

the independent validation set, but the effect size associated with FDG

uptake is overestimated if tumor volume is not appropriately accounted

for.

Author Disclosure: J.H. Rasmussen: None. I. Vogelius: None. J. Kjems:

None. B.M. Fischer: None. S.M. Bentzen: None. L. Specht: None.

230Dynamic Optical Contrast Imaging as a Novel Modality to RapidlyDistinguish Oral Squamous Cell Carcinoma From SurroundingNormal TissueZ. Taylor,1 I. Kim,1 J. Pesce,1 W. Grundfest,1 and M.S. St. John2; 1UCLA,

Los Angeles, CA, 2University of California, Los Angeles, Los Angeles, CA

Purpose/Objective(s): Head and neck cancers are debilitating diseases

for which patient prognosis depends heavily on complete tumor resection.

Currently, it is the surgeon’s fingers and eyes that determine the location of

tissue margins. An intraoperative instrument that can significantly improve

the accuracy of margin detection over current methods will improve out-

comes for cancer patients by minimizing removal of normal functional

tissue while also ensuring complete tumor removal. The objective herein is

to demonstrate the utility of dynamic optical contrast imaging (DOCI) in

reliably and accurately delineating tumor tissue from surrounding normal

tissues.

Materials/Methods: Oral squamous cell carcinoma (OSCC) specimens

and surrounding tissues from the surgical bed were collected; fluorescence

decay images were acquired using a wide-field DOCI system. Samples

from 55 patients were subsequently processed for standard histological

assessment by head and neck pathologists. Mean relative fluorescence

decay signatures were calculated for tumor, fat, muscle, and collagen

tissues. Statistical analyses were performed using the Wilcoxon signed

rank test.

Results: Qualitative analysis of DOCI images revealed microscopic

characterization sufficient for tissue type identification comparable to

histology. Quantitative analysis revealed a statistically significant differ-

ence (P<.05) between tumor and collagen among 10 of 10 wavelength

bands analyzed, between tumor and muscle in 10 bands, and between fat

and tumor in 2 bands.

Conclusion: This study demonstrates a novel imaging modality capable of

rapidly and significantly distinguishing OSCC from surrounding normal

tissue. Such an intraoperative tool would be transformative: allowing for

an intraoperative capacity to delineate tumor tissue from nontumor tissue,

thus maximizing the efficacy of tumor resection and minimizing damage to

adjacent structures, thus improving patient outcomes.

Author Disclosure: Z. Taylor: None. I. Kim: None. J. Pesce: None. W.

Grundfest: None. M.S. St. John: None.

International Journal of Radiation Oncology � Biology � Physics922

231Primary Tumor and Nodal Regression Rates: The Prognostic Value ofVolumetric Image Guided Radiation Therapy for Head and NeckCancerS.R. Campbell,1 A.S.R. Mohamed,2 J. Heukelom,3 M.J. Awan,4

A.S. Garden,2 G.B. Gunn,2 D.I. Rosenthal,2 and C.D. Fuller2;1Midwestern University- Chicago College of Osteopathic Medicine,

Downers Grove, IL, 2The University of Texas MD Anderson Cancer Center,

Houston, TX, 3Netherlands Cancer Institute, Amsterdam, Netherlands,4Case Western Reserve University, Cleveland, OH, United States

Purpose/Objective(s): The objective of this study is to evaluate weekly

primary tumor regression rates (PTRR) and nodal tumor regression rates

(NTRR) of head and neck cancers (HNC) during radiation (RT) as a

prognostic indicator of oncologic outcomes and survival. Image guided

radiation therapy (IGRT), specifically computed tomography (CT)-on-

Rails (CToR), increases the accuracy of daily RT and additionally affords

the opportunity for intratreatment response evaluation.

Materials/Methods: A single-institution retrospective review from 2008 to

2013 was completed for patients with HNC who received RT with CToR.

Forty-three patients with 70 measurable targets, 43 primary lesions and 27

metastatic lymph nodes, met inclusion criteria. Patients without radio-

graphically evident primary tumors and those with surgical intervention

prior to RT were excluded.

Results: The analysis included 43 patients with a median age of 56 years

(21-78), and 91% of them were male. The majory of patients were diag-

nosed with oropharynx cancers (63%), 26% nasopharynx and 11% sino-

nasal. Fifty-eight percent of patients received definitive chemoradiation,

and 26% received induction followed by chemoradiation. The mean pri-

mary and nodal gross tumor volume (GTV) pre-RTwas 38.5 mL (standard

deviation [SD] 34.9) and 13.6 mL (SD 10.3), respectively. PTRR of �25%

at fraction 15 (nZ26) was associated with superior 5-year local control

(LC) (100% vs 70%, PZ.003), relapse-free survival (RFS; 84% vs 45%,

PZ.007), and overall survival (OS; 88% vs 50%, PZ.005). On both

univariate and multivariate analysis, PTRR <25% was associated with

increased hazard of local failure compared to �25% (PZ.001 for both).

There was not a strong independent correlation of NTRR with regional

control, distant control or RFS. Further characterization of tracked lymph

nodes revealed that 70% (nZ19) were purely cystic or had a cystic

component. Many of the highly cystic nodes demonstrated an initial in-

crease in size, likely secondary to RT induced inflammation, before later

regression.

Conclusion: CToR enables accurate GTV tracking during treatment which

carries prognostic value. PTRR of �25% at midtreatment can be used as

an indicator for LC, RFS, and OS. NTRR does not appear to carry the same

prognostic value as PTRR due to variations in nodal architecture, an initial

paradoxical size increase, especially in cystic nodes, and treatment

response often continuing up to 4 months post-RT. The clinical implication

of TRR appears to be of most value when tracking the primary lesion, and

specifically if PTRR <25% treatment plan modification may be warranted

given a higher likelihood of treatment failure.

Author Disclosure: S.R. Campbell: None. A.S. Mohamed: None. J.

Heukelom: None. M.J. Awan: None. A.S. Garden: None. G.B. Gunn:

None. D.I. Rosenthal: None. C.D. Fuller: None.

232Comparison of Tumor Volume Delineation on Magnetic Resonance/Positron Emission Tomography Versus Standard ComputedTomography for Head and Neck Cancer: Is There Added Value?K. Wang,1 B. Mullins,2 A. Falchook,2 J. Lian,2 M.J. Dance,2 W. Lin,2

T. Sills,2 B. Huang,2 and B.S. Chera2; 1UNC Lineberger Comprehensive

Cancer Center, University of North Carolina School of Medicine, Chapel

Hill, NC, 2University of North Carolina Hospitals, Chapel Hill, NC

Purpose/Objective(s): Computed tomography (CT), positron emission

tomography/computed tomography (PET/CT), and magnetic resonance

imaging (MRI) are commonly used in head and neck radiation treatment

planning. Combined MRI and PET (MR/PET) offers the advantage of

simultaneous acquisition of multimodality biological information and

possibly improved spatial resolution. In this study we assessed the added

value of gross tumor volume (GTV) delineation on MR/PET compared to

the standard practice of GTV delineation on CT.

Materials/Methods: We prospectively enrolled patients with squamous

cell carcinoma of the head and neck treated with chemoradiation therapy in

an institutional review boardeapproved study. We performed 2 pretreat-

ment scans: contrast-enhanced planning CT and gadolinium-enhanced

MR/PET. Primary and nodal CT-based GTVs (GTV-CT) were delineated

on planning CT scans. Treatments were delivered based on these volumes.

Primary and nodal MR/PET-based tumor volumes were delineated retro-

spectively after treatment completion using anatomic MRI and metabolic

PET information to form a biologic tumor volume (MRBTV). MRBTVs

were delineated without knowledge of the GTV-CT (BM, AF). Accuracy

of MRBTV contours was assessed by a neuroradiologist (BH). MR/PET

scans were registered to planning CT scans and the MRBTV was trans-

ferred. The Dice similarity coefficient (DSC), a validated metric to eval-

uate spatial overlap, was calculated from the intersection and union of the

GTV-CT (planning CT) and MRBTV (MR/PET). A DSC of 0.7 or greater

was considered a “good” overlap. Doses received by the GTV-CT and

MRBTV were compared.

Results: Twelve patients completed a pretreatment planning CT and

MR/PET. Primary disease site was oropharynx in 11 patients and larynx in

1 patient. Ten pts had evaluable primary tumor GTVs and 8 had evaluable

nodal GTVs. For the primary GTV, mean GTV-CT was 12.9 mL and

MRBTV was 13.3 mL. Mean primary GTV/BTV intersection was 8.1 mL,

with a Dice ratio of 0.4. Dose received by at least 95% of the tumor

volume (D95%) was 65.8 Gy for the primary GTV-CT versus 65.7 Gy for

the primary MRBTV. For the nodal GTV, mean GTV-CTwas 29.7 mL and

MRBTV was 29.2 mL. Mean nodal MRBTV intersection was 19.7 mL

with a Dice ratio of 0.5. D95% was 62.9 Gy for the nodal GTV-CT versus

61 Gy for the nodal MRBTV.

Conclusion: MRBTV volumes were similar to GTV-CT but the spatial

overlap was moderate. Despite only moderate spatial overlap, radiation

dose received by the MRBTV was similar to that received by the GTV-CT.

Most patients in our study had oropharyngeal primaries, for which the

added sophistication of MRPET may not improve the quality of GTV

delineation. MR/PET may show added value for primary tumors that are

more difficult to delineate on CT, such as sinonasal, nasopharyngeal, and

base of skull tumors.

Author Disclosure: K. Wang: None. B. Mullins: None. A. Falchook:

None. J. Lian: None.M.J. Dance: None. W. Lin: None. T. Sills: None. B.

Huang: None. B.S. Chera: None.

233Radiation-Induced Mucositis Does Not Increase Radiotracer Uptakeon PET/CT Imaging in Head and Neck Cancer PatientsS. Song,1 J. Zweit,1 M.J. Fratkin,1 J.F. Williamson,2 and D. Holdford1;1Virginia Commonwealth University Health System, Richmond, VA,2Virginia Commonwealth University, Richmond, VA, United States

Purpose/Objective(s): To determine how radiation-induced mucositis

affects radiotracer uptake on positron emission tomography/computed to-

mography (PET/CT) scan during and immediately after radiation therapy

(RT) in head and neck cancer (HNC) patients.

Materials/Methods: Patients with locally advanced HNC were enrolled in

an institutional review boardeapproved phase 1 trial to determine the

feasibility of using PET/CT to guide treatment dose escalation. Patients

were treated with accelerated regimen, 2.27 Gy per fraction to 68.1 Gy in

30 fractions over 6 weeks. Concurrent cisplatin was given every 3 weeks.

PET scans using 2 types of tracers, [18F]FDG and [18F]FLT, were obtained

before RT, 2 weeks and 4 weeks during RT, and 2 weeks after RT. As part

of posttreatment evaluation, all patients also underwent FDG-PET at 2

months after RT. The use of FLT-PET was to avoid the interference of

Abstract 233; Table 1

Pre-RT 2 weeks 4 weeks 2 weeks post-RT

Avg mucositis grade 0 1.33 2.67 1.67FLT-PET SUVmax 10.5 3.25 1.55 2.15FDG-PET SUVmax 8.53 5.53 4.27 5.13

Volume 94 � Number 4 � 2016 Posters 923

mucositis in imaging interpretation, as FLT, unlike FDG, does not high-

light metabolic or inflammatory activity of the tissue. A total of 9 PET

scans were planned per patient (4 FLT-PET and 5 FDG-PET scans). Acute

mucositis was recorded according to Common Terminology Criteria for

Adverse Events version 4. PET scans were read by the same radiologist.

Maximal standard uptake value of FDG-PET at each timepoint was

compared with FLT-PET. Activity of both FLT- and FDG-PET was also

matched to the mucositis score to determine the correlation between PET

activity and the severity of mucositis.

Results: Three patients were enrolled in the first phase of the feasibility

study. A total of 26 PET scans were obtained (missing was an FLT scan in

1 patient at week 2). All patients completed treatment according to the

protocol. All patients developed mucositis, which nearly peaked at the end

of week 4. Both FDG and FLT PET activity significantly decreased at week

2 and reached the lowest level at week 4. FDG-PET did not show activity

in areas with the severe mucositis at and around the primary site, same as

FLT-PET (see table). Interestingly, FLT- as well as FDG-PET activity

slightly increased at week 2 after RT at the mucosal sites while mocositis

subsided to less than grade 2. At minimal follow-up of 1 year, all patients

were without treatment failure.

Conclusion: This study demonstrates that radiation-induced mucositis in

oral cavity or oropharynx does not increase FDG or FLT uptake during

radiation therapy.

Author Disclosure: S. Song: None. J. Zweit: None. M.J. Fratkin: None.

J.F. Williamson: None. D. Holdford: None.

234PET/CT in CT Simulation: Significance of a Standardized PositioningProtocol for Head and Neck Radiation Therapy PlanningG. Tolekidis,1 M. Choi,2 M. Mai,1 and A.Z. Diaz1; 1Rush University

Medical Center, Chicago, IL, 2Loyola University Medical Center,

Maywood, IL

Purpose/Objective(s): Diagnostic imaging scans, such as PET/CT and

MRI, provide a valuable tool for radiation therapy treatment planning

(RTP) for head and neck (H&N) cancer. Information from diagnostic scans

can be incorporated into the RTP process by performing a dedicated

diagnostic scan in the treatment position or by coregistering an existing

scan with the simulation scan. The former approach should minimize

differences in anatomical positioning that may hinder accurate coregis-

tration. The purpose of this study was to use the clival incline to quantify

differences in H&N positioning between patients undergoing diagnostic

PET/CTs positioned with versus without the RTP immobilization mask.

Materials/Methods: Twenty patients who received intensity modulated

radiation therapy for H&N cancer from 2011 to 2015 at our institution

were selected for this retrospective review. Ten patients underwent diag-

nostic PET/CT using the mask created during simulation (Group A) while

10 patients underwent PET/CT without the mask (Group B). A 5-point

Orfit Efficast thermoplastic mask, size B head support, base plate, and,

occasionally, a block or wedge were used for immobilization. Clival

incline was measured for each simulation and PET/CT group using

MIRADA software by a single user.

Results: Mean clival incline for Group B was 61.44� (standard deviation

[SD], 8.30�; standard error mean [SEM], 2.62�), while clival incline for

Group A was 72.25� (SD, 7.78�; SEM, 2.46�). Comparing the simulation

CT to the CT from the PET/CT, mean clival incline difference was 12.61�

in Group B (SD, 5.62�; SEM, 1.78�), and 1.48� (SD, 1.026�; SEM, 0.32�)in Group A. Both these differences between the groups were statistically

significant, PZ.008 and PZ.001, respectively, using t test analysis for the

equality of means.

Conclusion: Based on these results we reach 2 conclusions: (1) when no

mask is used (when PET/CT technologists position patients without

guidance from radiation oncology staff), there is a different approach to

positioning: PET/CT technologists favor a more neutral to flexion position,

while we favor a neutral to extended position. (2) Using the simulation

mask for the PET/CT greatly reduces the difference in head position when

compared to its respective simulation scan. This allows for a more robust

registration. Given these data, when possible, patients should have PET/CT

performed using the immobilization mask created for simulation. How-

ever, as this is not always feasible, dosimetric and clinical studies are

warranted to further explore the implications of a standardized neck

positioning protocol for both H&N simulations and PET/CT scans.

Author Disclosure: G. Tolekidis: None. M. Choi: Employee; Rush Uni-

versity Medical Center. M. Mai: None. A.Z. Diaz: Employee; Loyola

University Medical Center, Hines VA Hospital. cancer committee chair;

Rush University Medical Center.

235Does Routine Surveillance PET/CT at 3 Months Have Value AfterLaryngectomy (L) Followed by Radiation therapy (RT) forLocoregionally Advanced Laryngeal Cancer?T. Yuan,1 J.M. Watkins,2 S.L. Mott,2 M. Marquardt,2 A. Hoover,3 W. Sun,2

J. Buatti,2 N. Pagedar,2 and C.M. Anderson2; 1Cancer Center of

Guangzhou Medical University, Guangzhou, China, 2University of Iowa

Hospitals & Clinics, Iowa City, IA, 3The University of Kansas Hospital,

Kansas City, KS

Purpose/Objective(s): While positron emission tomography/computed

tomography (PET/CT) has demonstrated value following definitive che-

moradiation therapy in the setting of locoregionally advanced laryngeal

cancer (LC), its routine use in surveillance following laryngectomy (L) +

radiation therapy (RT) remains to be determined. This study explores the

diagnostic value of PET/CT 3 months following L+RT.

Materials/Methods: This institutional review boardeapproved retrospec-

tive review included patients who underwent curative-intent L with adju-

vant RT for initial presentation of LC, with 3-month post-RT surveillance

PET/CT performed in the absence of suspected active disease. The positive

and negative predictive values for locoregional disease (PPV/NPV) were

calculated based upon the 3-month PET/CT head and neck findings, as

correlated with clinical or pathologic outcomes during the 12 months after

the scan. Patients with PET/CT performed <60 or >130 days post-RT or

who had <12 months of potential post-PET oncologic follow-up were

excluded.

Results: Between 2004 and 2014, 18 patients were eligible for the present

analysis; 3 had partial laryngectomy and 15 had total laryngectomy. The

median age at diagnosis was 56 years (range 48-73), and pathologic stage

was III, IVA, and IVB for 3 (17%), 14 (78%), and 1 (6%) patients,

respectively. All patients started adjuvant RT at a median of 45 days post-L

(31-72), with median dose 66 Gy (60-70) and 10 patients (56%) also

received concurrent platinum-based chemotherapy. PET/CT was per-

formed at a median of 98 days post-RT completion (69-127), with 4 pa-

tients (22%) demonstrating positive locoregional findings. At a median

follow-up of 31.9 months (10.3-75.5) post-PET/CT, 11 patients were alive

(10 without recurrence, 1 with active disease) and 7 patients had died (5

with recurrent LC; 4 of 5 with distant recurrence only, 1 of 5 with regional

and distant recurrence). The sensitivity, specificity, PPV, and NPV for

surveillance 3-month PET for locoregional disease were 100%, 80%, 25%,

and 100%, respectively. Occult distant metastases were discovered on 3-

month PET/CT in 4 of 18 patients (22%), all of whom had lung metastases.

Bone and liver mets were additionally discovered in 1 of these 4 patients.

The remaining patient who died of LC recurred in the lungs 38.1 mo after

International Journal of Radiation Oncology � Biology � Physics924

L, with no signs of locoregional recurrence at death. For patients with

locoregionally negative 3-month PET/CT, the estimated 3-year disease-

free and overall survivals were 80% (95% confidence interval, 40%-95%)

and 78% (36%-94%), respectively.

Conclusion: Locoregional recurrence is rare 3 months after L+RT; thus,

most locoregional PET/CT findings are false positives in this setting. There

was enhanced detection of early distant metastatic disease through PET/

CT use. Surveillance CT chest at 3 months post-RT as an alternative to

PET/CT will be explored.

Author Disclosure: T. Yuan: observership, research; University of Iowa

Hospitals & Clinics. Research Grant; Chinese Scholarship Council. J.M.

Watkins: Employee; University of Iowa Hospitals & Clinics. S.L. Mott:

None. M. Marquardt: None. A. Hoover: Independent Contractor; Uni-

versity of Kansas Medical Center. W. Sun: None. J. Buatti: Research

Grant; NIH. N. Pagedar: Partner; Physicians Clinic of Iowa. C.M.

Anderson: Uncompensated, not related to this research; Galera Thera-

peutics, Inc.

236Dynamic Contrast-Enhanced MRI Detects Acute RadiationTherapyeInduced Alterations in Mandibular Microvasculature:Prospective Assessment of Imaging Biomarkers of Normal TissueInjuryV.C. Sandulache,1 B. Hobbs,2 A.S.R. Mohamed,3 S.J. Frank,3 J. Hazle,4

M.J. Awan,5 D.I. Rosenthal,3 A.S. Garden,3 G.B. Gunn,3 Y. Ding,6

J. Wang,6 R. Colen,2 K.A. Hutcheson,3 J.S. Lewin,3 M.S. Chambers,3

T.M. Hofstede,3 R.S. Weber,3 S. Lai,3 and C.D. Fuller3; 1UT MD Anderson

Cancer Center, Houston, TX, 2MD Anderson, Houston, TX, 3The University

of Texas MD Anderson Cancer Center, Houston, TX, 4MD Anderson

Cancer Center, Houston, TX, United States, 5Case Western Reserve

University, Cleveland, OH, United States, 6MD Anderson Cancer Center,

Houston, TX

Purpose/Objective(s): Osteoradionecrosis (ORN) is a devastating late

sequelae of external beam radiation therapy (EBRT). The ability to

detect EBRT-induced changes in bone vascularity represents an impor-

tant step in improving clinicians’ ability to successfully manage ORN

and its precursor states. We hypothesized that dynamic contract

enhanced magnetic resonance imaging (DCE-MRI) can provide in-

treatment information regarding EBRT-induced changes in mandibular

vascularity.

Materials/Methods: Thirty-two patients undergoing EBRT treatment for

head and neck cancer underwent prospective DCE-MRI imaging prior to,

midway through, and following chemoradiation therapy. DCE-MRI scans

were coregistered to dosimetric maps to correlate EBRT dose and change

in mandibular bone vascularity as measured by Ktrans.

Results: All patients had successful completion of treatment. We identified

3 patterns of vascularity changes associated with EBRT. One group of

patients demonstrated a dose-dependent increase in Ktrans midway during

(nZ9) and following completion (nZ5) of treatment. One group of pa-

tients demonstrated a dose-dependent decrease in Ktrans midway during

(nZ8) and following completion (nZ6) of treatment. One group of pa-

tients demonstrated no change in Ktrans midway during (nZ13) and

following completion (nZ20) of treatment. Overall, the frequency of dose-

dependent changes in mandibular vascularity abated from the midpoint to

the endpoint of the treatment regimen.

Conclusion: We demonstrate, in a prospective imaging trial, that DCE-

MRI can detect dose-dependent alterations in mandibular bone vascularity

during and following chemoradiation therapy and measures a biomarker of

acute mandibular injury and recovery temporal kinetics.

Author Disclosure: V.C. Sandulache: None. B. Hobbs: None. A.S.

Mohamed: None. S.J. Frank: None. J. Hazle: None. M.J. Awan: None.

D.I. Rosenthal: None. A.S. Garden: None. G.B. Gunn: None. Y. Ding:

None. J. Wang: None. R. Colen: None. K.A. Hutcheson: None. J.S.

Lewin: None.M.S. Chambers: None. T.M. Hofstede: None. R.S. Weber:

None. S. Lai: None. C.D. Fuller: Research Grant; GE/MDACC, Elekta/

MDACC.

237WITHDRAWN

238Pretreatment FDG-PET Textural Analysis in Laryngeal CancerPatients Treated With Definitive Radiation TherapyR. Kabarriti,1 P. Brodin,2 A.G. Berkowitz,3 A. Ingber,3 N. Ohri,4

K.P. McGovern,1 T.J. Ow,1 A. Tassler,1 S. Packer,5 B.A. Schiff,1

R.V. Smith,1 M. Haigentz,1 C. Guha,6 S. Kalnicki,4 W.A. Tome,4

and M.K. Garg4; 1Montefiore Medical Center-Albert Einstein College of

Medicine, Bronx, NY, 2Albert Einstein College of Medicine, Bronx, NY,3Montefiore Medical Center, Bronx, NY, 4Albert Einstein College of

Medicine, Montefiore Medical Center, Bronx, NY, 5Montefiore Medical

Center-Albert Einstein College of Medicine, Bronx, NM, 6Albert Einstein

College of Medicine, Montefiore Medical Center, New York, NY

Purpose/Objective(s): Statistical image features from computed tomog-

raphy (CT) and positron emission tomography (PET) scans are currently

being explored for their potential to predict clinical outcome and thus

provide indications for risk-adapted treatment strategies. Here, we analyze

tumor textural features from pretreatment 18F-fluorodeoxyglucose (FDG)

PET scans as related to clinical outcomes in patients treated with definitive

radiation therapy (RT) for laryngeal cancer.

Materials/Methods: Seventy-five consecutive patients with laryngeal

squamous cell carcinoma treated with definitive RT with available pre-

treatment PET/CT scans at our institution were retrospectively evaluated.

Disease and patient characteristics, treatment information, and clinical

outcomes data were tabulated for each patient. Pretreatment PET/CT scans

were used for image feature analysis on the primary tumor volume from

each scan. Multiple statistical image features were computed, along with

several measures related to the standard uptake value (SUV). Redundant

image features were excluded based on a strong correlation with param-

eters commonly reported as important such as metabolic tumor volume

and maximum SUV (SUVmax). The correlation between the remaining

image features and local control (LC), progression-free survival (PFS), and

overall survival (OS) was evaluated through univariate Cox regression, and

statistically significant features were combined with clinical parameters in

multivariate Cox proportional hazards regression models. Correction for

multiple testing was performed using the Benjamini-Hochberg method.

Results:With a median follow-up of 31.4 months, the 3-year LC, PFS, and

OS (with 95% confidence intervals) were 66% (53%-79%), 49% (36%-

62%), and 75% (64%-87%), respectively. Out of the calculated image

features, only SUVpeak (highest possible mean value of a 1 cm3 spherical

volume within the primary tumor) corrected for lean body mass appeared

to correlate with outcome. Interestingly, the SUVpeak was found to be a

statistically significant parameter in the model for PFS (PZ.006), but not

for LC, suggesting that the SUVpeak value may hold important informa-

tion about potential regional or distant spread. Dividing the patient pop-

ulation based on the median SUVpeak value shows a split of the Kaplan-

Meier survival curves, although not statistically significant with a log-rank

P value of .082. KPS>70 was the only parameter significantly related to

OS (PZ.0002).

Conclusion: The SUVpeak value from pretreatment PET scans may hold

important information about the risk of distant and regional spread for

laryngeal cancer patients. We are currently working on gathering addi-

tional patient data to validate these findings.

Author Disclosure: R. Kabarriti: None. P. Brodin: None. A.G. Berko-

witz: None. A. Ingber: None. N. Ohri: None.K.P. McGovern: None. T.J.

Ow: None. A. Tassler: None. S. Packer: None. B.A. Schiff: None. R.V.

Smith: None. M. Haigentz: None. C. Guha: None. S. Kalnicki: None.

W.A. Tome: None. M.K. Garg: None.

Volume 94 � Number 4 � 2016 Posters 925

239Prognostic Significance of FDG-PET/CT Metabolic Parameters AfterInduction Chemotherapy in Squamous Cell Carcinoma of the Headand NeckR.I. Chin,1 C.R. Spencer,2 J. Ley,2 T.A. DeWees,2 H.A. Gay,2

B.H. Haughey,2 B. Nussenbaum,2 J.S. Lewis, Jr,3 F. Dehdashti,2 B. Siegel,2

M.D. Daly,2 L. Michel,2 T. Wildes,2 D. Adkins,2 and W.L. Thorstad2;1Saint Louis University School of Medicine, St. Louis, MO, 2Washington

University School of Medicine, St. Louis, MO, 3Vanderbilt University,

Nashville, TN

Purpose/Objective(s): Previously, we reported that the absolute meta-

bolic tumor volume (MTV) measured on 18F-fluorodeoxyglucose positron

emission tomography/computed tomography (FDG-PET/CT) after 2 cycles

of induction chemotherapy (IC) and the percent reduction of MTV are

prognostic for tumor recurrence in patients with squamous cell carcinoma

of the head and neck (SCCHN). In this study, we updated the analysis to

include patients from an additional IC protocol to evaluate the prognostic

significance of MTV and related PET parameters in a larger cohort.

Materials/Methods: From 2009 to 2013, 60 consecutive patients with

SCCHN (T1 excluded) were enrolled in 2 single-institutional prospective

single-arm phase 2 trials. From 2009 to 2010, 30 patients were treated with

APF-C (nab-paclitaxel, cisplatin, 5-FU, and cetuximab x3 cycles). From

2012 to 2013, another 30 patients were treated with APF x3 cycles. IC was

scheduled to be followed by concurrent chemoradiationwith cisplatin and 70

Gy at 2 Gy per fraction for both trials. Of the 60 patients, 54 patients had

evaluable FDG-PET/CT studies obtained both before and after 2 cycles of

IC. The primary tumors were contoured using a gradient-based method

software tool. Total body weight was used to calculate SUVs. SUVmax,

SUVmean, MTV, and total glycolytic activity (TGAZSUVmean*MTV) of the

pre- and post-IC scans as well as the percent changes in these parameters

were calculated. Wilcoxon rank sum testing was used to compare the pa-

rameters of patients with any tumor recurrence to those without. A likeli-

hood-based approach was used to identify the optimal cutpoints with regard

to recurrence-free survival and determine the associated P values.

Results: The primary tumors were oropharynx (38; 31 of 37 were p16+ by

IHC), larynx (11), hypopharynx (4), and oral cavity (1). Tumor stages were

as follows: T2 (15), T3 (19), T4a (18), and T4b (2). At the median follow-

up of 2.9 years, 5 patients (9%) had tumor recurrence: 3 local, 2 regional,

and 2 distant metastases (1 patient had all 3 types of recurrences). Of the

patients with recurrence, 4 had oropharynx and 1 had larynx cancer.

Compared to the patients without recurrence, those with recurrence had

higher post-IC MTV (PZ.011) and TGA (PZ.008) and had smaller post-

IC reductions in MTV (PZ.008) and TGA (PZ.016). Factors that

increased the risk of tumor recurrence included a postinduction MTV of

more than 4 cm3 (P<.001, absolute risk [AR] 24% vs 0%), a postinduction

TGA of more than 16 SUVmean*cm3 (P<.001, AR 29% vs 0%), a decrease

in MTV of less than 63% (P<.001, AR 28% vs 0%), and a decrease in

TGA of less than 90% (PZ.002, AR 21% vs 0%).

Conclusion: MTVand the TGA of the tumor on FDG-PET/CT after IC, as

well as the percent reduction in these parameters after therapy, are prog-

nostic for tumor recurrence.

Author Disclosure: R. Chin: None. C.R. Spencer: None. J. Ley: None.

T.A. DeWees: None. H.A. Gay: None. B.H. Haughey: None. B. Nus-

senbaum: None. J.S. Lewis, Jr.: None. F. Dehdashti: None. B. Siegel:

None. M.D. Daly: None. L. Michel: None. T. Wildes: None. D. Adkins:

None. W.L. Thorstad: None.

240Redefining Osteoradionecrosis With the Use of Cone BeamTechnologyR. Hoffman,1 J. Kelly,2 E.G. Freymiller,3 S. Tetradis,3 and D.P. Copps4;1Valley Radiotherapy Associates, Burbank, CA, 2Mayo Clinic, Rochester,

MN, 3UCLA, Los Angeles, CA, 4Private Practice, Tarzana, CA

Purpose/Objective(s): Osteoradionecrosis (ORN) has been defined in

multiple ways; however, this diagnosis is generally made when exposed

bone in a previously irradiated field is documented. This review of head

and neck cancer patients who have been irradiated has incorporated the use

of cone beam computed tomography (CBCT) scans to uncover ORN that is

not visible on routine physical exam. Combining these radiographic

findings with other characteristics in patients who have been diagnosed

with ORN, there appear to be subsets of ORN.

Materials/Methods: The charts of 45 patients with head and neck cancer

who had received radiation therapy and been diagnosed with ORN were

retrospectively reviewed. Specific characteristics looked at included the

location of ORN, patient’s sex, dose to area of breakdown of bone, finding

on CBCT scan, and timing of diagnosis after radiation. Utilizing these 5

characteristics, it became clear that there were subsets of ORN that

correlated with a patient’s prognosis with respect to their ORN and the best

approach in terms of managing this condition.

Results: Three subsets of patients with ORN were identified: (1) spon-

taneous ORN (SpO), (2) subclinical ORN (SubO), and (3) traumatic

ORN (TO). Patients with SpO always developed ORN in the posterior

aspect of the mylo-hyoid ridge of the mandible, were male, received

between 6500 and 7000 centiGray, and occurred within 12 months after

radiation. Patients with SubO did not have exposed bone but, instead,

radiographically on CBCT scan, they had clear evidence of breakdown in

their mandibular bone, this was not specific to males or females, on

occasion they received less than 6500 centiGray, and occurred beyond 12

months out from radiation. Patients with TO developed ORN as a result

of iatrogenic medical means (ie, following a biopsy in the previously

irradiated field) or by an invasive dental procedure (ie, extraction or

peridontal surgery in an area previously irradiated). Depending on the

timing and management of one’s treatment for ORN, patients with SpO

had the best prognosis while those with TO had the worst. Prognosis of

those with SubO was in between.

Conclusion: This review suggests that rather than treating all patients with

ORN the same way, there appears to be a benefit with placing them into 1

of 3 subsets of ORN: SpO, SubO, or TO. Doing so will help the treating

radiation oncologist and dentist in caring for these individual’s symptoms

and counseling them in how best to deal with this potentially devastating

complication of therapy both emotionally and psychologically.

Author Disclosure: R. Hoffman: None. J. Kelly: None. E.G. Freymiller:

None. S. Tetradis: None. D.P. Copps: None.

241Anatomic and Dosimetric Changes in Patients With Head and NeckCancer Treated With a Tri-60Co Teletherapy/Magnetic ResonanceImaging DeviceG. Raghavan,1 A.U. Kishan,1 M. Cao,1 Y. Yang,1 and A.M. Chen2;1University of California, Los Angeles, Los Angeles, CA, 2University of

California, Los Angeles- David Geffen School of Medicine,

Los Angeles, CA

Purpose/Objective(s): Prior studies have quantified the anatomic changes

in head and neck cancer patients during a course of intensity modulated

radiation therapy (IMRT) by analyzing serial images acquired using on-

board computed tomography (CT). A novel tri-60Co teletherapy system

equipped with a 0.35 T magnetic resonance imaging (MRI) allows MRI-

based image guided IMRT. We sought to determine the feasibility of

quantifying the anatomic and dosimetric changes of the primary gross

tumor volumes (GTVs) and the parotid glands using daily pretreatment

MRIs among patients with head and neck squamous cell carcinoma

(HNSCC) treated with this MRI/IMRT device.

Materials/Methods: Six patients with HNSCC were treated with the

MRI/IMRT system to 66.00 to 69.96 Gy in 33 fractions. Pre-IMRT MRIs

on days 1, 5, 10, 15, 20, 25, 30, and 33 were imported into a contouring

interface. Primary GTVs and parotids were contoured on each MRI and

volumes were quantified. The center of mass (COM) shifts for these

structures were assessed relative to the COM on day 1 in the medial-

lateral direction. Daily dose delivery data were imported into the con-

touring interface, and doses to the GTVs and parotids were assessed

International Journal of Radiation Oncology � Biology � Physics926

based on daily contours. Statistical significance was determined by the

2-tailed t test, while correlations were assessed by the Spearman

coefficient.

Results: Primary GTVs decreased significantly in volume over the

course of IMRT (median % volume loss, 38.7%; range, 29.5-72.0%;

P<.05) at a median rate of 1.2% per fraction (range, 0.92%-2.2% per

fraction). Both ipsilateral and contralateral parotid glands underwent

significant volume loss (P<.05). Median percentage of volume loss of

ipsilateral parotid glands was 31.1% (2.3%-43.9%), and median rate of

volume loss was 0.97% per fraction (0.07%-1.4% per fraction).

Contralateral parotids experienced a median percentage of volume loss

of 21.8% (4.0%-40.5%) and median shrinkage rate of 0.68% per frac-

tion (0.13%-1.3% per fraction). Both ipsilateral and contralateral

parotids shifted medially during IMRT. Weight loss correlated signifi-

cantly with parotid gland volume loss and medial COM shift (P<.05).

Dose received by the primary GTVs and parotids did not significantly

change during treatment. Median difference between administered mean

dose at fraction 33 and planned dose per fraction was 0.008 Gy (-0.06-

0.18 Gy) for ipsilateral and 0.01 Gy (-0.12-0.27 Gy) for contralateral

parotids.

Conclusion: Integrated on-board MRIs from a tri-60Co teletherapy system

can be used to accurately contour and analyze primary GTVs and parotid

glands over the course of IMRT. COM shifts and significant volume re-

ductions of the primary tumors and parotids were observed, confirming the

results of CT-based studies. The enhanced resolution of the MRIs acquired

may facilitate online adaptive replanning in the future. Despite anatomic

changes during IMRT, significant dosimetric shifts were not always

apparent.

Author Disclosure: G. Raghavan: None. A.U. Kishan: None. M. Cao:

None. Y. Yang: None. A.M. Chen: None.

242Validation Is Crucial in Head and Neck Translational ImagingResearchC. Terhaard,1 E.A. Jager,2 M. Philippens,1 T. Schakel,1 J. Caldas

Magalhaes,3 H. Ligtenberg,1 N. Kasperts,1 F. Pameijer,1 S. Willems,1

and C.P. Raaijmakers1; 1UMC Utrecht, Utrecht, Netherlands, 2UMC

Utrecht, dep. Radiotherapy, Utrecht, Netherlands, 3University Medical

Center Utrecht, Utrecht 3584CX, Netherlands

Purpose/Objective(s): Delineation of the tumor, based on imaging, is the

weakest link in radiation therapy. The value of new and existing imaging

techniques should be demonstrated by comparison with the ground truth:

histopathology. Total laryngectomy (TLE) specimens are an appropriate

model, since the resected tissue remains cohesive.

Materials/Methods: For 27 patients who received primary TLE because of

T3/T4 laryngeal or hypopharyngeal cancer, 18F-fluorodeoxyglucose posi-

tron emission tomography (FDG-PET), computed tomography (CT), and

magnetic resonance imaging (MRI) images were made in a radiation mask

before surgery. The specimen was fixated and sliced, and whole-mount

hematoxylin-eosin (H&E) stained sections were obtained. These were used

to digitally reconstruct the specimen in 3 dimensions, which was registered

to the CT, MRI, and FDG-PET scan. H&E sections, CT, and MRI were

delineated by 3 observers. For the FDG-PET scan, automatic delineations

were obtained with a Gaussian mixture model. MRI was delineated in 2

sessions, the first according to clinical practice. The second session

involved excluding regions with high signal intensity on T1 weighted MRI

enhanced with gadolinium or T2 weighted images, based on guidelines for

cartilage invasion. The required extensions to the delineations to cover

95% of the pathologic derived tumor outer contour, including microscopic

extension, were determined.

Results: Comparative examination between 3 pathologists revealed a

slight variation with respect to tumor delineation. The median tumor

volume was 10.5 cm3. Variation between the 3 observers for the CT and

MRI delineations was much larger. The median delineated volume on CT

and on the first MRI delineation was 17.5 and 19.9 cm3, respectively.

Automatic delineations using FDG-PET had the least tumor over-

estimation, but also some underestimation. In the second MRI delineation,

the median volume reduced from 19.9 cm3 to 15.7 cm3, after consensus

based on the new guidelines for tumor definition by MRI. The required

median extension of the MRI delineation to cover 95% of the pathologic

delineation, including microscopic invasion, was similar for both MR

delineation sessions.

Conclusion: The current practice of tumor delineation using CT, FDG-

PET, and MRI results in significant tumor volume overestimation. Auto-

matic delineation on FDG-PET showed the smallest overestimation but at

cost of slightly larger missed volume. New guidelines for MRI delineation

could be validated with pathology, showing a decrease in overestimation.

The clinical target volume could be reduced by improving tumor definition

and, thereby, reduce the chance of complications.

Author Disclosure: C. Terhaard: None. E. Jager: None. M. Philippens:

None. T. Schakel: None. J. Caldas Magalhaes: None. H. Ligtenberg:

None. N. Kasperts: None. F. Pameijer: None. S. Willems: None. C.P.

Raaijmakers: None.

243Role of Positron Emission Tomography and Computed TomographyImaging for Detecting Disease Recurrence Following AdjuvantRadiation Therapy in Oropharyngeal CancerM. Shuja, M. Rosado, D.R. DeLone, E.J. Moore, R.L. Foote, and D.J. Ma;

Mayo Clinic, Rochester, MN

Purpose/Objective(s): The role of posttreatment imaging for oropha-

ryngeal cancer (OPC) patients treated with definitive chemoradiation is

well defined. However, the role of posttreatment surveillance for patients

undergoing adjuvant radiation therapy is unclear. The goal of this study is

to provide data on how surveillance modality (clinical surveillance vs

computed tomography [CT] vs positron emission tomography [PET]/CT)

impacts the detection of disease recurrence.

Materials/Methods: This is a retrospective, institutional review

boardeapproved analysis of 52 cases of OPC recurrence from the year

1997 to 2015 at Mayo Clinic Rochester, MN. These cases failed after being

treated with surgery followed by adjuvant radiation therapy with or

without chemotherapy. Time to failure was determined following adjuvant

radiation treatment. Local, nodal, or distant failure patterns were catego-

rized based on clinical or imaging evaluations.

Results: Among the 52 cases with disease recurrence, the mean time to

recurrence following adjuvant treatment was 10.7 months (median 8

months). Distant, regional, and local failure comprised 79%, 13%, and 8%

of the cohort, respectively. A total of 36 patients (69%) failed within the

first year following completion of treatment: 3 local, 5 nodal, and 28

distant. Among the patients that failed, 12 were detected clinically,

whereas 24 were detected by imaging modalities. CT detected 8 cases

while PET was successful in detecting 16 cases. The mean time to clinical

detection was 5 months (3 local [25%], 2 nodal [17%], 7 distant [58%]),

while the mean time to detection by imaging was 6.2 months (1 local

[4%], 5 nodal [21%], 18 distant [75%]). In the imaging category, the mean

time of recurrence detection was 7.5 months by CT scan and 5.5 months

by PET. Early local failures are generally detected by physical exam (25%

vs 4%). However, imaging exam was superior at detecting nodal and

distant failure (17% vs 21% in nodal and 58% vs 75% in distant). PET

detected distant failures earlier than CT by an average of approximately 2

months.

Conclusion: Imaging modalities remain the optimal tool for detecting early

distant and regional recurrence. Surveillance and follow-up in OPC pa-

tients undergoing adjuvant therapy warrants early imaging for regional and

distant disease detection.

Acknowledgment(s): *Authors M. Shuja and M. Rosado contributed

equally to this abstract.

Author Disclosure: M. Shuja: None. M. Rosado: None. D.R. DeLone:

None. E.J. Moore: None. R.L. Foote: None. D.J. Ma: None.

Volume 94 � Number 4 � 2016 Posters 927

244Functional Outcome of Total Glossectomy and Anterolateral ThighFree Flap (ALTF) Reconstruction: Long-Term Follow-up ResultY. Choi and M.S. Kim; Seoul St. Mary’s Hospital, Seoul, South Korea

Purpose/Objective(s): In advanced tongue cancer, wide excision and

reconstruction is indipensable. After reconstruction, functional problems

like speech and swallowing can occur, and tongue volume and height are

important factors related to this functional problem. The purpose of this

study was to evaluate the relationship between reconstructed tongue

volume and functional outcome in advanced tongue cancer patients who

received total or near total glossectomy.

Materials/Methods: From January 2004 to December 2013, tongue cancer

patients who received total or near total glossectomy and reconstruction

with anterolateral thigh free flap (ALTF) in St. Mary Hospital in Seoul,

Korea, were involved in this study. All patients had postoperative

computed tomography or magnetic resonance imaging scans at 6 months

and 2 years, and reconstructed tongue volume was calculated by summa-

tion of all 2-dimensional areas of the reconstructed tongue. Speech ability

was scored from 0 to 100 points by a professional speech language

pathologist. Swallowing ability was scored from 1 to 7 by the latest

Modified Barium Swallow test or esophagogram.

Results: From a pool of 25 patients, 19 patients who followed up were

involved in this study. Tongue volume was reduced about 19% from 6

month to 2 years following surgery. There was a statistical difference be-

tween tongue volume and 2 years following surgery (PZ.01), and more

tongue volume led to less weight change. Patients with a larger tongue

volume loss showed larger weight loss, but there was no statistical signif-

icance. Speech and swallowing ability were related to tongue volume at 6

months and 2 years following surgery, and this was statistically significant.

Conclusion: There was a statistically significant relation between recon-

structed tongue volume and postoperative functional outcome.

Author Disclosure: Y. Choi: None. M. Kim: None.

245The Prognostic Utility of 18F-FDG-PET Metabolic Tumor ResponseAfter Chemoradiation Therapy for Locally Advanced Head and NeckCancerF.L. Ampil,1 G. Caldito,1 S. Devarakonda,1 C. Reiser,1 C.A.O. Nathan,2

A. Takalkar,1 G. Mills,1 T. Lian,1 and V. Mehta1; 1Louisiana State

University Health, Shreveport, LA, 2LSU Health Sciences Center

Shreveport, Shreveport, LA, United States

Purpose/Objective(s): To characterize 18F-fluorodeoxyglucose positron

emission tomography (FDG-PET) metabolic tumor response (MTR)

findings and the associated prognosis following chemoradiation therapy

(CRT) for locally advanced head and neck cancer (HNC).

Materials/Methods: A retrospective study of the pre- and posttreatment

FDG-PET scans of 47 individuals who underwent CRT for stage III-IV

HNC between 2003 and 2011 was undertaken. FDG-PET was usually

performed before therapy and 2 to 3 months after completion of CRT.

Interpretation of images by visual analysis was either a positive or negative

scan indicating incomplete or complete MTR respectively.

Results: Thirty-nine patients (83%) exhibited complete resolution of

tumor and 8 individuals (17%) had incomplete or absent MTRs. At a

median follow-up of 36 months, the overall locoregional failure and distant

metastases rates were 17%. The 2-year survival rates in people with

complete and incomplete disappearance of the neoplasms were 92% and

57%, respectively (PZ.02); the corresponding 2-year disease-free survival

rates were 97% and 25% (P<.01). After adjusting for patient age, primary

and regional disease stage, site and volume of the neoplasm, number of

chemotherapy drugs used, and the presence or absence of significant co-

morbidity, tumor response was the only significant independent predictor

of prognosis.

Conclusion: FDG-PET imaging, aside from documenting MTR �2 months

after CRT, may also be useful for its prognosis predictive potential in

patients with locally advanced HNC.

Author Disclosure: F.L. Ampil: None. G. Caldito: None. S. Devar-

akonda: None. C. Reiser: None. C.O. Nathan: None. A. Takalkar: None.

G. Mills: None. T. Lian: None. V. Mehta: None.

246Imaging Treatment-Related Carotid Artery Inflammation in Headand Neck CancerC.C. Okoye, Y. Zheng, R.F. Muzic, Jr, M. Yao, and J.A. Dorth; University

Hospitals Case Medical Center, Cleveland, OH

Purpose/Objective(s): Radiation therapy (RT) is known to increase the

risk of carotid artery stenosis and resultant cerebrovascular events (CVEs)

in patients treated for head and neck cancer (HNC), but there are no

established prevention or treatment strategies. It is possible that early in-

creases in vascular inflammation due to RT may mediate long-term

vascular injury and be quantifiable using 18F-fluorodeoxyglucose positron

emission tomography/computed tomography (FDG-PET/CT) scanning.

We hypothesize that there will be an increase in carotid artery FDG uptake

between pre-RT and 3 months post-RT PET/CT scans. Herein, we present

an analysis of the first 5 patients out of 18 accrued patients to a prospective

study.

Materials/Methods: This prospective, institutional review boardeapprovedstudy enrolled patients with HNC treated with RT or chemoradiation with

curative intent and targeting 1 or more neck lymph node regions containing

the carotid artery. Patients were excluded if they had a history of carotid

stenosis, CVEs, autoimmune disease, or human immunodeficiency virus

infection. Patients completed pre-RTand 3 months post-RT PET/CTand CT

scans with intravenous contrast using a standardized protocol, wearing a

thermoplastic headholder. PET/CT scans were obtained 1 hour after injec-

tion of FDG. The carotid arteries were delineated from their origin in the

upper chest to the skull base using CTwith contrast images fused to PET/CT

images that were registered based on bony and soft tissue anatomy. Contrast-

enhanced regions for the common carotid, carotid bulb, and internal carotid

arteries were expanded by 1.5 mm to include the vessel wall in the carotid

contours. A volume containing the gross tumor volume plus a 5-mm

expansion was subtracted from each carotid contour. The maximum and

mean standardized uptake value (SUV) was measured within each carotid

contour and divided by themean SUVwithin the superior vena cava to obtain

the maximum and mean target-to-background ratio (TBR), respectively.

Results: Among 5 patients, the maximum TBR for the left and right ca-

rotid arteries was 1.71�0.47 (mean�standard deviation) and 1.80�0.52,

respectively, before treatment, compared with 1.75�0.18 and 1.63�0.13

after treatment. The mean TBR for the left and right carotid arteries was

1.11�0.17 and 1.17�0.21 before treatment, compared with 1.13�0.1 and

1.18�0.1 after treatment. No significant difference was detected.

Conclusion: Carotid artery FDG uptake was not different pre- versus 3

months post-RT in our small group of patients. Final conclusions will be

drawn from an analysis of additional enrolled patients. Future studies will

investigate the impact of extending the FDG uptake time beyond 60 mi-

nutes to potentially increase sensitivity.

Author Disclosure: C.C. Okoye: None. Y. Zheng: None. R.F. Muzic, Jr.:

None. M. Yao: Research Grant; Bristol Myers Squibb. J.A. Dorth: None.

247Dermal Backflow Seen Through Near-Infrared FluorescenceImaging: An Early Response to Cancer Treatments in Head and Neckfor the Detection of LymphedemaS.H.S. Naqvi, R.J. Karni, I.C. Tan, J. Rasmussen, M. Aldrich, J. Morrow,

and E. Sevick; The University of Texas Health Science Center- Houston,

Houston, TX

Purpose/Objective(s): Following surgical and/or radiological treatment,

up to 75% of head and neck (H&N) cancer survivors encounter lymphe-

dema (LE), characterized by swelling that can impede swallowing and

other vital functions. Early detection and treatment of H&N LE is

International Journal of Radiation Oncology � Biology � Physics928

important to prevent long-term changes in tissues occurring with LE.

“Seeing” lymphatic vessels and function could guide manual lymphatic

drainage therapy to alleviate fluid accumulation.

Materials/Methods: Lymphatic anatomy and function (pumping) were

visualized using near-infrared fluorescence lymphatic imaging (NIRFLI)

with a customized imaging system. Intraorbital and intradermal injections

of indocyanine green (ICG) delivered microdose amounts (less than 300

micrograms total per subject) of the fluorescent dye. Seven study subjects

were imaged before surgery, before radiation, and then longitudinally at

approximate 3-month intervals, for a total of 7 imaging sessions per study

subject. Of note, 1 subject received no radiation.

Results: Head and neck lymphatic vessels and were readily visualized in

all subjects. Dermal lymph backflow was noted in all subjects who

received radiation, 1 of these subjects developed temporary edema, and

another of these subjects developed lymphedema. The subject who

received no radiation did not develop dermal backflow or lymphedema.

Conclusion: Radiation treatment may impart a greater risk of H&N

lymphatic abnormalities. NIRFLI allowed early detection of changes in

lymphatic vessel usage and function after H&N cancer treatment.

Author Disclosure: S.H. Naqvi: None. R.J. Karni: None. I. Tan: None. J.

Rasmussen: None. M. Aldrich: None. J. Morrow: None. E. Sevick:

None.

248Combined Radiation and PD-L1 Blockade Improved Tumor Controlin Mouse Head and Neck Squamous Cell Carcinoma (HNSCC)D.T. Kamdem,1 J.C. Steel,2 T. Wise-Draper,1 W.M. Kassing,3 N. Hashemi

Sadraei,1 M.L. Mierzwa,4 and J.C. Morris5; 1University of Cincinnati,

Cincinnati, OH, 2The University of Queensland, Brisbane, Brisbane,

Australia, 3University Hospital, Cincinnati, OH, United States, 4University

of Michigan, Ann Arbor, MI, 5University of Cincinnati, Cincinnati, OH,

United States

Purpose/Objective(s): Expression of the negative immune regulatory

molecule, programmed death-ligand 1 (PD-L1) by tumor cells is associ-

ated with T cell dysfunction, attenuated antitumor immune responses, and

poor clinical outcomes. Blocking the interaction between PD-L1 and its

receptor, programmed death-1 (PD-1), expressed on activated CD8+

cytotoxic T lymphocytes (CTLs) enhances antitumor activity. Radiation

induces cell death and is known to result in the release of tumor-associated

antigens capable of triggering antitumor responses as well as enhancing

antigen presentation. Radiation therapy (RT) is a commonly used treatment

for locally advanced or recurrent HNSCC. In the current study, we

examined the effect of radiation on major histocompatibility complex class

I (MHC-I) and PD-L1 expression on HNSCC cell lines, and whether the

combination of RT and PD-L1 blockade will improve tumor control.

Materials/Methods: Seven human (UCP1SCC-90, UDSCC2, UMSCC-1,

UMSCC-6, UMSCC-22B, UMSCC-47, and 93W147T), and 3 mouse

(Meer, MTE, and PAM212) HNSCC cell lines were examined for expres-

sion of MHC-I antigens and PD-L1 by flow cytometry after irradiation with

1, 6, and 10 Gy, and at different timepoints (0, 12, 24, 48, 72, and 96 h, and 7

days). The antitumor effect of RT combined with anti-PD-L1 antibody was

evaluated in a subcutaneous mouse Meer HNSCC tumor model.

Results: Both human and mouse HNSCC cell lines constitutively

expressed PD-L1. Irradiation increased the expression of MHC-I and PD-

L1 in a dose and time-dependent manner. This trend was observed in all

cell lines. Irradiation induced both MHC-1 and PD-L1 expression 3 to 4

fold between days 4 and 7, compared to nonirradiated cells. In the mouse

model, treatment with combined RT and anti PD-L1 antibody resulted in

complete tumor eradication in 10 of 16 mice (62.5%). Median survival of

untreated control mice (nZ10) was 33 days (range, 25-41 days), that of

mice receiving XRT alone (nZ6) was 50 days (range, 36-71 days), and

animals treated with the combination of XRT and anti-PD-L1 (nZ8)

survived a median of 101 days (range, 67-101+ days).

Conclusion: Radiation increased expression of both MHC-I and PD-L1 on

tumor cells in a dose- and time-dependent manner. The combination of

radiation and anti-PD-L1 improved survival of mice bearing subcutaneous

Meer HNSCC compared to mice treated with radiation, or anti-PD-L1

alone. The combination radiation therapy and PD-L1/PD-1 blockade

deserve further investigation in HNSCC.

Author Disclosure: D.T. Kamdem: None. J.C. Steel: None. T. Wise-

Draper: None. W.M. Kassing: None. N. hashemi Sadraei: None. M.L.

Mierzwa: None. J.C. Morris: None.

249Low Pretreatment Lymphocyte and White Blood Cell Count Predictfor Disease-Specific Death Among Patients With HPV-PositiveSquamous Cell Carcinoma of the Head and Neck Treated byRadiation TherapyJ.J. Meshman,1 J. Wang,2 E. Abemayor,1 M. St. John,1

A. Mendelsohn,1 D. Wong,1 S. Bhuta,1 R.K. Chin,1 and A.M. Chen2;1University of California, Los Angeles, Los Angeles, CA,2University of California, Los Angeles- David Geffen School of

Medicine, Los Angeles, CA

Purpose/Objective(s): To investigate the prognostic value of pretreatment

hemoglobin (Hgb), white blood cell count (WBC), and circulating

lymphocyte count (CLC) among patients with human papillomavirus

(HPV)-positive and HPV-negative head and neck squamous cell carcinoma

treated by radiation therapy.

Materials/Methods: The medical records of 136 consecutive patients

with locally advanced, nonmetastatic squamous cell carcinoma of the

head and neck (109 oropharynx, 20 larynx, and 7 hypopharynx) with

known HPV status treated with radiation therapy at a single institu-

tion were reviewed. Pretreatment Hgb, WBC, and CLC were recorded

for each patient based on laboratory values obtained within a year

prior to institution of definitive therapy (104 primary radiation ther-

apy and 32 primary surgery followed by radiation therapy). Median

age was 62 years (range, 22-91). HPV status was evaluated for high-

risk and low-risk HPV genotypes by in situ hybridization. We eval-

uated p16INK4A (p16) protein expression by immunohistochemistry.

HPV positivity was defined by HPV high-risk status when available,

followed by p16 status. The Kaplan-Meier method was adopted to

estimate local control, distant control, and disease-specific survival. A

linear regression model was used to estimate the correlation of each

outcome with Hgb among all patients and with WBC and CLC

stratified by HPV status. Mean follow-up was 21 months (range,

3-69).

Results: One hundred and eleven patients had available pretreatment

laboratory data. Sixty-eight patients (61.3%) were HPV positive, and 43

patients (38.7%) were HPV negative. The HPV-positive patients had

similar pretreatment Hgb, WBC, and CLC median values in comparison to

the HPV-negative patients. Overall, 2-year local control, distant metas-

tasis-free survival, and disease-specific survival rates were 78% (95%

confidence interval [CI] 66-85), 92% (95% CI 84-96), and 90% (95% CI

80-95), respectively. In the HPV-positive cohort, distant metastasis

correlated with lower pretreatment WBC and CLC (mean difference: 1.98,

95% CI 0.48-3.47, PZ.01 and 0.5, 95% CI 0.03-0.97, PZ.04, respec-

tively). Disease-specific death also correlated with lower WBC (mean

difference: 1.74, 95% CI 0.16-3.32, PZ.03). However, in the HPV-

negative cohort, neither WBC nor CLC were associated with the outcomes

(P>.05, for all). Hgb did not significantly correlate with any of the out-

comes among all patients.

Conclusion: In this small cohort, low pretreatment CLC and WBC were

correlated with distant failure and disease-specific death among HPV-

positive patients. These associations were not observed in the HPV-nega-

tive cohort. These findings may have implications with respect to patient

counseling and the design of future clinical trials.

Author Disclosure: J.J. Meshman: None. J. Wang: None. E. Abemayor:

None. M. St. John: None. A. Mendelsohn: None. D. Wong: None. S.

Bhuta: None. R.K. Chin: None. A.M. Chen: None.

Volume 94 � Number 4 � 2016 Posters 929

250Synergistic Combination of PD-1 and EGFR Antibodies: A ProposedPhase 1 Clinical Trial Using Nivolumab and CetuximabB. Bastos and A. Zaharcu; Cleveland Clinic Florida, Weston, FL

Purpose/Objective(s): Cetuximab induces head and neck cancer cell

death through ADCC mediated by NK cells. NK cells when activated by

the Fc exposed region of cetuximab bound to EGFR increase production of

IFN-gamma. The high levels of IFN-gamma produced dendritic cell

maturation and also increase CD8 T cell expression of PD1. The combi-

nation of an EGFR mab with a checkpoint (PD1) inhibitor may be syn-

ergistic through an enhancement of CD8 T cell activation and dendritic cell

engagement which will increase cancer cytolysis. Based on this preclinical

data, we proposed a phase 1 clinical trial exploring the combination of

cetuximab plus nivolumab in recurrent head and neck cancer patients. The

goal of the trial is to identify the maximal tolerated dose (MTD) and the

recommended phase 2 dose (RP2D). Secondary endpoints would be the

assessment of response rate, progression-free survival (PFS), and overall

survival (OS).

Materials/Methods: Patients with recurrent head and neck cancers will be

eligible. We recommended a phase 1 clinical trial with a 3+3 dose esca-

lation design as described here: (1) dose Level 1: cetuximab 200 mg/m2 IV

loading dose followed by weekly cetuximab 125 mg/m2 IV nivolumab 1.5

mg/kg IV every 14 days; (2) dose Level 2: cetuximab 300 mg/m2 IV

loading dose followed by weekly cetuximab 187.5 mg/m2 IV nivolumab

2.25 mg/kg IV every 14 days; and (3) dose Level 3: cetuximab 400 mg/m2

IV loading dose followed by weekly 250 mg/m2 IV nivolumab 3 mg/kg IV

every 14 days.

Results: Since cetuximab as single agent has a response rate of 10% to

20% as a single agent in metastatic head and neck cancer, we anticipate a

response rate of 30% with the proposed combination therapy. Since the

agents do not have overlap toxicity, we do not anticipate major toxicities.

Conclusion: PD-1 and EGFR Ab therapy may be synergistic, and further

exploration in a clinical setting is proposed on this abstract.

Author Disclosure: B. Bastos: Advisory Board; Boehringer Ingelheim. A.

Zaharcu: None.

251Shorter Reirradiation Intervals for Head and Neck Cancer AreAssociated With Severe Long-Term ToxicityJ.Y. Lee, E. Sapir, and A. Eisbruch; University of Michigan, Ann Arbor, MI

Purpose/Objective(s): Reirradiation to the head and neck for recurrent

disease or a second primary malignancy is increasing in frequency. A

retreatment interval cutoff of 6 months is often cited when considering

reirradiation, as studies have demonstrated a correlation between retreat-

ment interval and survival. However, treatment-related toxicity can be

substantial after reirradiation, and little is known regarding its de-

terminants. We sought to identify factors that correlate with severe long-

term toxicity.

Materials/Methods: We reviewed treatment plans and outcomes of pa-

tients who underwent intensity modulated radiation therapy (IMRT)-based

reirradiation to the head and neck at a single institution from 2008 to 2015

with data available for both the initial and reirradiation course. Graded

(Common Terminology Criteria for Adverse Events version 4.0) treatment-

related acute and long-term toxicities were analyzed. Patient-, tumor-, and

treatment-related variables including retreatment interval were analyzed as

predictors of outcomes such as locoregional control, overall survival, and

toxicity. Univariate analysis was performed with 2-tailed t tests, covariates

were assessed with multiple regression modeling, and survival was

assessed by the Kaplan-Meier method.

Results: Seventy-one reirradiation patients were identified with a median

retreatment dose of 68 Gy. Reirradiation followed resection in 27 patients

(38%), and concurrent chemotherapy was delivered to 42 patients (59%).

After a median follow-up of 26 months, 21 (30%) were alive and free of

disease. For all patients, median overall survival was 22 months and me-

dian locoregional control was 31 months. Of the 62 of 71 patients that

survived longer than 3 months after reirradiation, 16 patients (26%)

experienced severe long-term toxicity (� grade 3), with 12 (75%) suffering

PEG tubeedependent dysphagia. The median treatment interval in patients

who experienced severe toxicity after reirradiation was 20 months (9-167)

compared to 53 months (8-304) in patients who did not (PZ.017). On

multivariable analysis, only retreatment interval was significant in pre-

dicting severe toxicity. Neither concurrent chemotherapy nor surgery prior

to reirradiation was associated with severe long-term toxicity. Using the

median retreatment interval at which patients experienced severe long-

term toxicity (20 months) as the threshold, median survival was 8 months

versus 27 months, favoring longer intervals (PZ.04).

Conclusion: In a cohort treated with modern IMRT-based reirradiation to

the head and neck, rates of severe long-term toxicity was 26%. To our

knowledge, this is the first report showing that shorter retreatment intervals

are significantly associated with severe long-term toxicities. The increase

in severe toxicity should also be considered in addition to poor survival in

patients with short reirradiation intervals.

Author Disclosure: J.Y. Lee: None. E. Sapir: None. A. Eisbruch: None.

252Patterns of Failure in Human Papillomavirus (HPV)-Positive VersusHPV-Negative Oropharyngeal CancerE. Dave,1 U. Ozbek,1 V. Gupta,2 E. Genden,1 B. Miles,1 M. Teng,1

E. Demicco,1 M. Posner,1 K. Misiukiewicz,1 and R.L. Bakst1;1Icahn School of Medicine at Mount Sinai, New York, NY,2Roswell Park Cancer Institute, Buffalo, NY

Purpose/Objective(s): Human papillomavirus positivity in oropharyngeal

squamous cell carcinomas (OSCC) generally confers a more favorable

prognosis. However, a limited number of patients still develop recurrent

and distant disease posttreatment. We sought to better understand failure

patterns and survival in HPV-positive (HP+) versus HPV-negative (HP-)

patients using data from a single institution’s experiences with OSCC.

Materials/Methods: We identified all OSCC patients in our head and neck

database of 546 patients who had developed local and/or distant failures,

and retrospectively analyzed their treatment, survival, and failure patterns

with regards to their HPV status. Thirty-three such patients were identified

and analyzed. The HPV status of all patients was determined either by p16

immunostaining or polymerase chain reaction. When available, the HPV

genotype was specified for a given patient. All patients were treated with

intensity modulated radiation therapy (IMRT) with/without chemotherapy,

or IMRT with/without chemotherapy adjuvantly with surgery. Primary

endpoints included locoregional failure, distant failure, sites of distant

failure, and overall survival (OS) since time of failure. The Kaplan-Meier

method was used to calculate survival among the patients.

Results: The mean age of our patients was 58.8 years. Thirty (91%) were

male, and 3 (9%) were female. Within this group, 20 (61%) patients were

HPV+ and 13 (39%) HPV-. HPV genotypes were available for 12 patients:

10 HPV16, 1 HPV33, and 1 HPV35. Primary sites included 15 base of

tongue (45.5%), 13 tonsils (39.3%), and 5 oropharynx nonspecified site

(15.1%). Median follow-up time since local or distant failure for all pa-

tients was 16.2 months. Median OS following local failure was not

significantly different between the 2 cohorts (17 months for HPV+ vs 14

months for HPV-, PZ.23). However, HPV+ patients who failed distantly

lived significantly longer than HPV- patients who failed distantly (median

42 months vs 11 months, PZ.004). Two-year OS after distant metastasis

was 65% for HPV+ patients versus 0% for HPV- patients. Notably, HPV+

patients were more likely to develop distant metastases to sites other than

the lung and bones (Table 1).

Conclusion: Following distant recurrences, HPV positivity indicates a

much more favorable prognosis with the potential for long-term survival.

By contrast, for patients with OSCC who fail locally, prognosis may not be

driven by HPV status. Lastly, in HPV+ patients who are undergoing

Abstract 252; Table 1

Distant metastasis site HPV+ HPV-

Lung 12 3Lymph nodes below the clavicle 8 3Bone 4 1Visceral organs 4 2Extremity/soft tissue 2 0Brain 1 0

International Journal of Radiation Oncology � Biology � Physics930

follow-up scans to assess for distant metastasis, imaging of the chest alone

may not be sufficient given the anatomic range of metastatic sites in pa-

tients who are HPV+.

Author Disclosure: E.E. Dave: None. U. Ozbek: None. V. Gupta: None.

E. Genden: None. B. Miles: None. M. Teng: None. E. Demicco: None.

M. Posner: None. K. Misiukiewicz: None. R.L. Bakst: None.

253Role of EphB4 in Radiosensitization of Head and Neck SquamousCell CarcinomaS. Bhatia,1 K. Hirsch,1 K. Kavanagh,1 A. Jimeno,2 X.J. Wang,3

and S. Karam4; 1University of Colorado Denver, Anschutz Medical

Campus, Aurora, CO, 2Department of Medical Oncology, University of

Colorado Denver, Aurora, CO, 3University of Colorado, Aurora, CO,

United States, 4Department of Radiation Oncology, University of Colorado

Denver, Aurora, CO

Purpose/Objective(s): Head and neck squamous cell carcinomas

(HNSCCs) afflict over half a million patients annually worldwide. In the

absence of disease at distant sites, salvage treatment may provide durable

disease control in only approximately 15% of such patients. Our research

goal is to improve radiation therapy for aggressive HNSCCs by identifying

novel targets for radiosensitization. The EphB4 receptor is ubiquitously

expressed in HNSCCs and has been shown to promote tumorigenic and

invasive properties of HNSCCs but the effect of EphB4 on cellular radi-

osensitization has not been investigated. We hypothesize that knockdown

of EphB4 receptor will enhance radiosensitization of HNSCCs by inhib-

iting EphB4 targets involved in radioresistance.

Materials/Methods: To fulfill our objective, we used EphB4-targeting

siRNA and performed clonogenic assays in HNSCC cell lines to determine

the in vitro radiosensitization effect following EphB4 knockdown. Effects

of EphB4-siRNA on cell cycle progression, DNA damage response, and

cell death pathways were also investigated.

Results: We observed a decrease in the survival fractions in HNSCC cell

lines following knockdown of EphB4 at increasing doses of radiation. Cell

cycle analysis showed an enhanced G2 arrest of HNSCC cells following

EphB4 knockdown and radiation exposure. In addition, we observed an

increase in the expression of p-H2AX, a DNA damage marker protein, in

HNSCC cells suggesting activation of DNA damage response pathway

following EphB4 knockdown and radiation exposure. This was further

accompanied by DNA fragmentation and modulation of key apoptotic

markers. Studies are currently underway to determine the effect of EphB4

inhibition on radiosensitization in an in vivo patient-derived xenograft

model of HNSCC.

Conclusion: Our findings support the hypothesis that EphB4 promotes

resistance of HNSCCs to ionizing radiation and its targeted inhibition will

therefore result in enhanced radiosensitization. In conclusion, the suc-

cessful completion of this study will allow us to design functional analyses

of EphB4 targets responsible for radioresistance. From translational point

of view, these targets could serve as the basis for development of combined

therapy, prognostic biomarkers, or patient selection strategies in future

clinical trials.

Author Disclosure: S. Bhatia: None. K. Hirsch: None. K. Kavanagh:

None. A. Jimeno: None. X. Wang: None. S. Karam: None.

254Phase 2a Study of Cetuximab and Dasatinib in Patients WithCetuximab-Resistant Recurrent/Metastatic Head and NeckSquamous Cell Carcinoma (R/M HNSCC)J.L. Geiger,1 A.M. Egloff,2 J. Ohr,3 L. Stabile,3 W. Gooding,3 J. Flaherty,3

M. Gibson,4 J. Grandis,5 and J.E. Bauman6; 1University of Pittsburgh

Cancer Institute, Pittsburgh, PA, 2University of Boston, Boston, MA,3University of Pittsburgh Cancer Institute, Pittsburgh, PA, 4University

Hospitals Seidman Cancer Center, Cleveland, OH, 5University of

Pittsburgh, Pittsburgh, PA, 6University of Pittsburgh Cancer Institute,

Pittsburgh, PA, United States

Purpose/Objective(s): Activation of Src family kinases (SFKs) leads to

resistance to EGFR inhibitors in preclinical models of HNSCC. Dual

EGFR-SFK targeting has been proposed as a way to overcome cetuximab

resistance in R/M HNSCC. We evaluated the combination of cetuximab

and SFK inhibitor dasatinib in patients (pts) with cetuximab-resistant R/M

HNSCC.

Materials/Methods: This was a single-arm, 2-stage phase 2a/2b design

evaluating the open-label combination of cetuximab (250 mg/m2/week)

and dasatinib (150 mg/day) in pts with cetuximab-resistant R/M HNSCC

defined as progressive disease (PD) after prior cetuximab in definitive or R/

M setting. An all-comers design was selected to target an improvement in

the overall response rate (ORR). At the end of the first stage there were no

responses among 12 pts, and the trial was stopped for futility. Baseline

serum was queried for candidate biomarkers mechanistically associated

with Src activation to identify preliminary relationships with clinical

benefit (defined as stable disease [SD] �12 weeks).

Results: Fourteen pts were enrolled from 2012 to 2013. Four pts (29%)

had human papillomavirus-positive oropharyngeal cancers. The median

age was 62 years (range 51-72); median time from previous cetuximab

treatment was 4.7 months (range 0-9.7). The most common grade 3

adverse events (AEs) were infection (14%) and pleural effusion (14%); no

grade 4 AEs were observed. There were no objective responses in the first

12 response-evaluable pts, and the trial was suspended after 14 accruals.

Biomarker analysis compared pts who experienced SD�12 weeks (36%,

NZ5) with pts who had PD (64%, NZ9). Among 5 predefined candidate

biomarkers, we observed a significant association between low baseline

serum interleukin 6 (IL6) and clinical benefit. IL6 is the ligand for JAK2

and a key activator of STAT3, a known mechanism of acquired resistance

to dasatinib. Five pts with SD (duration 5-15 months) had median serum

IL6 of 5 pg/mL versus 34 pg/mL for pts with PD (Wilcoxon test PZ.028).

We hypothesized that baseline IL6/JAK/STAT3 activation represents a de

novo mechanism of resistance to dasatinib. Based on this finding the

second stage of the trial was revised to a phase 2b biomarker-enrichment

design enrolling pts with undetectable serum IL6.

Conclusion: The combination of cetuximab and dasatinib in an unselected

population had no objective responses and was stopped early for futility.

However, a biomarker was discovered in pts who derived clinical benefit:

low baseline serum IL6. Baseline IL6/JAK/STAT3 activation may prevent

dasatinib rescue of clinical cetuximab resistance. Thus, the trial is now

proceeding as an independent, Bayesian, biomarker-selected phase 2b

study in pts with undetectable IL-6. Clinical trial information:

NCT01488318.

Author Disclosure: J.L. Geiger: None. A. Egloff: None. J. Ohr: None. L.

Stabile: None. W. Gooding: None. J. Flaherty: None. M. Gibson: None.

J. Grandis: None. J.E. Bauman: Consultant; Incyte. Partnership; Jaleva

Pharmaceuticals.

255Reirradiation of Recurrent and Second Primary Head and NeckCancer With Proton TherapyM.W. McDonald,1 O. Zolali-Meybodi,2 S.J. Lehnert,2 A.A. Cohen-Gadol,2

and M.G. Moore2; 1Winship Cancer Institute of Emory University, Atlanta,

GA, 2Indiana University School of Medicine, Indianapolis, IN

Volume 94 � Number 4 � 2016 Posters 931

Purpose/Objective(s): To report clinical outcomes of head and neck

reirradiation with proton therapy.

Materials/Methods: Between 2004 and 2014, 61 patients received cura-

tive-intent fractionated proton reirradiation, 90.2% for disease involving

skull base structures and 80.3% recurrent T4, at a median of 23 months

from the most recent prior course of radiation. The most frequent histol-

ogies were squamous cell (54.2%), adenoid cystic (11.0%), and undiffer-

entiated (8.2%) carcinoma. Fourteen patients (23.0%) had carcinomas of

cutaneous origin. Salvage surgery prior to reirradiation was undertaken in

47.5% of patients. Gross residual disease was present in 70.5% of patients.

For patients with microscopic residual disease, the median dose of reir-

radiation was 66 Gy (relative biological effectiveness [RBE]); for gross

disease, 70.2 Gy (RBE). The median cumulative lifetime dose of frac-

tionated radiation was 136 Gy (range 96-203.2 Gy). Concurrent chemo-

therapy was used in 27.9%.

Results: The median follow-up time was 15.2 months; it was 28.7 months

in patients remaining alive. The 2-year overall survival (OS) estimate was

32.7% (95% confidence interval: 20.2%-45.2%), and median OS was 16.5

months. The 2-year estimate of the risk of developing local failure was

23.6%, of developing regional nodal failure 4.0%, and of developing

distant metastases 42.0%. In multivariable analysis, the presence of a

gastrostomy tube prior to reirradiation, gross residual disease, and an

increasing number of prior courses of radiation therapy were associated

with greater hazard ratio for death, while the presence of a gastrostomy

tube prior to reirradiation, gross residual disease, and a dose of reirradia-

tion �60 Gy were associated with a greater hazard ratio for local failure.

The median percent weight loss at completion of reirradiation was 2%

(range 10% weight gain to 10% loss). Acute toxicity of maximum grade 2

occurred in 47.5%, grade 3 in 13.1%, and grade 5 in 1.6%. Late toxicity of

maximum grade 2 occurred in 22.6%, grade 3 in 15.1%, grade 4 in 5.7%,

and grade 5 in 3.8%. There were a total of 3 treatment-related deaths.

Conclusion: Despite a preponderance of unfavorable characteristics

including advanced-stage recurrent disease, a high frequency of macro-

scopic perineural invasion, and an absence of more favorable disease sites

such as the larynx, our cohort outcomes are comparable to other series of

head and neck reirradiation. Reirradiation with proton therapy, with or

without chemotherapy, provided reasonable locoregional disease control,

toxicity profiles, and survival outcomes for an advanced stage and heavily

pretreated population. The risk of distant metastatic disease was high.

Additional data are needed to identify which patients are most likely to

benefit from aggressive efforts to achieve local disease control and to

evaluate the potential benefit of proton therapy relative to other modalities

of reirradiation.

Author Disclosure: M.W. McDonald: None. O. Zolali-Meybodi: medical

student; Indiana University School of Medicine. S.J. Lehnert: medical

student; Indiana University School of Medicine. A.A. Cohen-Gadol:

None. M.G. Moore: None.

256Patterns of Failure After Salvage Surgery and Intensity ModulatedRadiation Therapy Reirradiation for Recurrent Head and NeckSquamous Cell CarcinomaG.V. Martin,1 V. Takiar,2 and J. Phan1; 1The University of Texas MD

Anderson Cancer Center, Houston, TX, 2University of Cincinnati,

Cincinnati, OH

Purpose/Objective(s): Surgical salvage is the preferred treatment for

patients with recurrent head and neck squamous cell carcinomas (HNSCC)

after radiation therapy (RT). Adjuvant reirradiation is reserved for those

with adverse features, but the optimal reirradiation volume remains unclear

given the potential for severe treatment toxicity. Here we evaluated the

patterns of locoregional failure (LRF) after salvage surgery and post-

operative intensity modulated RT (IMRT) reirradiation for patients with

recurrent HNSCC.

Materials/Methods: The records of recurrent HNSCC patients who un-

derwent surgical salvage followed by adjuvant IMRT reirradiation at our

institution were reviewed. Patterns of LRF were determined by 2 methods:

(1) visual comparison (“eyeball method”) of the IMRT treatment plan and

diagnostic computed tomography (CT) scan utilizing anatomic landmarks;

and (2) use of deformable image registration (DIR) between treatment

planning CT and diagnostic CT. Failures were defined as in-field if

completely within the high-dose clinical target volume (CTV-HD), mar-

ginal if <1 cm from the CTV-HD, and out of field if �1 cm from the CTV-

HD or subclinical dose CTV (CTV-SD). All recurrences were verified by

biopsy or radiological progression. Logistic and Cox regression analyses

were performed to identify predictors for LRF.

Results: Fifty-nine patients were analyzed with a median follow-up of 22

months after reirradiation (range 1-115 months) and median dose of 60 Gy

(range 60-70 Gy). Of these, 27 patients (46%) had documented LRF with a

median time to LRF of 5 months after reirradiation (range 1-16 months).

By visual comparison, 7 patients failed in-field (26%), 9 failed marginally

(33%), and 11 failed out of field (40%). Utilizing DIR, 57% of the in-field

failures received <95% of the CTV-HD prescription dose, and 78% of the

marginal failures received <95% of the CTV-SD prescription dose. Of the

in-field and marginal failures, 56% occurred within 1 cm of the surgical

flap. Of the out-of-field failures, 55% recurred in the first or second ech-

elon nodal station, and 27% recurred on the contralateral side. On

correlative analysis, a first recurrence in a pharyngeal mucosal site (vs

nodal, soft tissue, or bone sites) after initial radiation therapy predicted for

higher LRF risk after reirradiation (P<.05). The use of chemotherapy,

CTV dose or volume, or time to first recurrence did not correlate with LRF

risk, although a trend for increased LRF risk was observed for those with

PNI or positive margins after salvage surgery.

Conclusion: The majority of locoregional failures after surgical salvage

and adjuvant IMRT reirradiation for recurrent HNSCC were within or <1

cm from the CTV-HD and occurred within 6 months of reirradiation.

Author Disclosure: G.V. Martin: None. V. Takiar: None. J. Phan: None.

257Salvage Therapy for Locoregionally Recurrent Head and Neck CancerOccurring in a Previously Irradiated Field: A National Survey ofPractice PatternsJ.C. Rwigema,1 N. Lee,2 A.M. Chen,3 and D.E. Heron4; 1University of

California, Los Angeles, Los Angeles, CA, 2Memorial Sloan Kettering

Cancer Center, New York, NY, 3University of California, Los Angeles-

David Geffen School of Medicine, Los Angeles, CA, 4University of

Pittsburgh Cancer Institute, Pittsburgh, PA

Purpose/Objective(s): To evaluate the practices of radiation oncologists

in the United States in regard to salvage therapy for recurrent head and

neck cancer.

Materials/Methods: A customized anonymous survey, consisting of 2

deidentified clinical scenarios with a total of 86 questions was e-mailed to

1862 board-certified or board-eligible radiation oncologists (243 academic;

1619 community based) to evaluate the attitudes and practice patterns in

the care of locoregionally recurrent head and neck cancer occurring in a

previously irradiated field. Two by 2 contingency tables were used to

assess differences between survey responses of specific groups using

Fisher exact test. Multivariate logistic regression analyses were performed

to determine significant demographic predictors of treatment recommen-

dation. A P value of .05 was set for statistical significance.

Results: A total of 271 survey responses from 43 US states and the District

of Columbia were evaluable. Response rates were 37.9% and 11.1% for

academic and community practitioners, respectively. Overall, for the

unresectable recurrent head and neck cancer scenario (case 1), treatment

recommendations were reirradiation with curative intent in 72.7%, reir-

radiation with palliative intent in 4.1%, chemotherapy alone in 5.1%, and

referral to a tertiary center in 18.1%. For the postoperative recurrent head

and neck cancer scenario (case 2), recommendations were observation in

45.8%, reirradiation with curative intent in 32.5%, chemotherapy alone in

10.3%, and referral to a tertiary center in 11.4%. Treatment recommen-

dations changed significantly in both clinical scenarios, when respondents

International Journal of Radiation Oncology � Biology � Physics932

were asked to consider a shorter time interval between radiation courses

(ie, 6 months vs 2 years from previous radiation therapy). On multivariate

analysis, the number of cases treated per year and number of years in

practice significantly influenced treatment recommendations for unre-

sectable recurrent disease (PZ.009 and PZ.036, respectively). For a

postoperative salvage scenario (case 2), only the number of years in

practice significantly predicted treatment recommendations (PZ.016).

There was significant variation in treatment recommendation with regard

to radiation therapy technique, choice of systemic therapy, fractionation

schedules, radiation therapy doses, definition of radiation therapy target,

and frequency of image guidance.

Conclusion: Recommendations for salvage therapy of recurrent head and

neck cancer vary considerably in terms of reirradiation approaches and

systemic therapy recommendations among radiation oncologists in the US.

These findings may be of utility in patient counseling and the design of

prospective clinical trials comparing salvage therapy modalities.

Author Disclosure: J. Rwigema: None. N. Lee: None. A.M. Chen: None.

D.E. Heron: None.

258Association of a Claims-Based Marker of Functional ImpairmentWith Treatment Patterns and Cost in Metastatic Squamous CellCarcinoma of the Head and Neck (mSCCHN)J. Kubisiak,1 M. Chace,1 L. Bowman,2 E. Nash Smyth,2 L. Li,2

E. Kubisiak,1 D. Gilden,1 A.B. Lin,2 and D.M. Gilden1; 1JEN Associates,

Inc., Cambridge, MA, 2Eli Lilly and Company, Indianapolis, IN

Purpose/Objective(s): Information on patients’ performance status (PS)

is important to characterizing treatments and health care costs in the real

world setting; however, this information cannot be directly ascertained

from administrative claims data. The objective of this analysis was to

evaluate the JEN Frailty Index (JFI), a marker of functional impairment

that can be derived from claims data and potentially used as proxy for PS,

in order to describe outcomes among mSCCHN patients.

Materials/Methods: The JFI has been found to be significantly related to

concurrent and future need for long-term care services; it has been utilized

by Centers for Medicare & Medicaid Services and has been evaluated in

studies of multiple sclerosis and Alzheimer’s disease. In the current study,

treatment pattern data for patients with mSCCHN were evaluated using

categories of low, medium, and high JFI, with high JFI corresponding to

high impairment and by extrapolation, to poor PS. The associations be-

tween JFI and treatment patterns as well as monthly total Medicare costs

were assessed. This study population was derived from the 2005-2009

Surveillance, Epidemiology, and End Results (SEER) cancer registry,

linked to Medicare claims for 2002-2010. SCCHN was categorized as

metastatic based on either determination of stage IVC disease at diagnosis

or the existence of secondary or distant cancer diagnoses in Medicare

claims.

Results: A total of 4616 patients with mSCCHN were eligible for study

inclusion. Approximately 60%, 40%, and 41% of the total population

received radiation, surgery, or systemic therapy, respectively. Twenty-nine

percent of the total population had high JFI scores at 3 months post-

metastatic diagnosis. The percentage of patients with high JFI scores

fluctuated over time: 5% one year prior to diagnosis, 34% 6 to 8 months

after diagnosis, and 16% at 18 months postdiagnosis. In patients with

observed deaths, the proportion with high JFI expanded in the 36 months

prior to death, rising from 22% to 51% during the last observation month.

Of the 1902 patients who received systemic therapy, only 18% had high

JFI scores at the time of treatment initiation. There was a trend toward

administering cetuximab monotherapy rather than platinum-containing

regimens in first- and second-line treatment for patients with high JFI

scores. Multivariate regression analyses found a cost impact associated

with high JFI score during the observation period (9.13-fold increased cost

relative to low JFI, P<.0001).

Conclusion: The JFI was used as a proxy for PS. Patterns of systemic

therapy use and cost of care varied according to JFI strata in mSCCHN

patients. This study supports the opportunity and need for further valida-

tion of the JFI as a proxy for PS in observational claims-based oncology

research.

Author Disclosure: J. Kubisiak: contract work; Janssen. contract work; Eli

Lilly and Company. M. Chace: contract work; Eli Lilly and Company,

Janssen Pharmaceutical. L. Bowman: Stock; Eli Lilly. E. Nash Smyth:

Stock; Eli Lilly and Company. Oversee the management and direction,

financial development and stability, planning and policy decisions, and

evaluation of the organization; Cancer Support Community of Central

Ohio. L. Li: Stock; Eli Lilly. E. Kubisiak: contract work; Eli Lilly and

Company, Janssen Pharmaceutical. D. Gilden: contract work; Eli Lilly and

Company. A.B. Lin: Stock; Eli Lilly. D.M. Gilden: Independent

Contractor; Eli Lilly and Company, Janssen Pharmaceutical.

259Elective Neck Management for Squamous Cell Carcinoma Metastaticto the Parotid-area Lymph NodesM.P. Herman,1 R.J. Amdur,2 J.W. Werning,1 P.T. Dziegielewski,1

C.G. Morris,1 and W.M. Mendenhall1; 1University of Florida, Gainesville,

FL, 2University of Florida Hospitals, Gainesville, FL

Purpose/Objective(s): To determine if radiation therapy to the regional

lymphatics is a suitable alternative to elective neck dissection in patients

who undergo parotidectomy for squamous cell carcinoma metastatic to the

parotid lymph nodes.

Materials/Methods: We retrospectively reviewed the medical records of

107 patients consecutively treated from November 1969 to March 2012 for

squamous cell carcinoma metastatic to the parotid lymph nodes with a

clinically and radiographically node-negative neck. Primary therapy con-

sisted of parotidectomy in all cases. We compared regional control in 2

subgroups: 42 patients treated with elective neck dissection and radiation

therapy and 65 patients treated with elective neck radiation therapy alone.

Results: The median time of follow-up was 5.5 years (range, 0.3-30 years)

for all patients and 11 years for living patients (range, 1.8-26 years). There

was 1 neck recurrence in each subgroup: elective neck dissection and RT,

1/42 (2%); and elective radiation therapy alone, 1/65 (1.5%). No patient

experienced a complication related to neck radiation therapy.

Conclusion: Elective neck radiation therapy to a dose of approximately 50

Gy is a suitable alternative to elective neck dissection in patients with

squamous cell carcinoma metastatic to the parotid lymph nodes.

Author Disclosure: M.P. Herman: None. R.J. Amdur: None. J.W.

Werning: None. P.T. Dziegielewski: None. C.G. Morris: None. W.M.

Mendenhall: None.

260Linear AcceleratoreBased Stereotactic Ablative Radiation TherapyReirradiation for Unresectable Recurrent Head and Neck CancerJ. Phan,1 A.S. Garden,1 G.B. Gunn,1 C.D. Fuller,1 S.J. Shah,2 S.J. Frank,1

B.M. Beadle,1 W.H. Morrison,1 S. Garcia,1 H. Wang,1 C. Wang,1 S. Tung,1

M. Edson,1 V. Takiar,3 P.D. Brown,1 and D.I. Rosenthal1; 1The University

of Texas MD Anderson Cancer Center, Houston, TX, 2MD Anderson Sugar

Land, Sugar Land, TX, United States, 3University of Cincinnati,

Cincinnati, OH

Purpose/Objective(s): To report our early institutional experience uti-

lizing linear accelerator (Linac)-based fractionated stereotactic ablative

radiation therapy (HN-SABR) for reirradiation of unresectable recurrent

head and neck tumors.

Materials/Methods: From 2013 to early 2015, 26 patients with biopsy-

confirmed head and neck cancer recurrence after radiation therapy were

treated with HN-SABR and retrospectively analyzed. Patients with <3

months follow-up were excluded. Patients were treated for isolated

Volume 94 � Number 4 � 2016 Posters 933

retropharyngeal or upper neck nodal recurrence (nZ14), skull base

recurrence (nZ7), and pharyngeal mucosal recurrence (nZ5). All patients

were immobilized using an in-house, frameless custom cushion-mask-bite-

block system and underwent real-time image guidance during each frac-

tion. Planning target volume margin was typically 2 to 3 mm with >95%

coverage of prescribed dose and delivered using volumetric modulated arc

radiation therapy (VMAT). Treatment-related adverse events were docu-

mented according to Common Terminology Criteria for Adverse Events

version 4.0. The Kaplain-Meier method was used to estimate clinical

endpoints.

Results: Twenty-one patients (81%) had squamous cell carcinoma. All

patients received 5 fractions delivered every other day. Twenty-three pa-

tients (88%) received a prescribed dose of 45 Gy, 2 patients (8%) received

40 Gy, and 1 patient received 47.5 Gy. Twenty-two patients (86%)

received concurrent weekly cetuximab. With a median follow-up time of

6.2 months (range 3.1-20.7 months), the 6-month overall survival, disease-

free survival, and locoregional control rates were 79%, 73%, and 91%,

respectively. Five patients failed in the head and neck. One had persistent

disease in-field after tongue base treatment, 1 recurred in the soft tissue

adjacent (<1 cm) to the treated skull base, 1 recurred in an inferior

adjacent nodal station after isolated neck treatment, and 2 failed in the

ipsilateral neck after oropharynx treatment. No local failures were

observed in 9 patients treated for isolated retropharyngeal recurrence

(median follow-up of 7.8 months; range 4.2-19.4 months). There were 3

deaths, all due to metastatic disease. There were 5 acute grade 1 adverse

events (3 odynophagia, 1 mucositis, and 1 dysgeusia), 2 grade 2 events

(1 mucositis and 1 odynophagia) that required intermittent nonnarcotic

oral analgesics, and no acute grade 3 toxicity. Five patients (23%)

developed grade 1-2 cetuximab-related folliculitis. One patient developed

edema of his oropharyngeal flap 10 days after treatment of the bilateral

retropharyngeal nodes to 47.5 Gy.

Conclusion: Linac-based SABR utilizing VMAT for head and neck reir-

radiation appears safe, tolerable, and effective. Longer follow-up is needed

to assess late treatment toxicity and tumor control durability.

Author Disclosure: J. Phan: None. A.S. Garden: None. G.B. Gunn:

None. C.D. Fuller: None. S.J. Shah: None. S.J. Frank: None. B.M.

Beadle: None. W.H. Morrison: None. S. Garcia: None. H. Wang: None.

C. Wang: None. S. Tung: None. M. Edson: None. V. Takiar: None. P.D.

Brown: None. D.I. Rosenthal: None.

261Reirradiation for Recurrent and New Primary Head and Neck Cancer:A Single-Institutional ReportM.A. Velez,1 J.C. Rwigema,1 D. Veruttipong,1 J. Wang,2 M.S. St. John,1

E. Abemayor,1 and A.M. Chen2; 1University of California, Los Angeles,

Los Angeles, CA, 2University of California, Los Angeles- David Geffen

School of Medicine, Los Angeles, CA

Purpose/Objective(s): Reirradiation with curative intent for recurrent

head and neck cancer is clinically challenging. We conducted a retro-

spective review of our institutional experience to evaluate the feasibility

and toxicity of head and neck reirradiation for recurrent or second primary

head and neck cancer.

Materials/Methods: Clinical outcomes of 80 patients who underwent

reirradiation for recurrent or second primary head and neck cancer were

retrospectively reviewed and analyzed. Reirradiation was defined as any

overlap between the treatment fields of previous radiation and retreatment.

Disease upon reirradiation was classified as either primary tumor site

recurrence or neck recurrence. Late toxicity was evaluated and categorized

using the Radiation Therapy Oncology Group grading criteria. Longitu-

dinal estimates of survival were calculated using the Kaplan-Meier

method. Univariate analysis was performed using Cox proportional hazard

and logistic regression to determine predictors of outcomes and severe late

toxicity.

Results: From November 1998 to January 2015, 69 of 80 patients (median

age 56 [range 26-84], 26 male, 43 female) who underwent reirradiation at

our institution were evaluable. Intensity modulated radiation therapy

(IMRT) was used in 63 patients. The median follow-up for patients alive at

last follow-up was 11 months (range 2 to 142 months). The most common

site of recurrence or second primary disease was the neck (nZ43, 62%)

and the primary tumor site (nZ26, 38%). Twenty-two patients underwent

salvage surgery upon recurrence before reirradiaton, and 42 received

concurrent systemic therapy. Median dose of initial radiation therapy was

68 Gy (range 9 to 136 Gy). With a median interval from the initial radi-

ation therapy of 42 months (range 2 to 322 months), the median reirra-

diation dose for patients treated with conventional fractionation was 60 Gy

(range 14 to 70 Gy). The median tumor volume was 48.95 mL (range 0.7

to 219.64 mL). The 2-year overall survival and progression-free survival

rates were 54% and 33%, respectively. The 2-year locoregional and distant

control rates were 39% and 65%, respectively. Patients who underwent

salvage surgery prior to radiation treatment had significantly better

locoregional control, progression-free survival, and overall survival rates

(P<.05). Severe (grade 3+) late complications were observed in 21 patients

(30%). No significant prognostic factors for severe late toxicity were found

on univariate analysis.

Conclusion: Our results showed similar locoregional control rates to his-

torical controls for patients undergoing reirradiation of recurrent and

second primary head and neck cancer. The observed rates of severe late

toxicity were significant. Further studies are necessary to optimize

locoregional control while reducing treatment-related morbidity.

Author Disclosure: M.A. Velez: None. J. Rwigema: None. D.

Veruttipong: None. J. Wang: None. M.S. St. John: None. E. Abemayor:

None. A.M. Chen: None.

262Evaluation of Weekly Paclitaxel, Carboplatin, and Cetuximab inHead and Neck Cancer Patients With Incurable DiseaseL. Narveson,1 E. Kathol,2 M. Rockey,2 D. Henry,2 D. Grauer,3

and P. Neupane4; 1North Dakota State University, Fargo, ND,2The University of Kansas Hospital, Kansas City, KS, 3The University of

Kansas School of Pharmacy, Lawrence, KS, 4The University of Kansas

School of Medicine, Kansas City, KS

Purpose/Objective(s): Weekly paclitaxel, carboplatin, and cetuximab

(PCC) has been found to be efficacious and well tolerated in patients with

squamous cell carcinoma of the head and neck (SCCHN) with good per-

formance status (PS) when used as induction chemotherapy. Use of PCC in

incurable SCCHN in patients with poor PS or in a noninduction setting is

an area that warrants further evaluation. Current recommendations for

incurable disease consist of a platinum-based regimen with fluorouracil

and cetuximab. Studied in patients with PS of 0 to 1, the fluorouracil-based

regimens were associated with significant toxicities. Therefore, weekly

PCC may offer an appealing, less toxic alternative for incurable patients

with poor PS.

Materials/Methods: This retrospective analysis evaluated 41 patients with

very advanced or metastatic head and neck cancer who had received PCC

(paclitaxel 80 mg/m2, carboplatin AUC 2, and a cetuximab 400 mg/m2

loading dose, followed by 250 mg/m2 weekly) for up to 6 cycles between

April 2008 and September 2014. Maximal response achieved and pro-

gression-free survival (PFS), as well as dose intensity and adverse effects,

were evaluated.

Results: Of the 41 patients evaluated, baseline PS ranged as follows: PS of

2 (41%), PS of 1 (54%), and PS of 0 (5%). Patients received 2 to 6 cycles,

averaging 4 cycles. Thirty-one patients (76%) required treatment to be

held, delayed, or dose reduced, most commonly for hematologic toxicities.

Grades 3/4 neutropenia occurred in 16 patients (39%), grades 1/2 neu-

tropenia in 12 patients (29%), with grades 3/4 thrombocytopenia in 1

International Journal of Radiation Oncology � Biology � Physics934

patient (2%) and grades 1/2 thrombocytopenia in 2 patients (4%). No

patients developed febrile neutropenia or required hospitalization due to

treatment. Partial radiographic response occurred in 15 patients (37%),

complete radiographic response in 2 patients (5%), stable disease in 14

patients (34%), and progression in 8 patients (20%). PFS ranged from 1.6

to 45 months, with a median duration of 4.6 months, and median overall

survival of 5.25 months.

Conclusion: Analysis indicates that use of weekly PCC appears to be an

effective and well-tolerated treatment option for patients with incurable

squamous cell carcinoma of the head and neck, specifically with PS of

0 to 2.

Author Disclosure: L. Narveson: None. E. Kathol: None. M. Rockey:

None. D. Henry: None. D. Grauer: None. P. Neupane: None.

263Reirradiation With Simultaneously Integrated Boost (SIB) inPatients With Local Recurrence of Squamous Cell Carcinoma of theHead and Neck: Own ExperienceA.V. Mikhailov, N.A. Vorobyov, V.P. Sokurenko, E.V. Smirnova,

G.I. Andreev, D.D. Puchkov, and M.V. Rukhlenko; DTC IIBS named after

S.Berezin, Saint-Petersburg, Russia

Purpose/Objective(s): Locoregional recurrence is a major cause of death

in patients with squamous cell carcinoma of the head and neck (HNSCC).

At the moment, there are no clear recommendations and standards

regarding the timing, total doses, and dose tolerance of normal tissues to

re-exposure. Based on limited studies on the reirradiation with high total

doses, we evaluated the tolerability of high-dose reirradiation with

simultaneous integrated boost.

Materials/Methods: Fourteen patients with histologically confirmed

locoregional recurrence of HNSCC received reirradiation. Median time

after primary radiation therapy course was 60 months. The treatment

volumes and total doses were formed as follows: gross tumor volume

(primary lesion and involved lymph nodes, delineated on computed to-

mography [CT], magnetic resonance imaging, and 18F-fluorodeox-

yglucose positron emission tomography/CT) + clinical target volume

(0.5-1.0 cm) + planning target volume (PTV; 0.3-0.5 cm) was treated to the

total dose equivalent to 66 to 70 Gy of conventional fractionation, the

upper neck (if indicated, levels I-III + PTV 0.5 cm) to 60 Gy, the lower

neck (if indicated, levels IV-V + PTV 0.5 cm) e equivalent to 50 Gy.

Single doses to these volumes were 2.14 to 2.21 Gy, 2.0 Gy, and 1.8 Gy,

respectively. Radiation treatment was once a day, 5 days a week, 6 weeks

long (30 fractions). A treatment planning system was used (intensity

modulated radiation therapy [IMRT]), and patients were treated with 2

linear accelerator models. According to the literature, in a year after pri-

mary irradiation, almost complete recovery of normal tissue tolerances is

observed. Tolerances of the eye, lens, optic nerves and chiasm, brain stem,

spinal cord, parotid gland, intact mucosa of the mouth and pharynx were

not exceeded. Patient positioning accuracy was controlled by kV-imaging

daily and cone beam CT weekly.

Results: Three of 14 patients received the full course of radiation therapy

without a break. Radiation toxicity manifested with grade 2 oral and

pharyngeal mucositis and grade 2 radiation epidermitis. After 1 month,

almost complete relief of radiation mucositis and dermatitis was observed.

One patient took a break of 7 days after the 25th fraction due to the

development of grade 3 mucositis and grade 3 dysphagia.

Conclusion: Using the technique of SIB with IMRT during curative reir-

radiation of recurrent HNSCC is available with maintaining satisfactory

tolerability.

Author Disclosure: A.V. Mikhailov: None. N.A. Vorobyov: None. V.P.

Sokurenko: None. E.V. Smirnova: None. G.I. Andreev: None. D.D.

Puchkov: None. M.V. Rukhlenko: None.

264Outcomes of Patients With Recurrent and Metastatic Squamous CellCarcinoma of the OropharynxB.J. Debenham1 and R.N. Banerjee2; 1University of Alberta Edmonton,

AB, Canada, 2University of Calgary, Calgary, AB, Canada

Purpose/Objective(s): Oropharyngeal cancer is increasing in incidence inNorth America. There is little data to advise patients on their outcomes if

they present with metastatic disease or recurrent disease after primary

treatment. The objective of our study was to report the outcomes of these

patients from a large retrospective database.

Materials/Methods: This study selected all stage III-IV nonmetastatic

oropharyngeal squamous cell cancer patients treated with curative intent

and metastatic patients treated with palliative intent in our province be-

tween 2006 and 2012. Institutional ethics board approval was granted.

Patient data was extrapolated from a prospective registry and retrospective

review of medical records. Kaplan-Meier estimates of overall survival

(OS) and 95% confidence interval (CI) were obtained for recurrent and

metastatic patients. Logistic regression was used to explore the association

between patient, tumor, and treatment factors for the outcomes of these

patients.

Results: Four hundred thirty-three patients were identified from our reg-

istry. Of the 433 patients, 12 had metastatic disease on initial presentation

and 76 were identified as recurring after initial primary treatment with

surgery or radiation therapy, for a total group of 86 patients. Patients with

recurrent disease had a median survival time of 8 months and a 1-year OS

rate of 41% after being diagnosed with recurrence. Patients with metastatic

disease at presentation had a median survival time of 9.6 months and a

1-year OS rate of 17%. Only 5 out of 76 (6.6%) recurrences were

successfully salvaged. Thirty percent of patients received palliative

chemotherapy, and 40% received palliative radiation therapy. Receiving

more than 1 cycle of palliative chemotherapy was significant on univariate

analysis (HR 1.67, PZ.03) for survival, but not on multivariate analysis.

Age, comorbidity index, p16 status, and smoking status were not signifi-

cant in univariate or multivariate analysis.

Conclusion: Patients with recurrent or metastatic oropharyngeal cancer

have a poor prognosis. There may be a benefit to palliative chemotherapy,

although the percentage of patients receiving chemotherapy was small.

More effective strategies are needed to improve the salvage rate of

recurrent patients.

Author Disclosure: B.J. Debenham: None. R.N. Banerjee: None.

265Disease-Free Interval After Primary Chemoradiation Therapy CanHelp Predict the Utility of Salvage Surgery in Patients WithSquamous Cell Carcinoma of the LarynxT.K. Hamilton,1 D. Bush,1 S. Langevin,1 K. Casper,2 and J. Mark1;1University of Cincinnati, Cincinnati, OH, 2Univ of Michigan,

Ann Arbor, MI

Purpose/Objective(s): Each year, more than 12,000 people in the United

States are diagnosed with cancer of the larynx, and nearly 3700 die from

the disease. The 2 primary treatment options for patients with advanced

disease are chemoradiation (CRT) or surgery. The results of the Veterans

Affairs Laryngeal Study Group demonstrated equivalent 2-year overall

survival between these 2 approaches. Due to this, definitive therapy for

locally advanced laryngeal cancer has shifted to an organ-preservation

approach with surgery increasingly being reserved for salvage; however,

approximately 25% of the patients initially undergoing CRT will eventu-

ally require salvage total laryngectomy due to recurrent disease or an

incompetent larynx. Patients that require salvage surgery have the possi-

bility of long-term survival, but there is a subset of patients that do poorly,

Volume 94 � Number 4 � 2016 Posters 935

either from postoperative complications or early secondary recurrence. The

aim of the study was to identify clinical factors that might predict early

failure and/or worse outcome after salvage laryngectomy in patients

initially treated with definitive CRT for advanced laryngeal cancer.

Materials/Methods: This was a retrospective chart review of salvage

laryngectomies performed for advanced laryngeal squamous cell carci-

noma at a single academic institution. An existing prospectively collected

database of head and neck cancer patients was queried to identify patients

who had undergone salvage laryngectomies. Additional pertinent infor-

mation on the identified patients was gathered from manual chart review

for analysis after institutional review board approval was obtained. Patients

were separated into groups based upon how quickly they recurred after

primary therapy with CRT: group 1 was defined as patients that relapsed in

less than 1 year, and group 2 consisted of patients that recurred after the

1-year mark. Survival in each group was estimated using the Kaplan-Meier

method, and differences between curves were assessed by the log-rank test.

A multivariable Cox proportional hazards model was used to estimate the

hazard ratio, adjusted for age and initial stage at diagnosis.

Results: The 2 groups were homogeneous with respect to age, sex, race,

and disease location/severity. The Kaplan-Meier analysis indicated that

laryngeal cancer patients with a recurrence less than 1 year after definitive

chemoradiation (group 1) fared poorer than those with later recurrences

(PZ.06). After adjusting for patient age and stage at diagnosis, patients in

group 1 still experienced poorer outcomes, although the results were not

significant (hazard ratioZ2.30, 95% confidence interval: 0.79-6.73).

Conclusion: Patients with larynx cancer who initially undergo CRT who

relapse in less than a year have lower overall survival and are less likely to

benefit from a salvage total laryngectomy.

Author Disclosure: T.K. Hamilton: None. D. Bush: None. S. Langevin:

None. K. Casper: None. J. Mark: None.

266Novel Preclinical In Vitro and In Vivo Model Systems for AdenoidCystic Carcinoma of Salivary GlandS. Agarwal, C. Chen, S. Choudhury, X. Liu, E. Glasgow, and R. Schlegel;

Georgetown University Medical Center, Washington, DC

Purpose/Objective(s): Adenoid cystic carcinoma (ACC) is one of the

most malignant salivary gland cancers with a high incidence of both local

and distant (lung and bone) metastases. In order to study this tumor type, a

reliable model system is critical. Patient-derived mouse xenografts (PDX)

have been shown to maintain the histology and gene expression profile of

the primary tumor, making them a valid model system. However, PDX

models suffer from high cost for generation and maintenance, low take rate

(30%-50%), lack of manipulation, and high throughput capability. Lack of

authenticated cell line has compromised studies of ACC basic biology and

drug development.

Materials/Methods: A new conditionally reprogrammed (CRC) method

described previously combines the use of irradiated mouse fibroblasts and

a ROCK inhibitor to induce the rapid and long-term growth of normal and

tumor cells without any additional immortalization. We have used this

technology to establish cell cultures using PDX tissue materials from 5

different individuals. We also established an in vivo zebrafish models

system to study migratory and metastatic behavior of tumor cells using

2-day post fertilized embryos. We used transgenic zebrafish (flk:GFP)

expressing green reef coral fluorescent protein in the vascular endothelial

under the control of a VEGFR2 promoter, which makes it easy to follow

the tumor cell trafficking in real time. The injected cells (tumor or normal)

in the yolk sac were labeled with live dye (Mitotracker) allowed us to

follow it in real time using live imaging system.

Results: We have stable cultures from 5 individual PDX tumors using

CRC technology. One ACC cell line, ACC11, has shown to maintain the

MYB-NF1B translocation, mutations in FGFR2 and ATM genes and

overexpression of Myb-NF1B fusion protein similar to the tumor of origin

(PDX tumor). We also established an in vivo zebrafish xenograft model

system for ACCs, using ACC11 as a prototype, we showed that these cells

are migratory and possess metastatic potential as a fraction of tumor cells

moved from the yolk sac to the tail via vascular invasion within 3 days

postinjection. Similar results were obtained when PDX tissue material for

ACC11 was transplanted directly to the yolk sac. In contrast, injection of a

non-ACC salivary gland cell line did not show any movement of cells even

after 7 days of post injection.

Conclusion: We have successfully established a toolkit consisting of an in

vitro (cell line) and an in vivo (Zebrafish) model system for biological and

translational research.

Author Disclosure: S. Agarwal: Research Grant; Adenoid Cystic Carci-

noma Research Foundation. C. Chen: None. S. Choudhury: None. X.

Liu: None. E. Glasgow: None. R. Schlegel: Research Grant; National

Institute of Health. Royalties from HPV vaccine; Georgetown University.

Patent/License Fees/Copyright; Propagenix. Advise on scientific di-

rections; Propagenix.

267WITHDRAWN

268The Cancer Genome Atlas Data Suggest Only a Modest Role for TP53in Head and Neck Squamous Cell Carcinoma PrognosticationJ.W. Rocco1 and E.A. Mroz2; 1The James Cancer Hospital and Solove

Research Institute Wexner Medical Center at The Ohio State University,

Department of Otolaryngology-Head and Neck Surgery, Columbus, OH,2The James Cancer Hospital and Solove Research Institute, The Ohio State

University Wexner Medical Center, Columbus, OH

Purpose/Objective(s): Identifying combinations of clinical variables and

biomarkers best related to outcome is important for informing clinical

decisions and for designing trials. Early studies could not distinguish

human papillomavirus (HPV)-positive from HPV-negative head and neck

squamous cell carcinoma (HNSCC), so it is now important to determine

which combinations of prognostic variables are best related to outcome in

these types of HNSCC. The Cancer Genome Atlas (TCGA) provides a rich

set of clinical and molecular data on HNSCC; the mix of cases is repre-

sentative of higher stage disease. Taking HPV status into consideration, we

combined TCGA survival analysis with best-subset and LASSO variable

selection to identify combinations of prognostic variables most closely

associated with outcome.

Materials/Methods: Ten prognostic variables potentially related to

outcome had sufficient TCGA data for analysis (259 cases, 114 deaths):

age, gender, smoking history, T and N classifications, TNM stage, tumor

grade, TP53 mutation status, HPV status (32 HPV positivs including 1

TP53 mutant), and mutant-allele tumor heterogeneity (MATH). Best-

subset selection examined the relation of all combinations of variables to

overall survival in Cox proportional hazards models. LASSO selection

added individual variables sequentially as Cox model complexity was

allowed to increase. To avoid overfitting, model complexity was penalized

by the Akaike Information Criterion in best-subset analysis, and by the

sum of the absolute values of regression coefficients in LASSO.

Results: Surprisingly, best-subset selection omitted TP53 mutation status

as a prognostic variable. Age (P<.001), smoking history (PZ.001), MATH

(PZ.005), HPV status (PZ.008), and N classification (PZ.01) was the

combination of variables best related to outcome. The pattern of sequential

variable selection by LASSO provided an explanation for this omission.

TP53 mutation was the first variable selected by LASSO for inclusion in

the 259-case Cox model, followed by smoking history. As MATH, HPV

status, N classification, and age entered the model, however, the

International Journal of Radiation Oncology � Biology � Physics936

contribution of TP53 status diminished. Variable selection by LASSO was

particularly striking in HPV-negative cases, as 5 other variables (smoking

history, N classification, MATH, tumor grade, and age) were included in

the model before TP53.

Conclusion: Although TP53 mutation was identified long ago as a prog-

nostic variable in HNSCC, these results suggest that its prognostic value

might be modest outside of its acting as a surrogate for HPV status. These

results should be interpreted cautiously, as they are based on a single data

set and not from a predesigned clinical study. Nevertheless, despite its

importance in the biology of HNSCC, the role of TP53 mutation as a

prognostic variable in HPV-negative HNSCC may need to be reconsidered

in future studies, particularly as novel biomarkers are evaluated.

Author Disclosure: J.W. Rocco: Patent/License Fees/Copyright; Massa-

chusetts General Hospital (MGH). Development and discussion of clinical

trial ideas; NCI. comments and reviews on clinical trial ideas; NCI. E.A.

Mroz: Patent/License Fees/Copyright; Massachusetts General Hospital

(MGH).

269miR-203 Inhibits Human Papillomavirus Oral Tumor Growth bySuppressing Proliferation in Differentiated Tumor CellsM.T. Spiotto,1 J. Bechill,2 and R. Zhong2; 1University of Chicago,

Chicago, IL, 2The University of Chicago, Chicago, IL

Purpose/Objective(s): In order to facilitate viral replication, the human

papillomavirus (HPV) oncogene E7 stimulates differentiated keratinocytes

in the suprabasal epithelial layers to initiate DNA synthesis. Here, we

studied the extent to which HPV oncogenes caused proliferation in

differentiated tumor cells and the pathways governing this process.

Materials/Methods: Tamoxifen treatment of triple transgenic KHR mice

induced expression of the HPV oncogenes E6 and E7 and a mutant

KrasG12D oncogene to cause oral tumor development. As a control, we

used HPV-negative KR mice that developed oral tumors in which prolif-

eration was limited to the basal stem cell compartment. To study how a

regulator of epithelial differentiation, miR-203, impacted primary tumor

growth, we generated a dual inducible mouse model, KHR-203, where

doxycycline (DOX) induced miR-203 expression independent of HPV and

KrasG12D oncogenes. Gene expression was assessed using Affymetrix

Mouse Gene 1.0 ST arrays, quantitative RT-PCR, and in situ hybridization

(ISH). The proliferation markers Mcm7, Pcna, and BrdU as well as the

keratinocyte differentiation marker CK10 were assessed by immunohis-

tochemistry or immunofluorescence.

Results: Compared to HPV-negative KR tumors, HPV-positive KHR tu-

mors grew faster (tumor volume at d24: 387.1�74.3 mm3 for KHR tumors

vs 140.1�66.9 mm3 for KR tumors; P<.001). KHR oral tumors had more

cells expressing proliferation markers Mcm7 (81.5�2.2% of KHR vs

23.8�0.7% of KR tumors; P<.001) and Pcna (76.8�1.4% of KHR vs

31.9�1.4% of KR tumors; P<.001). Furthermore, KHR tumors, but not

KR tumors, had partially differentiated CK10+ cells incorporating BrdU

indicating that proliferation occurred in partially differentiated cells.

Expression profiling demonstrated that 112 genes and 1 miRNA, miR-203,

were differentially expressed between HPV-positive and HPV-negative

tumors. qRT-PCR confirmed that miR-203 was downregulated 1.9-fold in

HPV-positive tumors (P<.01), and ISH demonstrated that loss of miR-203

occurred primarily in the suprabasal layers. In KHR-203 oral tumors, DOX

treatment induced miR-203 expression and suppressed HPV-positive tumor

growth (tumor volume at d24: 147.3�18.9 mm3 for DOX-treated vs

426.1�25.3 mm3 for vehicle-treated tumors; P<.0001). The decrease

growth of DOX-treated tumors was associated with decreased proliferation

in differentiated tumor cells as measured by percentages of Mcm7-positive

cells (P<.001) and Pcna-positive cells (P<.001).

Conclusion: HPV oncogenes accelerated oral tumor growth by inducing

proliferation of partially differentiated tumor cells that was regulated by

miR-203. Our results suggest that restoring the miR-203 pathway in

primary tumors may antagonize HPV oncogenes to inhibit cell prolifera-

tion and tumor growth.

Author Disclosure: M.T. Spiotto: None. J. Bechill: None. R. Zhong:

None.

270A Regimen Combining the Wee1 Inhibitor AZD1775 With HistoneDeacetylase Inhibitor Vorinostat is Highly Active Against Head andNeck Squamous Cell Carcinoma Harboring High-risk TP53 MutationsA.A. Osman, N. Tanaka, N.L. Silver, N.N. Kalu, A.A. Patel, J. Wang,

L. Tang, M.J. Frederick, F.M. Johnson, S. Fu, and J.N. Myers; The

University of Texas MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): The cure rate for patients with advanced head and

neck squamous cell carcinoma (HNSCC) remains poor due to resistance to

standard therapy primarily consisting of chemoradiation. Since mutation of

TP53 in HNSCC occurs in 60% to 80% of nonhumanpapillomavirus

(HPV)-associated cases and is in turn associated with resistance to these

treatments, novel therapeutic approaches are needed to overcome drug

resistance and improve survival outcomes in patients with advanced

HNSCC. Wee-1 is a kinase that has been linked to DNA damage-induced

G2/M arrest, owing to its ability to inactivate cyclin-dependent kinase 1

(CDK1) through phosphorylation of the Tyr15 residue. Our laboratory has

shown that the Wee-1 kinase inhibitor AZD1775 sensitizes HNSCC cells

harboring high-risk p53 mutations to cytotoxic therapies both in vitro and

in vivo. Vorinostat is a small molecule inhibitor of histone deacetylase

(HDAC) that has been shown in vitro and in vivo to have promising

anticancer activity. Treatment with vorinostat alone shows preferential

cytotoxicity for mutant p53 HNSCC cells. This finding supports the

rationale to use vorinostat-based regimens to achieve maximum synthetic

lethality in HNSCC with p53 mutations. In this study, we evaluated the

efficacy of a regimen combining vorinostat and AZD1775 in HNSCC cells

with a variety of p53 mutations.

Materials/Methods: Clonogenic survival assays were performed to

examine the in vitro sensitivity of several TP53 mutant HNSCC cell lines

following treatment with vorinostat and AZD1775. Cell cycle and western

blotting analyses were performed to investigate cellular mechanisms. An

orthotopic mouse model of oral cancer and HNSCC patient-derived tumor

xenografts (PDX) were used to evaluate in vivo efficacy of the drugs.

Results: Vorinostat synergized with AZD1775 in vitro and reduced cell

survival of mutant p53 HNSCC cells. Interestingly, addition of vorinostat

had no effect on AZD1775 responses in the wild-type p53 HNSCC cells. It

appears that the reduction in cell survival with vorinostat or combination

treatment is mediated through apoptosis. Treatment of HNSCC cells with

vorinostat and AZD1775 increased p21 induction and promoter activity

independent of p53 expression. Interestingly, shRNA knock-down of p21

did not attenuate the lethal effect of combined treatment, suggesting that

p21 might be necessary but not sufficient for vorinostat-mediated cell

death in these cells. Finally, coadministration of vorinostat with AZD1775

resulted in significant tumor growth inhibition and prolonged animal sur-

vival in an orthotopic mouse model of oral cancer and HNSCC patient-

derived xenografts.

Conclusion: A regimen combining AZD-1775 with vorinostat is highly

active in preclinical models of p53 mutant head and neck cancer.

Author Disclosure: A.A. Osman: None. N. Tanaka: None. N.L. Silver:

None. N.N. Kalu: None. A.A. Patel: None. J. Wang: None. L. Tang:

None. M.J. Frederick: None. F.M. Johnson: None. S. Fu: None. J.N.

Myers: None.

271E2F1 Mediates Human Papillomavirus (HPV) Oncogene Toxicity andSuppresses HPV Oral Tumor GrowthM.T. Spiotto,1 R. Zhong,2 J.M. Melotek,1 and J. Bechill2; 1University of

Chicago, Chicago, IL, 2The University of Chicago, Chicago, IL

Volume 94 � Number 4 � 2016 Posters 937

Purpose/Objective(s): The human papillomavirus (HPV) expresses

the viral oncogene E7 that inhibits the retinoblastoma protein (RB1).

RB1 mediates contradictory cell growth and cell death pathways via

E2F family members. Here, we assessed the extent to which HPV on-

cogenes caused toxicity as measured by mouse survival and tumor

growth.

Materials/Methods: iHPV mice contained a LoxP-Stop-LoxP (LSL)-

iE6E7 transgene in which conditional E6E7 expression is regulated by Cre

recombinase. We constitutively expressed HPV oncogenes by breeding

iHPV transgenic mice to CMV-Cre transgenic mice expressing Cre

recombinase under a CMV promoter (CMV-HPV mice). We induced HPV

oncogene expression in adult mice using RosaHPV mice containing the

iHPV transgene and a Rosa-CreERtam transgene expressing a tamoxifen-

regulated (TAM) regulated Cre recombinase in all tissues. We studied

primary oral tumors using triple transgenic KHR mice containing a

K14-CreERtam transgene, a LSL-iE6E7 transgene and a LSL-KrasG12D

transgene that formed HPV-positive oral tumors after TAM treatment. We

assessed the role of E2f1 on survival and oral tumor formation using

RosaHPV-E2f1-/- mice and KHR-E2f1-/- mice, respectively, that contained

a homozygous E2f1deletion. HPV oncogene expression and E2f target

gene expression was assessed by quantitative RT-PCR.

Results: Induction of HPVoncogenes caused embryonic lethality as CMV-

HPV double transgenic mice were born at significantly lower frequencies

compared to mice carrying single transgenes (P<.0001). Tamoxifen

treatment of adult RosaHPV (RosaHPV+TAM) mice caused recombina-

tion of the LSL-E6E7 transgenes and HPV oncogene expression in all

organs tested. Furthermore, RosaHPV+TAM mice had decreased survival

compared to vehicle treated RosaHPV mice (median survival: 50d for

RosaHPV+TAM vs not reached for RosaHPV-TAM; P<.0001). Decreased

survival in RosaHPV+TAM mice was associated with focal necrosis in

hepatocytes and pancreatic tissues and the activation of the E2f target

genes. Deletion of E2f1 increased survival of RosaHPV+TAM mice indi-

cating that E2f1 mediated HPV oncogene toxicity (median survival: not

reached for RosaHPV-E2f1-/- mice vs 49 days for RosaHPV-E2f1+/- mice

vs 30 days for RosaHPV-E2f1+/+; P<.0001). Compared to tumors with

heterozygous loss of E2f1, KHR tumors with homozygous loss of E2f1

grew faster and had more proliferating tumor cells as measured by Pcna

immunohistochemistry (tumor volume at d18: 453.7 mm3 for KHR-E2f1-/-

vs 139.7 mm3 for KHR-E2f1+/-; PZ.0004).

Conclusion: Our results indicate that HPV oncogenes activated the E2f1

pathway to cause toxicity in normal mice and to suppress oral tumor

growth. These results suggest that selective modulation of the E2f1

pathway, which is activated in HPV tumors, may facilitate tumor

regression.

Author Disclosure: M.T. Spiotto: None. R. Zhong: None. J.M. Melotek:

None. J. Bechill: None.

272A Modular Polymer Platform That Delivers Recombinant Cytokinesand Cisplatin Allows for De-escalation of Radiation Therapy in anAnimal Model of Head and Neck Squamous Cell CarcinomaJ. Mallen St. Clair,1 D. Wong,2 A.M. Chen,3 S. Dubinett,1 S. Sharma,1

B. Wu,1 and M.S. St. John2; 1UCLA, Los Angeles, CA, 2University of

California, Los Angeles, Los Angeles, CA, 3University of California, Los

Angeles- David Geffen School of Medicine, Los Angeles, CA

Purpose/Objective(s): To evaluate the safety and efficacy of a novel

modular polymer platform for head and neck squamous cell carcinoma

(HNSCC). Fifty percent of HNSCC patients fail primary management, and

salvage of patients with recurrent disease is of paramount importance. We

had previously shown the antitumor efficiency of this novel polymer in

delivering chemokines (CCl21) and cisplatin in an animal model of

SCCHN. Here we evaluate the safety and efficacy of this polymer in

combination with radiation therapy (RT) in an effort to see if this com-

bination allows for a de-escalation of RT.

Materials/Methods: SCCVII/SF tumors were established in in C3H/HeJ

mice. Tumors were then treated with (1) no polymer; (2) plain polymer; (3)

CCl21 polymer; (4) cisplatin polymer; or (5) combination CCl21 and

cisplatin-secreting polymer. The mice were then treated with 3 different

doses of RT. Tumor size was measured every day until the mice were

euthanized. Four weeks later, necropsy was performed to evaluate for

vascular or nerve damage and to assess tumor size and weight.

Results: Cisplatin polymer, CCL21 polymer, and the combination CCl21-

cisplatin polymer effectively reduced SCCVII/SF tumors in the C3H/HeJ

mice by over 16-fold (P<.01) as compared to control and plain polymer

groups. In addition, treatment with Cisplatin polymer, CCL21 polymer,

and the combination CCl21-cisplatin polymer allowed for a 4-fold

reduction in the dose of RT required. Histopathology revealed no adverse

tissue effects when the cisplatin polymer was inserted in direct contact

with the carotid artery, jugular vein, or vagus nerve.

Conclusion: Our promising results indicate that this polymer may repre-

sent a new therapeutic modality for patients with HNSCC that is safe and

efficacious. Our data provide a strong rationale for further evaluation of

this polymer in deintensification of radiation therapy. Once this polymer

platform is further optimized we will plan for the ultimate validation in the

context of a prospective trial in patients with unresectable advanced or

recurrent HNSCC.

Author Disclosure: J. Mallen St. Clair: None. D. Wong: None. A.M.

Chen: None. S. Dubinett: None. S. Sharma: None. B. Wu: None. M.S.

St. John: None.

273Cetuximab Has Antiviral Activities in Human Papillomavirus(HPV)-Infected Cells and HPV-Associated TumorsA.M. Griego,1 M.A. Ozbun,1 P.F. Barraza,1 C. Hu,1 A. Dziduszko,1

B.K. Crawley,1 H.J. Hathaway,1 and J.E. Bauman2; 1University of New

Mexico School of Medicine and Comprehensive Cancer Center,

Albuquerque, NM, 2University of Pittsburgh Cancer Institute, Pittsburgh,

PA, United States

Purpose/Objective(s): Human papillomaviruses (HPVs) are small DNA

viruses with species specificity and strict tropism for squamous epithelium.

Oncogenic HPV16 is associated with up to 72% of HPV-positive (HPV+)

oropharyngeal squamous cell carcinomas (OPSCC). Patients with locally

advanced, HPV+ OPSCC have more favorable outcomes than those with

HPV-negative (HPV-) cancers when treated with radiation and cetuximab,

a monoclonal antibody against the epidermal growth factor receptor

(EGFR). EGFR overexpression is a hallmark of HPV- head and neck

cancer, whereas EGFR expression is low in HPV+ cancer. Thus, the mo-

lecular mechanisms underlying prognostic advantage in the context of

cetuximab radiation are unclear. This clinical paradox led us to hypothe-

size that HPV oncoproteins augment EGFR signaling independent of

EGFR expression level, establishing a positive feedback loop wherein

enhanced signaling upregulates viral gene transcription. As a corollary, we

postulated that interrupting this loop would blunt viral oncoprotein levels,

restore tumor suppressor functions, and sensitize the cells to apoptotic

stimuli from DNA damaging agents.

Materials/Methods: We employed syngeneic HPV- and HPV16+ cell

lines. To determine the effect of EGF stimulation on viral transcription,

cells were incubated with 10 ng/mL EGF. Viral E6, E7, and E1E4/E5

transcript levels were quantified by RT-qPCR. We pretreated cells with

cetuximab and measured (1) EGFR pathway inhibition by detection of

p-EGFR Y-1173 and Y-1068, and downstream p-ERK1/2; (2) early viral

transcripts by RT-qPCR; and (3) cellular surrogates of oncoprotein activity

(p53, p21, pRb and p16) by immunoblot. To determine whether cetuximab

enhances apoptotic response to DNA damage, we evaluated cell viability

after varying schedules of cetuximab and cisplatin. We studied cetuximab’s

activity in xenograft models using HPV16+ tumor cell lines.

Results: In cells harboring episomal HPV16 genomes, EGFR stimulation

significantly upregulated, while cetuximab suppressed viral oncogene

International Journal of Radiation Oncology � Biology � Physics938

levels. The latter was accompanied by restored tumor suppressor p53 and

pRb functions. Cell viability was not reduced by either cetuximab or

cisplatin alone. However, viability decreased when cetuximab and cisplatin

were combineddwith greatest effect when cetuximab was delivered prior

to cisplatin. Finally, cetuximab reduced viral oncogene transcript levels

and inhibited tumor growth in HPV+ xenografts.

Conclusion: Cetuximab displays anti-HPV activity, reducing viral onco-

gene expression, diminishing viral load, and restoring tumor suppressor

function. Moreover, cetuximab sensitized HPV+ cells to DNA damage by

cisplatin, which may be attributed, in part, to return of apoptotic function.

Cetuximab’s anti-HPV action may explain its effectiveness in the context

of definitive cetuximab radiation for locally advanced HPV-positive

OPSCC.

Author Disclosure: A.M. Griego: None. M.A. Ozbun: None. P.F.

Barraza: None. C. Hu: None. A. Dziduszko: None. B.K. Crawley:

None. H.J. Hathaway: None. J.E. Bauman: Partnership; Jaleva

Pharmaceuticals.

274The Relationship Between CD44, EGFR, and c-MET Expression inPatients With Locally Advanced p16-Positive and p16-NegativeHead and Neck Squamous Cell CarcinomaA.M. Baschnagel,1 N.Y. Tonlaar,2 G.D. Wilson,2 L. Williams,3

and M. Eskandari3; 1Dept. of Human Oncology, University of Wisconsin,

Madison, WI, 2William Beaumont Health System, Royal Oak, MI,3William Beaumont Hospital, Royal Oak, MI

Purpose/Objective(s): CD44 is a cell surface glycoprotein known to be a

cancer stem marker in head and neck squamous cell carcinoma

(HNSCC). It is involved in tumor growth, metastasis, and self-renewal. c-

MET is a marker of poor prognosis and may have stem cell capacity in

HNSCC. Combined CD44/c-MET signaling can drive repopulation in

cell culture. The relationship between c-MET and CD44 expression in

HNSCC patient tissue is unknown. We examined the association between

CD44 and c-MET in relation to p16 and the epidermal growth factor

receptor (EGFR) in a cohort of HNSCC patients treated with

chemoradiation.

Materials/Methods: Immunohistochemical staining of CD44, p16, EGFR,

and c-MET was performed on a tissue microarray consisting of tumors

from 103 locally advanced HNSCC patients treated with definitive che-

moradiation. Tissue cores were graded by 2 blinded pathologists. p16 was

graded as positive (3+ or 2+ in �50% of tumor cells) or negative. Positive

EGFR expression was defined as any 3+ or 1+/2+ in �50% of tumor cells.

High c-MET expression was defined as any 3+ and CD44 high expression

was defined as 2+/3+ in �50% of tumor cells. CD44 expression was

correlated with p16, EGFR, and c-MET, locoregional control (LRC),

distant metastases (DM), disease-free survival (DFS), and overall survival

(OS). Univariate and multivariate analyses (MVA) were performed.

Results: CD44 high expression was present in 37% of patients. Fifty-two

percent of patients were p16 positive, including 78% of oropharynx

tumors. Fifty-four percent of patients were positive for EGFR, and 36%

were classified as high c-MET expression. High CD44 expression was

associated with higher T stage (PZ.01), nonoropharynx primaries

(P<.001), p16-negative (P<.001), and EGFR-positive tumors (P<.001).

High CD44 expression was highly correlated with c-MET expression with

58% of CD44 high-expressing tumors having high c-MET expression

compared to 23% of CD44 low-expressing tumors (P<.001). Twenty-one

percent of all patients had both high CD44 and high c-MET expression,

and 17% of all patients had a combination of p16-negative, EGFR-positive,

high c-MET, and high CD44 expression. On univariate analysis, high

CD44 expression predicted for worse LRC (hazard ratio [HR]: 2.40; 95%

confidence interval [CI]: 1.14 to 5.05; PZ.021), DFS (HR: 2.50; 95% CI:

1.25 to 4.04; PZ.007) and OS (HR: 2.02, 95% CI: 1.02 to 4.00; PZ.044)

but not DM (PZ.69). Patients with both high CD44 and c-MET expression

had a dismal prognosis, with a 2-year DFS of 31% compared to 69% in the

rest of the cohort (PZ.003). On MVA, after adjusting for site, T stage,

smoking history, and EGFR status, high c-MET (PZ.040) and negative

p16 status (P<.001) predicted for worse DFS, while high CD44 expression

did not (PZ.80).

Conclusion: High CD44 expression is associated with high c-MET

expression, p16-negative tumors, and EGFR-positive tumors. The combi-

nation of these markers predict for poor prognosis in HNSCC patients

treated with chemoradiation.

Author Disclosure: A.M. Baschnagel: None. N.Y. Tonlaar: None. G.D.

Wilson: None. L. Williams: None. M. Eskandari: None.

275Genomic Instability in Larynx CancerA. Deeb,1 J. Beck,2 E. Schuetz,3 T. Nwizu,4 L. Romick-Rosendale,5

F. Lucas,1 R. Butler,6 T. Wise-Draper,1 M. Mirezwa,1 J.C. Morris,6

K. Casper,7 D.J. Adelstein,4 J. Grandis,8 E.M. Bahassi,1 H. Urnovitz,3

and N. hashemi Sadraei6; 1Univ of Cincinnati, Cincinnati, OH, 2Chronix

Biomedical, Gottingen, Germany, 3ChronixBiomedical, Gottingen,

Germany, 4Cleveland Clinic, Cleveland, OH, 5Cincinnati Children’s

Hospital, Cincinnati, OH, 6University of Cincinnati, Cincinnati, OH, 7Univ

of Michigan, Ann Arbor, MI, 8University of Pittsburgh, Pittsburgh, PA

Purpose/Objective(s): Defects in the Fanconi anemia (FA) DNA repair

pathway have been associated with head and neck cancers. Previous data

support an increased sensitivity to DNA-damaging therapy in FA-deficient

cells. Little is known about FA mutations and DNA instability in larynx

cancer patients undergoing salvage surgery.

Materials/Methods: Whole exome and targeted sequencing data from

tumor samples of patients with larynx cancer were evaluated for 18

FA-associated genes. Disease-related characteristics and outcome data

were collected retrospectively. To evaluate for genomic instability, DNA

sequences were compartmentalized, normalized, and compared to the

mean (�3 standard deviations) of controls to identify regions of chro-

mosomal number imbalance. Algorithms were used to determine genomic

instability index from selected regions of cancer genome.

Results: A total of 111 patients with laryngeal squamous cell carcinoma

were identified from institutional and TCGA cohorts with whole exome or

targeted sequencing data on FA-associated genes. Twenty-five patients

(22.5%) had at least one FA-associated gene mutation detected, and 7 had

multiple genes involved (6.3 %). The most prevalent mutations were seen

in FANCD1, FANCG, and RAD51C. Median overall survival in the wild

type versus the mutant group was 22 versus 34 months. Among a smaller

cohort of patients who underwent a salvage laryngectomy after radiation

failure (nZ12), median progression-free survival was shorter in those who

accumulated fewer chromosomal imbalances and had a low genomic

instability index; 7 versus 23 months. (Statistical comparison not signifi-

cant due to small numbers.)

Conclusion: Defects in FA-related DNA repair pathways are frequently

detected in larynx cancer patients. Additional studies need to confirm

whether FA erelated mutations and cumulative genomic instability may be

predictive of increased sensitivity to radiation and other DNA-damaging

therapies.

Author Disclosure: A. Deeb: None. J. Beck: None. E. Schuetz: employ-

ment; Chronix Biomedical. Stock Options; Chronix Biomedical. CMO;

chronix Biomedical. T. Nwizu: None. L. Romick-Rosendale: None. F.

Lucas: None. R. Butler: None. T. Wise-Draper: Research Grant; Merck.

M. Mirezwa: None. J.C. Morris: Speaker’s Bureau; Boehringer Engel-

heim. K. Casper: None. D.J. Adelstein: None. J. Grandis: None. E.

Bahassi: None. H. Urnovitz: CEO; Chronix Biomedical. Stock Options;

Chronix Biomedical. N. hashemi Sadraei: Research Grant; Merck.

Speaker’s Bureau; Genentech. Advisory Board; Genentech.

Volume 94 � Number 4 � 2016 Posters 939

276XPAeA Biomarker With Potential Prognostic Value in Patients WithOropharyngeal Head and Neck Squamous Cell CarcinomaS. Prochnow, A. Muenscher, R. Knecht, W. Wilczak, and T. Clauditz;

University Hospital Hamburg-Eppendorf, Hamburg, Germany

Purpose/Objective(s): Platinum-based chemotherapy resistance has been

under investigation for a long time, looking at the nucleotide excision

repair (NER) pathway, which is responsible for DNA adduct repair. One of

the participating proteins in that pathway is XPA. There is little infor-

mation about this protein regarding its prognostic value in patients with

head and neck squamous cell carcinoma (HNSCC). Therefore, we inves-

tigated XPA expression as a prognostic factor by retrospectively looking at

overall survival, time to recurrence, and correlation with clinical

parameters.

Materials/Methods: Tissue microarrays were constructed from 453 cases

of HNSCC including 222 oral (49%), 126 pharyngeal (27.8%), and 105

laryngeal (23.2%) tumors. Two hundred ninety-three tumor blocks were

evaluable for XPA immunohistochemistry. Expression levels were

dichotomized into a high and low XPA expressing group followed by a

comparison of age, gender, TNM status, grading, and UICC stage. Out-

comes for overall survival and time to recurrence were analyzed by using

the Kaplan-Meier method and performed for different subsites of the head

and neck.

Results: Analysis of overall survival and time to recurrence showed no

difference between both expression levels of XPA in the overall patient

cohort. However, superior overall survival in patients with oropharyngeal

SCC and a high XPA expression could be observed (PZ.0386). Looking

at SCCs of the oral cavity, a trend toward an inferior overall survival in

patients with a high XPA expression was seen, whereas investigations in

the hypopharynx and larynx showed no significant differences between

high and low XPA expressing tumors. Looking generally at gender, M

stage, and grading, no statistical correlation was found. Analyzing T and

N stage in all tumors, a trend toward a lower XPA expression in

advanced tumors (>pT4 PZ.0543 and >pN1 PZ.0546) could be seen.

This trend was confirmed by statistical lower expression of XPA in

patients with UICC stage IV looking at the overall patient cohort

(PZ.035).

Conclusion: The shown results suggest that XPA might be a novel pre-

dictive marker for overall survival in patients with oropharyngeal SCC

with a superior survival in tumors with a high XPA expression. Further-

more, this study shows that subsites in the head and neck will have to be

looked at separately in the future to determine the predictive value of

biomarkers for therapy outcome and pretherapeutic risk stratification of

patients. To increase statistical power of this study and to evaluate the

effect on platinum-based chemotherapy resistance, further studies with

even larger patient cohorts will be needed and are ongoing.

Author Disclosure: S. Prochnow: None. A. Muenscher: None.

R. Knecht: Advisory Board of Merck; Advisory Board of Merck.

W. Wilczak: None. T. Clauditz: None.

277MGMT HypermethylationdMissing Piece of the Puzzle in PrimaryHead and Neck Squamous Cell Carcinoma?S.G. Jhavar, D. Fink, A. Rao, and N. Deb; Baylor Scott & White

Healthcare Temple Clinic, Temple, TX

Abstract 277; Table 1

HPV+ (nZ48)

Tobacco+ (nZ33) Tobaccoe (nZ15)

MGMT -HM MGMT unmethylated MGMT-HM MGMT unmethylated M

9 (27%) 24 (73%) 8 (53%) 7 (47%) 1

Purpose/Objective(s): A secific mechanism explaining the higher che-

moradiation responsiveness of human papillomavirus (HPV)-positive head

and neck squamous cell carcinoma (HNSCC) and the modulation of this

effect by tobacco exposure remains unidentified. Promoter hyper-

methylation (HM) of the DNA-repair O6-Methylguanine (O6-MG)-DNA-

methyl-transferase (MGMT) gene is an independent predictor of response

to chemoradiation and survival in gliomas and lymphomas. We are

investigating the role of MGMT in HNSCC in the context of tobacco

exposure and HPV status.

Materials/Methods: After exclusion of patients with no pathology,

cytology only, and outside slides, primary HNSCC patients suitable for this

ongoing investigation were identified from the institutional cancer data-

base after we secured institutional review board approval. Tumor was

extracted from paraffin blocks using laser-capture microdissection. DNA

was extracted from the tumor followed by bisulphite treatment, methyl-

ation-specific PCR, and pyrosequencing using a commercially available

kit. In addition, HPV status was ascertained with the help of p16 immu-

nohistochemistry. Age, gender, site, stage, and tobacco use was extracted

from patient charts.

Results: As a part of this ongoing investigation, we provide the interim

report on 100 cases with bothMGMT and p16 status. MGMT-HM was seen

in 35% of patients (35 of 100). The distribution of HNSCC tumors with

respect to HPV status, tobacco use, and MGMT-HM is shown in Table 1.

Tobacco status did not influence HPV status (45% [33 of 73] of tobacco-

associated cancers vs 56% [15 of 27] of tobacco-unassociated cancers were

p16 positive; PZ.3; Fisher exact test). HPV status did not influence

MGMT status (35% [17 of 48] p16 positive cancers vs 35% [18 of 52] p16

negative cancers had MGMT-HM; PZ1.0; Fisher exact test). However,

tobacco use was found to be significantly associated with MGMT status

(29% [21 of 73] tobacco-associated cancers vs 52% [14 of 27] tobacco-

unassociated cancers had MGMT-HM; PZ.0369; Fisher exact test).

Furthermore, tobacco use influenced MGMT status particularly in the

HPV-positive subgroup (27% [9 of 33] tobacco-associated/p16-positive

cancers vs 53% [8 of 15] tobacco-unassociated/p16-positive cancers had

MGMT-HM; PZ.04; X2 test).

Conclusion: At least one-third of HNSCC patients have MGMT-HM. To

our knowledge, this is first study to report an association between tobacco

use and MGMT status, especially in HPV-positive HNSCC.

Author Disclosure: S.G. Jhavar: None. D. Fink: None. A. Rao: None.

N. Deb: None.

278Variations in Genome Structure Between Follicular Variant andHighly Aggressive Papillary Thyroid CancerD.V. Bann,1 K.E. Sheldon,2 K. Houser,2 L. Zhang,2 J. Broach,2

and D. Goldenberg1; 1Penn State Milton S. Hershey Medical Center,

Hershey, PA, 2Penn State College of Medicine, Hershey, PA

Purpose/Objective(s): Structural genome instability is a cardinal feature

of cancer. Genomic structural variations (SVs), including insertions, de-

letions, inversions, copy number alterations, and translocations, acquired

during malignant transformation may alter gene expression, resulting in

tumor growth, invasion, and metastasis. However, the short reads of 50-200

base pairs (bp) used by many next-generation sequencing technologies

lead to systematic errors in identifying and localizing SVs within the

genome. To understand how SVs contribute to invasion or metastasis we

HPV- (nZ52)

Tobacco+ (nZ40) Tobaccoe (nZ12)

GMT-HM MGMT unmethylated MGMT-HM MGMT unmethylated

2 (30%) 28 (70%) 6 (50%) 6 (50%)

International Journal of Radiation Oncology � Biology � Physics940

are using a novel high-throughput genome mapping technology to map

tumor genomes isolated from thyroid cancer patients. This system images

single DNA molecules up to >1,000,000 bp in size, which permits accu-

rate identification and localization of SVs in the genome. Using this

technology, we have constructed the first genomic map of follicular variant

papillary thyroid cancer (PTC) isolated from patient tissue. Our current

efforts focus on mapping the genome of a highly invasive, metastatic PTC.

We hypothesize that the invasive PTC genome will contain unique SVs

compared to the follicular variant PTC genome.

Materials/Methods: High molecular weight DNA was extracted from the

blood (germline), primary thyroid tumor, and lymph node metastases of a

patient with highly aggressive PTC. Similarly, DNA was extracted from

germline, thyroid tumor, and adjacent normal thyroid tissue from a patient

with follicular variant PTC. The DNAwas fluorescently labeled at specific

7-bp sequences, background stained, and loaded onto the genome mapping

instrument. Single DNA molecules ranging in length from 150,000 bp to

>1,000,000 bp were imaged and used to assemble genome maps via the

genome mapping analysis pipeline. Maps were assembled at >80x

coverage for germline and normal tissue and >100x coverage for tumor

tissue. SVs were visualized using software. Unique SVs in tumor samples

were identified.

Results: In the follicular variant PTC, we identified 185 putative insertionsand deletions unique to the tumor genome. The average size across the 69

putative deletions identified in the follicular variant PTC was 7120 bp with

a range of 982 to 70,602 bp. Among 116 putative insertions in the

follicular variant PTC the average size was 14,820 bp with a range of 1002

to 299,045 bp. Analysis of the invasive PTC genome map is forthcoming.

Conclusion: These data indicate that SVs may be a significant source of

variation in tumor genomes. Importantly, this study represents the first

effort to compare maps of patient-derived solid tumor genomes and will

increase to our understanding of how changes to the genome structure

mediate tumor growth, invasion, and metastasis.

Author Disclosure: D.V. Bann: None. K.E. Sheldon: None. K. Houser:

None. L. Zhang: None. J. Broach: None. D. Goldenberg: None.

279The DEK Oncogene Can Be Detected in the Plasma of Head and NeckCancer Patients and May Be Correlated With Tumor ImmuneResponse and PrognosisT. Wise-Draper,1 N. hashemi Sadraei,1 A. Sendilnathan,1 N. Pease,2

J. Qualtieri,1 R. Butler,1 K. Casper,3 M.L. Mierzwa,4 J.C. Morris,1

Y. Patil,5 K. Wilson,1 J. Mark,1 and L. Privette Vinnedge2;1University of Cincinnati, Cincinnati, OH, 2CCHMC, Cincinnati, OH,3Univ of Michigan, Ann Arbor, MI, 4University of Michigan, Ann Arbor,

MI, 5University of Cincinnati Medical Center/University of Cincinnati

College of Medicine, Cincinnati, OH

Purpose/Objective(s): Head and neck cancer remains the sixth most

common cancer worldwide. Although infection with human papilloma-

virus (HPV) has emerged as a favorable prognostic factor, no plasma

biomarkers currently exist to predict tumor response and/or relapse. One

candidate plasma biomarker is encoded by the human DEK gene. DEK is

an apoptosis and differentiation inhibitor and overexpression results in

oncogenesis. DEK knockdown results in decreased growth and apoptosis

of cancer cells. DEK mRNA and protein are highly up-regulated in tissue

specimens from several tumor types, including head and neck squamous

cell carcinoma (HNSCC), breast cancer, and melanoma, and antibodies to

DEK are detected in patients with autoimmune disease. High levels of

tumor DEK mRNA are correlated with advancing stage and poor survival.

However, our previous work has demonstrated that DEK protein is present

in HNSCC tissue specimens regardless of stage or HPV infection. Addi-

tionally, in vitro data have suggested that tumor-associated macrophages

secrete DEK protein. Therefore, we sought to determine whether plasma

DEK protein may have a different value than tissue levels leading to the

hypothesis that DEK may be present in the plasma of cancer patients and

may be correlated with patient outcome.

Materials/Methods: Peripheral blood was collected from patients with

newly diagnosed or untreated HNSCC and age-matched normal healthy

controls. Plasma was separated from the samples and subjected to DEK-

specific ELISA (Cusabio, Wuhan, China). Plasma DEK levels were

compared to normal controls, tumor stage, age, and smoking status. Plasma

DEK levels were also compared to inflammatory markers in the plasma

and tissue.

Results: DEK was indeed found to be present in the plasma of both

healthy control subjects and those with head and neck cancer. DEK was

decreased in head and neck cancer patients compared to healthy patients

and inversely correlated with IL-6 in the plasma. Immune infiltration

(defined by presence of CD8+ T cells) of the tumor also appeared to be

correlated with high DEK plasma levels.

Conclusion: Interestingly, although DEK expression is increased in head

and neck cancer tissue, plasma DEK levels are decreased in patients with

cancer compared to controls and are further decreased with advancing

stage. DEK plasma levels are inversely correlated with IL-6 levels, sug-

gesting that high plasma DEK levels may be correlated with a better

prognosis. Furthermore, high DEK levels in the plasma may predict su-

perior immune infiltration of tumors. Further analyses are ongoing to

determine whether DEK levels predict response to various treatment

modalities, correlate with the body’s immune response, and whether DEK

presence in the plasma will predict residual disease and/or early relapse.

These data will be important to verify DEK plasma measurements as a

clinically useful test and may give insight to future personalized and tar-

geted treatment strategies for HNSCC.

Author Disclosure: T. Wise-Draper: Research Grant; MERCK.

N. hashemi Sadraei: Research Grant; MERCK. Advisory Board; Gen-

nentech. A. Sendilnathan: None. N. Pease: None. J. Qualtieri: None.

R. Butler: None. K. Casper: None. M.L. Mierzwa: None. J.C. Morris:

None. Y. Patil: None. K. Wilson: None. J. Mark: None. L. Privette

Vinnedge: None.

280M3814, a DNA-dependent Protein Kinase Inhibitor (DNA-PKi),Potentiates the Effect of Ionizing Radiation (IR) inXenotransplanted Tumors in Nude MiceL. Damstrup,1 A. Zimmerman,1 C. Sirrenberg,1 F. Zenke,1

and L. Vassilev2; 1Merck Serono, Darmstadt, Germany,2EMD Serono, Billerica, MA

Purpose/Objective(s): Agents that generate breaks in DNA are among the

most widely used classes of cancer therapeutics. These agents induce

different forms of DNA damage, including double-strand breaks (DSBs),

which are the most lethal if left unrepaired. M3814 targets tumor cell

growth and survival by inhibiting a critical DSB DNA damage repair

mechanism. The antitumor effect of M3814 is dependent on the func-

tionality of DNA repair and checkpoint signaling in cancer cells, which

have a lowered ability to cope with DSBs, leading to cell death. Hence, the

rationale of DNA-PK inhibition is to increase the amount of DSB DNA

damage generated by IR. The objective of the series of nonclinical ex-

periments was to demonstrate that IR together with M3814 was more

efficacious than IR alone.

Materials/Methods: The efficacy of M3814 in combination with IR was

evaluated in 6 human xenograft models (HCT116, FaDu, NCI-H460,

A549, Capan-1, BxPC3) in mice representing 4 different cancer types

(colon, head and neck, lung, and pancreas). Tumor cells were injected s.c.

into nude mice, and treatment started when palpable tumors were estab-

lished (w100-200 mm3). M3814 was given orally at different doses

(25-300 mg/kg) 10 min prior to IR. IR was applied using a radiation

therapy device for small rodents calibrated to deliver 2 Gy. Autophos-

phorylation of DNA-PK (serine2056 ) in FaDu tumor lysates was measured

by immunoassay to assess pharmacological inhibition by M3814.

Volume 94 � Number 4 � 2016 Posters 941

Results: In combination with IR, M3814 showed efficacy in all of the 6

mouse models of human cancer as demonstrated by a strong potentiation

of the effect of IR. In all models, a dose of 2 Gy administered daily for 1

week in combination with M3814 induced statically significant tumor

growth inhibition compared to IR alone. In 2 models, FaDu and NCI-

H460, M3814 induced tumor regression. In the FaDu model, no tumor

regrowth (duration of the experiment >100 days) was observed in the

combination arm with IR (2 Gy per fractions, 6 weeks, 5 days per week:

total dose 60 Gy) at the doses of 25 and 50 mg/Kg. In the IR only arm, no

tumor responses were observed. These effects were a likely consequence

of inhibiting DNA-PK activity, as shown by measuring the autophos-

phorylation of DNA-PK in FaDu tumor tissue. M3814, alone or in com-

bination with IR, did not induce significant weight loss or visual signs of

toxicity in the mice in any study.

Conclusion: M3814 is active in nonclinical experiments in combination

with IR. Strong antitumor activity was observed in several xenograft

models with complete regressions of tumors upon application of the

established clinical IR schedule of 2-Gy fractions for 6 weeks in the FaDu

model (squamous cell carcinomas of the head and neck). Clinical evalu-

ation of M3814 is ongoing.

Author Disclosure: L. Damstrup: None. A. Zimmerman: None. C. Sir-

renberg: None. F. Zenke: None. L. Vassilev: None.

281A Phase 2 Study Evaluating Axitinib in Patients With Unresectable,Recurrent, or Metastatic Head and Neck CancerP. Swiecicki,1 E. Dickerson,5 A. Srinivasan,5 L. Zhao,1 E. Bellile,1

A.G. Sacco,2 D.B. Chepeha,1 I. Dobrosotskaya,3 M.E. Spector,1

A. Shuman,1 K. Malloy,1 J. Moyer,1 E. McKean,1 G.T. Wolf,1

A. Eisbruch,4 M. Prince,1 C. Bradford,1 T. Carey,1 and F.P. Worden1;1University of Michigan Comprehensive Cancer Center, Ann Arbor, MI,2University of California San Diego Moores Cancer Center, La Jolla, CA,3Henry Ford Hospital, Department of Oncology, Detroit, MI, 4University of

Michigan, Ann Arbor, MI, 5University of Michigan Cancer Center, Ann

Arbor, MI

Purpose/Objective(s): Documented response rates in unresectable,

recurrent, and/or metastatic head and neck squamous cell cancer (HNSCC)

range between 10% and 30% with single-agent regimens. Hence, new

rational therapies are needed. Axitinib is a small molecule tyrosine kinase

inhibitor with selective inhibition of VEGFR 1, 2, 3, as well as inhibition

of potential downstream effectors of the EGFR pathway. Given the upre-

gulation of EGFR and VEGFR in HNSCC, treatment with axitinib holds

promise as a rational targeted therapy. We conducted a phase 2 trial to

investigate the clinical activity of this agent in patients with unresectable

recurrent and/or distant metastatic HNSCC. Our hypothesis was that

treatment with axitinib would result in improvement in 6-month progres-

sion-free survival (PFS) compared to historical controls.

Materials/Methods: Patients with unresectable, recurrent, or metastatic

HNSCC were included in this open-label, single-arm, phase 2 trial. Pri-

mary endpoint was 6-month PFS. All patients received single-agent axi-

tinib with planned dose escalation based on tolerability. Treatment-related

adverse events were graded according to the Common Terminology for

Adverse Events version 3.0. Treatment response was evaluated by

Response Evaluation Criteria in Solid Tumors version 1.0. A planned

interim efficacy analysis was performed after enrollment of 30 patients. A

simulation study performed to evaluate efficacy as determined by PFS did

not demonstrate superiority of axitinib, and the trial was prematurely

closed.

Results: Forty-two patients were registered, of whom 30 were evaluable

for response. While treatment was well tolerated, with no severe bleeding

events, only 19 patients were able to achieve full planned dose. The dis-

ease control rate was 76.7%, with a median PFS of 3.7 months (95%

confidence interval [CI]: 3.5-5.7) and overall survival of 10.9 months

(95% CI: 6.4-17.8). Exploratory analysis demonstrated that patients with a

smaller sum of diameter of target lesions experienced improved response

rates and better PFS and overall survival. Exploratory correlative studies

were performed on patient serum, and it was noted that the change in IL-8

after the first dose of axitinib was associated with response to therapy

(PZ.04).

Conclusion: Although limited by premature closure based on a perceived

lack of response, single-agent axitinib should be considered for further

evaluation based on its acceptable toxicity profile and favorable median

overall survival compared to standard therapies. We observed cystic

degeneration and mild tumor volume increase after initiation, suggesting

likely treatment effect and clinical benefit. Future studies are needed to

define whether other radiologic definitions of response are more appro-

priate in defining response to TKI or anti-VEGF therapy in head and neck

cancer.

Author Disclosure: P.L. Swiecicki: None. L. Zhao: None. E. Bellile:

None. A.G. Sacco: None. D.B. Chepeha: None. I. Dobrosotskaya: None.

M.E. Spector: None. A. Shuman: None. K. Malloy: None. J. Moyer:

None. E. McKean: None. G.T. Wolf: None. A. Eisbruch: None. M.

Prince: None. C. Bradford: None. T. Carey: None. F.P. Worden: None.

282Morphoproteomics Identifies the EZH2 Pathways as a PotentialTherapeutic Target of Human PapillomaviruseAssociatedOropharyngeal CancerS.H.S. Naqvi, R. Brown, J. Buryanek, and R.J. Karni; The University of

Texas Health Science Center- Houston, Houston, TX

Purpose/Objective(s): Human papillomavirus (HPV) has been identified

as an etiopathogenetic factor in oropharyngeal cancer. The HPV E6 and E7

oncogenes are instrumental in leading to proliferation and a block in dif-

ferentiation and tumorigenesis. Although surgical intervention can remove

such tumors, the potential for an etiologic field effect with recurrent dis-

ease is real. The objective of this preliminary study was to identify a

known downstream effector of E7 oncoprotein, Enhancer of Zeste

Homolog 2 (EZH2), which is known to pose a block in differentiation and

proliferation and lead to HPV-induced malignant transformation. Such a

pathway is amenable to low toxicity therapies designed to promote dif-

ferentiation to a more benign state and prevent recurrent disease and

incorporation of HPV into the genome.

Materials/Methods: Morphoproteomic analysis of tumor from a patient

with p16INK4a-positive oropharyngeal cancer included the immunohis-

tochemical probes for the detection of EZH2. The expression level of this

protein analyte was assessed and integrated with the proliferation index

(Ki-67) and the mitotic activity. An additional 9 patients with p16INK4a-

positive oropharyngeal cancer were also assessed in a similar fashion.

Results: Representative sections of the patients’ tumors showed >90%

immunopositivity for EZH2 in tumoral nuclei (signal intensity at 2 to 3+

on a scale of 0 to 3+) and coincided with a proliferation index of >80%

with mitotic progression.

Conclusion: Enhancer of zeste homolog 2 (EZH2), a histone methyl

transferase, is tumorigenic by virtue of the fact that it inactivates tumor

suppressor genes and contributes to a state of differentiation and prolif-

eration in tumors and in facilitating their migratory potential. EZH2

expression is activated by HPV16 oncoprotein E7 at the transcriptional

level. In oral SCC, knocking down EZH2 decreased the proliferative ca-

pacity of the carcinoma, and inhibiting EZH2 also inhibited the migration

and metastasis of oral SCC cells. EZH2 can be downregulated by the

upregulation of miRNAs to include miR-26a and miR-101. Metformin and

curcumin have that ability, and Vacurin (a curcumin-based vaginal cream)

has been shown in a preclinical model to prevent progression of vaginal

HPV-associated cancer. Morphoproteomics has identified EZH2

International Journal of Radiation Oncology � Biology � Physics942

expression in p16INK4a-positive oropharyngeal cancer. Institutional re-

view board approval has been obtained to do a more expanded study with

HPV-associated oropharyngeal cancer. Moreover, this speaks to the effi-

cacy of metformin and curcumin as adjuvant therapies in such cases and to

their application as maintenance therapy in the prevention of etiologic field

effect in recurrent disease.

Author Disclosure: S.H. Naqvi: None. R. Brown: None. J. Buryanek:

None. R.J. Karni: None.

283DNA Methylation Regulates ANO1 Expression Through AlternateMechanisms at 3 Distinct CpG Islands in Head and Neck SquamousCell CarcinomaA. Finegersh,1 S. Kulich,2 and U. Duvvuri3; 1University of Pittsburgh

School of Medicine, Pittsburgh, PA, 2VA Pittsburgh Healthcare System,

Pittsburgh, PA, 3University of Pittsburgh Medical Center, Pittsburgh, PA

Purpose/Objective(s): Mechanisms underlying metastasis of head and

neck squamous cell carcinoma (HNSCC) are still poorly understood and

contribute to tumor recurrence. ANO1 is a calcium-activated chloride

channel whose expression was recently found to act as a switch between

tumor growth and metastasis in HNSCC. We hypothesized that changes in

DNA methylation of key CpG residues near the ANO1 promoter underlie

changes in gene expression in HNSCC.

Materials/Methods: We analyzed HNSCC samples from TCGA with

available DNA methylation, expression, and human papillomavirus (HPV)

status (nZ275). Using these samples, we studied 3 CpG islands annotated

by the UCSC Genome Browser near the ANO1 transcriptional start site

(TSS). We calculated Spearman rank correlations between DNA methyl-

ation and expression at each CpG and defined statistical significance as

P<.001. Effects of copy number variation, survival, and HPV status were

also studied.

Results: Of the 3 CpG islands studied, an enhancer w90 kb upstream of

the TSS had 6 of 8 CpGs positively correlated with ANO1 expression, a

canonical CpG at the TSS had 8 of 10 negatively correlated CpGs, and a

CpG island at the exon 2 promoter had 4 of 6 positively correlated CpGs.

HPV+ (nZ34) samples had significantly decreased expression of ANO1

compared to HPV-negative samples (nZ241; P<.01); additionally, HPV-

positive samples had significantly decreased DNA methylation at the 2

positively correlated CpG islands (P<.001) but no change at the negatively

correlated CpG island. Copy number amplification enhanced epigenetic

effects, with amplified samples having increased DNA methylation of

positively correlated CpGs and decreased DNA methylation of negatively

correlated CpGs relative to those with neutral copy number. Finally,

increased expression of ANO1 as well as increased average DNA

methylation across all CpGs studied were associated with decreased pa-

tient survival (P<.05), suggesting a greater component of positively

correlated CpGs to expression.

Conclusion: DNA methylation regulates ANO1 expression through 2

mechanismsehypomethylation of CpGs at the TSS increases expression

while hypomethylation of the enhancer and intragenic CpG island de-

creases expression. Notably, hypomethylation of positively correlated

CpGs was associated with decreased expression in HPV-positive samples,

suggesting these regions are more critical for regulation. Complex effects

of copy number variation potentiating epigenetic mechanisms were also

observed. Since differential expression of ANO1 regulates growth and

metastasis of HNSCC and is associated with patient survival, modulating

its expression may be a useful clinical target. Studies are currently ongoing

to investigate specific mechanisms of DNA methylation on the ANO1

promoter and better characterize the understudied role of positively

correlated CpGs on gene expression.

Author Disclosure: A. Finegersh: None. S. Kulich: None. U. Duvvuri:

None.

284A Novel Third Generation Thiosemicarbazone, COTI-2, Is HighlyEffective in Killing Head and Neck Squamous Cell Carcinomas(HNSCC) Bearing a Variety of TP53 MutationsN.L. Silver,1 A.A. Osman,2 A.A. Patel,2 N. Tanaka,2 L. Tang,2 G. Zhou,3

and J.N. Myers2; 1UT MD Anderson Cancer Center, Houston, TX,2The University of Texas MD Anderson Cancer Center, Houston, TX,3UT MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): The cure rate for patients with advanced head and

neck squamous cell carcinoma (HNSCC) remains in the 25% to 40% range

due to resistance to standard therapy, primarily consisting of platinum-

based chemoradiation therapy. Currently, there are no established molec-

ular biomarkers that predict response to chemotherapeutic agents in

HNSCC, although studies have shown that TP53 mutational status can

predict clinical response in patients treated with neoadjuvant chemo-

therapy. Since mutation of TP53 in HNSCC is associated with resistance to

these treatments, we are interested in novel therapeutic approaches to

overcome this resistance. COTI-2 (Critical Outcome Technologies Inc), a

novel third generation thiosemicarbazone, appears to restore wild-type like

functional activity to a wide range of p53 mutant tumor cells through a

mechanism that involves zinc chelation. In vitro studies have shown that

COTI-2 has nanomolar activity in multiple human cancer cell lines and has

shown growth inhibition in xenograft mouse models for several solid

tumors. We hypothesized that this drug would be effective in killing

HNSCC cells harboring various TP53 mutations.

Materials/Methods: Clonogenic survival assays were performed to

examine the in vitro sensitivity of several TP53 mutant HNSCC cell lines

following treatment with COTI-2. Cell cycle and western blotting ana-

lyses were performed to investigate cellular mechanisms. An orthotopic

mouse model of oral cancer was used to evaluate the in vivo efficacy of

the drug.

Results: COTI-2 reduced cell survival of both mutant p53 and wild-type

HNSCC cells. The reduction in cell survival is through both apoptosis and

senescence, which appear to occur via p53-dependent and p53-indepen-

dent mechanisms. Treatment of HNSCC cells with COTI-2 alone increased

p21 induction and promoter activity. COTI-2, as a single agent, resulted in

significant tumor growth inhibition in an orthotopic mouse model of oral

cancer.

Conclusion: COTI-2 inhibits in vitro and in vivo tumor growth in HNSCC

cells irrespective of TP53 status through the induction of apoptosis and

senescence.

Author Disclosure: N.L. Silver: None. A.A. Osman: None. A.A. Patel:

None. N. Tanaka: None. L. Tang: None. G. Zhou: None. J.N. Myers:

None.

285Characterizing the Phenotype and Biomolecular Activity of 2Clinically Relevant Mutations in TMEM16AN. Godse,1 N.I. Khan,2 C. Kemp,3 and U. Duvvuri3; 1University of

Pittsburgh, Pittsburgh, PA, 2Baylor College of Medicine, Houston, TX,3University of Pittsburgh Medical Center, Pittsburgh, PA

Purpose/Objective(s): TMEM16A is a calcium-activated chloride chan-

nel that is overexpressed or mutated in a variety of squamous cell carci-

nomas, including 30% of head and neck squamous cell carcinomas

(HNSCC). Two relevant missense mutations, R429P and R561L, have been

identified and appear to demonstrate enhanced oncogenic capacity. In this

study, we sought to (1) characterize in vitro behavior of these mutants and

Volume 94 � Number 4 � 2016 Posters 943

(2) identify biomolecular pathways that permit enhanced oncogenic

capacity in these mutants.

Materials/Methods: We transduced OSC19 cells and created daughter

cells that express low or high levels of wild-type TMEM16A as controls,

the R429P mutation, or the R561L mutation using a lentiviral vector. We

established the in vitro phenotype of these mutant cells using colony

formation assays and Cell Titer-Glo viability assays under a variety of

conditions. Biomolecular pathway activity was analyzed using western

blot. All data are reported as fold increase relative to the empty vector

control cells.

Results: The R429P and R561L mutants demonstrated significantly

increased growth in colony formation assay: 7.23- and 2.42-fold increase

(P<.05), respectively. Additionally, the mutants demonstrated enhanced

viability in anchorage-dependent (1.41- and 1.12-fold increase, respec-

tively; P<.05) and anchorage-independent (2.35- and 2.24-fold increase,

respectively; P<.05) Cell Titer-Glo viability assays at 48 hours.

TMEM16A mutants demonstrated increased activity of the pEGFR (1.6-

and 2.6-fold increase, respectively), pERK (6.9- and 9.1-fold increase,

respectively), and cyclin D1 (2.8- and 2.8-fold increase, respectively)

pathway on baseline western blot analysis.

Conclusion: Our results support the phenotype that cells expressing the

R429P and R561L TMEM16A mutations have enhanced growth and

proliferative capacity possibly mediated by increased activity of the

pEGFR/pERK/cyclin D1 pathways. Future work will aim to clarify

structure-function changes in mutant TMEM16A, mechanisms of pEGFR/

pERK/cyclin D1 overactivation, and in vivo growth patterns. Ultimately, a

thorough biomolecular description of these mutations could better inform

treatment and management options in patients expressing these mutations.

Author Disclosure: N. Godse: None. N.I. Khan: None. C. Kemp: None.

U. Duvvuri: None.

286Migration of Head and Neck Squamous Cell Cancer Cells isDependent on Tenascin-C ExpressionC.M. Thomas,1 M. Glogowska,2 S. Keysar,2 and A. Jimeno2;1Department of Otolaryngology, University of Colorado Denver, Aurora,

CO, 2Department of Medical Oncology, University of Colorado Denver,

Aurora, CO

Purpose/Objective(s): Tenascin-C (TNC) is a large extracellular matrix

glycoprotein that is expressed during embryogenesis, wound healing,

chronic inflammation, and in various cancers. Prior studies in oral head and

neck squamous cell carcinoma (SCCa) have found the expression of TNC

to be increased in metastatic oral SCCa. TNC is produced by both

epithelial tumor cells as well as cancer-associated fibroblasts (CAFs) and

deposited into the extracellular matrix (ECM) where it contributes to the

remodeling of the tumor microenviroment (TME) and the epithelial-to-

mesenchymal transition of tumor cells. The exact mechanism of action of

TNC is unknown; however, TNC does increase expression of matrix

metalloproteinases, which remodel the TME/ECM to facilitate cell

migration and metastasis. The goal of this study was to determine if TNC

was produced by head and neck SCCa cell lines isolated directly from

patient tumor samples, and if expression of TNC was necessary for cancer

cell migration.

Materials/Methods: Expression of TNC was confirmed in human head

and neck SCCa tissue samples and immortalized cancer cell lines via

RT-PCR, TNC-specific immunohistochemistry (IHC), and western blot

analysis. Location of primary tumor samples included tonsil, base of

tongue, hypopharynx, oral tongue, and floor of mouth. Treatment with

TNC-specific siRNA was utilized to reduce the expression of TNC in

cancer cell lines (CUHN013C and CUHN049C), which was confirmed by

RT-PCR. Matrigel invasion assays were performed comparing wild-type

CUHN013C and TNC siRNA-treated CUHN013C cells. For the invasion

assays, cells were plated in serum-free media on matrigel inserts placed in

serum-containing media and allowed to incubate for 24 hours. Inserts were

then stained, and cell invasion was quantified.

Results: TNC is present in the tumor microenvironment, especially at the

tumor-stroma interface, in head and neck SCCa primary tumors as well as

lymph node metastases. Immortalized cancer cell lines derived from pa-

tient tumor samples also express TNC, which can be significantly (P<.05)

decreased by TNC-specific siRNA treatment. Decreased expression of

TNC results in significantly (P<.05) reduced migration of CUHN013C

cancer cells in matrigel invasion assays.

Conclusion: TNC is expressed by head and neck SCCa cells from a variety

of primary sites. TNC expression is necessary for migration of head and

neck SCCa cells in an in vitro invasion assay. Therapies specifically tar-

geting TNC may reduce the invasiveness and metastatic potential of head

and neck SCCa.

Author Disclosure: C.M. Thomas: None. M. Glogowska: None.

S. Keysar: None. A. Jimeno: None.

287Promising Biomarkers of Thyroid Cancer: BRAF V600E Mutation,miRNAs, Integrins, and Their Extracellular Matrix LigandsS. Shevchenko; Novosibirsk State University, Novosibirsk, Russia

Purpose/Objective(s): Some of the most important issues in thyroid

cancer treatment are preoperative diagnosis of malignancy and prognosis

determination. For such tasks, molecular biomarkers can be successfully

used. The most frequent genetic alteration in thyroid cancer is V600E

mutation in BRAF gene, which occurs in 45% to 88% of papillary thy-

roid carcinoma and associates with aggressive tumor fenotype. Its

presence leads to changes of cell membrane integrin receptors expression

and their ligandsdextracellular matrix proteinsdosteopontin (OPN) and

thrombospondin-1 (TSP1). Such changes promote migration, invasion,

and metastasis of tumor cells. Other important mechanisms for regula-

tion of gene expression in tumor tissue are mediated by microRNAs.

Differences in MiRNAs expression has been demonstrated in variety of

cancers, presenting the potential for tumor typing. Thus, the objective of

the study was to compare the gene expression profile of miRNAs 21,

221, 222, and 155; integrins ITGA2, ITGA3, ITGAV, ITGA6, ITGA9,

ITGB1, and ITGB3; and their ligands OPNa, OPNb, and TSP1 in

different types of thyroid tumors to find some promising biomarkers for

clinical use.

Materials/Methods: Intraoperative thyroid tissue samples from patients

diagnosed with different types of thyroid carcinoma (nZ112) and diffuse

nodular nontoxic goiter (nZ120) were analyzed. To evaluate the

expression levels of the investigated genes and microRNAs, real-time

RT-PCR was used. Immunohistochemistry was conducted to confirm the

PCR results and to estimate the amount of protein products. The pres-

ence of BRAF V600E mutation was identified using allele-specific

amplification.

Results: In this study a significant increase (P>.05) in expression levels of

ITGA3, ITGAV, ITGB1, OPNb, and TSP1 (2.9 fold, 1.9 fold, 17.1 fold, 2.5

fold, and 3.2 fold, respectively) was observed in the PTC tissue samples in

comparison with visually unchanged thyroid tissue. For BRAF V600E-

positive tumors with potentially aggressive properties, high expression

levels of ITGA3 (80.7 fold, PZ.0473) and ITGAV (13.6 fold, PZ.0252)

were registered. Expression of miRNA 21, 221, 222, and 155 was

significantly increased in cancer samples, in comparison with benign

neoplasms (which ranged from 3.1 fold to 7,2 fold, P<.05).

Conclusion: Thus, the observed changes in the expression levels of the

studied genes indicate their potential role as biomarkers in thyroid tumor

typing and determination of thyroid cancer prognosis and treatment.

Author Disclosure: S. Shevchenko: None.

International Journal of Radiation Oncology � Biology � Physics944

288Human PapillomavirusePositive Oropharyngeal Squamous CellCarcinoma Expresses High Hydrogen Sulfide Synthesizing EnzymesR. Shackelford,1 R. Shanti,2 and G.E. Ghali3; 1LSU Health Shreveport,

Department of Pathology, Shreveport, LA, 2Oral & Maxillofacial Surgery,

Craniofacial & Cleft Surgery, Head & Neck Surgery, LSU School of

Medicine, Shreveport, LA, 3Gamble Professor & Chairman, Oral &

Maxillofacial Surgery, Craniofacial & Cleft Surgery, Head & Neck

Surgery LSU School of Medicine, Shreveport, LA

Purpose/Objective(s): Human papillomavirus (HPV)-positive oropha-

ryngeal squamous cell carcinoma (SCC) is associated with a positive

therapeutic response compared to HPV-negative SCC that tumors often

result from alcohol and tobacco exposures. The reason for the better

prognosis is currently unknown, although a wild-type/less mutated p53

gene status in HPV-positive tumors may play a role. The HPV E6

protein causes the degradation of miRNA34a, an event known to in-

crease nicotinamide phosphoribosyltransferase (Nampt) levels. Nampt in

turn is a known positive regulator of the enzymes that synthesize

hydrogen sulfide (H2S). Here we examined the levels of Nampt, the H2S

synthesizing enzymes, and the total H2S, acid labile, and bound sulfide

pools in HPV-positive and negative cells lines derived from oropha-

ryngeal SCC.

Materials/Methods: We used western blotting to examine the levels

of the 3 known enzymes involved in H2S synthesis; cystathionine

b-synthase (CBS), cystathionine g-lyase (CSE), and 3-mercaptopyruvate

S-transferase (3-MST), and Nampt in 1 HPV-positive and 2 HPV-nega-

tive cell lines derived from individuals with oropharyngeal SCC (scc47,

cal27, and cal33, respectively). High performance liquid chromatography

was used to measure total H2S and the acid labile and bound sulfide

pools. The colony-forming efficiency assay was used to measure cell

colony-forming ability following cell treatments with oxidative stress

and H2S inhibitors.

Results: Nampt and CBS, CSE, and 3-MST were all elevated in the HPV-

positive SCC cell line at least 4 fold (3-MST 10 fold) compared to the

negative cell lines. Additionally, the HPV-positive SCC cell line was

preferentially sensitive to the colony-forming inhibitory effects of t-butyl

hydroperoxide compared to the HPV-negative SCC cell linesdan event

enhanced by H2S synthesis inhibitor pre-exposure. Last, the total H2S, acid

labile, and bound sulfide pools were not increased in the HPV-positive

SCC cell line compared to the negative cell line, and all 3 cell lines

had very high H2S compared to other studies of intracellular H2S

concentrations.

Conclusion: Our findings suggest that HPV increases the dependence of

oropharyngeal SCC on H2S for tumor growth and oxidant exposure, which

would rapidly deplete H2S levels, exerting preferentially toxic effects on

HPV-positive SCC cells. The dysregulation of H2S in HPV-positive

oropharyngeal SCC may in part explain the vulnerability of these cancers

to oxidants, and radiation/chemotherapies, and in part explain the better

therapeutic response seen in HPV-positive oropharyngeal SCCs.

Author Disclosure: R. Shackelford: None. R. Shanti: None. G.E. Ghali:

None.

289Histologic Variation in High-Grade Oral Epithelial Dysplasia WhenAssociated With High-risk Human PapillomavirusB. Shumway,1 S. Khanal,1 P. Trainor,1 M. Zahin,1 S.J. Ghim,1 J. Joh,1

S.N. Rai,2 and A.B. Jenson1; 1University of Louisville, Louisville, KY,2University of Louisville School of Medicine, Louisville, KY

Purpose/Objective(s): Oral squamous cell carcinoma (OSCC), a subset

of head and neck cancer, often develops from the precursor high-grade

oral epithelial dysplasia (hgOED). Histologic features of near full-thick-

ness mitosoid and apoptotic cells within hgOED were recently attributed

to high-risk human papillomavirus (HR-HPV) infection, though

genotyping was not performed. While HPV DNA integration into the host

genome is a critical event in malignant progression, the intracellular

localization of HPV DNA in hgOED is largely unexplored. We hypothe-

sized that there are significant differences in HPV presence, p16 expres-

sion, and HPV integration status among hgOED cases showing HPV-like

histologic features than in hgOED cases showing focal or none of these

microscopic features.

Materials/Methods: Cases of hgOED showing generalized histologic

features of HPV infection (group 1, nZ16), those with focal features

(group 2, nZ14), and samples lacking these changes (controls, nZ10)

were compared by PCR-based HPV genotyping, p16 immunohistochem-

istry, and qRT (quantitative real time)-PCR-based HPV-DNA integration

status. Data analyses were conducted by SAS 9.4.

Results: Significant differences were noted in HPV presence; HR-HPV

was found significantly more in group 1 patients (81.3%) than in group 2

patients (57.1%) or controls (20%). HPV16 was present in 75%, 50%, and

20% of groups 1, 2, and controls, respectively. HPV was strongly asso-

ciated with group 1 (PZ.004) and marginally associated with group 2

(PZ.06) than controls. Low-risk HPV (HPV6) was detected only in group

2 (7.1%). Significantly stronger p16 staining was seen in group 1 (75.0%,

PZ.002) but not in group 2 (14.3%, PZ1.0) or controls (11.1%). Group 1

patients were relatively younger (51.1�10.3) than were group 2 patients

(59.0�12.3) and controls (59.3�8.5). Group 1 had solely or predominantly

integrated HPV (90.9%) compared to group 2 (28.6%) and controls

(0.0%), PZ.006.

Conclusion: Diffuse mitosoid and apoptotic figures in hgOED are

confirmed histologic indicators of the presence of HR-HPV (HPV16 pre-

dominant) and are associated with p16 positivity. When focally identified,

these features can be associated with HPV but not p16 and when missing in

hgOED, HR-HPV and p16 reactivity are typically absent. Presence of in-

tegrated HPV DNA in group 1 hgOED may confer higher transformation

than group 2 or controls, though more research is necessary. As the pro-

gression of HPV infection in premalignant lesions to HPV-associated

OSCC is unknown, clinical and histologic characterization of HPV-asso-

ciated hgOED and adjunctive analyses (eg, p16 IHC, integration status) are

needed to advance our understanding and to improve prevention and

treatment strategies in vulnerable populations.

Author Disclosure: B. Shumway: None. S. Khanal: None. P. Trainor:

None. M. Zahin: None. S. Ghim: None. J. Joh: None. S.N. Rai: None.

A.B. Jenson: None.

290S100A16 Functions as a Tumor Suppressor Protein in OralSquamous Cell CarcinomaD. Sapkota,1 O. Bruland,2 H. Parajuli,1 T. Osman,1 M.T. Teh,3

A. Johannessen,1 and D. Costea1; 1University of Bergen, Bergen, Norway,2Center of Medical Genetics and Molecular Medicine, Haukeland

University Hospital, Bergen, Norway, 3Queen Mary University of London,

London, United Kingdom

Purpose/Objective(s): Although differential expression of S100A16 has

been reported in human cancers, its precise functional roles in tumori-

genesis are not fully understood. The aim of the study was to investigate

the expression pattern and functional role of S100A16 in oral squamous

cell carcinoma (OSCC).

Materials/Methods: S100A16mRNA and protein levels were examined by

quantitative RT-PCR and immunohistochemistry in specimens of normal

human oral mucosa (NHOM), oral dysplastic lesions (ODL) and OSCCs.

S100A16 was overexpressed in OSCC-derived (CaLH3 and H357) cells by

using retroviral S100A16 expression construct and the subsequent func-

tional effects were examined by using established in vitro and in vivo

tumorigenesis assays.

Results: Both S100A16 mRNA and protein levels were found to be

gradually down-regulated from NHOM to ODL and OSCC. Low

S100A16 protein level in OSCC was found to be correlated with reduced

overall survival and poor differentiation grade. In vitro functional studies

Volume 94 � Number 4 � 2016 Posters 945

showed significant reduction in cell proliferation, sphere formation, and

invasive abilities of CaLH3 and H357 cells upon S100A16 over-

expression. S100A16 over-expression also suppressed tumorigenesis of

H357 cells in a mouse xenograft model. Molecular analysis revealed that

S100A16 over-expression led to induction of differentiation markers and

suppression of proliferation and invasion-related molecules both in vitro

and in vivo.

Conclusion: These results indicate that S100A16 might function as a

tumor suppressor protein in OSCC. S100A16 might be useful clinically

as a prognostic marker or as a novel therapeutic target for OSCC

treatment.

Author Disclosure: D. Sapkota: None. O. Bruland: None. H. Parajuli:

None. T. Osman: None. M. Teh: None. A. Johannessen: None.

D. Costea: None.

291Early Detection of Potential Cannibalistic Cells Before AnyMorphological Evidence of Cannibalism Using CD68 and Lysozymein Oral Squamous Cell CarcinomaS.D. Gawande,1 G.S. Sarode,2 S.C. Sarode,2 and S.S. Chaoudhary1; 1Dr. D.

Y. Patil Dental College and Hospital, Pimpri, Pune, India, 2Dr. D. Y. Patil

Dental College and Hospital, Pimpri, Pune, India

Purpose/Objective(s): Our institute is the first to report the cannibalism

phenomenon along with the concept of complex cannibalism in oral

squamous cell carcinoma (OSCC). The transformation of a cancer cell into

a cannibalistic cell is a process involving numerous sequential events.

Initial events could be associated with genetic expression of proteins

required for the execution of cellular cannibalism. Molecules that are

implicated in this process are CD68, lysozyme, caveolin-1, actin, ezrin,

cathepsin B, 9 transmembrane segment (TM9SF4), and vimentin, among

others. Genetic expression is later manifested in the form of engulfment

and digestion of adjacent cancer cells that are detectable on routine his-

topathology. We believe that identification of these markers in the tumor

cells can help us in recognizing the cells possessing potential for canni-

balistic behavior. The purpose of this investigation was to study expression

of CD68 and lysozyme for early detection of potential cannibalistic tumor

cells in OSCC before any morphological evidence of cannibalism with

histopathological correlation.

Materials/Methods: A total of 30 histopathologically diagnosed cases of

OSCC were subject to immunohistochemical staining using CD68 and

lysozyme antibodies. The expression was correlated with clinical and

histopathological grades.

Results: Out of 30 cases, 12 cases displayed positive expression for both

CD68 and lysozyme. For CD68, weak expression was reported in 8 cases,

followed by strong expression in 3 cases. Similarly for lysozyme, 10 cases

displayed weak expression, and 1 case showed strong positivity. There was

a significant difference in the expression of CD68 and lysozyme in

different histopathological grades (PZ.015 and .001).

Conclusion: Positive expression of CD68 and lysozyme in tumor cells of

OSCC could suggest the cannibalistic potential of the tumor cells before

any morphological evidence of cannibalism.

Author Disclosure: S.d. Gawande: None. G.S. Sarode: None. S.C.

Sarode: None. S.S. Chaoudhary: None.

292Extracellular Biosensors to Selectively Radiosensitize Head andNeck CancersS.J. Advani, S.R. Adams, H.Y. Yang, J.L. Crisp, J. Aguilera, Q.T. Nguyen,

and R.Y. Tsien; University of California, San Diego, La Jolla, CA

Purpose/Objective(s): Tumor resistance to concurrent chemotherapy and

radiation therapy remains a significant barrier to improving outcomes of

patients diagnosed with locally advanced head and neck squamous cell

cancers (HNSCC). The ability to deliver more potent chemotherapies with

radiation therapy is limited by normal tissue toxicity. We hypothesized that

potent drug-conjugated activatable cell-penetrating peptides (ACPP)

would function as biosensors to deliver highly potent radiosensitizing

drugs to tumors. ACPP are selectively cleaved and activated in the tumor

microenvironment through tumor-associated matrix metalloproteinases

(MMP). Once cleaved, they release drug-conjugated polycation cell-

penetrating peptides that are taken up by tumor cells.

Materials/Methods: In cell culture, we evaluated the tumoricidal and

radiosensitizing ability of AZD-7762 (a potent CHK 1/2 inhibitor) in a

panel of HNSCC cell lines (CAL-27, SCC-4, SCC-25, SCC-35, and SCC-

61) by Alamar Blue and clonogenic assays. In animal tumor models, we

tested the biosensor function of an MMP 2/9 cleavable ACPP to differ-

entiate radioresistant HNSCC tumors from those that are more radiosen-

sitive by using a ratiometric ACPP labeled with Cy5 and Cy7 that was iv

injected. Finally, we measured the efficacy of iv-delivered ACPP conju-

gated AZD-7762 with IR in HNSCC tumor xenografts.

Results: In a panel of HNSCC cells, AZD-7762 was more potent than

cisplatin. The IC50 for cisplatin was 3, 13, and 5 uM in CAL-27, SCC-

35, and SCC-61 cells and for AZD-7762 the IC50 was 0.3, 0.2, and 0.7

uM, respectively. In clonogenic assays, AZD-7762 significantly reduced

cell survival after 2 Gy, a clinically relevant IR dose. While 73% of

CAL-27 cells survived a dose of 2 Gy, survival was reduced to 45% in

cells treated with AZD-7762 and IR, P<.001. Mechanistically, AZD-

7762 reversed IR induced accumulation of cells in the G2/M phase of the

cell cycle. In HNSCC xenografts, tumors from relatively radioresistant

CAL-27 cells had increased cleavage of ratiometric ACPP (high Cy:Cy7

ratio) compared to relatively radiosensitive HNSCC SCC-4 tumors.

Finally in CAL-27 HNSCC xenografts, mice treated with tumor-targeted

ACPP conjugated AZD-7762 and IR had significantly prolonged tumor

xenograft regression compared to mice treated with nontargeted AZD-

7762 and IR, P<.01. Importantly, a noncleavable ACPP conjugated to

AZD-7762 did not radiosensitize tumors, thereby indicating the necessity

of tumor ACPP cleavage to release AZD-7762 for it to function as a

radiosensitizer.

Conclusion: In summary, we have demonstrated that ACPP conjugated

AZD-7762 provides a unique solution to tumor drug delivery by holding

the conjugated drug in an inert state until cleaved by tumor MMP 2/9.

ACPP therapeutically redirects tumor MMP 2/9 to amplify delivery of

potent radiosensitizers to HNSCC tumors.

Author Disclosure: S.J. Advani: None. S.R. Adams: None. H.Y. Yang:

None. J.L. Crisp: None. J. Aguilera: None. Q.T. Nguyen: Advisor;

Avelas. R.Y. Tsien: Advisor; Avelas.

293Prevalence and Prognostic Implications of Human Papillomavirusin Oral Cavity CancerD. Miller,1 S. Ko,1 Q. Zhai,1 Y. Habboush,1 K.S. Tzou,1 J.L. Peterson,1

L.A. Vallow,1 R.C. Miller,1 S.J. Buskirk,1 S.H. Patel,2 and R.L. Foote3;1Mayo Clinic, Jacksonville, FL, 2Mayo Clinic, Phoenix, AZ, 3Mayo Clinic,

Rochester, MN

Purpose/Objective(s): Over the last several decades, a dramatic rise in

the prevalence of human papillomavirus (HPV) among oropharyngeal

squamous cell carcinoma (OPSCC) has been seen. Prior studies have

shown the prognostic importance of p16 protein expression in OPSCC as a

surrogate marker for HPV infection. The interest in a possible corre-

sponding increase in HPV prevalence in non-OPSCC, particularly the oral

cavity, and its prognostic implications currently requires further investi-

gation. We attempt to address these questions through a retrospective

International Journal of Radiation Oncology � Biology � Physics946

review of patients with oral cavity squamous cell carcinoma (OCSCC) and

known HPV status.

Materials/Methods: The expression of p16 and high-risk HPV for patients

with OCSCC was determined by in situ hybridization (ISH) and immu-

nohistochemistry (IHC) from a single institution. The overall prevalence of

HPV in OCSCC was determined, as well as the median overall survival

time in months. All patients with a previous or concurrent diagnosis of

OPSCC were excluded from our review. These results were then compared

to other reported prevalence data.

Results: P16 expression was found to be positive in 6.6% (7 of 106)

of patients diagnosed with OCSCC from 1999 to 2015 at a single

institution. The results of our analysis are consistent with the majority

of prior smaller series showing the prevalence of high-risk HPV in the

oral cavity to be <10%, despite other recent publications suggesting

this could be higher. Our results did not show a positive prognostic

value for p16 and high-risk HPV with a median overall survival of 18

months (range, 6 to 37) for the HPV-positive group versus 70 months

(range, 3 to 150) for the HPV-negative group. The median follow-up

time for all patients was 23.5 months. The 3-year rates of overall

survival were 66.6% and 34.4% (PZ.097 by log-rank test),

respectively.

Conclusion: In contrast to patients with OPSCC in whom p16 expression

caries a positive prognostic value, our patient outcome data suggest that

this correlation does not appear to hold true for p16 and high-risk HPV

OCSCC patients. However, due to the low prevalence of HPV within the

oral cavity, it is difficult to predict the actual clinical impact of p16 and

high-risk HPV in OCSCC.

Author Disclosure: D. Miller: None. S. Ko: None. Q. Zhai: None.

Y. Habboush: None. K.S. Tzou: None. J.L. Peterson: None. L.A.

Vallow: None. R.C. Miller: None. S.J. Buskirk: None. S.H. Patel: None.

R.L. Foote: None.

294High-Grade Transformation of Acinic Cell Carcinoma: PotentiallyUnderrecognized and Inadequately TreatedA.V. Chintakuntlawar,1 W. Shon,2 M. Erickson-Johnson,3 E. Bilodeau,4

S. Jenkins,1 J. Davidson,1 M. Keeney,1 M. Rivera,1 D.L. Price,1

E.J. Moore,1 K.D. Olsen,1 J.L. Kasperbauer,1 R.L. Foote,1

K. Price,1 and J.J. Garcia1; 1Mayo Clinic, Rochester, MN,2University of Florida College of Medicine, Gainesville, FL,3Kailos Genetics, Inc., Huntsville, AL, 4University of Pittsburgh School of

Dental Medicine, Pittsburgh, PA

Purpose/Objective(s): Acinic cell carcinoma (ACC) is an uncommon

salivary gland tumor. We aim to describe the clinical and pathologic

characteristics of ACC with or without high-grade transformation

(HGT). Importantly, cases of mammary analogue secretory carcinoma

(MASC), a recently described histologic mimic of ACC, have been

excluded using cytogenetics and molecular studies for ETV6-NTRK3

fusion.

Materials/Methods: Archival tissue for patients (pts) diagnosed with ACC

at a single institution from 1990 to 2010 was obtained. Tumors with ETV6-

NTRK3 fusion transcript were excluded. Pathology was reviewed by 2

independent pathologists. Tumors with HGT were characterized by areas

with infiltrative growth, nuclear anaplasia, prominent nucleoli, brisk

mitotic rate, necrosis, and stromal desmoplasia. Relapse was defined as

nodal/local recurrence (locoregional), distant metastasis, or death.

Relapse-free survival (RFS), overall survival (OS), locoregional recur-

rence-free survival (LRFS), and distant metastasis-free survival (DMFS)

were calculated using the Kaplan-Meier method.

Results: Out of 54 pts, 6 were excluded due to positive ETV6-NTRK3

fusion transcript. ACC with HGT was seen in 8 of 48 pts (17%). Pts with

HGT were significantly older than pts without HGT (median 69 vs 54

years, PZ.04). Angiolymphatic invasion was more common in ACC

with HGT (PZ.02). The initial treatment in all pts consisted of surgery.

All 8 pts in the HGT group received total parotidectomy and neck

dissection. In the non-HGT group, 19 underwent parotidectomy alone,

and 21 had parotidectomy with neck dissection. Five pts (63%) in the

HGT group received adjuvant radiation in contrast to 5 pts (14%) from

the non-HGT group. None of the pts received adjuvant chemotherapy.

The HGT cohort showed a 5-year RFS of 25% (95% confidence interval

[CI]: 0-55) when compared to 82% (95% CI: 68-95) in the non-HGT

group (hazard ratio [HR] 10.4, 95% CI 3.4-31.5; P<.0001). The 5-year

OS was 43% (95% CI: 6-80) in the tumors with HGT in comparison to

97% (95% CI: 92-100) for the non-HGT group (HR 9.3, CI 2.6-33.0;

P<.0001). Locoregional recurrence was seen in 2 (of 8) and 9 (of 40) pts

with HGT and without HGT, respectively. Both pts in HGT who had

locoregional recurrence had not received adjuvant radiation after primary

resection. In the non-HGT group, 8 (of 9) pts who developed locore-

gional recurrence had not received adjuvant radiation. The 5-year LRFS

was 75% (95% CI: 33-100) for the HGT group and 84% (95% CI: 71-97)

for the non-HGT group (HR 3.44, CI 0.66-17.9; PZ.12). Six pts

developed distant metastasis in HGT tumor group versus none in the

non-HGT group. The 5-year DMFS was 38% (95% CI: 4-71) for the

HGT group and 100% (CI cannot be calculated) for the non-HGT group

(P<.0001).

Conclusion: The prevalence of HGT in cases of ACC is higher than pre-

viously suggested once cases of MASC are excluded. Prognosis for OS and

DMFS for ACC pts with HGT is significantly worse than for pts without

HGT in this large cohort.

Author Disclosure: A.V. Chintakuntlawar: None. W. Shon: None.

M. Erickson-Johnson: None. E. Bilodeau: None. S. Jenkins: None.

J. Davidson: None. M. Keeney: None. M. Rivera: None. D.L. Price:

None. E.J. Moore: None. K.D. Olsen: None. J.L. Kasperbauer: None.

R.L. Foote: None. K. Price: None. J.J. Garcia: None.

295Nomograms for Predicting Survival and Recurrence in Patients WithAdenoid Cystic CarcinomaI. Ganly; Memorial Sloan Kettering Cancer Center, New York, NY

Purpose/Objective(s): Due to the rarity of adenoid cystic carcinoma

(ACC), information on outcome is based on small retrospective case series.

The aim of our study was to create a large multi-institutional international

dataset of patients with ACC in order to design predictive nomograms for

outcome.

Materials/Methods: ACC patients managed at 10 international centers

were identified. Patient, tumor, and treatment characteristics were

recorded and an international collaborative dataset created. Multivariable

competing risk models were then built to predict the 10-year recurrence-

free probability (RFP), distant recurrence-free probability (DRFP),

overall survival (OS), and cancer-specific mortality (CSM). All pre-

dictors of interest were added in the starting full models before selection,

including age, gender, tumor site, clinical T stage, perineural invasion,

margin status, pathologic N status, and M status. Stepdown method was

used in model selection to choose predictive variables. An external

dataset of 99 patients from 2 other institutions was used to validate the

nomograms.

Results: Of 438 ACC patients, 27.2% (119 of 438) died from ACC and

38.8% (170 of 438) died of other causes. Median follow-up was 56 months

(range 1-306). The nomogram for OS had 7 variables (age, gender, clinical

T stage, tumor site, margin status, pathologic N status, and M status) with a

concordance index (CIN) of 0.71. The nomogram for CSM had the same

variables, except margin status, with a CIN of 0.70. The nomogram for

RFP had 3 variables (gender, tumor site, and margin status; CIN, 0.67).

The nomogram for DRFP had 6 variables (gender, clinical T stage, tumor

site, pathologic N status, perineural invasion, and margin status; CIN,

Volume 94 � Number 4 � 2016 Posters 947

0.64). Concordance indexes for the external validation set were 0.76, 0.72,

0.67, and 0.70, respectively.

Conclusion: Using an international collaborative database, we have createdthe first nomograms that estimate outcome in individual patients with

ACC. These predictive nomograms will facilitate patient counseling in

terms of prognosis and subsequent clinical follow-up. They will also

identify high-risk patients who may benefit from clinical trials on new

targeted therapies for patients with ACC.

Author Disclosure: I. Ganly: None.

Abstract 297; Table 1 Baseline characteristics

NZ26 No. (%)

Age, yMedian 58Range 37 - 84

GenderMale 11 42Female 15 58

ECOG0 10 381 16 62

Primary siteParotid 13 50Minor Salivary 6 23Other 7 27

HistologyAdenoid cystic 12 46Salivary duct 5 19Acinic cell 3 12Adenocarcinoma 2 8Mucoepidermoid 1 4Poorly differentiated carcinoma 3 12

296Role of Postoperative Radiation Therapy in Stage IA Merkel CellCarcinoma of the Head and NeckS.R. Takagishi,1 T. Marx,1 C. Lewis,1 E. Shantha,1 I. Juhlin,1 T. Fu,2

A. Blom,1 J. Iyer,1 J.J. Liao,1 Y.D. Tseng,1 P. Nghiem,1

and U. Parvathaneni1; 1University of Washington, Seattle, WA, 2Stanford

University, Redwood city, CA

Purpose/Objective(s): Merkel cell carcinoma (MCC) is a rare and often

aggressive skin cancer. The head and neck (HN) is a common location

for MCC, and approximately 50% of patients present with stage I dis-

ease. Typically, surgery is the primary treatment. Postoperative radiation

therapy (PORT) is often recommended to improve local control, as wide

margins are generally not achievable in the HN region and MCC is a

radioresponsive disease. It is unclear whether PORT may be safely

omitted in selected low-risk stage IA cases. From our clinical observa-

tions, we hypothesized that PORT would reduce the risk of local

recurrence (LR) in HNMCC, even in patients with low-risk

characteristics.

Materials/Methods: We conducted a retrospective analysis of 46 low-risk

HNMCC cases treated with primary resection, identified from our re-

pository of 1171 patients enrolled between 2006 and 2015. Inclusion

criteria were (1) primary tumor �2 cm in maximum dimension, (2)

negative margins on final pathology, (3) negative sentinel lymph node

biopsy, and (4) no immunosuppression. We used the Kaplan-Meier method

to estimate LR and overall survival (OS). Local recurrence was defined as

tumor recurrence within 2 cm of the primary surgical bed. The cumulative

incidence of disease-specific death (DSD) was estimated using death from

non-MCC causes as a competing risk. The Fisher exact test was used for

group comparisons.

Results: Low-risk HNMCC patients were treated with and without PORT

(nZ23, for both groups). No patients received adjuvant chemotherapy.

There were no significant differences between the 2 groups in terms of sex,

race, age at diagnosis, tumor size, depth of invasion, lymphovascular space

invasion, and width of surgical margins. The median follow-up time was

3.55 years for the surgery alone group and 5.34 years for the surgery +

PORT group. There was a significant difference in LR between the groups

treated with and without PORT (PZ.02). Among the 23 patients treated

with surgery alone, 6 (26%) recurred locally. The median time to local

recurrence was 11 months. Five of the LRs occurred in patients with

primary tumors �5 mm in diameter, and the sixth was in an 8-mm tumor.

Out of the 23 patients treated with both surgery and PORT, no patients

recurred locally. The median RT dose was 50 Gy (range, 16-66 Gy). In

general, the surgical bed with a minimum margin of 3 to 5 cm was irra-

diated with electrons/photons. Significant late toxicity (>grade 2 by

Common Terminology Criteria for Adverse Events version 4) was not

reported in patients who received PORT. There were 2 regional relapses in

the surgery only group, and 1 in the surgery + PORT group. There was no

difference in OS or DSD. Zero events in 1 group precluded multivariate

analysis.

Conclusion: PORT was associated with a significantly lower risk of local

recurrence in patients with stage IA MCC of the head and neck.

Author Disclosure: S.R. Takagishi: None. T. Marx: None. C. Lewis:

None. E. Shantha: None. I. Juhlin: None. T. Fu: None. A. Blom: None.

J. Iyer: None. J.J. Liao: None. Y.D. Tseng: None. P. Nghiem: None.

U. Parvathaneni: None.

297Genomic Profiling and Matched Therapy for Recurrent or MetastaticMalignant Salivary Gland Tumors (MSGT): Preliminary ResultsD. Day, R. Jang, A. Spreafico, E. Chen, T. Stockley, S. Kamel-Reid,

I. Weinreb, B. Perez-Ordonez, L. Siu, A.R.A. Razak, and A. Hansen;

Princess Margaret Cancer Centre, Toronto, ON, Canada

Purpose/Objective(s): MSGT are rare, clinicopathologically heteroge-

neous tumors with limited systemic therapeutic options. This single-

institution, prospective study in advanced MSGT involves 2 phases:

genomic profiling (profiling phase) and then treatment with either

genomically-matched or unmatched therapy (treatment phase), if clinically

indicated and available. This ongoing trial (NCT02069730) will compare

response rates in patients (pts) with MSGT for matched and unmatched

therapy.

Materials/Methods: Pts with recurrent/metastatic MSGT with archived

paraffin tumor samples were enrolled in the profiling phase. Following

pathology review and DNA extraction, targeted next-generation

sequencing (NGS) was performed using the Illumina Miseq TruSeq

Amplicon Cancer Panel (48 genes, 212 amplicons) in a CLIA certified

laboratory. Immunohistochemistry (IHC) and fluorescence in situ hybrid-

ization (FISH) were performed for androgen receptor (AR), HER2, and

ALK. Successfully profiled pts could continue onto the treatment phase to

receive an opportunistically matched therapy, accessed via early phase

clinical trials or from approved agents. If no actionable mutation was

found or no matched therapy was available, pts could receive selinexor

(KPT-330), the oral selective inhibitor of nuclear export (SINE) compound

that inhibits XPO1 (unmatched therapy).

Results: From July 2014, 26 pts were enrolled in the profiling phase.

Demographic and tumor details are listed in Table 1. Currently, 8 tumor

samples are awaiting processing and analysis, 2 had incomplete analysis,

and 16 have results returned. Of these, 7 pts (44%) had at least 1 actionable

aberration. Overall, 12 aberrations were found in AR by IHC (3), HER2 by

FISH (1), PIK3CA (3), EGFR (1), c-KIT (1), BRAF (1), HRAS (1), and

TP53 (1) by targeted NGS. Four pts (57%) with actionable aberrations

started genotype-matched treatment; 3 with androgen deprivation and 1

with a PI3K inhibitor on a phase 1 trial. Nine pts (56%) had no mutations,

and of these, 3 have received unmatched therapy and 1 screen-failed.

Median time from consent to report of genomic profile was 53 days (range

27-101). Eight pts are being managed expectantly.

Conclusion: Genomic profiling can be integrated into clinical care for pts

with MSGT, and the majority of pts with actionable mutations received

matched therapy. These results are preliminary, and further analysis is

needed to determine the clinical benefit of genomically selected treatment

for MSGT.

International Journal of Radiation Oncology � Biology � Physics948

Author Disclosure: D. Day: None. R. Jang: None. A. Spreafico: None.

E. Chen: None. T. Stockley: None. S. Kamel-Reid: None. I. Weinreb:

None. B. Perez-Ordonez: None. L. Siu: Research funding (clinical trial);

Novartis, Karyopharm. A.R. Razak: Research support; Karyopharm.

A. Hansen: None.

298A Multi-institutional Comparison of Outcomes ofImmunocompromised and Immunocompetent Patients TreatedWith Surgery and Radiation Therapy for Cutaneous Squamous CellCarcinoma of the Head and NeckB. Manyam,1 A.A. Garsa,2 R.I. Chin,3 C.A. Reddy,1 B. Gastman,1

A.T. Vidimos,1 W.L. Thorstad,3 S.S. Yom,2 B. Nussenbaum,3 S. Wang,2

and S. Koyfman4; 1Cleveland Clinic, Cleveland, OH, 2University of

California San Francisco, San Francisco, CA, 3Washington University

School of Medicine, St. Louis, MO, 4Cleveland Clinic Foundation,

Cleveland, OH

Purpose/Objective(s): Patients (pts) who are chronically immunosup-

pressed (IS) have higher rates of cutaneous squamous cell carcinoma of the

head and neck (cSCC-HN), one of the most aggressive forms of skin

cancer. This multi-institutional study analyzes the effect of immune status

on disease outcomes in pts with primary or recurrent stage I-IV cSCC-HN

treated with surgery and postoperative radiation therapy (RT).

Materials/Methods: Pts who received surgical resection and postoperative

RT between 1995 and 2015 at Cleveland Clinic, Washington University St.

Louis, and University of California San Francisco were identified in each

institution’s respective institutional review boardeapproved registry. Pts

were categorized as IS if they were diagnosed with chronic hematologic

malignancy, HIV, or were treated with immunosuppressive therapy for

organ transplantation �6 months prior to diagnosis. Overall survival (OS),

locoregional recurrence-free survival (LRRFS), and disease-free survival

(DFS) were calculated using the Kaplan-Meier method. Univariate (UVA)

and multivariate (MVA) analysis was performed using Cox proportional

hazards regression to identify patient, tumor, and treatment-related vari-

ables associated with locoregional recurrence (LRR).

Results: In this study of 205 pts, 138 (67.3%) were immunocompetent

(IC) and 67 (32.7%) were IS. Poor differentiation (PD) (55.2% vs

36.2%; PZ.009) was more frequent in IS pts, but perineural invasion

(PNI) (53.7% vs 46.4%; PZ.323) and nodal extracapsular extension

(77.3% vs 64.2%; PZ.664) were numerically, but not statistically

different between groups. LRRFS (47.3% vs 86.1%; P<.0001) and DFS

(43.1% vs 79.7% P<.0001) were significantly lower in IS pts compared

to IC at 2 years. OS in IS pts at 2 years was not statistically different

(60.9% vs 78.1%; PZ.135). UVA on all pts demonstrated that IS status

(hazard ratio [HR] 3.82; P<.0001), recurrent disease (HR 1.90;

PZ.025), PD (HR 2.30; PZ.002), and PNI (HR 1.78; PZ.03) were

significantly associated with higher rates of LRR. On MVA, IS status

(HR 3.79; P<.0001), recurrent disease (HR 2.67; PZ.001), PD (HR

2.08; PZ.006), and PNI (HR 2.05; PZ.009) remained significantly

associated with higher rates of LRR.

Conclusion: This multi-institutional analysis demonstrates that IS pts with

cSCC-HN more frequently present with high-risk pathologic features and

have dramatically lower DFS and LRRFS at 2 years compared to IC pts. IS

status, recurrent disease, PD, and PNI were significantly associated with

higher LRR. Intensification strategies for IS pts merit investigation.

Author Disclosure: B. Manyam: None. A.A. Garsa: None. R. Chin:

None. C.A. Reddy: None. B. Gastman: None. A.T. Vidimos: None. W.L.

Thorstad: None. S.S. Yom: Research Grant; Genentech. B. Nussenbaum:

None. S. Wang: None. S. Koyfman: None.

299Molecular Attributes of Metastatic Thyroid CancersE.S. Kim,1 R. Feldman,2 D. Carrizosa,1 Z. Gatalica,2 S. Reddy,2

and G. Yoo3; 1Levine Cancer Institute - Carolinas HealthCare

System, Charlotte, NC, 2Caris Life Sciences, Phoenix, AZ,3Barbara Ann Karmanos Cancer Institute - Wayne State

University, Detroit, MI

Purpose/Objective(s): Thyroid cancers are generally considered low-risk

tumors, but when recurrent or metastatic, can be challenging to treat. As

molecular abnormalities have been described, more treatment options have

emerged, especially targeting pathways of BRAF, VEGF, and RET. Mo-

lecular differences between primary and metastatic tumors may help

identify additional treatment strategies. We inquired a database of multi-

platform molecular profiling data to ascertain molecular events that may

contribute to metastasis of a generally indolent disease and how those

molecular changes may provide treatment options for patients with later

stage disease.

Materials/Methods: Two hundred forty-nine samples from patients with

thyroid cancers (included follicular [FTC], medullary [MTC], papillary

[PTC], Hϋrthle, anaplastic [ATC], poorly differentiated [PDTC], mixed,

and other histotypes) referred to Caris Life Sciences between 2010 and

2015 were evaluated. Specific testing was performed per physician request

and included a combination of sequencing (Sanger, NGS), protein

expression (IHC), and gene amplification (CISH/FISH). Fisher exact test

(P<.05), JMPv10.0 (SAS Institute Inc., Cary, NC) was utilized for statis-

tical analysis.

Results: Profiling was performed on 95 (38%) and 154 (62%) primary

(P) and metastatic (M) specimens, respectively. Multidrug resistance

proteins, MRP1, but not BCRP or PGP, demonstrated statistically sig-

nificant differences between M and P (74% vs 45% protein over-

expression, P<.05). Other proteins that demonstrated differences in

expression, however, did not reach statistical significance and included

TOPO1 and TUBB3. Interestingly, expression of programmed death-1

(PD1) on tumor infiltrating lymphocytes (TIL) was expressed at higher

rates in P versus M (52% vs 26%, trending toward significance), whereas

the expression of PD-L1 in P and M thyroid cancer was consistent at

43%. Of 47 genes tested by targeted next generation sequencing, only

PTEN mutation rates were significantly different among P versus M

(13%; 5/38 vs 0%; 0 of 57, P<.01). Driver mutations commonly asso-

ciated with thyroid cancers identified in both P and M subgroups

included BRAF (45% and 35%, respectively), RET (5% and 7%

respectively), KRAS (6% and 3%, respectively) and NRAS (13% and

16%, respectively).

Conclusion: Multidrug resistance proteins may contribute to a chemo-

therapy-resistant phenotype in advanced, metastatic thyroid cancers.

Thyroid-immune interactions may play a role in the reduced PD1 TILs

observed in the metastatic stages of thyroid cancer. Biomarker identifica-

tion is an important tool for evaluating and treating patients with thyroid

cancers. Immune checkpoint inhibitors may be reasonable therapy options

to assess in clinical trials for patients with thyroid cancers based on

consistent levels of PD-L1 expression.

Author Disclosure: E.S. Kim: None. R. Feldman: None. D. Carrizosa:

None. Z. Gatalica: None. S. Reddy: None. G. Yoo: None.

Abstract 301; Table 1 Survival analysis and prognostic factors univariateanalysis.

OS DSS LRFS DRFS

Median survival (months) 32 50 43 123-y survival (%) 48 (37-60) 51 (40-64) 57 (44-72) 38 (26-49)5-y survival (%) 31 (18-43) 40 (27-54) 36 (19-53) 28 (16-39)Hazard ratios (multivariate analysis)Subsite (nasal) 3.41* 3.45*MMHN T classification 0.43* 0.43* 0.43*Radiation therapy 2.81*Post-RT PET positive 0.24* 0.21* 0.17* 0.46

*P<.05, P<.1 for all reported hazard ratios

Hazard ratio >1 indicated improved survival with factor.

Volume 94 � Number 4 � 2016 Posters 949

300Radiation Therapy Improves Survival With Merkel Cell Carcinoma ofthe Head and NeckT. Strom, V.K. Sondak, J.S. Zager, J.L. Messina, A.O. Naghavi, J.F. Torres-

Roca, T.A. Padhya, A. Trotti, J.J. Caudell, and L.B. Harrison; H. Lee

Moffitt Cancer Center and Research Institute, Tampa, FL

Purpose/Objective(s): Merkel cell carcinomas (MCC) are aggressive,

small round blue cell tumors of the skin often found in a head and

neck location. Given the frequent cosmetic and functional limitations

of head and neck surgery, wide circumferential margins are often

difficult to obtain. We hypothesized that postoperative radiation

therapy (RT) would reduce the risk of locoregional recurrence and

improve survival.

Materials/Methods: A single-institution institutional review board-

eapproved study was performed including 144 patients with MCC of the

head and neck without distant metastatic disease treated from 1989 to

2012. Most patients were treated with wide excision (nZ140, 97.2%) �sentinel lymph node biopsy (SLNB) � neck dissection (LND). Post-

operative RT (nZ99, 68.9%) was delivered to the primary tumor bed �draining lymphatics. Patient, tumor, and treatment characteristics were

compared based on receipt of RT. Patients were treated with RT to a

median total dose of 5000 cGy (range 2,500-6,600 cGy). The primary

outcomes were local control (LC), locoregional control (LRC), disease-

free survival (DFS), and overall survival (OS). Kaplan-Meier (KM) ana-

lyses with log-rank tests and Cox multivariate models were created for the

outcomes of interest.

Results: Median follow-up of surviving patients was 26 months. Receipt

of RT was associated with more advanced nodal status (cN1 status 10.8%

vs 2.4% and pN1 status 15.7% vs 11.9%, respectively; PZ.003), and a

lower median age (76 vs 78, PZ.02). On KM univariate analysis, post-

operative RT was associated with improved 3-year LC (89.3% vs 67.6%;

PZ.001), LRC (68.1% vs 28.5%; P<.001), DFS (56.6% vs 20.9%;

P<.001), and OS (75.0% vs 65.3%; PZ.003). Similarly, on Cox MV

analysis, RT was associated with improved LC (hazard ratio [HR] 0.19,

95% confidence interval [CI] 0.07-0.51; PZ.001), LRC (HR 0.22, 95% CI

0.11-0.41; P<.001), DFS (HR 0.24, 95% CI 0.14, 0.41; P<.001), and OS

(HR 0.41, 95% CI 0.24-0.72; PZ.002). The only other variable associated

with OS on MV analysis was a tumor size >2 cm (HR 3.21, 95% CI 1.51-

6.81; PZ.002). A benefit in median LRC time was seen with RT among

node-negative patients (nZ111, 187 mo vs 10 mo, respectively; P<.001)

and node-positive patients (nZ33, median not reached vs 6 mo, respec-

tively; P<.001). The 3-year LC rates of patients with negative margins

treated with surgery and RT (nZ87) compared with patients with positive

margins or gross disease treated with definitive RT (nZ15) were 88.8%

and 92.3%, respectively (PZ.88). Similarly, the 3-year regional control

rates of node-positive patients treated with LND � RT (nZ25) compared

with SLNB positive and definitive RT (nZ6) were 79.2% and 100%,

respectively (PZ.25).

Conclusion: RT is associated with improved LC, LRC, DFS, and OS in

patients with MCC of the head and neck.

Author Disclosure: T. Strom: None. V.K. Sondak: Independent

Contractor; Amgen, BMS, Merck, Navidea, Novartis, Provectus. J.S.

Zager: Independent Contractor; IGEA, Delcath Systems, Inc.. J.L. Mes-

sina: Independent Contractor; Durect Corporation, Glaxo Smith Kline.

A.O. Naghavi: None. J.F. Torres-Roca: Partner; Cvergenix, Inc.. Patent/

License Fees/Copyright; Cvergenix, Inc.. T.A. Padhya: None. A. Trotti:

None. J.J. Caudell: None. L.B. Harrison: None.

301Localized Sinonasal Mucosal Melanoma: Outcomes and AssociationsWith Stage, Radiation Therapy, and Positron Emission TomographyResponseR. Samstein,1 C.A. Barker,1 R.D. Carvajal,2 M.A. Postow,1

M.K. Callahan,1 A.N. Shoushtari,1 S.G. Patel,1 and N. Lee1; 1Memorial

Sloan Kettering Cancer Center, New York, NY, 2Clumbia University, New

York, NY

Purpose/Objective(s): Sinonasal mucosal melanoma (SNMM) is a rare

neoplasm with a poor prognosis. Management of localized disease consists

of surgery with or without adjuvant radiation therapy (RT).

Materials/Methods: Retrospective analysis was conducted on 78 patients

with localized SNMM surgically resected and treated at 1 institution be-

tween 1998 and 2013. Demographic, tumor, molecular pathology, imaging,

and treatment factors were recorded and survival and disease-control

outcomes were analyzed. Multivariate Cox proportional hazard analysis

was performed on factors to determine the association with recurrence and

survival.

Results: Median age was 68 years, 49% of patients were male, and 88%

were white. Tumor was located in the nasal cavity in 52 patients (67%) and

clinical stage was T2, T4a, and T4b in 39 (50%), 29 (37%), and 8 patients

(10%), respectively. Sixty-four patients (82%) received RT, and 41 had

positron emission tomography (PET) imaging after RT. Median overall

survival (OS) and disease-specific survival (DSS) were 32 and 50 months,

with 5-year OS of 31% and DSS of 40%. Median local recurrence-free

survival (LRFS) and distant recurrence-free survival (DRFS) were 43 and

12 months, respectively, with distant recurrence ultimately observed in

66% of patients. Univariate and multivariate analysis demonstrated greater

OS and DSS in tumors of the nasal cavity and of earlier T stage. RT was

associated with significantly greater LRFS (5-year LRFS, 35% vs 59%) but

no difference in OS. KIT, BRAF, NRAS, and GNAQ were altered in 2 of 29

(7%), 1 of 27 (4%), 6 of 24 (25%), and 0 of 13 (0%) cases and were not

associated with tumor subsite, disease recurrence, or survival. No associ-

ation was observed between preoperative or postoperative PET, and any

survival metric but post-RT PET response was associated with greater

LRFS, DRFS, OS, and DSS on multivariate analysis with median OS of 56

months compared with 14 months.

Conclusion: Distant metastatic recurrence is the predominant mode of

recurrence in SNMM, but local recurrence is also common. Radiation

therapy is associated with improved local control but no benefit in survival.

The prognostic value of postradiation therapy PET imaging warrants

further investigation.

Author Disclosure: R. Samstein: None. C.A. Barker: None. R.D. Car-

vajal: None. M.A. Postow: None. M.K. Callahan: None. A.N. Shoush-

tari: None. S.G. Patel: None. N. Lee: None.

302The Role of External Beam Radiation in the Adjuvant Setting forStage IV Differentiated Thyroid CancerV. Mehta,1 C.A.O. Nathan,1 S.R. Katz,2 and J. Flores3; 1LSU Health

Shreveport, Shreveport, LA, 2Willis-Knighton Cancer Center, Shreveport,

LA, 3Johns Hopkins School of Public Health, Baltimore, MD

Purpose/Objective(s): There is a relative paucity of well-powered, pro-

spective studies examining the role of adjuvant radiation in the survival of

patients with stage IV differentiated thyroid cancer (DTC). This analysis

expands current understanding of long-term survival in follicular thyroid

cancer (FTC) and papillary thyroid cancer (PTC) subgroups that receive

International Journal of Radiation Oncology � Biology � Physics950

adjuvant therapy with radioactive iodine, external beam irradiation, or

neither.

Materials/Methods: The survival of 3162 patients from the National

Cancer Data Base (representing 14,440 person-years at risk) was analyzed

using fully parametric, multilevel survival time models. Patients were

stratified by type of cancer (FTC vs PTC) and by grade within each cancer

(well, moderately, or poorly differentiated). The relationship between ra-

diation therapy and survival was compared between 3 groups (external

beam radiation [EBRT] and radioactive iodine therapy [RAI] vs no radi-

ation). All models are adjusted for age, gender, income, lymphatic/vascular

tumor invasion, and extent surgical resection.

Results: The mean age (and standard deviation) of the sample was 63.59

(11.89) years, and 57% were female. Follicular thyroid cancer: For

patients with well or moderately differentiated FTC, 5- and 10-year

mortality rates for EBRT and RAI were not statistically significantly

different compared to those for patients receiving no radiation. For cases

with poorly differentiated FTC, the RAI subgroup was associated with

improved 10-year survival in univariate analysis (hazard ratio [HR]

0.61, 0.48-0.79), but this protective association was no longer signifi-

cant after statistical adjustment. Papillary thyroid cancer: In univariate

analysis, RAI therapy was associated with improved 10-year survival

compared to patients receiving no radiation therapy in all PTC sub-

groups (well differentiated: HR 0.52, 0.38-0.70; moderately differenti-

ated: HR 0.60, 0.36-0.98; poorly differentiated: HR 0.49, 0.35-0.70).

This protective association was no longer significant for well-differen-

tiated PTC cases, was marginally associated for poorly differentiated

PTC cases (HR 0.52, 0.26-1.01, PZ.056), and remained statistically

significant for moderately differentiated PTC patients (HR 0.31, 0.19-

53, P<.0001).

Conclusion: In this large prospective cohort, the 5- and 10-year survival of

patients with FTC did not differ significantly between patients receiving

adjuvant EBRT or RAI compared to patients who did not receive either

treatment. However, adjuvant RAI therapy might have a more pronounced

role in the 5- and 10-year survival of patients with moderately and poorly

differentiated PTC.

Author Disclosure: V. Mehta: None. C.O. Nathan: None. S.R. Katz:

None. J. Flores: None.

303Bioradiation Therapy for High-Risk Cutaneous Squamous CellCancer of the Head and Neck: A Propensity Score AnalysisJ.D. Palmer,1 J. Strasser,2 A. Hanlon,3 J. Silberg,1 E.A. Mauer,3

M.F. Dzeda,2 N. Hockstein,2 R. Witt,2 C. Schneider,2 and A. Raben2;1Sidney Kimmel Medical College at Thomas Jefferson University,

Philadelphia, PA, 2Christiana Care Health System, Newark, DE,3University of Pennsylvania, Philadelphia, PA

Purpose/Objective(s): Nonmelanoma skin cancer is the most common

adult malignancy; with the growing elderly population, the incidence of

high-risk cutaneous squamous cell carcinoma (CSCC) is likely to increase.

The objective of this study was to investigate the safety, tolerability, and

preliminary efficacy of radiation therapy (RT) plus cetuximab in high-risk

CSCC patients.

Materials/Methods: Patients with high-risk CSCC diagnosed between

2006 and 2013 were analyzed. Patients were divided into 2 groups: RT

alone versus RT plus cetuximab. Among 68 patients meeting study criteria,

we identified 29 treated with cetuximab plus RT and 39 with RT alone.

Primary analysis examined disease-free and overall survival and freedom

from local and distant recurrence in the propensity scoreematched cohort.

Propensity score analysis was performed with weighted factors, including

Charlson comorbidity index score, age, KPS, primary location, T and N

stage, recurrent status, margin status, LVSI, PNI, and grade. Toxicity was

assessed using the Common Terminology Criteria for Adverse Events

version 4.0.

Results:Median follow-up for living patients was 30 months. Patients in

the cetuximab group were more likely to have advanced N stage,

positive margins, and recurrent disease. After propensity score match-

ing, the groups were well balanced. Six patients experienced > grade 3

acute toxicity in the cetuximab group. The 1-year, 2-year, and 5-year

progression-free survival (PFS) for patients in the cetuximab group were

86%, 72%, and 66%, respectively, and for the RT alone group were

77%, 53%, and 29%, respectively. The 1-year, 2-year, and 5-year overall

survival for patients in the cetuximab group were 98%, 80%, and 80%,

respectively, and for the RT alone group were 81%, 73%, and 61%,

respectively.

Conclusion: Although limited by small numbers, the combination of

cetuximab and RT in CSCC appears well tolerated; there were more long-

term survivors and fewer distant metastases in the cetuximab group. These

promising findings warrant further studies.

Author Disclosure: J.D. Palmer: None. J. Strasser: None. A. Hanlon:

None. J. Silberg: None. E.A. Mauer: None. M.F. Dzeda: None. N.

Hockstein: None. R. Witt: None. C. Schneider: None. A. Raben:

Speaker’s Bureau; Bristol Myers Squibb.

304Anaplastic Thyroid Cancer (ATC): Prognostic Factors, Patterns ofCare, and Overall SurvivalS.M. Glaser,1 S.F. Mandish,2 B.S. Gill,1 G. Balasubramani,3 D.A. Clump,

II,1 and S. Beriwal1; 1University of Pittsburgh Cancer Institute, Pittsburgh,

PA, 2Marshall University School of Medicine, Huntington, WV, 3University

of Pittsburgh School of Public Health, Pittsburgh, PA

Purpose/Objective(s): ATC represents a rare yet highly aggressive ma-

lignancy. We sought to analyze factors predictive for overall survival and

use of high-dose radiation therapy (HDRT) as well as patterns of care for

patients with ATC in a population-based cohort. In doing so, we aim to add

to the body of knowledge regarding patient selection for aggressive

treatment versus supportive care.

Materials/Methods: Using the National Cancer Data Base from 1999 to

2012, we identified 3552 patients with ATC. Factors associated with use of

HDRT (�59.4 Gy) were evaluated using a parsimonious multivariate lo-

gistic regression model. From this, a propensity score for receipt of high-

dose RTwas generated and incorporated into a multivariate Cox regression

analysis for overall survival. Three-month conditional landmark analysis

was performed.

Results: Median overall survival was 3.5 months (inerquartile range

[IQR], 1.5-8.8). RT use was 59.2% overall and 72.7% for patients who

lived more than 3 months, while for chemotherapy (CT) the rates were

42.4% and 56.5%, respectively. Both RT and CT were given to 37.3% of

patients (75.0% concurrent). Median time from diagnosis to RT initia-

tion was 27 days (IQR, 13-46) and 30 days for CT (IQR, 15-47). RT

lasted a median of 35 days (IQR, 17-48). Death within 30 days of

starting RT occurred in 18.2% of patients and within 60 days for 35%.

Median RT dose was 45 Gy (�36 Gy Z 32.1%, �59.4 Gy Z 38.1%).

Factors predictive for HDRT use were later year of diagnosis, private

insurance, rural residential setting, higher residential area median in-

come, no reported metastasis, negative surgical margins, receipt of

surgery, and receipt of chemotherapy. Factors associated with improved

overall survival were age <65 years, lower Charlson-Deyo comorbidity

score, treatment at a community/comprehensive community hospital,

facility volume >5 cases, primary confined to the thyroid, total thy-

roidectomy, other surgery, RT, and CT (hazard ratios [HR]Z0.71, 0.74,

0.82, 0.79, 0.71, 0.35, 0.48, 0.59, and 0.69, respectively, all P<.01).

Factors associated with decreased overall survival were clinical or

pathologic lymph node involvement, metastasis, tumor size >6 cm, and

positive surgical margins (HRZ1.22, 1.82, 1.38, and 1.53, respectively,

all P<.01). On conditional landmark analysis, the improved survival

seen with CT and surgery other than total thyroidectomy was lost but

persisted for total thyroidectomy and HDRT (HRZ0.85, PZ.082;

HRZ0.95, PZ.608; HRZ0.74, PZ.003; HRZ0.71 unadjusted and

0.72 propensity adjusted, P<.0005), while lower RT doses were not

associated with a survival advantage.

Volume 94 � Number 4 � 2016 Posters 951

Conclusion: This study has identified prognostic factors and patterns of

care for patients with ATC. High-dose RT was associated with an overall

survival benefit.

Author Disclosure: S.M. Glaser: None. S.F. Mandish: None. B.S. Gill:

None. G. Balasubramani: None. D.A. Clump: None. S. Beriwal: None.

305Improving Local Control for Unresectable/Incompletely ResectedSinonasal Cancer With Hyperfractionated Proton Therapy andConcurrent ChemotherapyR. Dagan, C.M. Bryant, and W.M. Mendenhall; University of Florida

Health Proton Therapy Institute, Jacksonville, FL

Purpose/Objective(s): Optimal delivery of local therapy for sinonasal

cancers has been limited by adjacent anatomic structures resulting in local

control (LC) rates of 20% to 40%. The unique physical characteristics of

proton therapy (PT) and hyperfractionation can potentially improve out-

comes. The purpose of this study is to report outcomes of patients with

unresectable/incompletely resected sinonasal cancer treated with hyper-

fractionated PT and concurrent chemotherapy.

Materials/Methods: Twenty-four patients were treated with PT

involving the following sinonasal sites: nasal/ethmoid (nZ16), maxil-

lary (nZ5), sphenoid (nZ2), and frontal (nZ1). Histologies included

squamous cell carcinoma (nZ9), adenoid cystic carcinoma (nZ8), ol-

factory neuroblastoma (nZ2), and other (nZ5: mucoepidermoid car-

cinoma, sinonasal undifferentiated carcinoma, neuroendocrine

carcinoma, polymorphous low-grade adenocarcinoma, and small cell

carcinoma). Twenty patients had intracranial tumors and 9 had orbital

invasion. Fourteen patients were treated after biopsy alone, 8 after

subtotal resection, and 2 for recurrence after gross total resection.

Median PT dose was 73.8 CGE (range, 56.8-74.4 CGE) at 1.2 CGE/

fraction twice daily with elective nodal therapy. All but 1 patient

received chemotherapy (mostly concurrent low-dose weekly cisplatin).

Median follow-up was 2.4 years (range, 0.4-6.7 years) for all patients

and 2.7 years for living patients (range, 0.6-6.7 years). LC, local-

regional control (LRC), freedom from distant metastases (FFDM), dis-

ease-free survival (DFS), and overall survival (OS) were estimated using

the Kaplan-Meier method. Grade 3 or higher toxicities were reported

using the Common Terminology Criteria for Adverse Events, version

4.0.

Results: Outcomes are shown in the table. There were 10 local recurrences

occurring at a median of 0.9 years after PT (range, 0.23-2.6 years). One

patient did not complete PT due to hospitalization for severe pneumonia.

The remaining received >70 CGE. One patient, whose tumor was unre-

sectable for bilateral medial canthus invasion, was coded as a local failure

for persistent disease. However, after a favorable partial response, the re-

sidual tumor was amenable to resection with bilateral orbital preservation;

this patient remains disease-free 6 months after salvage surgery. Grade 3+

toxicities occurred in 5 patients, including 1 fatal central nervous system

necrosis.

Conclusion: Hyperfractionated PT and concurrent chemotherapy for

unresectable/incompletely resected sinonasal cancers results in LC rates

superior to historic data with acceptable toxicity. Additional follow-up is

needed to further validate these results.

Abstract 305; Table 1

Outcome

1 year 1 year 2 year 2 year

% At risk % At risk

Local control 75 17 65 11Local-regional control 74 17 64 11Freedom from distant metastases 96 21 96 14Disease-free survival 70 17 61 11Overall survival 91 21 71 14

Author Disclosure: R. Dagan: Research Grant; Elekta. C.M. Bryant:

None. W.M. Mendenhall: None.

306Survival Impact of Local Therapy for Thyroid Carcinoma: ASurveillance, Epidemiology, and End Results AnalysisW. Sumner,1 A. Amini,1 J. McDermott,2 J. Eagles,2 A. Jimeno,2

D.W. Bowles,2 and S. Karam1; 1Department of Radiation Oncology,

University of Colorado Denver, Aurora, CO, 2Department of Medical

Oncology, University of Colorado Denver, Aurora, CO

Purpose/Objective(s): Thyroid carcinoma accounts for 2% of all human

malignancies. Approximately 5% present with distant disease at the time

of diagnosis. While radioactive iodine (RAI) is commonly used in meta-

static cases, the role of local treatment (surgery and/or radiation) and its

effect on survival is largely unknown. This study evaluates the impact of

local therapy (LT) on survival in metastatic thyroid using the Surveillance,

Epidemiology and End Results (SEER) database.

Materials/Methods: The SEER database was queried for patients diag-

nosed with metastatic thyroid carcinoma treated from 2004 to 2012; pa-

tients who died within 1 month of diagnosis were excluded. A comparative

analysis was performed between the presence or absence of LT. LT was

defined as receiving surgery and/or external beam radiation therapy

(EBRT); RAI alone or observation was considered as no LT. Overall

survival (OS) and cancer-specific survival (CSS) were estimated using the

Kaplan-Meier method and use of local therapy was analyzed using Cox

regression.

Results: A total of 1925 patients were included; 409 (21%) received no

local therapy, 949 (49%) surgery alone, 284 (15%) EBRT alone, and 283

(15%) surgery plus EBRT. Median follow-up was 16 months (range, 1-

107). Unadjusted median OS between no LT and LTwas 34 months versus

66 months, respectively (PZ.101). When accounting for age, gender, race,

histology, T and N stage, grade, marital status, residence, and percent

county below poverty, LT improved both OS (hazard ratio [HR], 0.57; 95%

confidence interval [CI], 0.49-0.67; P<.001) and CSS (HR, 0.62; 95% CI,

0.51-0.75; P<.001). On subgroup analysis, LT improved OS for papillary

(HR, 0.50; 95% CI, 0.37-0.67; P<.001), follicular (HR, 0.42; 95% CI,

0.26-0.67; P<.001), and anaplastic histology (HR, 0.50; 95% CI, 0.37-

0.67; P<.001) under MVA; no OS improvement with LT was appreciated

for medullary tumors (HR, 0.80; 95% CI, 0.50-1.29; PZ.361) under MVA.

Further, LT improved OS for high-grade (HR, 0.61; 95% CI, 0.48-0.78;

P<.001) but not low/intermediate grade (HR, 1.09; 95% CI, 0.35-3.42;

PZ.881) disease under MVA.

Conclusion: Local therapy with surgery and/or EBRT improves OS and

CSS for metastatic thyroid carcinoma patients in the SEER database.

Author Disclosure: W. Sumner: None. A. Amini: None. J. McDermott:

None. J. Eagles: None. A. Jimeno: None. D.W. Bowles: None. S. Karam:

None.

307Characterizing Programmed Death Ligand 1 (PD-1L) Expression inMerkel Cell CarcinomaG.J. Hanna,1 H.J. Grote,2 V. Vergara,1 B. Brunkhorst,3 G. Rabinowits,1

M. Thakuria,1 N.R. LeBoeuf,1 J.M. Cuillerot,3 J.A. DeCaprio,1

and J. Lorch1; 1Dana Farber Cancer Institute, Boston, MA, 2Merck

Pharmaceuticals, Darmstadt, Germany, 3EMD Serono, Rockland, MA

Purpose/Objective(s): Merkel cell carcinoma (MCC) is a rare neuro-

endocrine skin neoplasm thought to arise within the basal layer of the

epidermis, associated with Merkel cell polyomavirus (MCPyV). Immu-

nosuppressed patients are at higher risk for MCC, suggesting a critical role

for immune mechanisms. We sought to characterize PD-L1 expression

in MCC.

Materials/Methods: Clinical data and archival tissue from 59 patients with

MCC were gathered and PD-L1 expression was analyzed using immuno-

histochemical staining with a previously characterized antibody. Presence

and distribution of PD-L1 was analyzed using a semi-quantitative H-score.

International Journal of Radiation Oncology � Biology � Physics952

A two-sample Wilcoxon rank-sum (Mann-Whitney) test was used to detect

a median survival difference based on PD-L1 status.

Results: The median age at diagnosis was 70 (45 to 95 years). Twenty-four

(40.7%) patients were female. The extremity was the primary site of dis-

ease in the majority of patients. Median overall survival for the entire

cohort was 49 months (< 1 to 131 months), 34 months among PD-L1

positive patients. Of 67 evaluable tumors, 15 (22.4%) were PD-L1 positive,

with an H-score maximum of 6. Fourteen of those samples had membrane-

associated tumor staining for PD-L1, while all 15 demonstrated cyto-

plasmic PD-L1 staining. 62 (92.5%) samples had PD-L1 positive immune

infiltrates of varying size and density at the infiltrative tumor margin.

There was no median survival difference based on tumor PD-L1 status

(p Z 0.43).

Conclusion: PD-L1 was rarely expressed in tumor cells but was frequently

found in tumor infiltrating immune cells. These data provide a rationale for

immunotherapy targeting PD-L1 in MCC.

Author Disclosure: G.J. Hanna: None. H. Grote: Stock Options; Merck

KGaA. V. Vergara: None. B. Brunkhorst: Stock Options; EMD Serono.

G. Rabinowits: Consultant; Onyx. to DFCI as an institution; Exelixis. M.

Thakuria: None. N.R. LeBoeuf: None. J. Cuillerot: Stock Options; EMD

Serono. J.A. DeCaprio: None. J. Lorch: Research support; Novartis,

Millennium.

308WITHDRAWN

309External Beam Radiation Therapy (EBRT) for pT4-DifferentiatedThyroid Carcinomas: A National Cancer Data Base (NCDB) AnalysisB.S. Gill,1 S. Beriwal,1 U. Duvvuri,2 G. Balasubramani,3 D.E. Heron,1

and D.A. Clump, II1; 1University of Pittsburgh Cancer Institute,

Pittsburgh, PA, 2University of Pittsburgh Medical Center, Pittsburgh, PA,3University of Pittsburgh School of Public Health, Pittsburgh, PA

Purpose/Objective(s): Indications for external beam radiation therapy

(EBRT) in differentiated thyroid carcinomas (DTC) remain poorly defined.

In order to evaluate practice patterns and the impact of EBRT, the National

Cancer Data Base (NCDB) was queried for patients within this cohort at

high risk of local relapse.

Materials/Methods: Patients with pathologic T4 (pT4) DTC diagnosed

from 2004 to 2012 were extracted from the NCDB. Logistic regression was

conducted to determine factors associated with EBRT use with creation of

propensity scores with inverse probability of treatment weighting (IPTW).

Log-rank test and multivariable Cox regression were completed to estab-

lish factors impacting overall survival.

Results: Of 4776 patients, 89.8% underwent total thyroidectomy, and

66.4% received radioactive iodine (RAI). Surgical margin status was

known in 74.9%, of which 59.4%, 35.1%, and 5.5% underwent R0, R1,

and R2 resections, respectively. EBRT use was low at 6.8% and rela-

tively unchanged over time (2004-2006: 7.2% vs 2010-2012: 8.1%,

PZ.152). Patients who were older (>45 years of age: odds ratio [OR]

4.87, P<.001), had Medicare (OR 2.11, PZ.035), with adverse histol-

ogies (papillary tall cell: OR 2.28, P<.001; Hurthle cell: OR 4.70,

P<.001), with tumors >5 cm (OR 3.08, P<.001) or with positive sur-

gical margins (R1: OR 1.55, PZ.003; R2: OR 3.18, P<.001) were more

likely to receive EBRT. Lower EBRT use was seen in patients who

received RAI (OR 0.05, P<.001), underwent total thyroidectomy (OR

0.38, P<.001), live in the West North Central region (OR 0.37, PZ.012),

or are female (OR 0.52, P<.001). With 50.7-month median follow-up, 5-

year overall survival (OS) rates for all patients and patients with R0, R1,

and R2 resection were 80.4% and 85.2%, 80.2%, and 62.2%, respec-

tively (P<.001). The 5-year unadjusted OS rates without and with EBRT

were 82.9% and 43.4% (P<.001). On multivariable Cox regression

analysis, in addition to advanced age, higher comorbidity score, Medi-

care insurance, living >20 miles from a treating facility, tumor size >5

cm, R1-2 resection, and no receipt of RAI, patients receiving EBRT had

an over 2-fold increase in the risk of death (hazard ratio 2.31, P<.001),

which was upheld after IPTW adjustment and use of a 3-month condi-

tional landmark. A persistent detriment was seen in all subsets,

including R1 or R2 resection, older age, various histologies, or node-

positive patients.

Conclusion: Despite selection of patients at high risk of local relapse (pT4)and propensity adjustment, EBRT failed to improve survival, even in high-

risk subsets. Conversely, margin status remains a clear prognostic factor.

Negative findings here may allude to a lack of survival benefit from local

control with EBRT or uncaptured confounders (ie, RAI uptake, DNA

aneuploidy). Suboptimal outcomes in patients with R2 resection suggest

the need for more effective therapies.

Author Disclosure: B.S. Gill: None. S. Beriwal: None. U. Duvvuri: None.

G. Balasubramani: None. D.E. Heron: None. D.A. Clump: None.

310Radiation Therapy Regimens in Patients With Nonmelanoma Headand Neck Skin CancersY. Dundar, R. Cannon, M. Monroe, J. Hunt, G. Suneja, and Y.J. Hitchcock;

Huntsman Cancer Hospital, University of Utah, Salt Lake City, UT

Purpose/Objective(s): Radiation therapy (RT) has been used for treat-

ment of skin cancer either in the definitive or adjuvant setting. There are

limited data to compare different RT regimens, particular for very elderly

patients. Our goal is to assess the effectiveness and outcomes of different

RT regimens for nonmelanoma head and neck skin cancers.

Materials/Methods: We conducted a retrospective review of patients un-

dergoing definitive or adjuvant radiation for nonmelanoma head and neck

skin cancers at a single institution between 2004 and 2014. Patients were

categorized into 4 subgroups according to RT dose per fraction (�200

cGy/fraction, 240-250 cGy/fraction, 300-400 cGy/fraction, and 500-600

cGy/fraction). Demographics and tumor characteristics, and disease local-

regional control rates, were analyzed.

Results: There were 107 patients with 174 disease sites. We assessed the

results for both treated sites and patient based. Of those, 77.5% were

squamous cell carcinoma, 13.1% were basal cell carcinoma, and 9.3%

were other nonmelanoma subtypes. The mean age at diagnosis was 72.2

years old, 82.2% of patients were male, and mean follow-up was 23.2

months. Preauricular (18.7%) was the most common location, followed

by temple (13.1%), scalp (12.1%), cheek (11.2%), and forehead

(11.2%). Fourteen patients underwent definitive RT, 2 underwent palli-

ative RT, and 91 underwent adjuvant RT. The overall local and regional

control rates were 87.8 % and 89.7 %, respectively by patients, and 91.3

% and 86.2%, respectively by primary treatment sites. Age, primary

tumor location, T stage, N stage, group stage, perineural invasion, skull

base invasion, and RT subgroup were significantly associated with dis-

ease-specific and overall survival (OS) (P<.05). Histology and gender

were not significantly associated with survivals. Local and regional

control rates are not statistically different among the RT dose subgroups;

however, disease-free survival and OS varied among the subgroups when

analyzing by patient and by primary sites treated (P<.05). The mean

survival was longer in patients treated with 240 to 250 cGy per fraction

(78.4 months) and �200 cGy per fraction (62.4 months) and was shorter

in patients treated with 300 to 400 cGy per fraction (18.9 months), which

could associate with patient age or comorbidities. The mean survival

was 47.3 months in patients treated with 500 to 600 cGy per fraction.

There was no statistically significant difference for RT toxicity in the

subgroups.

Conclusion: Definitive or postoperative short-term RT regimens (with high

dose per fraction) for patients with locally or regionally advanced head and

neck nonmelanoma skin cancers appear feasible and effective with no

reportable severe toxicities, particularly in elderly patients or those who

cannot complete the typical standard regimen.

Author Disclosure: Y. Dundar: None. R. Cannon: None. M. Monroe:

None. J. Hunt: None. G. Suneja: None. Y.J. Hitchcock: None.

Volume 94 � Number 4 � 2016 Posters 953

311Predictors of Survival in Large Cell Undifferentiated Carcinoma ofthe Major Salivary GlandsT.P. Schrank, K. Zhan, and E.J. Lentsch; Medical University of South

Carolina, Charleston, SC

Purpose/Objective(s): To our knowledge, presented herein is an analysis

of the largest case series of large cell undifferentiated carcinoma of the

major salivary glands available in the literature. The great histological

diversity of salivary gland cancers poses a constant clinical challenge and

an opportunity to improve therapy. Among the rare salivary malignancies

are histologically undifferentiated types, which include lymphoepithelial

carcinoma as well as large cell undifferentiated carcinoma (LCUC). LCUC

has been associated with a grave prognosis, but little is known about the

disease in terms of the demographics, tumor factors at presentation, suc-

cess of treatment in terms of modality, or predictors of survival. Our

objective, therefore, is to better define the above issues with the purpose of

providing a platform for improved treatment and further investigation.

Materials/Methods: Data from the National Cancer Data Base including

cases diagnosed from 1998 to 2012 were analyzed. Two hundred forty-

seven records of LCUC were identified. Extracted variables include but

were not limited to age, gender, Charlson score, surgical margin status,

tumor size, type of surgery, grade, use of radiation therapy, use of

chemotherapy, extraglandular spread, clinical and pathologic stage, and

overall survival. Survival data were analyzed using the Kaplan-Meier

method, log-rank test, and multivariate COX regression where appropriate.

Results: LCUC comprised 0.68% of all major salivary gland cancers in the

data set, with 89.5% of these originating in the parotid gland. A total of

69.8% of patients presented with advanced-stage disease. The incidence of

occult nodal disease was 39.1%. A combination of surgery and radiation

was the most common treatment provided. Five- and 10-year overall

survival rates were 36% and 21%, respectively. On multivariate analysis,

treatment modality did not predict survival. However, Charlson score,

distant metastasis, and positive surgical margins were found to be inde-

pendent predictors of overall survival.

Conclusion: LCUC is a rare salivary malignancy with a poor prognosis.

This analysis of a large cancer database reveals that occult nodal disease is

common, and negative margins on resection are associated with improved

survival. Therefore, our data support complete surgical excision as well as

consideration of elective neck dissection. Although commonly used,

further research is needed to define the role of adjuvant therapy.

Author Disclosure: T.P. Schrank: None. K. Zhan: None. E.J. Lentsch:

None.

312Local Therapy for Metastatic Salivary Adenoid Cystic Carcinoma: ASurveillance, Epidemiology, and End Results and National CancerData Base AnalysisJ. McDermott,1 A. Amini,2 S. Karam,2 and D.W. Bowles1; 1Department of

Medical Oncology, University of Colorado Denver, Aurora, CO,2Department of Radiation Oncology, University of Colorado Denver,

Aurora, CO

Purpose/Objective(s): Approximately 10% of salivary gland adenoid

cystic carcinomas (ACC) present with metastatic disease. Systemic therapy

has only shown modest activity and the role of locoregional therapy (LT)

in this setting is unclear. This study evaluates overall survival (OS) out-

comes for patients with metastatic salivary ACC who received LT (surgery

alone, radiation therapy [RT] alone, or surgery plus RT) versus no

locoregional therapy (NLT).

Materials/Methods: The Surveillance, Epidemiology, and End Results

database (SEER) and the National Cancer Data Base (NCDB) were

queried for patients with T0-4, N0-3, or M1 salivary ACC. Patients were

diagnosed between 1973 and 2011 for the SEER and between 1998 and

2011 for the NCDB. Demographic and treatment information was obtained

on each patient, and analysis was based on the presence, absence, and type

of LT. The Kaplan-Meier method was used to estimate OS. Cox

proportional hazard models were used for univariate and multivariate an-

alyses (MVA).

Results: Two hundred seventy-seven patients from the SEER and 158

patients from the NCDB were identified. From the SEER, 221 underwent

LT (70 RT, 66 surgery, 85 both surgery and RT), and 56 had NLT. Patients

who received any LT had nonsignificant increased median OS (30 months)

compared to NLT (16 months) (hazard ratio [HR] 0.71, 95% CI 0.50-1.02,

PZ.063). When subdividing by treatment type, the median OS difference

became significant with NLT at 16 months, RT at 17 months (HR 1.21,

95% CI 0.81-1.84), surgery at 50 months (HR 0.58, 95% CI 0.38-0.89),

and surgery plus RT at 50 months (HR 0.56, 95% CI 0.37-0.84) (P<.001).

These differences persisted on landmark analyses excluding patients that

survived <6 and <12 months and MVA. Other factors impacting OS

included age <70 years, primary tumor site, geographic location, year of

diagnosis, and N stage. From the NCDB, 105 patients underwent LT (10

RT, 63 surgery, 32 surgery and RT) and 53 had NLT. Patients receiving LT

had a significantly longer median OS (26.2 months) compared to NLT

(14.5 months) (HR 0.56, 95% CI 0.39-0.82, PZ.003). The differences

between treatment subtypes were also significant with NLT at 14.5 months,

RT at 23.8 months (HR 0.92, 95% CI 0.42-1.79), surgery at 20.7 months

(HR 0.53, 95% CI 0.35-0.81), and surgery plus RT at 28.8 months (HR

0.52, 95% CI 0.31-0.86) (PZ.013). These survival differences persisted

with MVA. Landmark analysis excluding patients surviving <6 months

was similar, but analysis excluding patients surviving <12 months only

showed improved survival of patients receiving surgery alone. Other fac-

tors impacting OS included T and N stage and receipt of chemotherapy.

Conclusion: Locoregional therapy, specifically involving surgery, is asso-

ciated with longer OS in metastatic salivary ACC based on analysis of the

SEER and NCDB patient populations.

Author Disclosure: J. McDermott: None. A. Amini: None. S. Karam:

None. D.W. Bowles: None.

313Expression of Focal Adhesion Kinase Molecule byImmunohistochemical Localization in Odontogenic Epithelium ofAmeloblastoma and Dental FollicleS.M. Patil; Dr. D Y Patil Dental College and Hospital, Pimpri, India

Purpose/Objective(s): To study the expression of FAK (focal adesion

kinase) in odontogenic epithelium of ameloblastoma and dental follicle.

Ameloblastoma is a benign, solid, locally invasive, highly destructive

tumor of the jaws, with a high recurrence rate even following radical

surgery. This local biological behavior of solid multicystic ameloblastoma

is an argument for including it in the group of low-grade malignant tumors.

The invasion of surrounding healthy tissues by tumor cells is one of the

essential steps in tumor progression. But the exact molecular mechanism

of invasion in ameloblastoma has not been well elucidated. Thus, identi-

fication of prognostic markers for the biologic behavior of ameloblastoma

is of considerable importance to determine the most appropriate thera-

peutic approach and establish the prognosis of patients. Extensive studies

have indicated that focal adhesion kinase (FAK) is an important mediator

of cell invasion and migration in various malignancies.

Materials/Methods: Thirty-four cases of ameloblastoma and 17 cases of

dental follicles histopathologically diagnosed were retrieved from the

archival specimens for the study of immunohistochemical localization of

FAK in tumor cells.

Results: All the 34 cases of ameloblastoma showed expression of FAK in

the peripheral and central cells of follicles. Strong expression of FAK in

the peripheral cells was observed in 34 cases of ameloblastoma, while

central cells showed negative expression in 17 cases followed by strong

expression (12) and weak expression (05). All the cases of dental follicle

showed negative expression of FAK. FAK expression in the peripheral and

central cells of ameloblastoma was significant (P<.0001). Comparison of

expression of FAK in ameloblastoma (central cells) and dental follicle

were significant as well (PZ.0017). The comparison of FAK expression in

central cells of multicystic and unicystic ameloblastoma (PZ.7133) and

International Journal of Radiation Oncology � Biology � Physics954

comparison of FAK expression in central cells of recurrent and nonre-

current cases of ameloblastoma (PZ.3692) was found to be insignificant.

Conclusion: FAK expression in odontogenic epithelium of ameloblastoma

suggests its role in the locally aggressive behavior. FAK could be

responsible for cell invasion and migration in ameloblastoma.

Author Disclosure: S.M. Patil: None.

314Early T Stage Salivary Duct Carcinoma: Outcomes and ImplicationsA. Sharma,1 N.C. Schmitt,2 M.R. Gilbert,3 and S. Kim4; 1Division of

Otolaryngology - Head and Neck Surgery, Southern Illinois University,

Springfield, IL, 2Department of Otolaryngology - Head and Neck Surgery,

JohnsHopkinsUniversity, Baltimore,MD, 3Department of Otolaryngology -

Head and Neck Surgery, University of Pittsburgh Medical Center,

Pittsburgh, PA, 4University of Pittsburgh Medical Center, Pittsburgh, PA

Purpose/Objective(s): Salivary duct carcinoma (SDC) is a rare salivary

malignancy that often presents at advanced stage. However, some patients

are diagnosed early; the objective of this study is to describe the clinico-

pathologic presentation and survival outcomes of early T stage (Tis, T1,

and T2) SDC.

Materials/Methods: A single institution’s head and neck pathology data-

base was reviewed for all cases of histopathologically diagnosed early T

stage SDC between January 1995 and October 2014. We reviewed the

electronic medical record to determine patient demographics, tumor data,

treatment, and outcome, including disease-free and overall survival (DFS

and OS, respectively).

Results: Twenty-eight patients with early T stage SDC were identified.

Most tumors (75%) were in the parotid gland. All patients underwent

surgical resection, and 64% underwent neck dissection as well. Most pa-

tients (71%) received adjuvant treatment. Perineural, vascular, and

extracapsular invasion were associated with N stage. Median DFS and OS

were 3.2 and 4.7 years, respectively. Parotid primary site, vascular inva-

sion, extracapsular invasion, and facial nerve sacrifice were associated with

worse survival. Adjuvant therapy was not associated with survival.

Conclusion: Early T stage SDC is an aggressive malignancy with median

DFS of 3.2 years and median OS of 4.7 years. Predictors of survival

include primary site, vascular invasion, facial nerve sacrifice, and extrac-

apsular invasion. Despite low T stage and aggressive treatment, prognosis

is poor for these patients.

Author Disclosure: A. Sharma: None. N.C. Schmitt: None. M.R.

Gilbert: None. S. Kim: None.

315Regionally Metastatic Cutaneous Squamous Cell Carcinoma of theHead and Neck: Survival and High-Risk FeaturesM. Amoils,1 C.S. Lee,2 J. Sunwoo,2 S.Z. Aasi,2 W. Hara,3 J. Kim,2

D. Sirjani,2 A.D. Colevas,3 A.L.S. Chang,2 and V. Divi3; 1Stanford

Otolaryngology, Stanford, CA, 2Stanford, Stanford, CA, 3Stanford

University, Stanford, CA

Purpose/Objective(s): To describe treatment outcomes and clinicopath-

ologic correlates for patients treated for regionally metastatic cutaneous

squamous cell carcinoma of the head and neck (cHNSCC).

Materials/Methods: The study design was a retrospective chart review.

CPT and ICD-9 codes were used to identify patients treated for regionally

metastatic cHNSCC at our institution. Charts were reviewed to ensure

appropriate inclusion and extract demographics, clinicopathologic infor-

mation, and survival outcomes. The effect of various clinicopathologic

variables on overall survival was investigated with a Cox analysis. Kaplan-

Meier survival curves were constructed and compared with the log-rank

test.

Results: Eighty patients were included with a mean age of 73 years

(standard deviation 11.5). The majority presented with recurrent disease

(66%). The most common high-risk tumor features were perineural inva-

sion (41%) and extracapsular spread (ECS) (31%). Patients were treated

with surgery alone (16%), surgery + XRT (45%), and surgery + cXRT

(39%). On multivariate regression the only variables found to be signifi-

cantly associated with overall survival were cutaneous primary >2 cm

(PZ.03) and ECS (PZ.01). Location of regional metastasis (neck vs

parotid vs both) had no effect on overall survival (PZ.2). Survival was

also unaffected by the presence of a cutaneous primary at the time of

presentation (PZ.9). Overall survival at 5 years was 59% with a 51% rate

of recurrence.

Conclusion: Regionally metastatic cHNSCC is an uncommon but

aggressive disease associated with high recurrence rates. Patients with

tumors >2 cm and ECS have poorer overall survival despite adjuvant

therapy.

Author Disclosure: M. Amoils: None. C.S. Lee: Junior Board member;

Western Orthopedic Association. J. Sunwoo: None. S.Z. Aasi: None. W.

Hara: None. J. Kim: None. D. Sirjani: None. A.D. Colevas: Committee

member; AJCC. Panel member; NCCN. Board Member; ANCO. A.L.

Chang: None. V. Divi: None.

316Outcomes and Patterns of Failure for Sinonasal UndifferentiatedCarcinoma (SNUC): The Mayo Clinic ExperienceM.E. Gamez,1 D. Lal,2 M.Y. Halyard,1 W.W. Wong,2 C. Vargas,3

K. Curtis,1 S. Ko,4 R.L. Foote,5 and S.H. Patel1; 1Mayo Clinic, Scottsdale,

AZ, 2Mayo Clinic, Phoenix, AZ, 3Mayo AZ, Phoenix, AZ, 4Mayo Clinic,

Jacksonville, FL, 5Mayo Clinic, Rochester, MN

Purpose/Objective(s): Sinonasal undifferentiated carcinoma (SNUC) is a

rare aggressive disease arising in the nasal cavity and paranasal sinuses

with often dismal outcomes and unclear optimal management. This study

was performed to determine the outcomes and patterns of failure for SNUC

at the Mayo Clinic.

Materials/Methods: We identified for the present institutional review

boardeapproved analysis a subset of 40 patients (pts) treated at the Mayo

Clinic from 1990 to 2014. The median age at presentation was 56.7 years

(range, 29-82). Twenty-four pts (60%) were male. The primary site was

nasal cavity in 20 pts (50%) and ethmoid sinus in 10 pts (25%). The

majority of the pts, 32 (80%), presented with T4 disease. Twenty-four pts

(60%) received trimodality therapy in the form of chemotherapy, radiation

therapy (RT), and surgery. Most pts (60%) were treated with intensity

modulated RT, and 16 pts (40%) received total doses �60 Gy. The most

frequent chemotherapy regimen utilized for the trimodality approach was

cisplatin and etoposide in 11 pts (27.5%).

Results: The median follow-up for surviving pts was 6.9 years (range, 1.1-

17). A total of 16 pts (40%) experienced recurrent disease, 5 local re-

currences (12.5%), 1 regional recurrence (2.5%), and 10 (25%) distant

metastasis. The 5-year overall survival (OS), recurrence-free survival

(RFS), and locoregional control (LRC) were 44%, 39%, and 71%,

respectively. With respect to treatment modality, pts treated with multi-

modality therapy had improved OS, RFS, and LRC compared to pts treated

with single modality therapy. Improved OS was noted in patients receiving

intensity modulated RT and also in patients receiving �60 Gy. The most

common cause of death was distant metastasis. Acute radiation toxicity

was radiation dermatitis grade 1-2 (G1-2) (85%), mucositis G1-2 (60%),

and fatigue (50%). Long-term toxicity included nasal dryness (30%),

xerostomia (10%), retinopathy and optic neuropathy in (5%), and radio-

necrosis (2.5%).

Conclusion: SNUC is an aggressive malignancy that frequently presents at

a locally advanced stage with a high tendency to metastasize. The best

outcomes in our series were obtained with a multimodality approach.

Modern RT techniques and doses �60 Gy were associated with improved

OS. Proton therapy may potentially further improve outcomes by allowing

dose escalation with better sparing of normal organs. Optimal sequencing

of multimodality therapy still needs to be defined.

Author Disclosure: M.E. Gamez: None. D. Lal: None. M.Y. Halyard:

None. W.W. Wong: None. C. Vargas: None. K. Curtis: None. S. Ko:

None. R.L. Foote: None. S.H. Patel: None.

Volume 94 � Number 4 � 2016 Posters 955

317Facilitating Anaplastic Specialized Treatments (FAST)dAMultidisciplinary Anaplastic Thyroid Cancer (ATC) TeamM.E. Cabanillas, G.B. Gunn, M.D. Williams, N.L. Busaidy, W.N. William,

Jr, C. Lu, and S.Y. Lai; The University of Texas MD Anderson Cancer

Center, Houston, TX

Purpose/Objective(s): ATC is an exceedingly rare disease (w800 cases/

year in the US) that is rapidly progressive with a median overall survival of

5 months. Only 20% of patients (pts) are alive at 1 year after diagnosis.

Recently, there has been promising research with novel therapies. How-

ever, due to the rarity and poor prognosis, no trials have completed

enrollment. We sought to streamline the process for access into our

institution where several trials are available for ATC pts. We completed a

quality improvement project (QIp) to reduce time from referral to dispo-

sition (time pt is given appointment) and referral to appointment, with the

endpoint of facilitating rapid entry into the institution.

Materials/Methods: We completed and implemented this QIp, called

FAST, by August 2014. The FAST team members worked with the business

center (BC) to streamline access. We gave the BC the common synonyms

for ATC and asked that all pts be processed immediately with the same

standard imaging and labs. The ATC physicians reserved several appoint-

ment slots for these pts so that the BC could offer appointments sooner.

From September 1, 2014, to August 31, 2015, we collected data regarding

time from referral to disposition and time from referral to appointment. Our

historical data showed that the mean referral to disposition time was 8.7

business days and referral to appointment was 11.5 days.

Results: During the study period, 31 pts were referred to our institution fora diagnosis of ATC. Twenty-six of these were put into the FAST pathway.

The mean referral to disposition time was 0.5 business days (94% decrease

compared to historical data). Mean referral to appointment time was 8.7

business days (24% decrease compared to historical). Table 1 shows the

breakdown for pts referred and diagnosed with ATC during the study

period. The total number of ATC referrals was 28 (with 2 patients pending

pathologic confirmation). This represents an 86% increase in ATC pts

when compared to 2012 (nZ15 in 2012).

Conclusion: Since the implementation of the FAST program, the access

time has decreased and the number of referrals for ATC has increased

significantly. Establishment of similar fast track programs for aggressive

thyroid cancers at major referral centers could improve accrual to clinical

trials in ATC and related diseases, leading to improved survival as well as

better understanding of the biology of these diseases.

Author Disclosure: M.E. Cabanillas: None. G.B. Gunn: None. M.D.

Williams: None. N.L. Busaidy: None. W.N. William Jr.: None. C. Lu:

None. S.Y. Lai: None.

Abstract 317; Table 1

Referred for ATC 31Confirmed with ATC 23Unknown if ATC* 3Final diagnosis was not ATC 5

Other referral diagnosis**, but confirmed as ATC at our institution 5Total number of new ATC patients seen 28

*2 are pending confirmation^2 had thyroid lymphoma, and 3 were

poorly differentiated or papillary thyroid cancer **2 with thyroid mass;

1 each: thyroid cancer, poorly differentiated thyroid cancer, and

mucoepidermoid carcinoma

318Surgical Management of Primary Carcinoma of Eyelids andPeriorbital Skin: Review of 76 patients.A.H. Mebed and I.S. Fayek; National Cancer Institute - Cairo University,

Cairo, Egypt

Purpose/Objective(s): To review the clinical and pathologic features,

treatment, and outcomes of primary carcinoma of eyelids and periorbital

skin.

Materials/Methods: In this prospective study, 76 patients with primary

carcinomas of the eyelids and the periorbital skin were treated between

January 2009 and December 2014. We conducted an analysis of epide-

miologic data and clinicopathologic criteria of all lesions. General anes-

thesia was used for all patients, and frozen section examinations were done

only in eyelids and inner canthus lesions; the cutaneous margin including

the resected parts of the lid margins and the deep soft tissue margin were

confirmed negative intraoperatively. The follow-up period ranged from 8

months to 5 years. We recorded and analyzed the surgical complications

and their management, the functional and cosmetic results, and the

recurrence rate. Exclusion criteria included recurrent cases and those with

familial cancer syndromes.

Results: Male to female ratio was 1.7:1. Age of the patients ranged from

36 to 81 years with a mean of 66 years. BCC represented 93.5% of the

lesions (71 patients), SCC represented 5% of the lesions (4 patients), and

meibomian gland carcinoma of the upper eyelid occurred in 1 patient

(1.5%). The most common clinical variant of BCC was the nodular type,

and maximum diameter of the lesions ranged from 4 mm to 24 mm. Inner

canthus was the most common location for BCC, followed by the outer

canthus. SCC occurred only in the lids. Two of them were treated by

orbital exentration, one by wide local excision and median glabellar flap,

and the fourth by excision and primary closure. The patient with meibo-

mian gland carcinoma was treated by subtotal upper lid excision and lid

switch flap for lid reconstruction in addition to cervico-facial lymphade-

nectomy. All safety margins were confirmed negative in the paraffin sec-

tions. The most common method of repair was primary closure done in 35

patients (47%) followed by paramedian glabellar flap in 33 patients (43%).

The rate of postoperative complications was 23.5% (18 patients), and the

recurrence rate was 1.3% (1 patient only).

Conclusion: Primary cutaneous carcinoma of the periorbital region is a

curable disease, and most of the patients present early in the disease

course. Negative margins are easily obtained with conventional frozen

section techniques, but local recurrences still can occur. Functional com-

plications are inevitable. Their correction is an integral part of surgical

treatment.

Author Disclosure: A.H. Mebed: None. I.S. Fayek: None.

319Primary Salivary Gland Malignancies Treated With AdjuvantRadiation Therapy With or Without Concurrent Chemotherapy:Prognostic Factors and Treatment OutcomesB.J. Gebhardt,1 J. Ohr,1 J.E. Bauman,2 R.L. Ferris,3 U. Duvvuri,3 S. Kim,3

J.T. Johnson,3 D.E. Heron,1 and D.A. Clump, II1; 1University of Pittsburgh

Cancer Institute, Pittsburgh, PA, 2University of Pittsburgh Cancer

Institute, Pittsburgh, PA, United States, 3University of Pittsburgh Medical

Center, Pittsburgh, PA

Purpose/Objective(s): Radiation therapy (RT) is indicated for patients

with salivary gland malignancies with risk factors for recurrence following

surgical resection. We analyzed patients treated with resection and adju-

vant RT in a large, integrated National Cancer Institute-Designated

Comprehensive Cancer Center Network to determine the impact of prog-

nostic and treatment factors.

Materials/Methods: A retrospective analysis was performed of 132 pa-

tients treated with resection followed by intensity modulated RT (IMRT).

The median RT dose to the primary was 66 Gy (range 45-70.2 Gy), and

80.3% of patients received nodal RT to a median dose of 50 Gy (range 0-

54 Gy). Thirty-one (23.7%) patients were treated with concurrent

chemotherapy (ChemoRT) including cisplatin (6.1%), carboplatin and

paclitaxel (11.4%), or other regimens (6.1%). Outcomes included locore-

gional recurrence-free survival (LRRFS), defined as time from surgery to

locoregional recurrence or death, and progression-free survival (PFS),

defined as time to any recurrence or death. The Kaplan-Meier method was

used to estimate rates of LRRFS and PFS, and association with patient and

treatment factors were assessed with the log-rank test. Multivariable Cox

regression analysis was performed to evaluate factors associated with

LRRFS and PFS.

International Journal of Radiation Oncology � Biology � Physics956

Results: Median follow-up for the entire cohort was 35 months. The

overall 3-year rates of LRRFS and PFS were 87% and 81.8%, respectively.

Factors significantly associated with worsening LRRFS and PFS on uni-

variate analysis were high T stage, lymph node involvement, salivary

ductal carcinoma (SDC) histology, positive surgical margins (SM), extra-

nodal extension (ENE), and use of ChemoRT. Lymphovascular space in-

vasion was associated with decreased LRRFS, and high grade and

perineural invasion were associated with decreased PFS. Multivariable

analyses revealed that predictors of decreased LRRFS were SDC histology

(hazard ratio [HR] 4.57, 95% confidence interval [CI] 2.00-10.44) and

positive SM (HR 2.29, 95% CI 1.01-5.18). ENE significantly predicted

lower PFS (HR 3.81, 95% CI 1.90-7.66). Patients treated with ChemoRT

had significantly higher rates of SDC histology (PZ.047) and ENE

(PZ.009). The overall acute toxicity rates were 30.3% grade 1, 51.5%

grade 2, 11.4% grade 3, and 0.8% grade 4. There was no difference in rates

of grade 3+ acute toxicity with the use of RT alone versus ChemoRT

(PZ0.183). No late grade 3+ toxicities were reported.

Conclusion: This large retrospective analysis of patients treated with

adjuvant IMRT demonstrated LRRFS and PFS rates consistent with pre-

viously published rates. The use of ChemoRT was well tolerated. While

associated with decreased LRRFS or PFS on univariate analysis, Che-

moRTwas no longer significant in multivariable analysis, which was likely

due to significantly higher rates of adverse pathologic risk factors in this

group, resulting in selection of these patients for treatment intensification.

Author Disclosure: B.J. Gebhardt: None. J. Ohr: None. J.E. Bauman:

None. R.L. Ferris: None. U. Duvvuri: None. S. Kim: None. J.T. John-

son: None. D.E. Heron: None. D.A. Clump: None.

320Postoperative Radiation Therapy for Salivary Gland MalignanciesM.M. Mathew and M. Choi; Loyola University Medical Center, Maywood,

IL

Purpose/Objective(s): This study sought to review a single-institution

experience with the management of patients with salivary gland malig-

nancies treated with surgery and postoperative radiation therapy (RT),

focusing on outcome and patterns of failure.

Materials/Methods: We conducted a retrospective review of patients with

salivary gland malignancies who were treated with postoperative RT at our

institution from 2004 to 2014. All patients underwent surgery and post-

operative RT. RT was delivered using 3-dimensional conformal RT

(3DCRT) or intensity modulated RT (IMRT). The indications for post-

operative radiation included close or positive margins, lymph node spread,

extranodal extension, perineural invasion, and T4 disease. Locoregional

failure-free survival (LRFFS) and overall survival (OS) were calculated

using the Kaplan-Meier method.

Results: A total of 54 patients were analyzed. Median age was 60 years

(range, 13 to 84 years). Thirty-five patients (65%) were men. Common

histologic types included acinic cell carcinoma, adenocarcinoma, adenoid

cystic carcinoma, and squamous cell carcinoma. Eighty-five percent had a

tumor in the parotid gland, 13% in the submandibular gland, and the

remaining 2% involving the minor salivary glands. Twenty-six percent had

T1 disease, 30% T2 disease, 13% T3 disease, and 28% T4 disease. Thirty-

five percent had node-positive disease, 46% had perineural invasion, 15%

had extra nodal invasion, and 67% had close/positive margins. Fifty-nine

percent were treated with 3DCRTwhile 37% received IMRT. Median total

fractionated RT dose was 60 Gy (range, 48 to 66.6 Gy). Two patients were

treated with intraoperative RT (7.5 to 8 Gy). Twenty-two percent received

concurrent chemotherapy. With a median follow-up time of 30 months

(range, 0 to 194 months), 2-year rates were 76% for LRFFS and 81% for

OS. Locoregional failure was observed in 10 patients (18%) and distant

failure in 6 patients (11%). Local failure was most commonly observed at

skull base and skin and neck nodes.

Conclusion: Surgery and postoperative RT in salivary gland malignancies

at our institution resulted in local control comparable with other reported

series. Based on this review, we recommend postoperative RT for patients

with resected salivary gland tumors with adverse features, as achieving

locoregional control in these patients is critical. Further follow-up is

needed to determine long-term outcomes.

Author Disclosure: M.M. Mathew: None. M. Choi: None.

321Treatment Outcomes of Advanced Sinonasal Adenoid CysticCarcinomaE.D. Miller,1 D. Blakaj,2 S.A. Walston,1 A.D. Bhatt,2 V.M. Diavolitsis,2

and J.C. Grecula2; 1The Ohio State University Wexner Medical Center,

Columbus, OH, 2The James Cancer Hospital and Solove Research

Institute, Wexner Medical Center at The Ohio State University, Department

of Radiation Oncology, Columbus, OH

Purpose/Objective(s): Sinonasal adenoid cystic carcinoma (SNACC) is a

rare cancer that typically presents with nonspecific symptoms and at an

advanced stage. The purpose of this study is to evaluate the treatment

outcomes and complications of patients treated for advanced SNACC at

our institution.

Materials/Methods: The medical records of 14 patients with SNACC

treated between 1994 and 2012 were reviewed. Thirteen patients (93%)

had advanced disease (T3, T4) at diagnosis and 1 patient had metastatic

disease. The maxillary sinus (50%) was the most common primary tumor

site. The most common presenting symptoms were pain (46%) and eye

symptoms (38%). All patients received primary surgery and 11 patients

(79%) received postoperative radiation therapy (PORT) with 1 patient

receiving concurrent chemotherapy. PORT doses ranged from 40 to 68 Gy

(median 59.7 Gy). The median follow-up time for all patients and for

living patients was 37.5 months and 37 months, respectively.

Results: The 3- and 5-year actuarial survival outcomes were 70% and

40%, and the 3- and 5-year progression-free survival outcomes were 54%

and 43% for all patients, respectively. Seven patients (50%) failed primary

treatment with 3 local failures (21%), 2 distant failures (14%), and 2

locoregional and distant failures (14%). The median time to progression

was 26 months (range, 3-61 months). In the patients who developed

recurrent disease, the majority had T4 disease (85%) and positive surgical

margins (85%). All of the patients who developed local recurrence

received a PORT dose of less than 60 Gy. Three of the patients who

developed a first recurrence had a second recurrence with a median of 25

months (range, 1-30 months) after the initial recurrence. Of the 7 patients

who failed primary treatment, 1 patient (7%) was alive at last follow-up.

Six patients (43%) have been disease-free since primary treatment and

were alive at last follow-up. In this group, 4 of the patients had T4 disease

(67%), and 3 of the patients (50%) had positive surgical margins. All of the

patients who have been disease-free since primary treatment received a

PORT dose of 60 Gy or more, although this did not reach statistical sig-

nificance (PZ.104). Three patients (21%) have long-term complications

from treatment including a nasocranial fistula, osteoradionecrosis of the

frontal bone requiring hyperbaric oxygen therapy, and nasolacrimal duct

obstruction.

Conclusion: Surgery and PORT is the current standard for treatment of

SNACC. Patients who have negative surgical margins and receive a PORT

dose of 60 Gy or more had a more favorable prognosis than patients who

received a PORT dose of less than 60 Gy. Long-term complications were

observed in 21% of patients following treatment. We are currently eval-

uating p16/human papillomavirus status and prognostic significance in our

SNACC patients, as human papillomavirus has been shown to be associ-

ated with sinonasal tract tumors.

Author Disclosure: E.D. Miller: None. D. Blakaj: None. S.A. Walston:

None. A.D. Bhatt: None. V.M. Diavolitsis: None. J.C. Grecula: None.

322Angiotensin Blockade Reduces Quantitative and QualitativeRadiation-Induced Neck Fibrosis: Interim Analysis of ProspectiveTrialD. Okwan-Duodu, X. Yang, D.S. Yu, K. Patel, J. Switchenko, J.J. Beitler,

W.J. Curran, Jr, and T. Liu; Winship Cancer Institute, Emory University,

Atlanta, GA

Volume 94 � Number 4 � 2016 Posters 957

Purpose/Objective(s): Growing evidence suggests that the renin angio-

tensin system promotes proinflammatory and profibrotic processes. The

aim of our study was to investigate the influence of angiotensin blocker

(ACEi) use on patient, physician, and quantitative ultrasound evaluation of

radiation-induced neck fibrosis, a common sequala of radiation therapy

(RT) to the head and neck.

Materials/Methods: Eighty patients who had received RT for head and

neck cancers were enrolled in an institutional review boardeapproved

study of radiation-induced fibrosis. The median follow-up time was 20

months (range: 10-22 months). All participants had patient-reported scores

determined for neck stiffness, pain, and induration. Forty participants had

physician-based assessments of neck fibrosis performed according to the

Radiation Therapy Oncology Group (RTOG) late morbidity scoring

scheme. Forty patients had received ultrasound scans of the neck, and

Nakagami probability density function (PDF) was calculated to quantify

the severity of neck fibrosis. Patients with collagen vascular disease were

excluded.

Results: Twenty-two of 80 patients (27.5%) were on ACEi. For patient-

reported outcomes, patients in the ACEi group showed significantly

reduced neck stiffness compared to the no ACEi group. Using the

Cochran-Armitage trend test, patients in the ACEi group were more likely

to demonstrate low-grade or no fibrosis, whereas those in the no ACEi

group were more likely to develop higher grade fibrosis (91% vs 70%

grade 0, 9% vs 30% grade 1-3, ACEi vs no ACEi, respectively, P<.01). For

physician-reported outcomes, paients in the ACEi group had 62.5% grade

0, 37.5% grade 1, and no grade 2 fibrosis, compared to 40% grade 0, 30%

grade 1, and 30% grade 2 fibrosis in the no ACEi group (P<.09). On ul-

trasound, there was a significant difference in Nakagami PDF (ACEi 5.38 x

10-5 vs no ACEi 4.59 x 10-5, PZ.03, t test), respectively representing a

21.8% versus 33% decrease compared to control patients.

Conclusion: ACEi use was associated with reduced patient-reported,

physician-scored, and quantitative Nakagami PDF score of radiation-

induced neck fibrosis. These promising data suggest potential use of ACEi

for fibrosis prophylaxis in at-risk head and neck cancer patients.

Author Disclosure: D. Okwan-Duodu: None. X. Yang: None. D.S. Yu:

None. K. Patel: None. J. Switchenko: None. J.J. Beitler: None. W.J.

Curran: None. T. Liu: None.

323Cost-Coping Strategies and Perceived Social Isolation in LocallyAdvanced Head and Neck CancerS. Kung, J. O’Connor, B.J. Yap, and J.A. de Souza; University of Chicago,

Chicago, IL

Purpose/Objective(s): Locally advanced head and neck cancer (LAHNC)

has high morbidity and is expensive to treat, leading patients to adopt cost-

coping strategies as a result of their cancer and care. The use of these

strategies has been measured as a proxy for financial burden. Also,

perceived social isolation, defined as the lack of social support and

increased loneliness, has been previously shown to be a barrier to care in

other cancers. Little is known about the cost-coping strategies used by

LAHNC patients, as well as the effects of perceived social isolation on

health care use in LAHNC.

Materials/Methods: Patients with treatment-naive LAHNC diagnosed

from May 2013 to Nov 2014 at a single institution participated in a pro-

spective longitudinal study over 6 months. Monthly surveys assessed

lifestyle-altering cost-coping strategies, out-of-pocket costs, loss of pro-

ductivity, medication compliance, and baseline perceived social isolation.

Sociodemographics, wealth, household income, tumor type, and health

care use data (inpatient length of stay [LOS] and missed appointments)

were collected. Characteristics associated with increased use of cost-

coping strategies and perceived social isolation were assessed using

multivariable regression models.

Results: Seventy-three patients with LAHNC were recruited. Most pa-

tients were male (nZ57, 78%), Caucasian (nZ54, 74%), and had private

health insurance (nZ40, 54.8%). Overall, 50 participants (69%) used at

least 1 lifestyle-altering coping strategy. Strategies used included spending

savings (62%), borrowing money (42%), selling possessions (25%), and

having family members work more hours (23%). In multivariate analyses,

Medicaid patients had odds of 42.3 (95% confidence interval [CI], 4.19-

428, PZ.005) of using more coping strategies compared to privately

insured patients. Increased out-of-pocket costs and decreased wealth were

also independently associated with using more cost-coping strategies

(P<.01). Perceived social isolation was identified in 7 patients (9.5%) at

baseline, prior to treatment. Unemployment (PZ.02) and divorced/wid-

owed status (P<.001) were associated with high perceived social isolation.

These patients were more likely to take less medication than prescribed

(21.4 days missing medications vs 5.45, PZ.02) and to miss appointments

(7 vs 3, PZ.007). There was also a trend for them to have longer inpatient

LOS over the 6-month period (32.7 vs 27.6 days, PZ.17).

Conclusion: LAHNC patients extensively use lifestyle-altering strategies,

which can be considered a morbidity of head and neck cancer. High

perceived social isolation prior to treatment was identified as an inde-

pendent predictor of ineffective health care use, and social interventions

should be further investigated in this high-risk group.

Author Disclosure: S. Kung: None. J. O’Connor: None. B.J. Yap: None.

J.A. de Souza: Speaker’s Bureau; AstraZeneca.

324WITHDRAWN

325Prospectively Collected, Tooth-Specific Dosimetry Correlated WithAdverse Dental OutcomesA.T. Monroe,1 D. Flesher-Bratt,1 C.G. Morris,2 and A.V. Peddada3;1Penrose Cancer Center, Colorado Springs, CO, 2University of Florida,

Gainesville, FL, 3Penrose Cancer Center, Colorado Springs, CO, United

States

Purpose/Objective(s): Head and neck cancer survivors often experience

dental complications after radiation therapy (RT). Existing dose-response

data for dental complications has been based on relatively crude metrics.

We hypothesized that increasing dose, as measured prospectively to each

tooth, would correlate with adverse dental outcomes such as osteor-

adionecrosis (ORN) and periodontal disease, and that this information

could have important implications for surveillance and prevention of

dental complications.

Materials/Methods: Eighty-nine patients had dose to specific tooth

bearing portions of the mandible and maxilla prospectively collected

during treatment planning, resulting in 2490 tooth-specific data points.

Patients underwent a comprehensive dental intake evaluation to include

measurement of pocket depths and were then followed with serial dental

evaluations for a median of 2.5 years (0.2-6.9 years). Periodontal disease

was defined as an increase in pocket depth resulting in a final measurement

>6 mm. Patients were treated with intensity modulated RT +/- chemo-

therapy to a median prescription dose of 7000 cGy (5800-7200 cGy).

Results: At the patient level, the 3-year risks of ORN and periodontal

disease were 2.5% and 36.6%, respectively. For any individual tooth,

the risks of ORN and periodontal disease were 0.1% and 5.1% at 3

years. Dose to individual tooth bearing portions of bone was correlated

with ORN development (PZ.0165). Below 5000 cGy, the proportion

of teeth with ORN was 0.1% compared with 1.2% above 5000 cGy.

Periodontal disease also demonstrated a significant, but more gradual

dose response with an earlier inflection point around 2500 cGy

(PZ.0395).

Conclusion: We present the first tooth-specific dosimetric correlation be-

tween dose and adverse dental events. Attentive RT planning has the

potential to reduce dental complications, potentially sparing toxicity and

saving cost in the growing population of head and neck cancer survivors.

Moderate doses that were previously thought to be associated with low

risk may place patients at increased risk of periodontal disease.

Author Disclosure: A.T. Monroe: None. D. Flesher-Bratt: None. C.G.

Morris: None. A.V. Peddada: None.

Abstract 327; Table 1

Proton-IMPT IMRT

P valueMean (SD) (nZ35) Mean (SD) (nZ46)

Baseline 1.11 (1.32) 1.41 (2.15) .48Acute 5.97 (2.53) 6.44 (1.98) .36Subacute 5.15 (2.66) 6.58 (1.98) .01Chronic 3.66 (2.15) 4.14 (1.90) .43

International Journal of Radiation Oncology � Biology � Physics958

326Prognostic Value of Pretreatment Serum Inflammatory Markers inPatients Receiving Radiation therapy for Oropharyngeal Cancer(OPC)A. Kanwar,1 A.J. Hayes,2 A.S.R. Mohamed,3 C.C. French,4 G.B. Gunn,3

B.M. Beadle,3 J. Phan,3 H.D. Skinner,3 S. Lai,3 M.S. Kies,3 R.S. Weber,3

W.H. Morrison,3 D.I. Rosenthal,3 C.D. Fuller,3 and A.S. Garden3; 1Texas

Tech University Health Sciences Center School of Medicine, Lubbock, TX,2University of North Texas Health Science Center, Fort Worth, TX, 3The

University of Texas MD Anderson Cancer Center, Houston, TX, 4The

University of Texas Medical School at Houston, Houston, TX

Purpose/Objective(s): To assess the prognostic value and utility of pre-

treatment systemic inflammatory markers, namely neutrophil-to-lymphocyte

ratio (NLR) and platelet-to-lymphocyte ratio (PLR), in risk stratification of

patients receiving definitive radiation therapy (RT) for OPC.

Materials/Methods: We identified patients from an institutional review

boardeapproved institutional database of OPC outcome data who had pre-

RT WBC and platelet counts collected within 6 weeks prior to initiation of

RT. NLR and PLR were calculated by dividing absolute neutrophil and

platelet counts by absolute lymphocyte counts, respectively. Recursive

partitioning analysis (RPA) was performed to find specific cut points

associated with mortality in NLR and PLR. Kaplan-Meier and log-rank

tests were used to evaluate overall survival (OS) in identified cohorts.

Univariate and multivariate analyses were conducted using the following

variables: age, smoking, T and N stage, sex, chemotherapy, and pre-RT

NLR and PLR; the Cox proportional hazards model was used to identify a

relationship between OS and pre-RT NLR and PLR.

Results: A total of 448 patients were analyzed with a median follow-up of

87 months. Of these patients, 89% were men with a median age of 56

years. The majority had stage IVa disease (64%). Two hundred eighty-two

patients (63%) were treated using concurrent chemoRT, and 166 (37%)

received RT alone. RPA yielded an NLR cut point of 4.3 and a PLR cut

point of 183. Univariate analysis at these thresholds showed a statistically

significant association between NLR �4.3 and mortality (hazard ratio

[HR] 2.215; 95% confidence interval [CI] 1.57-3.13, P<.0001). A similar

association between PLR �183 and death was encountered (HR 2.27; 95%

CI 1.60-3.26, P<.0001). The 5-year OS was significantly better for patients

with NLR <4.3 (nZ276) compared to those with NLR �4.3 (84% vs 66%,

P<.0001). Likewise, patients with PLR <183 (nZ235) had better 5-year

OS compared to those with PLR �183 (85% vs 69%, P<.0001). Patients

with elevations of both NLR and PLR beyond the identified cut points

fared the worst compared to those without either (87% vs 66%, P<.0001).

In multivariate analysis, both NLR and PLR proved to be independent

prognostic factors for OS; NLR �4.3 (HR 1.55; 95% CI 1.06-2.27,

PZ.023) and PLR �183 (HR 1.89; 95%CI 1.28-2.85, PZ.0012).

Conclusion: Pretreatment systemic inflammatory markers are independent

prognostic factors for survival in OPC patients treated with RT. Future

investigations to validate the identified cut points and to develop risk-

adaptive treatment strategies are needed.

Author Disclosure: A. Kanwar: None. A.J. Hayes: None. A.S. Mohamed:

None. C.C. French: None. G.B. Gunn: None. B.M. Beadle: None. J.

Phan: None. H.D. Skinner: None. S. Lai: None. M.S. Kies: None. R.S.

Weber: None. W.H. Morrison: None. D.I. Rosenthal: None. C.D. Fuller:

Research Grant; National Institutes of Health/National Cancer Institute,

SWOG/Hope Foundation, Elekta AB/MD Anderson, MD Anderson Cancer

Center. In-kind Donation; General Electric Healthcare/MD Anderson. A.S.

Garden: None.

327Intensity Modulated Proton (IMPT) Versus Photon (IMRT)Radiation Therapies: Comparing Patient-Reported Outcomes (PRO)in Patients With Oropharyngeal Cancer Undergoing ChemoradiationT.T.W. Sio,1 H.K. Lin,2 Q. Shi,2 G.B. Gunn,3 C.S. Cleeland,3

P. Blanchard,4 N.G. Thaker,5 J. Phan,3 D.I. Rosenthal,3 A.S. Garden,3

W.H. Morrison,3 C.D. Fuller,3 T.R. Mendoza,3 X.S. Wang,6

and S.J. Frank3; 1Mayo Clinic, Scottsdale, AZ, 2Department of Symptom

Research; the University of Texas MD Anderson Cancer Center, Houston,

TX, 3The University of Texas MD Anderson Cancer Center, Houston, TX,4GUSTAVE ROUSSY INSTITUTE, Villejuif, France, 5MD Anderson Cancer

Center, Houston, TX, 6Department of Symptom Research, The University of

Texas M. D. Anderson Cancer Center, Houston, TX

Purpose/Objective(s): Treatments with IMPT for oropharyngeal cancer

are becoming increasingly common and may translate into producing better

patient (pt)-reported outcome (PRO) due to its superior physical beam

properties sparing organ at-risk structures, as compared with photon-based

IMRT. We hypothesized that pts treated with proton radiation therapy (RT)

would have less symptom burden as measured by PRO surveys.

Materials/Methods: From 2006 to 2013, through prospective registries, pts

with oropharyngeal cancer treated definitively with either concurrent chemo-

IMPT or chemo-IMRT completed the 22-item MD Anderson Symptom

Inventory-Head and Neck Cancer (MDASI-HN) symptom module at various

time intervals, including baseline, weekly during treatments (acute phase),

0 to 3 months after RT completion (subacute phase), and at scheduled

follow-ups thereafter (chronic phase, up to 24 months). Each individual item

in the MDASI-HN is scored from 0 to 10 (10 being worst). Individual and

top 5 and 11 of most affected symptoms were calculated and retrospectively

compared by descriptive and Student t test analyses.

Results: PRO data from 35 and 46 consecutive pts from the proton and photon

groups were collected and analyzed, respectively. The overall top 5 symptoms

were food taste (mean 4.91), dry mouth (4.49), swallowing difficulties (4.26),

lack of appetite (4.08), and fatigue (4.00). These top 5 symptoms averaged

3.89�2.96 for IMPT versus 4.66�2.98 for IMRT (PZ.04). The remaining

symptoms included in the top 11 symptom cluster were (in descending order)

mucus production, pain, disturbed sleep, drowsiness/sleepiness, mouth/throat

sores, and distress/upset. By overall scoring, the individual scores of mucus

production (mean, 3.37 vs 4.21), lack of appetite (3.43 vs 4.52), nausea (1.40 vs

2.14), vomiting (0.66 vs 1.22), and shortness of breath (0.59 vs 0.99) all favored

proton therapy (P<.05); none of the remaining individual items statistically

favored IMPT or IMRT. The respective top 11 scores were 3.50�2.74 versus

3.99�2.60 (PZ.14). By phases, the comparative PRO results for the top 5

symptoms were summarized (see Table 1), with a statistically significant dif-

ference favoring IMPT in the subacute phase.

Conclusion: Our PRO results demonstrated less symptom burden in pts

treated with IMPT in the top 5 symptoms cumulatively, as well as 5 in-

dividual symptom items. The greatest decrease in symptom burden was

seen with IMPT during the subacute phase, which represents a crucial

period of pts’ experience with some of the greatest symptom burden. The

use of proton therapy for oropharyngeal cancer is encouraging, and we are

currently validating this in a prospective clinical trial.

Author Disclosure: T.T. Sio: None. H. Lin: None. Q. Shi: None. G.B.

Gunn: None. C.S. Cleeland: None. P. Blanchard: None. N.G. Thaker:

None. J. Phan: None. D.I. Rosenthal: None. A.S. Garden: None. W.H.

Morrison: None. C.D. Fuller: None. T.R. Mendoza: None. X.S. Wang:

None. S.J. Frank: None.

328Comparison of Head and Neck Cancer Populations by Cause ofDeathS. Massa,1 N. Osazuwa-Peters,1 K. Christopher,1 R. Walker,2

and M. Varvares3; 1Saint Louis University School of Medicine, Saint Louis,

MO, 2Saint Louis University School of Medicine, St Louis, MO, 3Harvard

Medical School, Boston, MA

Volume 94 � Number 4 � 2016 Posters 959

Purpose/Objective(s): Patients with head and neck squamous cell car-

cinoma (HNSCC) are at risk for death from causes related and unrelated to

their cancer. The risk of death from noncancer causes is termed competing

risk from the perspective of cancer treatment and research. These patients

have been found to have elevated competing risks compared to the general

population and other cancer populations, likely related to high rates of

substance use and comorbidities. We aim to identify the sociodemographic

and tumor characteristics that distinguish the group of HNSSC patients

who die from their primary cancer to those that die of competing causes.

Materials/Methods: Deceased adult patients with histologically proven,

mucosal HNSCC diagnosed between 2004 and 2012 were identified from

the Surveillance, Epidemiology and End Result database. Tumor charac-

teristics, demographic variables, survival, and cause of death were

collected. Patients with insufficient data, including unknown cause of

death, and previous cancers were excluded. The competing causes of death

for HNSCC are presented with comparisons to thyroid and salivary gland

cancer patients and the general population. Standardized incidence ratios

(SIR) are calculated to compare death rates between the HNSSC cohort

and the general population. HNSSC patients who died of causes related to

their primary tumor are then compared to those deceased from other causes

using a chi-squared analysis.

Results: Of 64,598 patients meeting study criteria, 17,460 were

deceased, including 29.0% from noncancer causes. The 5 most common

causes of death not attributable to the primary tumor were cardiac

(26.4%), COPD (8.8%), lung cancer (6.4%), other cancers (5.3%), and

stroke (5.3%). These are also the 5 leading causes in a general popu-

lation cohort matched for age, race, and gender. Excluding cancer-

related mortality, the SIR was elevated among HNSCC patients for all

causes of death (11.3); most drastically for chronic liver disease (38.4),

suicide (36.1), and second HNSCC primaries (24.5). Compared to pa-

tients who died of their cancer, patients who died from competing

causes are older and more often white with early-stage and laryngeal

tumors (P<.001).

Conclusion: As a group, HNSCC patients are at substantially increased

risk of dying from all causes, not only their cancer. Compared to the

general population, these patients have greatly elevated risk of dying

from all causes, especially some rare causes like suicide. However, the

majority of noncancer-related deaths in this group are attributable to

causes also common in the general population, like cardiovascular

disease and stroke. While treating patients with HNSCC for their cancer,

screening and treatment of other risk factors should not be neglected,

especially in those with curable cancers at risk for mortality from other

causes.

Author Disclosure: S. Massa: None. N. Osazuwa-Peters: None. K.

Christopher: None. R. Walker: None. M. Varvares: None.

329Predictors of Dysgeusia in Patients Treated With Chemoradiationfor Head and Neck CancerE. Sapir, S. Samuels, F.Y. Feng, K.A. Vineberg, and A. Eisbruch;

University of Michigan, Ann Arbor, MI

Purpose/Objective(s): Dysgeusia is a significant morbidity for patients

receiving chemoradiation for head and neck cancer. Factors affecting the

severity of dysgeusia remain uncertain. In this study we investigated the

relationships between patient-reported dysgeusia, doses to the oral cavity,

salivary production (required to dissolve food particles), and patient-re-

ported xerostomia.

Materials/Methods: Seventy-nine patients with stage III-IVoropharyngeal

cancer (OPC) (nZ73) or nasopharyngeal cancer (NPC) (nZ6) participated

in a prospective, longitudinal study of quality of life (QOL), including

assessment of patient-reported gustatory function by taste-related ques-

tions from the Head and Neck QOL instrument (HNQOL) and the Uni-

versity of Washington Head and Neck-Related QOL instrument (UWQOL)

prior to therapy and periodically posttreatment. At the same time intervals,

patients also completed a validated 8-item self-reported xerostomia-

specific questionnaire (XQ) and had unstimulated and stimulated major

salivary glands flow rate measurements. Treatment consisted of definitive

intensity modulated radiation therapy (IMRT) aimed at sparing the salivary

glands, oral cavity, and swallowing structures and was delivered concur-

rent with chemotherapy in all patients. Regression models were used to

evaluate the effect of covariates (radiation dose, salivary output, and XQ

scores) on QOL scores.

Results: At 1, 3, and 12 months after treatment, severe taste dysfunction

was reported by 49%, 37%, and 21% of evaluable patients, and mean

stimulated salivary flow was 17%, 18%, and 30% of pretreatment levels,

respectively. Significant associations were found between patient-reported

taste dysfunction and radiation dose to the anterior tongue and oral cavity

(both P<.05), but not the posterior tongue, when pooled across all time-

points. A 10-Gy increase in oral cavity dose was associated with a 2.8-fold

increase in the odds of severe taste impairment (P<.05). While measured

salivary output was not significantly associated with postradiation taste

dysfunction, patient-reported XQ summary scores and xerostomia while

eating scores were significantly correlated with changes in taste at each

timepoint (P<.001).

Conclusion: Postradiation therapy taste impairment is significantly

correlated with mean radiation dose to both the anterior tongue and oral

cavity. Although decreased salivary gland output had no significant

correlation with changes in taste function, patient-reported xerostomia

was significantly correlated with dysgeusea. This discrepancy may be

explained by reduced oral cavity minor salivary glands mucins after RT,

not accounted for by major salivary glands output. Reducing dose to the

oral cavity with IMRT, if possible, may significantly contribute to

improving QOL by decreasing long-term patient-reported xerostomia,

which may translate into better and faster rehabilitation of taste

function.

Author Disclosure: E. Sapir: None. S. Samuels: None. F.Y. Feng: None.

K.A. Vineberg: None. A. Eisbruch: None.

330From Patient-Reported Outcomes to Quantitative Health States:Characterization of Head and Neck Cancer Patient SurvivorshipUtilities Using Prospective Longitudinal Assessment With theMDASI-HNC.C. Hansen,1 G.B. Gunn,2 A.S.R. Mohamed,2 D.I. Rosenthal,2 T. Brown,2

C.S. Cleeland,2 T.R. Mendoza,2 S.J. Frank,2 A.S. Garden,2 S. Lai,2

and C.D. Fuller3; 1Texas Tech University Health Science Center School of

Medicine, Lubbock, TX, 2The University of Texas MD Anderson Cancer

Center, Houston, TX, 3MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): For patients receiving treatment for head and neck

cancers (HNC), patient-reported outcomes (PROs) are routinely acquired,

but rarely is an equivalent conversion capable of translating these patient

experiences (as quantified by quality of life [QOL] measures) into a

defined utility. The aims of this study are to (1) explore the relationship

between various symptom scales and subsequent alteration in patient-

derived utilities, as assessed by EQ-5D, (2) measure an assessment of the

natural history and kinetics of PRO-derived utilities over the course of the

survivorship period, and (3) generate testable hypotheses and/or specific

interventional studies designed to maximize symptom-derived utilities for

future head and neck studies.

Materials/Methods: Two hundred and sixty-five HNC patients

completed both the MD Anderson Symptom Inventory-Head and Neck

Cancer (MDASI-HN) and EQ-5D during the survivorship period after

we obtained both institutional review board approval and informed

consent from all participants. Patients were contacted and undertook a

phone interview with administration of the MDASI-HN and EQ-5D. The

MDASI-HN has been validated across multiple languages and shown

high interpatient validity and repeatability across patients receiving

solely radiation or chemoradiation. Simultaneously, patients were

queried with the EQ-5D, which is recognized as the standard instrument

for conversion of nationally specific, age-adjusted utility states. These

can then be used as utilities for QOL.

International Journal of Radiation Oncology � Biology � Physics960

Results: The EQ-5D mean utility index was higher than anticipated at

0.84�0.2, given a mean VAS of 74�21. The mean of the MDASI-

HNComposite sum was 41�32, indicating a breadth of patient-reported

symptom burden in this study population. Dry mouth and oral/throat

mucus problems were the only mean MDASI-HN symptom item scores

greater than 3 out of 10, with scores of 3.86�3.53 and 3.10�3.27,

respectively. Utility scores placed in bins according to length of follow-

up were not collectively different from each other (PZ.165); however,

the “12 months and less” was significantly different from “12 to 24

months” (PZ.043) and “24 to 36 months” (PZ.007). The major

symptom drivers impacting QOL, according to the BICminimum model,

were pain (P<.0001), sadness (PZ.0003), shortness of breath

(PZ.0001), and upset (PZ.005). Interestingly, none of the head and

neck symptom items significantly contributed to QOL measures.

Conclusion: Our data demonstrate a very strong relationship between

PROs, assessed by MDASI-HN, and patient-reported QOL, as determined

by EQ-5D. In fact, the summation of all MDASI-HN symptom items

strongly correlates with the EQ-5D utility scores and can be used to

accurately predict a patient’s QOL.

Author Disclosure: C.C. Hansen: None. G.B. Gunn: None. A.S.

Mohamed: None. D.I. Rosenthal: None. T. Brown: None. C.S. Cleeland:

None. T.R. Mendoza: None. S.J. Frank: None. A.S. Garden: None. S.

Lai: None. C.D. Fuller: Research Grant; the Paul Calabresi Clinical

Oncology Award Program, NIH/NCI Clinician Scientist Loan Repayment

Program. Honoraria; the SWOG/Hope Foundation Dr. Charles A. Coltman,

Jr., Fellowship in Clinical Trials, a General Electric Healthcare/MD

Anderson Center for Advanced Biomedical Imaging In-Kind Awardan

Elekta AB/MD Anderson Department of Radiation Oncology Seed Grant:

the Center for Radiation Oncology Research at MD Anderson Cancer

Center, the MD Anderson Institutional Research Grant Program.

Abstract 332; Table 1 Kaplan-Meier estimates of freedom from failure(FFF) of the p16+ OPSCC patients subdivided by the median of their prognosticindex.

p16+ OPSCC patients

FFF (y) All (%) CI (%)Favorablegroup (%) CI (%)

Unfavorablegroup (%) CI (%)

1 89.1 (83.8-94.4) 95.9 (91,3-100) 81.2 (71.6-90.8)2 82.4 (75.9-88.9) 93.2 (87.5-98.9) 69.6 (58.2-81.0)3 79.1 (72.2-86.0) 89.8 (82.5-97.1) 66.3 (54.5-78.1)

331Financial Toxicity in Thyroid CancerdAn Analysis From the NorthAmerican Thyroid Cancer Survivorship StudyJ.A. de Souza,1 R. Grogan,2 and B. Aschebrook-Kilfoy2; 1University of

Chicago, Chicago, IL, 2The University of Chicago, Chicago, IL

Purpose/Objective(s): Financial toxicity has been associated with worse

health-related quality-of-life (HRQOL), compliance to treatment, and even

survival in cancer patients (pts). Measuring financial toxicity and under-

standing its predictors are of paramount importance when planning inter-

vention strategies, the value of care, and health care policies. We report

financial toxicity and its predictors in a large cohort of thyroid cancer pts

and survivors.

Materials/Methods: Pts with thyroid cancer were surveyed in the North

American Thyroid Cancer Survivorship Study. Financial toxicity was

assessed by the previously validated COmprehensive Score for financial

Toxicity (COST) patient-reported outcome (PRO), as well as by questions

related to financial distress (out-of-pocket costs, loss of income, and

bankruptcy). Data on sociodemographics, income, type of disease, time

since diagnosis, and prior therapies were collected. Predictors of financial

toxicity were assessed in multivariate analyses, controlling for potential

confounders, such as HRQOL (as measured by the thyroid cancerespecific

City of Hope instrument), type of treatment received, and time since

diagnosis.

Results: A total of 591 pts with thyroid cancer within the past 6 years

were surveyed in 2 countries: 553 (93.5%) in the US and 38 (6.5%) in

Canada. Most pts were women (nZ518 pts, 88%), Caucasian (nZ531,

89.9%), and had at least a graduate or professional degree (nZ206,

35.5%). The median time since diagnosis was 857 days (range 105-2176

days), and 430 pts (72.8%) had papillary thyroid cancer. There were 210

pts with stage I (35.7%), and 61 (10.3%) with stage IV disease. In total,

568 pts (96.1%) had a thyroidectomy, 447 pts (75.6%) received radio-

active iodine, and 11 pts (1.9%) were on tyrosine kinase inhibitors. In

total, 234 pts (39.5%) stated that their out-of-pocket costs were higher

than previously thought; 207 pts (35%) felt their disease resulted in loss

of income; 44 pts (7.4%) were unable to meet their monthly expenses;

and 7 pts (1.2%) declared bankruptcy after diagnosis. The median

COST value was 24 (range 0-44). In multivariate analyses, the inde-

pendent predictors of worse financial toxicity, as measured by COST-

PRO values, were lower income (P<.001), female gender (PZ.01),

lower educational level (PZ.002), health care delivery in the US,

(PZ.002), and worse HRQOL (P<.001).

Conclusion: A significant proportion of thyroid cancer pts experience

financial toxicity. We identified pt characteristics (gender, education, in-

come), as well as geographical differences (health care delivery in the US)

as predictors of financial toxicity in thyroid cancer pts and survivors, in-

dependent of their HRQOL.

Author Disclosure: J.A. de Souza: Speaker’s Bureau; AstraZeneca. R.

Grogan: None. B. Aschebrook-Kilfoy: None.

332A Prognostic Model for Time-to-Failure in p16-Negative PatientsProvides Useful Risk Stratification in p16-Positive PatientsG.B. Rasmussen,1 J.H. Rasmussen,2 B.M. Fischer,3 J.T. Friborg,1

M.H. Therkildsen,4 L. Specht,1 S.M. Bentzen,5 and I. Vogelius1;1Department of Oncology, Section of Radiotherapy, Rigshospitalet,

University of Copenhagen, Copenhagen, Denmark, 2Department of

Otorhinolaryngology, Head & Neck Surgery and Audiology,

Rigshospitalet, University of Copenhagen, Copenhagen, Denmark,3Department of Clinical Physics, Nuclear Medicine & PET, PET &

Cyclotron Unit, Rigshospitalet, University of Copenhagen, Copenhagen,

Denmark, 4Department of Pathology, Rigshospitalet, University of

Copenhagen, Copenhagen, Denmark, 5Division of Biostatistics and

Bioinformatics, University of Maryland Greenebaum Cancer Center, and

Department of Epidemiology and Public Health, University of Maryland

School of Medicine, Baltimore, WA

Purpose/Objective(s): Low event rates in p16-positive (p16+) patients

reduces the statistical power of multivariable statistical modeling. Here, we

test whether a prognostic model developed in p16-negative (p16-)/un-

known patients provides useful risk stratification of patients with p16+

oropharyngeal squamous cell carcinoma (OPSCC).

Materials/Methods: An institutional cohort of 600 consecutive head and

neck squamous cell carcinoma (HNSCC; all subsites) treated with intensity

modulated radiation therapy between 2005 and 2012 was analyzed. p16

status was considered a surrogate marker of human papillomavirus (HPV)-

related OPSCC. A Cox proportional hazards model stratified for subsite

within the head and neck was fitted to the p16-/unknown patients. Any

treatment failure was considered as an event; patients alive with no evi-

dence of disease were censored at last follow-up. A previous model

included more variables (eg, TNM), but the following factors captured the

variation and were prognostic: prescription of cisplatin, smoking status

(never, former, current), gross tumor volume (GTV), and maximal stan-

dardized uptake value (SUVmax). Variables were selected using backward

elimination. The final model was applied to assign each p16+ OPSCC

patient a prognostic index, calculated as their individual hazard ratio (HR)

for failure from the p16- model.

Results: The p16+ OPSCC group (nZ138) had 28 events, and the p16-/

unknown group (nZ435) had 187 events. Concomitant cisplatin

(HRZ0.51 [0.37-0.69]), GTV (HRZ1.01 [pr cm3] [1.00-1.01]), and

smoking status (HRZ1.61 [1.22-2.12]) were significant in the p16-/

Volume 94 � Number 4 � 2016 Posters 961

unknown patients (all P�.001). According to the p16- model, HR was a

highly significant prognostic in p16+ patients (log-rank PZ.0005).

Kaplan-Meier estimates of freedon from failure (FFF) according to HR

above versus below median are given in Table 1.

Conclusion: HPV+ OPSCC patients could be risk stratified using a prog-

nostic model developed in patients with p16- disease. This stratification

may be useful for clinical management as well as selecting patients for

clinical trials of novel/modified therapy.

Author Disclosure: G.B. Rasmussen: Employee; Novo Nordisk. J.H.

Rasmussen: None. B.M. Fischer: None. J.T. Friborg: None. M.H.

Therkildsen: None. L. Specht: Research Grant; Varian Medical Systems.

Honoraria for teaching courses held (to department); Varian Medical

Systems. S.M. Bentzen: None. I. Vogelius: Research Grant; Varian

Medical Systems. Honoraria for teaching courses held (to department);

Varian Medical Systems.

333Does Being Married Independently Predict Survival in PatientsWith Head and Neck Cancer? Results From a Single InstitutionN. Osazuwa-Peters,1 K. Christopher,2 S. Massa,2 L. Cass,2 A. Hussaini,2

A. Behera,2 R. Walker,3 and M. Varvares4; 1Saint Louis University Cancer

Center; Saint Louis University School of Medicine, Saint Louis, MO,2Saint Louis University School of Medicine, Saint Louis, MO, 3Saint Louis

University School of Medicine, St Louis, MO, 4Harvard Medical School,

Boston, MA

Purpose/Objective(s): An important factor in cancer survivorship is so-

cial support, often coming from a spouse. There is emerging literature on

the role of spousal support in head and neck cancer (HNC) outcomes;

however, most are based on results from national databases. While more

generalizable, these studies may not always capture the unique variation in

different patient populations. This single-institution study aimed to

describe the association between marital status and outcomes of HNC and

to determine if marital status independently predicts survival in a local

patient population.

Materials/Methods: We identified 460 patients aged 20 to 91 years (59.31

� 11.42) diagnosed with squamous cell carcinoma of the head and neck at

an academic tertiary referral center between 1997 and 2012 in this retro-

spective cohort study. Cox proportional hazards model assessed the effect

of marital status on survival. Results are based on the final model con-

structed after accounting for covariates in the data.

Results: Our study population was made up of 73% males and 82.2%

whites. We found an association between marital status and HNC survival.

Unmarried HNC patients had a 66% increase in the hazard of death

compared to married HNC patients (hazard ratio [HR] 1.66, 95% confi-

dence interval [CI] 1.23-2.23). This was after controlling for covariates,

which included sociodemographic variables (age, race, sex, and health

insurance status), social habits (tobacco and alcohol), primary anatomical

subsite (oral cavity, oropharyngeal, laryngeal, and others), stage at pre-

sentation (early vs late stage), and treatment modality (surgery, surgery

with adjuvant therapies, other single modality therapy, and palliative care).

Other factors found to be associated with an increased hazard of death

were age (�50 years), current tobacco use, late stage of presentation,

palliative care, and laryngeal subsite.

Conclusion: Marital status is associated with head and neck cancer

outcomes, and being married is an independent predictor of survival

among patients. This result, found in previous national studies, held true

in our local patient population. This underscores the need for the

multidisciplinary HNC team to recognize this aspect of survivorship and

to emphasize the need for social support among unmarried HNC pa-

tients. It could be that it is necessary to add social support to the clinical

practical guidelines for managing head and neck cancer beyond pallia-

tive care.

Author Disclosure: N. Osazuwa-Peters: None. K. Christopher: None. S.

Massa: None. L. Cass: None. A. Hussaini: None. A. Behera: None. R.

Walker: None. M. Varvares: None.

334Patient- Versus Physician-Reported Quality of Life AmongSurvivors of Head and Neck Cancer After Chemoradiation:Prospective Evaluation of Screening MethodologiesA.M. Chen, C. Felix, S. Hsu, J. Garst, and J. Wang; University of

California, Los Angeles- David Geffen School of Medicine, Los Angeles,

CA

Purpose/Objective(s): To compare quality of life (QOL) as determined

by patient- versus physician-assessment methods among survivors of head

and neck cancer treated by definitive chemoradiation.

Materials/Methods: Two validated QOL instruments, the University of

Washington Quality of Life questionnaire (UW-QOL) and the Functional

Assessment of Cancer Therapy for Head and Neck questionnaire (FACT-

H&N), were administered to patients returning for routine follow-up after

previously completing chemoradiation for squamous cell carcinoma of the

head and neck. The median radiation dose was 70 Gy (range, 60 to 72 Gy).

Only patients who had been clinically without evidence of disease for

longer than 6 months were sampled. Scores from these questionnaires were

compared to physician-reported QOL (rated as outstanding, very good,

good, fair, or poor) and toxicities, which were recorded blindly and

independently after each patient visit, using the National Cancer Institute’s

Common Toxicity Criteria (version 4.0). Two by 2 contingency tables were

constructed to assess differences between patient- and physician-reported

responses using Fisher exact test.

Results: A total of 50 patients (35 males; 15 females) completed both

instruments. The median time from completion of chemoradiation was 14

months (range, 12 to 79). The median age was 51 years (range, 35 to 89).

While 66% of patients’ QOL was rated as outstanding or very good by

physicians, only 32% of patients reported their QOL as outstanding or very

good using the self-reported UW-QOL (P<.001). Using the FACT-H&N,

only 40% of patients reported they are “enjoying life” either “quite a bit”

or “very much.” Significant discordance was also observed between pa-

tient- and physician-reported toxicities with respect to the domains of

swallowing, speech, appearance, pain, activity/energy, and mood (P<.05,

for both UW-QOL and FACT-H&N).

Conclusion: Significant discordance exists between patients and physi-

cians in the assessment of late toxicities for long-term survivors of head

and neck cancer treated by chemoradiation. Our findings suggest that

QOL, while subjective, may be frequently misevaluated by physicians and

point to a need for improved screening methodologies.

Author Disclosure: A.M. Chen: None. C. Felix: None. S. Hsu: None. J.

Garst: None. J. Wang: None.

335Insurance Status Has Differential impacts Among Patients WithSquamous Cell Carcinoma Arising From Different Head and NeckPrimary SitesdA SEER AnalysisB. Li1 and S. Pollack2; 1Weill Cornell Medical College New York, NY,2Department of Computer Information System and Decision Sciences, St.

John’s University, Jamaica, NY

Purpose/Objective(s): As a result of the Affordable Care Act (ACA), a

60% to 70% reduction in the number of underinsured or uninsured

Americans by 2019 is estimated. Given the significantly higher proportion

of Medicaid or uninsured patients with head and neck (HN) cancer, the

authors used the Surveillance, Epidemiology and End Results (SEER)

database to investigate the differential impacts of insurance status on

cancer stage at diagnosis, receipt of radiation treatment (RT), and survival

among patients with squamous cell carcinoma arising from different HN

primary sites.

Materials/Methods: The SEER database was queried from 2007 to 2011

for squamous cell carcinoma in HN cancers arising from oral cavity

(11,770), oropharynx (7201), larynx (6685), hypopharynx (1342), and

nasopharynx (1762). Patients were stratified by their insurance status,

including uninsured or Medicaid, and non-Medicaid (insured or insured/no

specifics). Covariates include age, gender, race, marital status, percentage

International Journal of Radiation Oncology � Biology � Physics962

of county below a 150% federal poverty level, TNM stage, and receipt of

cancer-directed surgery and/or RT. A Cox proportional hazards regression

was used to estimate the effect of insurance status on cancer stage, RT

received, and cause-specific survival.

Results: The proportion of Medicaid or uninsured patients was approxi-

mately 25% in oral cavity, oropharynx, and nasopharynx, 33% in larynx,

and 36% in hypopharynx. Overall, patients were more likely to have

localized diseased (T1-2N0) if they had non-Medicaid insurance, rather

than if they had Medicaid or were uninsured. The odds ratio (OR) for

localized disease was 2.7 in larynx, 2.0 in nasopharynx, and approximately

1.6 in the other sites (P<.001 for all sites). Similarly, patients were more

likely to receive RT with non-Medicaid insurance, with OR ranging from

2.9 in oral cavity to 1.7 in larynx (P<.001 for all sites). After adjustment

for all factors, patients were more likely to die of their cancer if they had

Medicaid or were uninsured. The hazard ratio ranged from approximately

2.0 in oral cavity, oropharynx, and nasopharynx to 1.4 in larynx (P<.001

for all sites).

Conclusion: Among patients with HN cancers from different primary sites,

there is difference in the association between uninsured/underinsured

status and more advanced disease stage, more suboptimal cancer treatment,

and worse survival. However, the level of the differential impact is inter-

mediate and the general trend is similar across all HN sites.

Author Disclosure: B. Li: None. S. Pollack: None.

336Cancer-Independent Loss of Life Expectancy in the Head and NeckCancer PopulationS. Massa,1 N. Osazuwa-Peters,1 K. Christopher,1 R. Walker,2

and M. Varvares3; 1Saint Louis University School of Medicine, Saint Louis,

MO, 2Saint Louis University School of Medicine, St Louis, MO, 3Harvard

Medical School, Boston, MA

Purpose/Objective(s): Patients diagnosed with head and neck cancer

(HNC) lose life expectancy due to cancer-related mortality, but further

longevity is lost to increased prevalence of comorbidities, substance use,

and other factors inherent to this group. The life expectancy and modifying

factors for this group have not been quantified. This study aims to asses the

life expectancy of HNC patients not attributable to their primary cancer.

We hypothesize that HNC patients will have shorter life expectancies than

the general population, independent from mortality related to their primary

cancer and after controlling for age, sex, and race.

Materials/Methods: Patients aged 35 to 100 years were selected from the

Surveillance, Epidemiology, and End Results database with a first cancer

diagnosis from 2004 to 2012 of head and neck mucosal sites, thyroid

gland, or salivary glands. Patients were excluded due to insufficient data or

death within 3 months of initial diagnosis to limit the effect of treatment-

related mortality. Life tables of survival versus age were created which are

left-truncated prior to enrollment in the database and right-censored for

cancer-related deaths. These life tables approximate a similar group of

patients who either did not develop cancer or whose cancers were cured.

Data were stratified by stage, site, histology, race, sex, insurance type, and

income quintile and compared to the United States population, matched for

gender and race distributions. The covariates were also analyzed within the

HNC cohort using univariate and multivariate logistic regression to assess

the hazard ratio of death from noncancer causes.

Results: A total of 15,703 patients were identified with an overall 5-year

reduction in median life expectancy compared to the general population

matched for sex and race despite excluding cancer-related mortality.

Larger losses in cancer-independent life expectancy were associated with

advanced stage disease (-12 years for stage IV), hypopharyngeal (-18

years), and laryngeal (-14 years) tumors, black race (-9 years), male sex (-7

years), squamous histology (-11 years), Medicaid insurance (-19 years),

and lower income quintile (-19 years for 1st quintile). Conversely, patients

with stage I disease and thyroid cancer had increased longevity compared

to the general population (+1 and +4 years, respectively). Multivariate

logistic regression found analogue trends among these variables regarding

changes in the mortality risk from noncancer causes. All reported results

are statistically significant (P<.05).

Conclusion: Patients with HNC, particularly poor, black, males with

mucosal squamous cell carcinoma, have a decreased life expectancy in

comparison with the general population despite excluding cancer-related

mortality. In addition to cancer treatments, this population may benefit

from improved prevention and management of other comorbidities and risk

factors.

Author Disclosure: S. Massa: None. N. Osazuwa-Peters: None. K.

Christopher: None. R. Walker: None. M. Varvares: None.

337Patient-Reported Distress in Head and Neck Cancer PatientsReceiving Radiation TherapyK. Gaines, L. Hollant, B. Smart, M. Single, Y. Habboush, K.S. Tzou,

S. Ko, H. Biers, A. Day, K. Wert, and R.C. Miller; Mayo Clinic,

Jacksonville, FL

Purpose/Objective(s): Patients undergoing head and neck (H&N) radia-

tion therapy (RT) experience distress due to the nature of their underlying

malignancies and the serious side effects associated with treatment. In our

clinic, patients receiving RT are prospectively screened for distress.

Retrospective review of this assessment was performed to identify patterns

in patient-reported distress (PRD).

Materials/Methods: H&N cancer patients treated at our institution were

eligible for inclusion in this study if they had completed a 30-question

PRD survey at the beginning of and/or during RT. Each question

concentrated on a particular possible cause of distress. Patients could rate

distress on a scale of 1 to 5, corresponding to a low and high distress level,

respectively. The assessment also included the National Comprehensive

Cancer Network (NCCN) Distress Thermometer scale that ranges from 1

(low overall distress) to 10 (high overall stress). We retrospectively

reviewed H&N PRD data from patients undergoing RT from April 9, 2012

to August 5, 2015.

Results: A total of 185 patients completed PRD screening forms. Ages

ranged from 23 to 93 years old with a median of 67 years. There were 152

(82.2%) male and 33 (17.8%) female patients. Definitive external beam RT

was used to treat all patients. The primary tumor histology was squamous

cell carcinoma in 131 patients (70.8%). Chemotherapy was given to 132

patients (71.4%), and 144 (77.8%) underwent a surgical intervention.

NCCN distress thermometer results ranged from 0 to 10, with a median of

4.6. Mean distress scores ranged from 1.44 to 2.83. The top 5 concerns

were “How will I feel during treatment” (2.83), “Pain that affects daily

functioning” (2.69), “Out of pocket medical costs” (2.63), “Fatigue”

(2.48), and “Sleep difficulties” (2.50). The least concerning complaints

were “A loved one relying on me for their physical care” (1.44), “Trans-

portation” (1.49), “Spirituality” (1.50), “Family communication about my

illness” (1.54), “My job” (1.55), and “Controlling my anger” (1.55).

Conclusion: Patients receiving definitive H&N RT in this population were

most worried about symptomatic changes due to treatment side effects, as

well as tumor and treatment-related pain. Notably, financial burdens of

treatment ranked in the top 5 concerns. These distresses should be assessed

in patients who receive H&N RT.

Author Disclosure: K. Gaines: None. L. Hollant: None. B. Smart: None.

M. Single: None. Y. Habboush: None. K.S. Tzou: None. S. Ko: None. H.

Biers: None. A. Day: None. K. Wert: None. R.C. Miller: Chief editor;

ASTRO. Board of Trustees; Mayo Clinic Health System Albert Lea

Austin.

338Prospective MD Anderson Dysphagia Inventory Outcomes AfterNonsurgical Treatment of Locoregionally Advanced OropharyngealCarcinomaR.P. Goepfert,1 J.S. Lewin,1 M.P. Barrow,1 G.B. Gunn,1 C.D. Fuller,1

B.M. Beadle,1 A.S. Garden,1 D.I. Rosenthal,1 M.S. Kies,1

V. Papadimitrakopoulou,1 S. Lai,1 D.L. Schwartz,2 and K.A. Hutcheson1;

Volume 94 � Number 4 � 2016 Posters 963

1The University of Texas MD Anderson Cancer Center, Houston, TX,2University of Texas Southwestern Medical Center, Dallas, TX

Purpose/Objective(s): Long-term swallowing outcomes including gas-

trostomy tube and aspiration rates are well described among oropharyngeal

cancer (OPC) survivors but long-term, prospective data on patient-reported

dysphagia have not been reported for patients treated with split-field in-

tensity modulated radiation therapy (IMRT) and laryngeal shielding.

Materials/Methods: A pooled analysis was conducted from 3 single-

institution clinical trials for advanced stage (III/IV) head and neck carci-

noma. Patients treated with definitive, split-field IMRT and chemotherapy

for OPC were sampled. Inclusion criteria for analysis included baseline

and at least 1 posttreatment MD Anderson Dysphagia Inventory (MDADI).

Patients with definitive surgical management of their primary tumor were

excluded. Prospectively collected MDADI composite and subgroup scores

were analyzed at baseline, 6, 12, and 24 months after treatment. Pairwise

comparisons were analyzed with bonferroni correction for multiple com-

parisons, and multilevel mixed effects linear regression models were fit to

evaluate predictors of MDADI scores.

Results: A total of 116 patients were included. Thirty-nine percent of

patients had tonsil and 61% had base of tongue primary tumors. Thirty-five

percent had T3/4 tumors, and 90% were N2b or greater. Human papillo-

mavirus and/or p16 status was known in 75% of patients; one or both were

positive in 89%. Eighty-four percent of patients underwent induction, and

56% received concurrent chemotherapy. Nineteen percent had post-IMRT

neck dissection. Mean baseline MDADI composite score was 88.3, drop-

ping to 73.8 at 6 months (P<.0001) and rising to 78.6 and 83.3 by 12

(P<.0001) and 24 months (P<.0001), respectively. This trend was mirrored

in subgroup scores. Mean baseline composite was higher in patients with

T1/2 tumors than T3/4 tumors (PZ.003, 95% confidence interval -13.8 to

-2.9), but this difference was not found at any posttreatment interval. No

significant differences in composite MDADI were noted by age, primary

site, use of concurrent and/or induction chemotherapy, neck dissection, or

pack-years of smoking at baseline or posttreatment intervals (P>.05).

Conclusion: Overall, patients report clinically meaningful dysphagia early

after modern split-field IMRT for locoregionally advanced OPC that re-

mains apparent 6 months after treatment. While MDADI scores recover

slowly thereafter, they remain depressed at 24 months compared to base-

line suggesting only partial recovery of perceived swallowing function.

Despite better baseline function, patients with smaller primary tumors have

a greater relative decrement in composites during treatment that does not

recover in the long-term compared to patients with larger primary tumors.

These data inform treatment counseling, clinical trial design, and inter-

pretation of long-term patient-reported outcomes.

Author Disclosure: R.P. Goepfert: None. J.S. Lewin: None. M.P.

Barrow: None. G.B. Gunn: None. C.D. Fuller: None. B.M. Beadle:

None. A.S. Garden: None. D.I. Rosenthal: None. M.S. Kies: None.

V. Papadimitrakopoulou: None. S. Lai: None. D. Schwartz: None. K.A.

Hutcheson: None.

339Predictors of Survival in Parotid Adenocarcinoma (Not OtherwiseSpecified): A National Cancer Data Base Study of 3155 Patients.K. Zhan,1 S.F. Khaja,1 A.T. Huang,1 D. Bell,2 and T. Day1; 1Medical

University of South Carolina, Charleston, SC, 2MD Anderson, Houston, TX

Purpose/Objective(s): Using the world’s largest cancer database, we

sought to better characterize and identify prognosticators for parotid

adenocarcinoma, not otherwise specified (PANOS), a rare salivary histo-

pathology with a poor prognosis.

Materials/Methods: A retrospective review of all cases of PANOS in the

National Cancer Data Base (NCDB) from 1998 to 2012 was performed.

Relevant demographic and disease characteristics were extracted and

analyzed. COX regression analysis was used for identifying predictors of

survival.

Results: A total of 30,728 parotid cancers were identified, and 10% were

PANOS. Median age was 67 years, and 62.8% of cases were in males. The

majority of patients had high-grade (67.2%) or clinically advanced disease

(56.4% stages III-IV). Regional metastasis was common (35.9%), while

occult nodal metastasis (20.2% overall) was less frequent in non-high-

grade lesions (8.5% vs 31.6%, P<.001). Distant metastasis was rare

(7.9%). Increasing age was significantly correlated with increasing grade

(P<.001). Extraglandular spread (EGS) was found in half of patients and

was significantly more frequent for high-grade than for low-grade disease.

Most patients were treated with surgery and radiation therapy (43.0%),

followed by surgery alone (26.0%), then triple-modality surgery with

chemoradiation (9.6%). Overall 5-year survival (OS) was 47%. Presence of

regional metastasis conferred a significant drop in survival. In multivariate

analysis, age, regional metastasis, distant metastasis, high grade, and high

T stage were significant, independent predictors of worse survival. Patients

with stage III-IV disease who received both surgery and radiation therapy

had a significantly better OS than those receiving surgery alone.

Conclusion: PANOS is an aggressive disease with frequent regional

metastasis and poor overall survival. Patients typically present with high-

grade and advanced-stage disease. Numerous clinical and pathologic var-

iables are associated with poorer survival outcomes.

Author Disclosure: K. Zhan: None. S.F. Khaja: None. A.T. Huang: None.

D. Bell: None. T. Day: None.

340Long-Term Patient-Reported Dysphagia in Low-Risk OropharyngealCarcinoma After Intensity Modulated Radiation TherapyR.P. Goepfert,1 J.S. Lewin,1 M.P. Barrow,1 G.B. Gunn,1 C.D. Fuller,1

B.M. Beadle,1 A.S. Garden,1 D.I. Rosenthal,1 M.S. Kies,1

V. Papadimitrakopoulou,1 S. Lai,1 N. Gross,1 D.L. Schwartz,2

and K.A. Hutcheson1; 1The University of Texas MD Anderson Cancer

Center, Houston, TX, 2University of Texas Southwestern Medical Center,

Dallas, TX

Purpose/Objective(s): Clinical trials involving treatment of oropharyn-

geal carcinoma (OPC) increasingly incorporate patient-reported swallow-

ing outcomes as endpoints. The MD Anderson Dysphagia Inventory

(MDADI) represents the most widely used metric for patient-reported

dysphagia. Prospectively collected long-term MDADI results after stan-

dard intensity modulated radiation therapy (IMRT) at prescribed doses

around 70 Gy are lacking for patients with low-risk (human papillomavirus

[HPV]-associated, non-T4/N3) OPC who would be eligible for enrollment

in current national trials (eg, E3311, HN002).

Materials/Methods: A pooled analysis was conducted from 3 single-

institution clinical trials for advanced-stage (III/IV) head and neck carci-

noma. Patients treated with definitive, split-field IMRT with chemotherapy

for OPC were sampled from trial databases. Inclusion criteria were clinical

stage III/IV (T1-2/N1-2b, T3/N0-2b), and MDADI results at baseline and

at least 1 posttreatment timepoint. Patients were excluded if they were

HPVor p16 negative or HPV and p16 unknown with a smoking history of

>10 pack-years such that all included patients were sampled to represent

likely HPV-associated disease. Prospective MDADI composite scores were

analyzed at baseline, 6, 12, and 24 months after treatment. Composite

scores were interpreted as follows: “optimal” >80, “adequate” >60, and

“inadequate” <60. Pairwise comparisons were analyzed with bonferroni

correction for multiple comparisons.

Results: Forty-six patients were included: 41% had tonsil and 59% had

base of tongue primary tumors. Seventy-two percent were HPVand/or p16

positive, and 70% of those were positive for both. All received bilateral

neck irradiation with a mean therapeutic dose of 68.8 Gy (range: 66 to 72

Gy). Seventy-two percent received induction chemotherapy, 57% received

concurrent chemotherapy, and 28% received both. Overall, the mean

baseline MDADI composite score was 90.1, dropping to 74.6 at 6 months

(P<.0001) and rising to 78.5 (P<.0001) and 83.1 (PZ.002) by 12 and 24

months, respectively, representing a clinically meaningful drop in MDADI

scores at 6 months that partially recovers by 24 months. MDADI scores did

not significantly differ at baseline or posttreatment intervals between pa-

tients with T1-2/N1-2b and T3/N0-2b disease (P>.05). Overall, compos-

ites at baseline, 6, 12, and 24 months were optimal for 87%, 40%, 44%,

International Journal of Radiation Oncology � Biology � Physics964

and 62% of patients; adequate for 9%, 49%, 41%, and 29% of patients; and

inadequate for 4%, 11%, 15%, and 9% of patients, respectively.

Conclusion: Low-risk patients with OPC treated with modern split-field

IMRT with laryngeal/esophageal-inlet shielding and chemotherapy are

highly likely to report recovery of adequate or optimal swallowing func-

tion in long-term follow-up. These data serve as a benchmark future trial

design and endpoint interpretation.

Author Disclosure: R.P. Goepfert: None. J.S. Lewin: None. M.P.

Barrow: None. G.B. Gunn: None. C.D. Fuller: None. B.M. Beadle:

None. A.S. Garden: None. D.I. Rosenthal: None. M.S. Kies: None. V.

Papadimitrakopoulou: None. S. Lai: None. N. Gross: None. D.

Schwartz: None. K.A. Hutcheson: None.

341Incidence of Long-Term Gastrostomy Feeding Tube Dependence byPrimary Treatment Modality Among Patients With Squamous CellCarcinoma of the Head and NeckJ.J. Meshman,1 J. Wang,2 E. Abemayor,1 M. St. John,1 A. Mendelsohn,1

D. Wong,1 S. Bhuta,1 R.K. Chin,1 and A.M. Chen2; 1University of

California, Los Angeles, Los Angeles, CA, 2University of California, Los

Angeles- David Geffen School of Medicine, Los Angeles, CA

Purpose/Objective(s): To evaluate the incidence of long-term gastro-

stomy feeding tube dependence according to primary treatment modality

among patients with squamous cell carcinoma of the head and neck treated

at a single institution.

Materials/Methods: The medical records of 136 consecutive patients

treated with definitive intent for locally advanced, nonmetastatic squamous

cell carcinoma of the oropharynx (80%), larynx (14%), and hypopharynx

(6%) were reviewed. The median age was 62 years (range, 22-91). Sev-

enty-eight patients (57%) were human papillomavirus (HPV) positive, and

58 patients (43%) were HPV negative. Patients with gastrostomy feeding

tubes placed prior to treatment were excluded. Gastrostomy feeding tube

dependence rates were calculated at 6 months and 1 year after treatment.

The primary treatment modality was radiation therapy for 104 (76%) pa-

tients to a median dose of 70 Gy (range, 41-70 Gy), of whom 86 (83%)

received concurrent chemotherapy, typically with a cisplatin-based

regimen. The primary treatment modality was surgery for 32 patients

(24%), of whom 20 (63%) were treated with transoral laser microsurgery

(TLM) or transoral robotic surgery (TORS). All surgery patients received

adjuvant radiation therapy to a median dose of 60 Gy (range, 60-69 Gy), 14

(44%) of whom received concurrent chemotherapy. The proportion of

patients who were gastrostomy feeding tube dependent were compared

using the t test, with statistical significance set with a P value of .05.

Results: Overall gastrostomy feeding tube dependence rates at 6 months

and 1 year, respectively, were 15% and 10% after primary chemoradiation,

0% and 8% after primary radiation therapy alone, 0% and 0% after TORS/

TLM, and 42% and 50% after all other surgery. When the analysis was

limited to the subset of patients with HPV-positive tumors, the respective

rates at 6 months and 1 year were 15% and 8% after primary chemo-

radiation, 0% and 0% after primary radiation therapy alone, 0% and 0%

after TORS/TLM, and 60% and 40% after all other surgery (P<.05).

Similarly, for HPV-negative tumors, the respective rates at 6 months and 1

year were 14% and 14% after primary chemoradiation, 0% and 11% after

primary radiation therapy alone, 0% and 0% after TORS/TLM, and 29%

and 67% after all other surgery (P<.05).

Conclusion: Gastrostomy feeding tube dependence rates after definitive

treatment for squamous cell carcinoma of the head and neck appeared to be

lowest for patients treated by primary TLM/TORS and primary radiation

therapy alone. The use of open surgical techniques and chemoradiation

was associated with significantly higher rates of swallowing dysfunction.

While these findings may be limited by selection bias, they may be useful

in patient counseling. Ongoing efforts to deintensify therapy for patients

with HPV-positive tumors appear particularly salient.

Author Disclosure: J.J. Meshman: None. J. Wang: None. E. Abemayor:

None. M. St. John: None. A. Mendelsohn: None. D. Wong: None. S.

Bhuta: None. R.K. Chin: None. A.M. Chen: None.

342Pretreatment Complete Blood Count Improves Prognostic ModelSurvival Prediction in Oropharyngeal Cancer Patients Treated WithRadiation TherapyA.S.R. Mohamed,1 S. Shoultz-Henley,2 A.S. Garden,1 T. Sheu,2 M. Kroll,2

D.I. Rosenthal,1 G.B. Gunn,1 J. Phan,1 S. Lai,1 M.S. Kies,1 K.A. Gold,3

E.M. Sturgis,1 and C.D. Fuller1; 1The University of Texas MD Anderson

Cancer Center, Houston, TX, 2The University of Texas, MD ANderson

Cancer Center, Houston, TX, 3MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): To evaluate the potential prognostic value of

pretreatment complete blood count (CBC) in oropharyngeal cancer pa-

tients receiving concurrent chemoradiation and to determine the potential

added value of laboratory assessment as a potential covariate of univariate

and multivariate overall survival (OS) for pretherapy risk stratification.

Materials/Methods: Oropharyngeal cancer patients (OPC) treated with

intensity modulated radiation therapy (IMRT) with concurrent chemo-

therapy between 2002 and 2012 were included under an approved insti-

tutional review board protocol. CBC data were extracted. Platelet and

hemoglobin from the last phlebotomy (PLTpre-chemoRT, Hgbpre-chemoRT)

before start of treatment were identified. Univariate and multivariate Cox

proportional hazards assessments were performed to determine whether

the following variables were correlated with OS: age, sex, Dahlstrom-

Sturgis schema (American Joint Committee on Cancer TNM staging,

smoking status, age, and site of disease), human papillomavirus (HPV)

status, subsite (tonsil vs nontonsil), platelet elevation, and anemia. After

multivariate analysis, best categorical model selected based on experi-

mental Bayesian information criteria (BIC) were converted into visual

nomograms of 3-, 5-, and 10-year OS prediction.

Results: Four hundred thirty-three OPC patients were identified. The median

follow-up was 69 months. OS was decreased for patients with a PLT pre-

chemoRT value of �350 x 109/L. Actuarial 5-year OS was better for patients

with normal platelet counts by comparison (76% vs 57%; P<.0001). Non-

anemic patients (hemoglobin �13.5 for males and �12.3 for females) like-

wise exhibited comparatively better OS at 5 years (84% vs 56%, P<.0001).

Multivariate analysis showed that platelets�350 x 109/L was an independent

predictor of dismal survival (hazard ratio 1.9; 95% confidence interval 1.2-

2.9; P<.006), in addition to anemia, Dahlstrom-Sturgis category, and HPV

status. On BIC analysis, those 4 variables achieved best performance and

were used to develop nomograms predicting 3-, 5-, and 10-year OS.

Conclusion: Pretreatment CBC data is a promising predictor of survival in

OPC patients treated with radiation therapy. We recommend continued

routine evaluation by care providers of pretreatment platelets and hemo-

globin levels and increased surveillance as appropriate for these patients

who may be at substantively increased risk of mortality and judicious risk

assessment using provided nomograms as an exploratory tool for valida-

tion in other datasets.

Author Disclosure: A.S. Mohamed: None. S. Shoultz-Henley: None. A.S.

Garden: None. T. Sheu: None. M. Kroll: None. D.I. Rosenthal: None.

G.B. Gunn: None. J. Phan: None. S. Lai: None. M.S. Kies: None. K.A.

Gold: None. E.M. Sturgis: None. C.D. Fuller: Research Grant; National

Institutes of Health/National Cancer Institute’s Paul Calabresi Clinical

Oncology Award Program (K12 CA088084-06), a General Electric

Healthcare/MD Anderson Center for Advanced Biomedical Imaging In-

Kind Award; an Elekta AB/MD Anderson Department of Radiation

Oncology Seed Grant.

343Survival Benefit of Chemotherapy in Oropharyngeal Cancer PatientsTreated With Surgery and Postoperative RadiationF.M. Makki, H. Zhang, L. Puttagunta, D. O’Connell, J. Harris, H. Seikaly,

and V.L. Biron; University of Alberta, Edmonton, AB, Canada

Purpose/Objective(s): The benefit of chemotherapy in the postsurgical

treatment of advanced-stage oropharyngeal squamous cell carcinoma is un-

clear in the current literature. This study aims to compare patients treated

Volume 94 � Number 4 � 2016 Posters 965

with primary surgery (PS) followed by radiation (RT) or chemoradiation

(CRT) to determine whether the addition of chemotherapy has a significant

survival advantage. We hypothesized that chemotherapy could have a sur-

vival advantage dependent on p16 status and tobacco smoking history.

Materials/Methods: Advanced-stage OPSCC patients diagnosed and

treated at a single tertiary cancer care center between 1998 and 2009 were

reviewed from a prospectively collected database. Patients who received

PS followed by either RT (PS+RT) or CRT (PS+CRT) were included in the

study. A review of patient records was performed to obtain patient clinical

characteristics as well as detailed pathologic, treatment, and survival data.

P16 status was obtained by immunohistochemistry of a tissue microarray

of surgical tumor specimen. Comparative analyses of survival were per-

formed between patients who received PS+RT and PS+CRT, stratified

according to human papillomavirus/p16 status and tobacco smoking

history.

Results: One hundred seventy-one patients were treated with PS+RT or

PS+CRT, of which 138 had p16 typing and smoking status for inclusion in

our analyses (PS+CRTZ67, PS+RTZ71). In patients treated with

PS+CRT, 73.5% received cisplatinin and 26.5% received carboplatin. In

p16-negative and p16-positive nonsmoking patients, no significant differ-

ences in survival were seen between patients treated by PS+RT versus

PS+CRT. In p16-positive smokers, patients treated with PS+CRT had

significantly higher 5-year disease-specific survival (88.2%) compared to

PS+RT (59.7%). Multivariate analysis showed these survival differences

regardless of patient covariates and factors.

Conclusion: In advanced-stage OPSCC patients treated with PS and RT,

the addition of platinum-based chemotherapy may have improved disease-

specific survival in select patients. Further prospective trials that include

p16 status would be recommended verify this hypothesis.

Author Disclosure: F.M. Makki: None. H. Zhang: None. L. Puttagunta:

None. D. O’Connell: None. J. Harris: None. H. Seikaly: None. V.L.

Biron: None.

344Analyis of Prognostic Factors and Outcomes FollowingParotidectomy for Cutaneous Squamous Cell Carcinoma Metastaticto the Parotid GlandS. Appachi,1 J. Scharpf,2 S. Koyfman,1 J.F. Greskovich, Jr,3

D.J. Adelstein,2 T. Nwizu,2 B.B. Burkey,2 D. Chute,2 M. Fritz,1

and C.A. Reddy2; 1Cleveland Clinic Foundation, Cleveland, OH,2Cleveland Clinic, Cleveland, OH, 3Cleveland Clinic Florida, Weston, FL

Purpose/Objective(s): To determine survival outcomes and prognostic

factors for recurrence of cutaneous squamous cell cancer (cSCC) met-

astatic to the parotid gland following parotidectomy at a tertiary care

center.

Materials/Methods: A retrospective chart review of patients diagnosed

with cSCC metastatic to the parotid gland who subsequently underwent

parotidectomy between 1992 and 2014 was performed. Demographic in-

formation, treatment information, and tumor characteristics were reviewed.

Risk of disease recurrence was analyzed using a competing risks regres-

sion model. Overall survival was analyzed using Cox proportional hazards

regression. Rates of disease recurrence were calculated using cumulative

incidence, and the Kaplan-Meier method was used to calculate rates of

overall survival.

Results: One hundred and six patients were included in this institutional

review boardeapproved analysis. Median follow-up after surgery was

37.0 months (range 0.3-233.8). Ninety-three patients (87.7%) were

male with a median age of 74 years (range 31-93). Nineteen patients

(17.9%) underwent radical parotidectomy, 28 (26.4%) received super-

ficial parotidectomy with facial nerve (FN) preservation, and 59

(55.7%) received total parotidectomy with FN preservation. Seventy-

six patients (71.7%) underwent adjuvant radiation, and 11 (10.4%)

received adjuvant chemoradiation. Nineteen patients (17.9%) suffered

locoregional recurrence, 13 (12.3%) had distant metastases, and 5

(4.7%) had both. One-year and 3-year disease recurrence rates were

32.3% and 37.1%, respectively, and 1-year and 3-year actuarial overall

survival rates were 70.6% and 49.1%, respectively. On multivariate

analysis, FN sacrifice (PZ.047, hazard ratio [HR] 2.0) and nodal

extracapsular extension (PZ.018, HR 2.2) were significantly correlated

with locoregional recurrence or distant metastasis. Age �75 years was

significantly correlated with a worse overall survival (P<.001, HR 2.7)

on multivariate analysis. Other variables, including immunocompro-

mised status, smoking or alcohol history, neck dissection, and nodal

involvement, were not significantly correlated with disease-free or

overall survival.

Conclusion: This study demonstrates the negative prognostic associations

of FN sacrifice and extracapsular extension with disease recurrence

following parotidectomy for metastatic cSCC. These 2 variables may serve

as proxies for aggressiveness of the parotid tumor.

Author Disclosure: S. Appachi: None. J. Scharpf: None. S. Koyfman:

None. J.F. Greskovich: None. D.J. Adelstein: None. T. Nwizu: None.

B.B. Burkey: None. D. Chute: None.M. Fritz: None. C.A. Reddy: None.

345Trismus Dose-Response Assessment Using Quantitative VolumetricMeasures in Head and Neck Radiation TherapyB.W. Warren, A.S.R. Mohamed, K.A. Hutcheson, G.B. Gunn, J.S. Lewin,

S. Lai, D.I. Rosenthal, A.S. Garden, and C.D. Fuller; The University of

Texas MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): The objective of this study was to investigate a

potential dose-toxicity relationship between radiation delivered to muscles

of mastication and subsequent development of trismus in patients receiving

intensity modulated radiation therapy (IMRT) for oropharyngeal squamous

cell carcinoma (OPSCC).

Materials/Methods: Retrospective data from 324 patients treated with

IMRT for OPSCC at 1 institution between 2004 and 2011 were analyzed via

an institutional review boardeapproved protocol. For each patient, clinical

data were extracted from electronic medical records, and incidence of trismus

was recorded as a binary variable. Regions of interest (ROIs), namely

masseter (M), medial pterygoid (MP), and lateral pterygoid (LP) muscles,

were segmented on planning computed tomographic images using an auto-

matic segmentation algorithm and defined as ipsilateral (I) or contralateral

(C) to the primary tumor. Radiation dose plans were then retrieved and dose-

volume histograms (DVHs) generated for ROIs. Dose-response assessment

was performed using logistic regression and Wilcoxon rank sum test with

Bonferroni correction for multiple comparisons. Cumulative group DVHs

were generated to visualize differences in continuous dose distributions be-

tween trismus and asymptomatic patient groups. Dose-threshold candidates

were selected via a bootstrap forest technique.

Results: Twenty-five of 324 patients (7.7%) treated with IMRT developed

trismus. Patient characteristics are as follows: median age, 56 years;

subsites, base of tongue (nZ179; 55.2%) and tonsillar complex (nZ139;

42.9%); chemotherapy agent, cisplatin (nZ282; 87.0%). Cumulative

group DVHs dichotomized by the trismus variable allowed for visualiza-

tion of differences between groups for each ROI across all doses. The top

dose-threshold candidate determined by bootstrap forest was IM V59

(volume of the ipsilateral masseter receiving 59 Gy).

Conclusion: Trismus is a well-known sequela of radiation therapy

treatment for head and neck cancers that negatively impacts quality of

life and presents a problem to both patients and health care providers.

Dosimetric parameters associated with increased probability of trismus

present potential opportunities, through further investigation, for risk

stratification and dose modification to specific structures when

possible.

Author Disclosure: B.W. Warren: None. A.S. Mohamed: None. K.A.

Hutcheson: None. G.B. Gunn: None. J.S. Lewin: None. S. Lai: None.

D.I. Rosenthal: None. A.S. Garden: None. C.D. Fuller: Research Grant;

NIH/National Cancer Institute, SWOG/Hope Foundation, Elekta AB/MD

Anderson Department of Radiation Oncology. In-kind Donation; General

Electric Healthcare/MD Anderson Center for Advanced Biomedical

International Journal of Radiation Oncology � Biology � Physics966

Imaging. Associate Director of MR Programmatic Development; The

University of Texas MD Anderson Cancer Center.

346Comparison of Symptom Interference of Quality of Life inPostradiation Treatment in Early-Stage Versus Late-StageLaryngeal Cancer PatientsB. Temple,1 L.C. Cooksey,2 K. Preston,2 C.D. Fuller,3 A.S.R. Mohamed,3

and G.B. Gunn3; 1Xavier University, New Orleans, LA, 2Abilene Christian

University, Abilene, TX, 3The University of Texas MD Anderson Cancer

Center, Houston, TX

Purpose/Objective(s): Quality of life studies are the benchmark for how

we expect patients to live after they receive treatment. Analyzing trends in

symptoms, particularly after radiation treatment, helps to determine how

treatment implementation should be improved and altered to limit any

long-term symptoms or discomforts. Our lab is attempting to analyze the

quality of life after receiving radiation treatment at 1 institution for

laryngeal cancer by going through a series of questionnaires with laryngeal

cancer patients that will improve treatment delivery and decrease the

likelihood of the treatment causing severe long-term symptoms. Our hy-

pothesis is as follows: The symptoms interfering with quality of life for

postradiation treatment late-stage laryngeal cancer patients are signifi-

cantly higher than are those for postradiation treatment early-stage

laryngeal cancer patients.

Materials/Methods: A symptom inventory questionnaire was collected

from 230 head and neck cancer patients 22 to 24 months following

treatment. Each patient was asked to rate the severity of their symptoms on

a scale from 0 (“not present”) to 10 (“worst possible”). The questionnaire

was administered over the phone and took 10 to 15 minutes to complete

with the each patient. Upon completion, the responses were compiled and

analyzed using Microsoft Excel.

Results: Based on the data, there is a slight difference in symptoms

interfering with quality of life among early-stage versus late-stage post-

treatment cancer patients 22 to 24 months after treatment. Late-stage pa-

tients did show an increase in symptom interference with quality of life.

However, the increase was not as significant as hypothesized. Based on the

symptom inventories conducted, analysis showed dry mouth to be the most

common symptom that interfered in both early-stage and late-stage pa-

tients, while vomiting was the least common.

Conclusion: The overall quality of life for early-stage and late-stage

laryngeal cancer patients was significantly high with a relatively low

measure of symptom severity. The results will continue to be further

analyzed in order to develop more effective treatment plans for head and

neck cancer patients. As the issue of dry mouth, in particular, persists in a

large number of these patients, there is a continuous need of refinement in

this area. Improving quality of life is an important consideration while

treating patients, and this study is part of a continuation of that process.

Author Disclosure: B. Temple: None. L.C. Cooksey: None. K. Preston:

None. C.D. Fuller: None. A.S. Mohamed: None. G.B. Gunn: None.

347Feeding Tube Use in Patients With Salivary Gland MalignancyK.A. Hutcheson,1 E.M. Sturgis,1 L. Xu,1 G.B. Gunn,1 C.D. Fuller,1

A.S.R. Mohamed,1 S. Lai,1 J.S. Lewin,1 and D. Chen2; 1The University of

Texas MD Anderson Cancer Center, Houston, TX, 2Baylor College of

Medicine, Houston, TX

Purpose/Objective(s): This study was undertaken to evaluate feeding

tube use in patients with salivary gland malignancies (SGM).

Materials/Methods: Patients were sampled from an epidemiologic SGM

registry during a 12-year period (August 2001 to November 2013).

Feeding tube history was reviewed from the medical records. Patients with

outside locoregional therapy or palliative treatment were excluded. Route

of enteral feeding (nasogastric versus gastrostomy) and length of depen-

dence were analyzed as a function treatment modality of site of SGM.

Analysis of organ-at-risk (OAR) dose-volume effects on gastrostomy

dependence is underway.

Results: Eighty of 286 patients (28%) required temporary nasogastric

tube feeding during treatment for SGM (median duration: 13 days;

maximum duration: 45 days). Of those 80 patients, 29 (10% of total

SGM cohort) required conversion to percutaneous endoscopic gastro-

stomy tube (PEG; 20 of 29 had temporary PEG). Median PEG duration

was 5.7 months (range: 1-138 months). PEG placement was only

necessary in patients receiving multimodality therapy (P<.001), and

50% of patients with SGM arising from pharyngeal/laryngeal sites

required PEG placement compared to 8% to 19% of SGM arising from

all other sites (PZ.001). At a median follow-up of 2.4 years, 9 (3%) of

all SGM patients were PEG dependent, but 14% (3 of 22) of patients

with laryngeal/pharyngeal SGM subsites treated with multimodality

therapy remained chronically PEG dependent.

Conclusion: While almost 30% of SGM survivors require a temporary

nasogastric tube during treatment, PEG use is uncommon in roughly

10% of SGM overall. PEG use appears exclusive to patients treated with

multimodality therapy, and chronic gastrostomy remains high (14%) in

patients with minor gland cancers arising in the pharynx and larynx,

suggesting impetus for dysphagia prophylaxis in these higher risk

subsets.

Author Disclosure: K.A. Hutcheson: None. E.M. Sturgis: None. L. Xu:

None. G.B. Gunn: None. C.D. Fuller: None. A.S. Mohamed: None. S.

Lai: None. J.S. Lewin: None. D. Chen: None.

348Mental Disorders Are Associated With Poor Disease-SpecificOutcomes for Elderly Patients With Squamous Cell Carcinoma of theLarynxT.P. Schrank, Y. Mikhaylov, and E.J. Lentsch; Medical University of South

Carolina, Charleston, SC

Purpose/Objective(s): Previous studies have found patients with colon

and breast cancer as well as mental disorders to present with later stage

disease, have increased risk of no treatment, and have poorer survival.

Considering the inherent communication difficulties faced by laryngeal

cancer patients and their need for complex/multidiscipliary care, we hy-

pothesize that such patients with mental disorders may be at particularly

high risk for poor outcomes. A paucity of information is available on this

subject; therefore, our objective is to investigate the incidence of mental

disorders in laryngeal cancer as well as their relationship to outcomes.

Materials/Methods: The Surveillance, Epidemiology and End

ResultseMedicare linked database was analyzed. A total of 11,640 cases

met the inclusion criteria of squamous cell carcinoma (SCCa) of the lar-

ynx, diagnosed from 1988 to 2005 in patients aged >65 years. Considering

the etiological role of alcohol and tobacco use in laryngeal cancer, these

disorders were not analyzed as psychiatric conditions. Extracted variables

included age, race, gender, TMN staging, surgical treatment, radiation

therapy, overall survival (OS), and disease-specific survival (DSS). Sur-

vival data were analyzed using the Kaplan-Meier method, log-rank test,

and COX regression where appropriate.

Results: The rate of nonalcohol- and tobacco-related mental disorders was

found to be 22%. Mood disorders were the most common, followed by

psychotic, dementia, and substance disorders. Patients with mental disor-

ders were more likely to be female, 25% versus 17% (P<.001). Patients

with mental disorders also present with higher stage disease. Forty-six

percent versus 51% presented with stage I cancer (PZ.002). In patients

without mental disorders, 44% received surgery, and 74% received radi-

ation. Treatment was similar in patients with mental disorders. Differences

were not significant. OS was decreased in patients with mental disorders

on multivariate analysis including stage, treatment, and available de-

mographic factors; hazard ratio (HR)Z 1.35 (P<.001). Individually, mood

disorders, psychotic disorders, and substance abuse disorders portended

poor outcomes as well. DSS was also found to be decreased with patients

with mental disorders on multivariate analysis, HR Z 1.28 (PZ.02).

Volume 94 � Number 4 � 2016 Posters 967

Conclusion: Mental disorders are common in laryngeal cancer patients.

While controlling for tumor and demographic factors, mental disorders

were associated with poorer DSS in elderly patients with SCCa of the

larynx. Although further studies are needed to elucidate the origin of this

disparity we feel that our data support a role for psychiatric treatment as

well as increased vigilance and mobilization of social resources to ensure

timely complete cancer treatment of this at-risk population.

Author Disclosure: T.P. Schrank: None. Y. Mikhaylov: None. E.J.

Lentsch: None.

349Device Life of the Tracheoesophageal Voice Prosthesis RevisitedJ.S. Lewin, L.M. Baumgart, M.P. Barrow, D.A. Barringer,

and K.A. Hutcheson; The University of Texas MD Anderson Cancer

Center, Houston, TX

Purpose/Objective(s): Voice prosthesis (VP) device life is a limiting

factor of tracheoesophageal (TE) voice restoration that drives patient

satisfaction, health care costs, and overall burden. Historic data suggest

that TE VPs have an average device life of 3 to 6 months, but these data are

typically derived from small samples using only 1 or 2 devices. We sought

to re-examine device life in a large, contemporary US hospital that uses a

wide assortment of VPs.

Materials/Methods: A retrospective study was completed using the re-

cords of 392 patients who underwent a total laryngectomy (TL) with a TE

puncture (TEP) from July 1, 2003, to December 31, 2013 at MD Anderson

Cancer Center (MDACC). Duration was examined using Kaplan-Meier

analysis. Disease, treatment, and patient-specific factors were analyzed as

predictors of duration.

Results: A total of 3643 VPs were placed in the 392 patients. Nearly half

of the patients (46%) underwent salvage laryngectomy after RT failure,

and 29% required pharyngeal reconstruction. Sixty-four percent underwent

primary TEP. Indwelling prostheses accounted for more than half (56%) of

the devices placed (55%: 20 Fr diameters; 33%: 8 mm). More than two-

thirds of prostheses (69%) were removed because of leakage, while the rest

were removed due to other complications or causes. Median device life

after TL due to leakage was 61 days for all prostheses, with a median of 70

days for indwelling types and 38 days for nonindwelling VPs. Neither

radiation history nor extent of surgery significantly impacted device life

(P>.05).

Conclusion: Our data suggest that VP duration demonstrates overall a

slightly lower durability than historically reported. The etiology of this is

unclear. A lower range of VP durability may be a function of different

scientific methodologies but more likely reflects the intensification of

current treatment regimens in an already complicated TEP population

within an era of organ preservation.

Author Disclosure: J.S. Lewin: None. L.M. Baumgart: None. M.P.

Barrow: None. D.A. Barringer: None. K.A. Hutcheson: None.

350Cognitive Function and Patient-Reported Memory ProblemFollowing Radiation Therapy for Cancers at the Skull Base: ASurvivorship Study Using the Telephone Interview for CognitiveStatus and the MDASI-HNC.C. Hansen,1 K. Chrane,2 G.B. Gunn,3 A.S.R. Mohamed,4

D.I. Rosenthal,4 J.S. Wefel,4 J. Phan,4 S.J. Frank,4 A.S. Garden,4

B. Smith,4 H. Eichelberger,5 C. Anderson,6 C. McCoy,2 M. Horiates,7

C. Patrick,6 S. Floris,2 C. French,5 B.M. Beadle,4 W.H. Morrison,4

S.Y. Su,4 E.Y. Hanna,4 C.M. Lewis,4 H.D. Skinner,4 S. Lai,4

and C.D. Fuller3; 1Texas Tech University Health Science Center School of

Medicine, Lubbock, TX, 2Abilene Christian University, Abilene, TX, 3MD

Anderson Cancer Center, Houston, TX, 4The University of Texas MD

Anderson Cancer Center, Houston, TX, 5The University of Texas Medical

School, Houston, TX, 6The University of Texas, Austin, TX, 7The University

of Texas, MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): Objective cognitive decline is a well-recognized,

radiation-attributable sequela of intracranial tumor (where brain paren-

chyma is a de facto target) irradiation and prophylactic cranial treatment.

Conversely, data are comparatively sparse for extracranial tumors of

nasopharynx and skull base origin. For these tumors, radiation therapy,

while increasingly conformal, still brings the potential for substantial

“beam path” dose to adjacent nontarget CNS structures. Using 2 accepted

and validated instruments, we sought to quantify the relative incidence and

severity of detectable objective cognitive symptoms observed using a

prospectively collected cross-sectional survey in a robust dataset

comprised of patients with tumors physically abutting nasopharyngeal or

skull-base structures that received radiation therapy.

Materials/Methods: Head and neck cancer survivors were recruited to

participate in this institutional review boardeapproved prospective cross-

sectional study. All participants gave informed consent, completed radia-

tion treatment at MD Anderson Cancer Center between January 2001 and

December 2013, and were subsequently administered the Telephone

Interview for Cognitive Status (TICS), a reliable cognitive impairment

assessment tool, and the MD Anderson Symptom Inventory-Head and

Neck module (MDASI-HN), a patient-reported multisymptom assessment

tool, simultaneously. Patient, disease, and treatment characteristics along

with TICS and MDASI-HN results were analyzed using nonparametric

statistics.

Results: A total of 131 survivors participated in this study. Fifty-six

percent were male, median age was 56 years, and median follow-up was 66

months. Most were nasopharyngeal cases (47%), followed by paranasal

sinuses (25%), nasal cavity (16%), and skull base (7%). The majority

(58%) of patients treated with modern-era radiation techniques showed no

frank detectable cognitive impairment; however, a minority of patients

showed either ambiguous (36%) or frank (6%) impairment on TICS

assessment, suggesting more granular evaluative tests may be indicated.

Additionally, mean MDASI-HNmemory was 3.4 for the entire cohort (range

0-10; SD � 2.71; SEM � 0.24).

Conclusion: MDASI-HNmemory was observed to be a clinically relevant

screening instrument to screen for frank cognitive impairment. Further

research in this area is warranted.

Author Disclosure: C.C. Hansen: None. K. Chrane: None. G.B. Gunn:

None. A.S. Mohamed: None. D.I. Rosenthal: None. J.S. Wefel: None. J.

Phan: None. S.J. Frank: None. A.S. Garden: None. B. Smith: None. H.

Eichelberger: None. C. Anderson: None. C. McCoy: None.M. Horiates:

None. C. Patrick: None. S. Floris: None. C. French: None. B.M. Beadle:

None. W.H. Morrison: None. S.Y. Su: None. E.Y. Hanna: None. C.M.

Lewis: None. H.D. Skinner: None. S. Lai: None. C.D. Fuller: Research

Grant; the Paul Calabresi Clinical Oncology Award Program, NIH/NCI

Clinician Scientist Loan Repayment Program. Honoraria; the SWOG/Hope

Foundation Dr. Charles A. Coltman, Jr., Fellowship in Clinical Trials, a

General Electric Healthcare/MD Anderson Center for Advanced

Biomedical Imaging In-Kind Awardan Elekta AB/MD Anderson Depart-

ment of Radiation Oncology Seed Grant: the Center for Radiation

Oncology Research at MD Anderson Cancer Center, the MD Anderson

Institutional Research Grant Program.

351A Community Hospital’s Approach to Addresssing Lymphedema inHead and Neck Cancer PatientsR. Hoffman,1 A. Sember,2 and C. Pranskevich3; 1Valley Radiotherapy

Associates, Burbank, CA, 2Providence Saint Josephs Medical Center,

Burbank, CA, 3Providence Saint Josephs, Burbank, CA

Purpose/Objective(s): A recent retrospective study estimated that greater

than 50% of treated head and neck cancer patients will develop lymphe-

dema and suggests that this condition is both underrecognized and

undertreated. Presented here is one community hospital’s approach to

addressing this treatment-related morbidity that illustrates how early

intervention makes a difference.

Materials/Methods: The medical records of 286 head and neck cancer

patients who have been treated by this program between June 2008 and

International Journal of Radiation Oncology � Biology � Physics968

June 2015 were reviewed. The following information was extracted from

each patient’s chart: whether they developed lymphedema, tumor stage,

whether they had surgery, dose of radiation received, type of chemotherapy

given, smoking history, and whether they had a neck dissection. All pa-

tients with documented lymphedema received complete decongestive

therapy. Patient response was assessed wtih digital photographs taken at

each visit and more recently incorporated the Neck Disability Index.

Results: Of the 286 patients, 27% were referred for lymphedema man-

agement. Of those referred, 67% had stage 3 or 4 cancer, 49% had not had

surgery (illustrating the fact that lymphedema occurs even in the absence

of surgery), and 85% had received in excess of 6000 cGy. Neither the type

of chemotherapy nor the patient’s smoking status correlated with the risk

of lymphedema. There was an increased risk of lymphedema with neck

dissection. Also evident was that increased awareness of this condition

among health care providers resulted in more appropriate patients being

seen earlier as part of a prehabilitation program.

Conclusion: Head and neck cancer patients with stage 3 and 4 disease,

those who have received in excess of 6000 cGy, and those who have had a

neck dissection are at an increased risk for developing lymphedema.

Digital photography and Neck Disability Index have been helpful in terms

of assessing each patient’s response, but its subjective nature points to a

glaring need for a more effective tool for measurement. With only 27% of

our patients being referred for lymphedema management, there is an

obvious need for better education in terms of identifying and treating

patients at risk for this common treatment-related side effect. Such

increased awareness will lead to patients being evaluated and treated

earlier as a component of a prehabilitation program, which will likely

result in better treatment outcomes.

Author Disclosure: R. Hoffman: None. A. Sember: None. C. Pranske-

vich: None.

Abstract 353; Table 1 Reported survival and tumor response outcomes bytreatment regimen component.a

Patientpopulation Cetuximab Docetaxel Methotrexate Paclitaxel

Platinum-naıveMedian OS,

months9.2 7.4 NR 8.0e11.1

(nZ2 trials)Median PFS,

months4.2 3.8 NR NR

RECIST-ORR, % NR NR NR NRPlatinum-refractoryMedian OS,

months4.3e11.8(nZ4 trials)

6.7 5.2 4.3e11.7(nZ2 trials)

Median PFS,months

1.4e3.8(nZ3 trials)

3.1 2.1 2.0e4.9(nZ1 trial;

2 populations)RECIST-ORR, % 9.7e10.7

(nZ2 trials)10.7 1.1 NR

NR, not reported. a Unless otherwise noted, median OS, median

PFS, and ORR data were based on individual trials.

352Lower Cranial Neuropathy After Oropharyngeal IntensityModulated Radiation Therapy (IMRT): A Prospective Study andLiterature ReviewK.A. Hutcheson,1 H. Lin,1 M.M. Yuk,1 R.M. Hubbard,1 M.P. Barrow,1

G.B. Gunn,1 C.D. Fuller,1 A.S. Garden,1 S. Lai,1 D.L. Schwartz,2

M.S. Kies,1 E.Y. Hanna,1 R.P. Goepfert,1 and J.S. Lewin1; 1The University

of Texas MD Anderson Cancer Center, Houston, TX, 2University of Texas

Southwestern Medical Center, Dallas, TX

Purpose/Objective(s): Lower cranial neuropathies (LCNP) are a rare, butfunctionally devastating late effect of radiation therapy typically reported

in nasopharyngeal cancer survivors. Despite a growing number of long-

term survivors, limited data examine these neuropathies after oropharyn-

geal cancer (OPC). The objective of this paper was to examine incidence,

time to event, and details of LCNP in long-term OPC survivors treated

with intensity modulated radiation therapy (IMRT) at standard prescribed

doses around 70 Gy. A supporting literature review was also conducted.

Materials/Methods: A pooled dataset was analyzed from 2 single-insti-

tution clinical trials with institutional review boardeapproved prospective

functional outcomes assessment. Patients treated with definitive IMRT for

locoregionally advanced OPC were sampled from trial databases. Pro-

spective functional analysis included radiographic swallow studies, clinical

cranial nerve examination, and questionnaires before, 6, 12, and 24 months

after treatment. Distributions of time-to-event LCNP events were esti-

mated by the Kaplan-Meier method. A literature review was conducted

between 1977 and 2015 to summarize published LCNP outcomes.

Results: Fifty-nine OPC survivors with a minimum of a 2-year disease-

free clinical follow-up after IMRT (66-72 Gy) with systemic therapy were

included. Three patients developed delayed hypoglossal palsy ipsilateral to

the index oropharyngeal tumor with a median latency of 6.7 years (range:

4.6-7.6 years). At a median follow-up of 5.7 years, cumulative incidence of

delayed lower cranial neuropathy was 5%; 5- and 7-year LCNP rates were

2.1% and 6.1%, respectively. Delayed LCNP preceded progressive

dysphagia (per fluoroscopy) in all 3 cases. Results of 27 articles reporting

LCNP as a result of definitive radiation therapy or chemoradiation therapy

for head and neck cancer were summarized (12 case reports, 6 case series,

and 9 retrospective cohorts). Median incidence rate of LCNP was 10.5% in

long-term survivorship of NPC. No OPC-specific incidence studies were

identified in published data.

Conclusion: While rare, the potential for late onset cranial neuropathies

precipitating swallowing deterioration highlights the importance of long-

term functional surveillance in OPC survivorship.

Author Disclosure: K.A. Hutcheson: None. H. Lin: None. M.M. Yuk:

None. R.M. Hubbard: None. M.P. Barrow: None. G.B. Gunn: None.

C.D. Fuller: None. A.S. Garden: None. S. Lai: None. D. Schwartz:

None. M.S. Kies: None. E.Y. Hanna: None. R.P. Goepfert: None. J.S.

Lewin: None.

353Survival and Tumor Response Outcomes in Recurrent or MetastaticSquamous Cell Carcinoma of the Head and Neck (SCCHN): ASystematic Literature ReviewJ. Signorovitch,1 N. Li,1 C. Zhang,1 J.W. Shaw,2 H. Dastani,2

and L. Orsini2; 1Analysis Group, Boston, MA, 2Bristol-Myers Squibb,

Lawrenceville, NJ

Purpose/Objective(s): Patients with recurrent or metastatic SCCHN,

particularly those with platinum-refractory disease, have treatment options

with limited efficacy. This study reviewed and summarized results from

systematically identified clinical studies reporting overall survival (OS),

progression-free survival (PFS), or objective response rate (ORR) among

patients with recurrent or metastatic SCCHN treated with commonly

available therapies.

Materials/Methods: Eligible English-language publications (published

between January 2005 and October 2014) were identified using EMBASE,

MEDLINE, and the Cochrane Library. Studies included randomized and

single-arm clinical trials and observational studies. Data were extracted

into standardized tables by 2 reviewers.

Results: A total of 38 publications reporting clinical outcomes from 35

studies across all lines of therapy were identified. OS, PFS, and ORR

measured using Response Evaluation Criteria in Solid Tumors (RECIST)

were reported in 35, 22, and 16 studies, respectively. When stratified by

prior treatment, median OS ranged from 7.4 to 11.1 months for platinum-

naıve patients and 4.3 to 11.8 months for patients with platinum-refractory

disease. Median PFS ranged from 2.7 to 4.2 months for platinum-naıve

patients and 1.4 to 8.6 months for patients with platinum-refractory dis-

ease. ORR was not reported for platinum-naıve patients and ranged from

1.1% to 10.7% for patients with platinum-refractory disease. OS, PFS, and

ORR outcomes stratified by commonly used treatment regimens, including

regimens containing cetuximab, docetaxel, methotrexate, and paclitaxel,

are summarized in Table 1.

Volume 94 � Number 4 � 2016 Posters 969

Conclusion: In clinical studies of patients receiving treatment for recurrent

or metastatic SCCHN, OS, PFS, and ORR outcomes have varied across

study populations. However, median OS was consistently shorter than 1

year, and median PFS was consistently shorter than 5 months. Median OS

and median PFS did not differ systematically among populations receiving

regimens containing cetuximab, docetaxel, methotrexate, or paclitaxel,

indicating an unmet need for treatment options that can delay progression

and prolong survival.

Author Disclosure: J. Signorovitch: Research Grant; Bristol-Myers

Squibb. N. Li: Research Grant; Bristol-Myers Squibb. C. Zhang: Research

Grant; Bristol-Myers Squibb. J.W. Shaw: Stock; Bristol-Myers Squibb. H.

Dastani: None. L. Orsini: None.

354A Pilot Study of Electronic Quality of Life Assessments UsingTablet Devices During and After Treatment of Head and NeckCancersE. Wang,1 E. Pollom,1 T. Bui,1 G. Ognibene,1 R. von Eyben,1 V. Divi,2

J. Sunwoo,2 M. Kaplan,2 A.D. Colevas,3 Q.T. Le,1 and W. Hara1; 1Stanford

Radiation Oncology, Stanford, CA, 2Stanford Otolaryngology, Stanford,

CA, 3Stanford Medical Oncology, Stanford, CA

Purpose/Objective(s): Electronic qualify of life (QOL) data collection is

increasingly used to improve efficiency, accuracy, and accessibility of

information. Efficient QOL assessment is particularly crucial in head and

neck oncology, where practice is constantly advancing to improve not only

cancer survival but also QOL, and the use of tablets is appealing given

their easy mobility and larger screen size compared to hand held devices.

However, because there is likely significant variation of patient comfort

with this platform, it is important to assess its relative usability and

feasibility when used for QOL assessments, especially in head and neck

cancer patients who may have social challenges.

Materials/Methods: Patients at our institution who were 18 years or older

with a pathological diagnosis of head and neck cancer were prospectively

enrolled. Each patient completed 2 questionnaires (EORTC-QLQ-C30 and

EORTC-QLQ-H&N35) administered on an iPad at initial consult, during

treatment, at the completion of treatment, and at each subsequent follow-

up visit for 1 year after treatment (spaced approximately 3 months apart).

Descriptive analysis was performed on patient and disease characteristics,

time for survey completion, and survey responses. The time for survey

completion data was analyzed in a repeated-measures model to account for

within-patient correlations. Statistical analysis was performed using SAS

with P values <.05 considered significant.

Results: A total of 50 patients were included in this study. All patients

underwent radiation treatment for their head and neck cancers, with the

majority of patients (66%) also receiving concurrent chemotherapy.

Although all patients completed the surveys at the initial consult, 86% of

enrolled patients completed surveys at the end of radiation treatment, and

48% of enrolled patients completed surveys by the fourth follow-up visit.

Average time to complete the survey for all patients over all timepoints

was 9.8 minutes (standard deviation, 6.1). Age as a continuous variable

was significantly associated with time for survey completion (P<.001),

with older age associated with longer survey completion times. Patients

who were 70 years or older had a mean survey completion time of 12.0

minutes (standard deviation, 7.6 minutes). There was no significant dif-

ference in time for survey completion between veterans and nonveterans

(PZ.9).

Conclusion: QOL assessment using tablet devices in head and neck cancer

patients is feasible in our endpoints of time to completion of survey and

compliance rates. However, elderly patients may benefit from more

assistance with electronic forms and should be allotted more time for

completing tablet-based QOL surveys. Our results highlight the increased

efficiency and accuracy of tablet-based electronic QOL data collection and

its potential utility in clinical trial design in head and neck cancer.

Author Disclosure: E. Wang: None. E. Pollom: None. T. Bui: None. G.

Ognibene: None. R. von Eyben: None. V. Divi: None. J. Sunwoo: None.

M. Kaplan: None. A.D. Colevas: None. Q. Le: None. W. Hara: None.

355Does Frequent Follow-up After Radical Treatment of OropharyngealCancer Improve Outcomes?A. Sundaramurthy,1 K.L. MacLennan,2 and I. Fragkandrea-Nixon2;1Edinburgh Cancer Centre, Edinburgh EH4 2XU, United Kingdom,2Edinburgh Cancer Centre, Edinburgh, United Kingdom

Purpose/Objective(s): Radiation therapy (RT) remains the standard

treatment for oropharyngeal cancers (OPC) both in terms of survival as

well as functional outcome. The highest risk of recurrence is within the

first 2 years, and hence the rationale for very close follow-up. We aim to

report on our patients who relapsed following radical radiation-based

treatment and intend to find whether a close follow-up improves survival.

Materials/Methods: We looked at an unselected group of 142 patients

treated with either concurrent chemoradiation therapy (CRT) or radiation

therapy alone (RT) for OPC at Edinburgh Cancer Centre (ECC) during

January 2006 to December 2010. Relapsed patients in this cohort were

identified and analyzed.

Results: There were a total of 13 relapses, of whom 6 were local and 7 had

distant disease. Nine patients had CRT, while 4 had RT. All but 1 patient

had locally advanced disease. Neoadjuvant chemotherapy (NAC) was used

in only 1 of the patients. The median time to relapse was 10.6 months

(range, 8.1-16.8). In this cohort, 5 patients either never smoked or were ex-

smokers, and 8 were current smokers. The current smokers had an early

relapse duration with a median of 8.0 months (range, 7.5-25) compared

those who never smoked or ex-smokers (median time to relapse 13.1

months; range, 8.5-50.5). The median RT duration was 45 days, with 9 out

of 13 patients completing radiation within 46 days. Patients who

completed radiation within 46 days had a longer median time to relapse

compared the group who took longer (10.6 months vs 8.5 months, c2

0.7436, PZ.38). One out of 6 local relapsed patients was eligible for

salvage surgery and lived for 9.5 months after surgery. The median survival

for all patients from diagnosis of relapse was 3.7 months (95% confidence

interval, 1.1-7.7).

Conclusion: Though the low numbers preclude a significant statistical

analysis, the trends suggest early relapse in smokers and those who had

prolonged radiation therapy duration. All the relapses happened within the

first 2 years of treatment. Though a very close follow-up of 142 patients

resulted in a longer survival for the 1 patient who had salvage surgery, the

impact of best supportive measures the other relapsed patients benefitted

from could not be measured from these data. Hence, we recommend

continuing the current practice of close follow-up as well as further

widening the number of patients to make robust decisions.

Author Disclosure: A. Sundaramurthy: None. K.L. MacLennan: None.

I. Fragkandrea-Nixon: None.

356Holistic Wellness Intervention for Head and Neck Cancer (HNC)Patients and Caregivers: Mayo Clinic Rochester (MCR) PilotK. Price, D.L. Price, L.M. Dierks, D.J. Rhodes, J.A. Ferguson,

S.M. Peterson, J.T. Hein, A.L. Murad, M.M. Clark, and K.D. Olsen; Mayo

Clinic, Rochester, MN

Purpose/Objective(s): Patients treated for HNC experience a high burden

of long-term treatment-related toxicity and medical comorbidities, yet few

programs have included interventions for healthy living, which have the

potential to improve overall survival and reduce treatment-related side

effects. This is a report of a pilot wellness program for patients with HNC

and their caregivers.

Materials/Methods: The program was developed and implemented

through collaboration with physicians specializing in HNC and the staff of

the Dan Abraham Healthy Living Program (DAHLP), a state-of-the-art

wellness center at MCR. The multidisciplinary team consisted of medical

oncologists, HNC surgeons, internists, dietitians, physical therapists,

health psychologists, and certified wellness coaches. The 1-day program

took place in the wellness facilities of the DAHLP at no cost to

International Journal of Radiation Oncology � Biology � Physics970

participants. Pre- and postintervention surveys were conducted to assess

the needs of the population and the overall satisfaction with the program.

Results: The pilot program took place on May 18, 2015. Target enrollment

was 30 patients and caregivers; 30 people registered and 28 attended (15

men, 13 women; 20 survivors, 8 caregivers). Most of the HNC survivors

were >2 years from diagnosis and >1 year from their last cancer treat-

ment. The key components of the program included (1) motivational

welcome by physician staff; (2) physical activity programs including

gentle yoga, NEAT (non-exercise activity thermogenesis), and a physical

therapy session entitled “Open & Release Your Neck & Shoulders”; (3)

resiliency participatory group sessions for both survivors and caregivers;

and (4) a hands-on nutrition/cooking session demonstrating strategies to

increase consumption of vegetables and fruits by making whole food

smoothies and soup. Twenty-one participants responded to the pre-

intervention survey. The top 2 reasons for participation were to improve

overall health and increase energy. Other common reasons included to

increase strength, increase flexibility, improve quality of life or longevity,

and reduce weight or improve body composition. Twenty-six of the 28

participants responded to the postintervention survey. Sixty-one percent

and 33% of patients were very satisfied and satisfied, respectively, and 1

person (6%) was neutral. Eighty-eight percent and 12% of caregivers were

very satisfied and satisfied, respectively. No participants were dissatisfied.

On the postintervention survey, the following components were rated most

valuable: gentle yoga, NEAT, and the nutrition session. Physician referral

was cited as an important motivator to participation.

Conclusion: A comprehensive HNC-specific wellness intervention is

feasible in this underserved patient population, and further programs for

HNC patients should be developed.

Author Disclosure: K. Price: None. D.L. Price: None. L.M. Dierks:

None. D.J. Rhodes: None. J.A. Ferguson: None. S.M. Peterson: None.

J.T. Hein: None. A.L. Murad: None. M.M. Clark: None. K.D. Olsen:

None.

357Osteoradionecrosis After Radiation Therapy for Salivary GlandMalignanciesJ.R. Tucker, L. Xu, E.M. Sturgis, T.M. Hofstede, M.S. Chambers,

G.B. Gunn, C.D. Fuller, A.S.R. Mohamed, S. Lai, and K.A. Hutcheson;

The University of Texas MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): The present study was undertaken to evaluate

osteoradionecrosis (ORN) in patients with salivary gland malignancies

(SGM) after treatment with radiation therapy.

Materials/Methods: The medical records of 172 patients treated with

conformal radiation therapy for SGM during a 12-year period (August

2001 to November 2013) were reviewed. Incidence, time to event, staging,

and management of ORN were analyzed. Analysis of mandibular dose-

volume effects on ORN is underway.

Results: Seven of the 172 patients (4%) developed ORN (median latency:

19 months, range: 4-72 months). Of those 7 patients, 4 required major

surgery, 1 required hyperbaric oxygen therapy (HBO), 1 required minor

debridement, and 1 required conservative management. Total radiation

dose varied from 50 Gy (1 case) to 70 Gy (1 case) among those patients

who developed ORN. Three of the 7 cases of ORN occurred after trau-

matic injury to the bone. Of the 7 patients who developed ORN, 3 had

SGM of the major glands, 3 had other sites of the oral cavity or

oropharynx, and 1 had a sinonasal location.

Conclusion: While the rate of ORN associated after radiation therapy for

SGM was lower (4%) than previously published data on patients with

squamous cell carcinomas of the head and neck treated with radiation

therapy (8% to 14%), clinically significant ORN necessitating major sur-

gery should not be ignored as a possible late effect of radiation therapy in

SGM survivors. Almost all cases that developed ORN had SGM of the

major glands or minor gland arising in the oral cavity or oropharynx.

Author Disclosure: J.R. Tucker: None. L. Xu: None. E.M. Sturgis: None.

T.M. Hofstede: None. M.S. Chambers: None. G.B. Gunn: None. C.D.

Fuller: None. A.S. Mohamed: None. S. Lai: None. K.A. Hutcheson:

None.

358Development of a Head and Neck Multidisciplinary PatientEducation BinderV.M. Diavolitsis,1 V. Heinke,2 E. Lucarelli,3 T. Teknos,4 E. Ozer,2

A. Agrawal,2 M. Old,2 A.D. Bhatt,1 D. Blakaj,1 L. Moore,3 L. Earich,3

D. Niekro,3 A. Krall,3 L. Arrese,3 L. Ray,3 J. Hendershott,2 M. Gillison,5

and J.C. Grecula1; 1The James Cancer Hospital and Solove Research

Institute, Wexner Medical Center at The Ohio State University, Department

of Radiation Oncology, Columbus, OH, 2The Ohio State University Wexner

Medical Center, Columbus, OH, 3Ohio State University, Columbus, OH,4The James Cancer Hospital and Solove Research Institute, Wexner

Medical Center at The Ohio State University, Department of

Otolaryngology-Head and Neck Surgery, Columbus, OH, 5The Ohio State

University, Columbus, OH, United States

Purpose/Objective(s): To improve the patient education resources, con-

tinuity of care, adherence to treatment plan, and interdisciplinary

communication for head and neck cancer patients at our institution.

Materials/Methods: Through surveys of head and neck cancer patients

under treatment, survivors, and the multidisciplinary team, a need was

identified for improved systematic patient education materials. A multi-

disciplinary group including physicians, social workers, speech language

pathologists, dietitians, nurse practitioners, pharmacists, physical thera-

pists, and nurses was convened in May 2013 to execute the project. Using

focus groups comprised of head and neck survivors and support persons, a

patient education binder and tote bag were determined to be the ideal

distribution method. Patient education materials were covered by the pa-

tient education group at our medical center. Total cost of binders and tote

bags were $10.50 per patient plus the staffing time spent educating

patients.

Results: After an inventory of existing patient education materials for this

population was made, current resources were revised and additional tools

were created. Materials were grouped into the following tabs: General

Information, Surgery, Radiation, Chemotherapy, Nutrition, Support and

Resources, and Survivorship. Business card holders and binder pockets

were also included. Staff members were educated on how to distribute the

binder and educate patients on the binder materials. Patient and staff

feedback were overwhelmingly positive. A random sample of patients

given these patient education materials was surveyed to evaluate the

effectiveness of this tool. All surveyed patients indicated that the binder

was organized in a way that was easy to read. Ninety percent of patients

felt that the binder enhanced communication with the medical team. All

patients surveyed responded positively when asked if they were satisfied

with the information given in the binder. Seventy percent of patients felt

that the binder was useful in managing their own health. To date, >250

binders and tote bags have been distributed with patients routinely refer-

encing the binders throughout the course of their treatment and survivor-

ship as well as using the tote bag for transportation of supplies and

nutritional supplements.

Conclusion: Creation and use of a patient education binder and tote bag

was helpful for patients undergoing head and neck cancer treatment at our

institution.

Author Disclosure: V.M. Diavolitsis: Employee; ohio state university. V.

Heinke: None. E. Lucarelli: None. T. Teknos: None. E. Ozer: None. A.

Agrawal: None. M. Old: None. A.D. Bhatt: Employee; Ohio State Uni-

versity. D. Blakaj: None. L. Moore: None. L. Earich: None. D. Niekro:

None. A. Krall: None. L. Arrese: Employee; Ohio Health. L. Ray: None.

J. Hendershott: None. M. Gillison: Employee; ohio state university. J.C.

Grecula: None.

International Journal of

Radiation Oncology

biology physics

www.redjournal.org

INDEX OF PRESENTING AUTHORS

A

Adelstein, David 162Adeniran, Oladapo 150Adkins, Douglas 9Advani, Sunil 292Agarwal, Seema 266Ahn, Peter 132Albergotti, William 114Allan, Eric 111, 131Alsafadi, Nabil 186Amini, Arya 100Amoils, Misha 315Ampil, Federico 245Amsbaugh, Mark 158, 205Anderson, Carryn 11, 168Appachi, Swathi 344

B

Bahig, Houda 226Bann, Darrin 278Baschnagel, Andrew 274Bastos, Bruno 250Bodmann, Joanna 195Britt, Christopher 165Brown, Elizabeth 134

C

Cabanillas, Maria 317Campbell, Shauna 231Cappello, Zachary 194Cass, Lauren 333Caudell, Jimmy 130Cavanaugh, Brittany 216Chadha, Juskaran 149Chang, John 4Chen, Allen 334Chera, Bhisham 5, 152Chintakuntlawar, Ashish 294Choi, Mehee 320Choi, Yongsug 244Churilla, Thomas 12Cohen, Ezra 10Cooksey, Luke 346Cracchiolo, Jennifer 103

D

Dagan, Roi 305Day, Daphne 297Damstrup, Lars 280Dave, Eesha 252de Souza, Jonas 331Debenham, Brock 264

Diavolitsis, Virginia 358Divi, Vasu 207Dong, Tuo 157Dorius, Timothy 198Dorth, Jennifer 246Dua, Bharat 189Dunlap, Neal 190

E

Elliott, David 128Enderling, Heiko 180

F

Fakhry, Carole 101, 106Fayek, Ihab 318Ferris, Robert 8Finegersh, Andrey 283Frakes, Jessica 6

G

Gaines, Katherine 337Gamez, Mauricio 133, 316Ganly, Ian 295Gawande, Shailesh 291Gebhardt, Brian 319Geiger, Jessica 254Giambattista, Joshua 184Gilden, Daniel 174, 258Gill, Beant 309Glaser, Scott 304Godse, Neal 285Goepfert, Ryan 338, 340Greskovich, John 140Gunn, G. 330

H

Hakansson, Katrin 176Hamilton, Thomas 265Hanna, Glenn 307Harr, Bridgett 155hashemi Sadraei, Nooshin 275Herman, Michael 259Hessel, Amy 137Hilbert, Ashley 200Hoffman, Rex 240Hu, Kenneth 175Hutcheson,Katherine 347, 352

I

Inokuchi, Haruo 171

J

Jackson, Matthew 104James, Joshua 147

Jegadeesh, Naresh 151Jhavar, Sameer 277Johnson, Catherine 160Joshi, Nikhil 113

K

Kabarriti, Rafi 238Kamdem, Donatien 248Kanwar, Aasheesh 326Karam, Sana 253Karni, Ron 247, 282Kass, Jason 170Khan, Adeel 177Khan, Nazir 156Kil, Whoon Jong 197Kim, Edward 299Kjems, Julie 105Koyfman, Shlomo 179Kung, Sunny 323

L

Lee, Irwin 145Lee, Jae 251Lee, Victor 204Lewin, Jan 349Li, BaoQing 335Lukens, John 118Luna-Ortiz, Kuauhyama 201

M

Makki, Fawaz 343Manyam, Bindu 298Mao, Yanping 112Margalit, Danielle 102, 208Martin, Daniel 143Martin, Geoffrey 256Massa, Sean 328, 336McDermott, Jessica 312McDonald, Mark 255Mehta, Vikas 302Melotek, James 3, 7, 116, 120Meshman, Jessica 199, 249,

341

Messer, Jay 224Mikhailov, Alexey 173, 263Miller, Daniel 293Miller, Eric 321Mohamed, Abdallah 236, 342,

350

Molitoris, Jason 154Monroe, Alan 325Moreira, Jonathan 187

Mowery, Yvonne 227Murphy, Colin 110

N

Narveson, Lisa 262Nedzi, Lucien 125Neskey, David 169Niska, Joshua 166Nocon, Cheryl 188

O

O’Donnell, Barrett 210Okwan-Duodu, Derick 322Orton, Andrew 219Osman, Abdullah 270Ozbun, Michelle 273

P

Palmer, Joshua 303Parvathaneni, Upendra 167,

296

Patil, Snehal 313Patil, Yash 141, 218Perez, Cesar 161Perni, Subha 220Pham, Brian 163Phan, Jack 260Pinheiro, A. Daniel LB1Platek, Alexis 123, 181, 182,

211

Porosnicu, Mercedes 192Price, Katharine 356Prochnow, Sebastian 276

R

Rabinowits, Guilherme 129Raghavan, Govind 241Rao, Shyam 153Rashid, Yasmin 217Rasmussen, Gregers 332Rasmussen, Jacob 229Roberts, Thomas 203Rocco, James 268Rosado, Miguel 243Rosko, Andrew 222Russo, James 121Rwigema, Jean-Claude 117,

257

S

Samstein, Robert 301Samuels, Stuart 124, 138Sanfilippo, Nicholas 127

International Journal of Radiation Oncology � Biology � Physics972

Sapir, Eli 136, 329Sapkota, Dipak 290Scaraficci, Ana 214Schoenfeld, Jonathan 2Schrank, Travis 311, 348Schwartz, David 142Sember, Andrew 351Shackelford, Rodney 288Shah, Jennifer 148Shaikh, Talha 126Shameem, Raji 215Sharma, Arun 314Shevchenko, Sergey 287Shumway, Brian 289Signorovitch, James 353Silver, Natalie 284Sio, Terence 327Skladowski, Krzysztof 119

Smith, William 172Song, Shiyu 233Spiotto, Michael 269, 271St. John, Maie 230, 272Stokes, William 109Strom, Tobin 300Sumner, Whitney 306Sundaramurthy,

Aravindhan 355Suppiah, Somu 223Swegal, Warren 178Swiecicki, Paul 281

T

Takiar, Vinita 159Terhaard, Chris 225, 242Thomas, Carissa 286

Thorstad, Wade 239Tolekidis, George 234Tucker, J. 357

V

VanderWalde, Noam 135Velez, Maria 261Verma, Vivek 193

W

Wald, Patrick 164Wang, Ellen 354Wang, Kyle 144, 232Ward, Matthew 107Warren, Benjamin 345

Wieland, Aaron 191Wise-Draper, Trisha 279Wojcieszynski, Andrzej 196Wong, Kee 228Wu, Szu-Yuan 108Wygoda, Andrzej 185

Y

Yuan, Taize 235Yusuf, Mehran 183

Z

Zevallos, Jose 1Zhan, Kevin 206, 212, 213,

339

Zumsteg, Zachary 115