Annals of Otology. Rhinology & Laryngology l20(10):627-634.© 2011 Annals Publishing Cotnpany. All rights reserved.

Local Injection of Bevacizumab (Avastin) and AngiolyticKTP Laser Treatment of Recurrent Respiratory Papillomatosis

of the Vocal Folds: A Prospective Study

Steven M. Zeitels, IVID; Anca M. Barbu, MD; Tali Landau-Zemer, MD;Gerardo Lopez-Guerra, MD; James A. Burns, MD; Aaron D. Friedman, MD;Mason W. Freeman, MD; Yuan-Di Halvorsen, PhD; Robert E. Hillman, PhD

Objectives: Photoangiolytic laser treatment of recurrent respiratory papillomatosis (RRP) is effective, but does not reli-ably prevent recurrence. Therefore, sublesional injections of the antiangiogenic agent bevacizumab (Avastin) were givento assess the adjunctive effect on disease recurrence. Since bevacizumab is a new therapeutic modality for RRP, therewere also primary safety objectives to determine whether there was a negative impact on the voice and whether therewere local or systemic complications.

Methods: A prospective open-label investigation was conducted in 20 adult patients with bilateral vocal fold RRP. Thepatients underwent planned 532-nm pulsed KTP laser photoangiolysis of bilateral glottal di.sea.se 4 times with an approxi-mately 6-vveek interval between procedures. At each planned laser procedure, the vocal fold that on initial presentationhad a greater volume of disease also underwent 4 serial sublesional bevacizumab injections (7.5 to 12.5 mg in 0.3 to 0.5mL). A sham injection with saline solution was administered to the other vocal fold as a control. Disease resolution wascompared between subjects' vocal folds, and objective measures of vocal function (acoustic, aerodynamic), as well aspatients' self-assessments of vocal function (Voice-Related Quality of Life survey), were obtained.Results: All 20 patients completed the study, and there were no local or systemic complications. After 4 injections, 3 ofthe 20 patients had no discernible disease in either vocal fold. Ofthe remaining 17 subjects, 16 had less disease in the bev-acizumab-treated vocal fold despite starting with more disease. Only I ofthe 17 had more disea.se in the bevacizumab-treated vocal fold after 4 injections. Moreover, 7 of the 20 patients (35%) did not require a laser procedure in the vocalfold that had received 4 bevacizumab injections, as compared with 3 ofthe 20 vocal folds (15%) that were treated withlaser alone. All of the vocal function measures displayed statistically significant posttreatment improvements, except foraverage fundamental frequency in the 3 female patients, in whom it fell below the normal range.Conclusions: This prospective investigation provided evidence that bevacizumab injections enhanced KTP laser treat-ment of glottal papillomatosis without systemic or local complications. Coupling the antiangiogenesis agent bevaciz-umab with KTP laser photoangiolysis is conceptually synergistic and scientifically promising since the mechanisms ofaction are complementary.

Key Words: Avastin, bevacizumab, glottis, KTP laser, larynx, laser, papilloma, papillomatosis, recurrent respiratory pap-illomatosis, RRP, vocal cord, vocal fold, voice.

INTRODUCTION

Recurrent respiratory papillomatosis (RRP) hasbeen an omnitemporal clinical problem in laryngol-ogy since the origin of the subspecialty 150 yearsago.' The mainstay of treatment through the past 2centuries has been transoral surgical removal to re-store airway and vocal function. Depending on theseverity of the disease, the treatment would com-monly result in permanent vocal dysfunction and.

rarely, could produce airway stenosis.

Hooper noted in 1882 that "The vascularity ofthe growth (papillomatosis) is perhaps the mostimportant factor in determining the rapidity of itsreappearance." '̂P- '̂*'̂ In essence, he was recognizingneoplastic angiogenesis, which would be clinicallydescribed again in 1966 by Jako and Kleinsasser-^and formally conceptualized as a watershed heuris-tic theory by Folkman in the

From the Departments of Surgery (Zeitels, Barbu. Landau-Zemer, Lopez-Guerra. Burns, Friedman, Hillman) and Medicine (Freeman,Halvorsen), Harvard Medical School, and the Center for Laryngeal Surgery and Voice Rehabilitation (Zeitels. Barbu, Landau-Zemer,Lopez-Guerra, Bums, Friedman, Hillman). the Translational Medicine Group, Center for Computational and Integrative Biology (Free-man, Halvorsen), and the Massachusetts General Hospital Institute of Health Professions (Hillman), Massachusetts General Hospital,Boston, Massachusetts. This work was supported in part by the Eugene B. Casey Foundation, the V Foundation, the Institute of Laryn-gology and Voice Restoration, and the Harvard Catalyst Program.

Correspondence: Steven M. Zeitels, MD, Center for Laryngeal Surgery and Voice Rehabilitation, Massachusetts General Hospital,One Bowdoin Square, 1 Ith Floor, Boston, MA 02114.

627

628 Zeiteis et al, Bevacizumab & KTP Laser Treatment of Papillomatosis

Remarkably, this philosophy has been the under-pinning of our treatment philosophy^" "̂ of RRP overthe past decade. Angiolytic laser treatment of RRPcomprises targeting the microvasculature of the neo-plasm. It was done by means of microlaryngoscopyand general anesthesia initially, and a decade ago, anoffice management procedure was created that uti-lized topical anesthesia with a 585-nm pulsed-dyelaser.^ Office-based angiolytic laser treatment ofRRP was a key innovation because it obviated theneed for ongoing general anesthetic interventions ina majority of adult patients with RRP. This proce-dure was made substantially easier by the introduc-tion of the 532-nm pulsed KTP (potassium titanylphosphate) laser because of its enhanced precision,reliability, and versatility.^"''^

On the basis of the possible role of vascular en-dothelial growth factor (VEGF) in the neoplasticprogression of RRP,'^ and on the success of local-ized injecrion of bevacizumab (Avastin) for oph-thalmologic neovascular growth,'"* we initiated lo-calized sublesional injection of bevacizumab as ananriangiogenesis agent to enhance angiolytic lasertreatment of RRP.' Bevacizumab is a recombinanthumanized monoclonal immunoglobulin Gl anti-body that binds to and inhibits the biologic activ-ity of human VEGF. It has a molecular weight ofapproximately 149 kd. Bevacizumab is formulatedas a clear to slightly opalescent, colorless sterile so-lution, and is supplied in lOO-mg preservative-freesingle-use vials at a concentration of 25 mg/mL.

Our retrospective initial report on the off-labeluse of localized injection of bevacizumab involved10 parients with laryngeal RRP. This study providedpreliminary evidence that bevacizumab injectionsenhanced photoangiolytic laser treatment of glot-tal papillomatosis with associated improvements inphonatory function. From our review of the litera-ture, this publication appears to be the first descrip-tion of localized injection of a pharmacologie anti-angiogenesis agent to treat a neoplasm, as distinctfrom treating aberrant blood vessel growth.''*'-''

Because of the success noted in the inirial study,we embarked on a collaborative effort with the Trans-lational Medicine Group at Massachusetts GeneralHospital (MGH) and initiated the prospecrive clini-cal trial reported herein. We obtained investigationalnew drug approval from the US Food and Drug Ad-ministration, as well as Institutional Review Boardapproval from MGH/Partners Healthcare.

The primary therapeutic objective of this studywas to determine whether bevacizumab treatment re-duces or eliminates the need for recurrent angiolyticKTP laser photoangiolysis surgery to control the pro-

gression of RRP of the vocal folds during the treat-ment period. The primary safety objectives were todetermine whether there was a negative impact onthe voice and whether there were local or systemiccomplications associated with bevacizumab.

MATERIALS AND METHODS

Study Design. A phase I single-center open-labelstudy was performed involving 20 patients in orderto assess the efficacy, safety, and tolerability of bev-acizumab for treating patients with bilateral vocalfold RRP. The investigation was funded from patient-based US nonprofit foundations or organizations, aswell as pilot grant support from Harvard Catalyst,the federally funded Clinical and Translational Sci-ence Center grant. The study protocol, informed con-sent form, and relevant supporting information werereviewed and approved by the Institutional ReviewBoard before the study was initiated. The subjectsfor this study were recruited from a patient popula-tion presenting to MGH Voice Center with dysphoniawho received a diagnosis of RRP of the vocal foldsand would normally receive KTP laser treatment. Re-cruitment for the study was posted on the clinical tri-als website (http://clinicaltrials.gov). The study waslimited to patients over 18 years of age who had dis-ease involving both the right and left vocal folds.

Subjects who enrolled in the protocol were inject-ed with bevacizumab in one vocal fold, which wasthe one that retained more disease on examinationbefore the first injection (video strobolaryngoscopyand/or microlaryngoscopy). A sham injection withsaline solution was administered to the other vo-cal fold as a control. This therapy was administeredonce every 6 ± 1 weeks for 6 months; the same vo-cal fold selected for the initial treatment always re-ceived the bevacizumab injection. The surgeon alsotreated any visible RRP lesions of both vocal foldsusing 532-nm pulsed KTP laser photoangiolysis. Ateach return visit, the surgeon reviewed the strobolar-yngoscopy images and compared the disease statusto the disease at baseline. T'he subjects underwent4 injections and were examined at approximately 6weeks after the fourth injection, which was approxi-mately 6 months after the first procedure.

Subject Enrollment. Study participation and en-rollment were initiated from patients who were con-sidering standard photoangiolytic KTP laser surgi-cal treatment for bilateral vocal fold papillomatosis.The staff surgeons (S.M.Z., J.A.B., and A.D.F.) ini-tially discussed consent for participation with eachpatient at an outpatient visit in the MGH Voice Cen-ter before the first surgical procedure. A patient had24 hours from the time of the initial discussion todecide whether he or she would like to sign the con-

Zeileis et ai Bevacizumah & KTP Laser Treatment of Papillomatosis 629

sent form for the first procedure. Patients could re-verse their decision at any time. It was explained toall patients that their decision about study participa-tion did not in any way affect the standard of carefor their disease.

All 20 patients met the following inclusion cri-teria to be enrolled in this study. The patients (maleor female) had dysphonia and a diagnosis of RRPof both vocal folds. They were between 18 and 60years of age. Female subjects were surgically ster-ilized, postmenopausal, or using adequate contra-ception. (The criteria for menopause were surgicalmenopause [hysterectomy, oophorectomy] or ageof more than 45 years with absence of menses formore than 12 months and an elevated level of se-rum follicle stimulating hormone, ie, more than 25mlU/mL. Tubal ligation with menses within the past12 montbs was not considered to be surgical steril-ization.) The patients were willing and able to com-ply with the investigational nature of the study andable to communicate well with the investigators.They had the ability to comprehend and willingnessto provide written informed consent in accordancewith institutional and regulatory guidelines.

Patients who met any of the following criteriawere excluded from the study: known sensitivity tobevacizumab; systemic treatment with bevacizumabwithin 4 months (approximately 6 half-lives); a he-matocrit level of less than 30%, a prothrombin timeor a partial tbromboplastin time more than 1.2 timesthe upper limit of normal, or absolute platelet countsbelow the lower limit of normal; an absolute neutro-phil count below 1 ,500/|JL; a urine protein-to-creati-nine ratio greater than 1.5 (a urine dipstick test posi-tive for a proteinuria level of at least 3+ was not anexclusion criterion, but necessitated spot urine col-lections); a history of congestive heart failure, angi-na that remained symptomatic despite medications,myocardial infarction or a coronary artery disease-related procedure within the past 6 months (includ-ing angioplasty, stent placement, bypass graft sur-gery), or a prosthetic heart valve; significant uncon-trolled hypertension (systolic blood pressure above160 mm Hg and/or diastolic blood pressure above100 mm Hg); surgery within 1 month or an unhealedsurgical incision at screening; a history of throm-boembolic events, gastrointestinal ulcer or bleeding,or wound debiscence; or nursing or plans to nurseduring the possibly 18-month treatment period.

Subject Evaluation. The following procedureswere performed before the first treatment (baseline)and before each subsequent treatment, with the finalassessment occurring approximately 6 weeks afterthe fourth treatment.

In terms of medical testing, we performed elec-trocardiography; vital signs testing including pulse,blood pressure, respiration rate, and body tempera-ture; serum chemistry; complete blood count; physi-cal examination; urinalysis; and pregnancy testing(female subjects of childbearing potential only).

Endoscopie imaging was also performed. Stan-dard transnasal flexible videolaryngoscopy was per-formed at each visit before treatment as a basis fordetermining the amount of papilloma on the vocalfolds. A Kay PENTAX 9200-C videostroboscopicsystem was used to make digital recordings as theendoscope was centered on the glottis and slowlyadvanced and withdrawn to acquire images over arange of distances while the vocal folds were abduct-ed during quiet breathing. Image distortion related toscope optics (barrel distortion) was measured on anoptical bench and was corrected for in postprocess-ing with a validated Matlab-implemented algorithm.KayPENTAX system software was used to selectthe still image for each endoscopie assessment thatshowed the best exposure of the entire glottis (ad-ducted vocal folds). The still images were saved injpeg bitmap format, corrected for image distortionwith the Matlab-implemented routine, and digitallyprocessed to equate all of the images with respectto the size and orientation of the glottis. The cursorfunction in ImageJ software was then used to delin-eate the approximate area of the disease in squarecentimeters for the' right and left vocal folds in eachstill image to assist in making comparisons based onvisual judgments. Eor this initial report, only the im-aging data from the last assessment were examinedas a basis for making the simple determination aboutwhich vocal fold had the most recurrent disease after4 treatments. This approach to the analysis actuallycreated a bias against finding a positive treatment ef-fect, because in each patient the vocal fold with theworst disease was injected with bevacizumab.

Noninvasive clinical voice assessment includedcompletion of the Voice-Related Quality-of-Lifesurvey (V-RQOL), on which 0 indicates the lowestvoice-related quality of life and 100 indicates thehighest voice-related quality of life,'^' and objectiveacoustic and aerodynamic measures of vocal func-tion. The details of the protocol for obtaining acous-tic and aerodynamic measures of vocal function havepreviously been described.'"' We focused on the sameobjective acoustic and aerodynamic measures thatwere used in an earlier pilot study of bevacizumabas an adjunct to KTP laser treatment of papilloma';1 ) average fundamental frequency (Fo) during read-ing of a standard passage (vocal pitch); 2) the noise-to-harmonies ratio during sustained vowels (voicequality); and 3) the ratio between vocal sound près-

630 Zeitels et al, Bevacizumah & KTP Laser Treatment of Papillomatosis

COMPARISON OF PRETREATMENT AND POSTTREATMENT VOCAL FUNCTION MEASURES

Measure

V-RQOL scoreNHRSPL/SG-pressureFo (male patients)Fo (female patients)

PretreatmentValue

(Mean±SD)

42.9 ±20.10.407 ±0.265

5.9 ±2.1174.2 ±43.2 Hz219.6 ± 19.6 Hz

PosttreatmentValue

(Mean ±SD)

83.3 ±16.80.266 ± 0.247

7.9 ±3.7129.7 ±34.9 Hz162.3 ±8.0 Hz

Mean Post-PreDifference

+40.4-0.141

+2.1^ 4 . 5 Hz-57.3 Hz

Normal Range

Total range. Oto 100<0.190>8.6

82 to 138180 to 230

Paired t-Test

-9.565 ((//= 19); p< 0.0012.449 (£?/'= 19);p = 0.024

-3.206 idf= 18);p = 0.0053.950 (i//= 15);p = O.OOI8.504 (d/= 2); p = 0.014

V-RQOL — Voice-Related Quality of Life survey; NHR — noise-to-harmonics ratio; SPL/SG-pressure — vocal sound pressure level divided byaverage subglottic air pressure. Results for average fundamental frequency (Fo) are shown separately for male and female patients because of itsgender dependence. See text for description of measures.

*Statistically significant (p < 0.05).

sure level and average subglottic air pressure duringproduction of standard syllable strings (in decibelsper centimeter of water; vocal efficiency). Group-based differences between pretreatment (baseline)and posttreatment average values for vocal functionmeasures were assessed for statistical significance (p< 0.05) by use of simple r-tests for paired compari-sons. The average values for vocal function measureswere also compared to historical norms, as has beentypically done in previous investigations.' Because11 of the 20 patients had undergone previous assess-ment and ongoing management for their disease atour facility before accrual for the present study (in-cluding KTP laser treatment), the decision was madeto use voice data from the very first evaluation thatthese patients had in order to more accurately char-acterize their pretreatment baseline function.

Treatment Protocol. The first treatment sessiontypically involved microlaryngoscopy and generalanesthesia. Bevacizumab (7.5 to 12.5 mg) was in-jected into the sublesional mucosa of the vocal foldthat was judged to have worse disease. The tech-nique was described previously.' Saline solutionwas injected into the other vocal fold in a compar-able volume and location.

Photoangiolysis of disease was done on both vocalfolds with the 532-nm KTP laser. The bevacizumabinjection of the vocal fold was done either before orafter the laser procedure according to nuanced indi-vidual anatomic considerations of each patient so asto most effectively deliver an accurate dose.

Three additional treatment sessions were carriedout at approximately 6-week intervals. A majorityof these treatments were done in the office with lo-cal topical anesthesia. At each session, bevacizumabwas reinjected into the previously injected vocalfold. The KTP laser was used for photoablation ofrecurrent disease on either vocal fold.

RESULTS

All 20 patients who enrolled in the study com-pleted it successfully, and none had a local or sys-

temic complication. At the end of the treatment pe-riod, 3 of the 20 patients had no discernible diseasein either vocal fold. Of the 17 remaining subjects,16 had less disease in the bevacizumab-treated vo-cal fold despite starting with more disease. Only 1 ofthe 17 had more disease in the bevacizumab-treatedvocal fold after 4 injections. This reduced appear-ance of disease in the bevacizumab-treated vocalfold significantly exceeds what may be expected bychance (2-tailed binomial calculation, p = 0.00027).Moreover, 7 of the 20 bevacizumab-treated vocalfolds (35%) did not require laser treatment at the6-week checkpoint subsequent to the 4 injections, ascompared to 3 of the 20 saline-injected vocal folds(15%), which had less disease initially, but this dif-ference did not achieve statistical significance (X'̂test, p = 0.114). Similarly, in the 80 bevacizumab-injected vocal folds, KTP laser treatment was re-quired 57 times, as compared with 69 times after the80 saline injections. Twenty-five of the 80 proce-dures (31 %) were done by means of microlaryngos-copy and general anesthesia, and 55 of the 80 (69%)were done in the office with topical anesthesia.

The results for the comparisons between pretreat-ment and posttreatment vocal function measures aresummarized in the Table. All of the measures dis-played statistically significant posttreatment changes,which included increases in the V-RQOL rating andvocal efficiency (vocal sound pressure level dividedby average subglottic air pressure), and decreases inacoustic noise levels (noise-to-harmonics ratio) andvocal pitch (Fo). However, posttreatment improve-ments in group-based averages for vocal efficiencyand acoustic noise level did not attain the normalrange. For the 17 male subjects, the average Fo de-creased from an abnormally high pretreatment valueto within normal limits after treatment. Remarkably,the 3 female subjects displayed an average pretreat-ment Fo value that was within normal limits but de-creased to below the normal range after treatment.

DISCUSSION

Treating RRP by coupling the antiangiogenesis

Zeiteis et al, Bevacizumab & KTP Laser Treatment of Papillomatosis 631

Tlii.s 49-yeai-c)l(J patient regularly sustained substantial recurrence of respiratory papilloma every 6 to 8 weeks before being re-ferred. He was first patient to undergo sublesional Avastin injection. A) On presentation to us and before initiation of Avastin. hehad diffuse transglottal disease, as seen on office laryngoscopy. B) Microlaryngoscopic view also demonstrates diffuse disease.C) Microlaryngoscopic view after KTP laser treatment but without Avastin injection. D) Office laryngoscopy shows extensiverecurrence approximately 6 weeks after initial KTP microlaryngo.scopic laser treatment. E) Office-based injection of Avastinwhile patient was beginning to demonstrate improvement In disease in left vocal fold. F) Office laryngoscopy at patient's mostrecent visit and after conclusion of investigation. Patient has been receiving office-based injections of Avastin to both tnie andfalse vocal folds (approximately 50 mg) at 8- to 10-week intervals for past 7 months. There is very limited laser use.

agent bevacizumab with KTP laser photoangiolysisis conceptually attractive and scientifically sound,since the mechanisms of action are complementary.'Observations from this prospective second investi-gation have provided further evidence that pharma-cologie antiangiogenesis by local injection is effi-cacious in treating laryngeal RRP (see Figure). The

invesrigation herein also included key safety moni-toring that demonstrated a lack of local or systemiccomplications, which will assist clinicians who areconsidering initiating this promising new manage-ment strategy.

Folkman-̂ '**'"^ established that angiogenesis is a

632 Zettels et al, Bevacizumah & KTP Laser Treatment ofPapilloniatosis

multiStep process primarily initiated through VEGF,which initiates a cascade of intracellular signal trans-duction pathways resulting in the formation of newvessels.-^ Given the microcirculatory histology ofRRP,^' there is evidence that VEGF plays an impor-tant angiogenic role in the pathogenesis of RRR'"*After sublesional injection of bevacizumab, the lo-cal soft tissue concentration is postulated to exceedthe concentration required to inhibit VEGF-inducedendothelial cell proliferation and neoplastic growth.

It has been well demonstrated that photoangiolyt-ic lasers target the microcirculation of RRP and ef-fectively involute glottal disease wbile preservingthe superficial lamina propria of the phonatory mu-cosa. However, similar to all surgical interventionsthat remove RRP disease, this approach does not re-liably prevent recurrence. Therefore, it was hypoth-esized that a sublesional injection of an antiangio-genic drug such as bevacizumab at the time of anangiolytic laser removal of laryngeal papillomatosiswould be a clinically attractive approach that mightprevent or slow disease recurrence.

Based on encouraging preliminary findings,' theprotocol herein was developed to better define theeffectiveness of the adjuvant bevacizumab-lasertreatment. Given the variation in surgical technique,a single-center open-label study was done to com-pare disease recurrence between drug-treated andnon-drug-treated vocal folds longitudinally in thesame patient. Tbis contralateral control study designin the same subject minimizes confounding factorsand maximizes the cbance of observing differenc-es in a relatively small patient population, particu-larly when the recurrence rate varies significant-ly from patient to patient. Given general repetitivegeographic patterns of disease recurrence, injectingbevacizumab into the vocal fold with tbe worse dis-ease produced a more conservative (negatively bi-ased) test of treatment effectiveness.

The laryngeal dosage selected was based in parton the ophthalmologic experience. Ocular doseshave typically been between 1.25 mg (0.05 mL) and2.5 mg (0.1 mL). Tbese doses were partially predi-cated on the desire not to have to remove vitreousfluid from the eye to accommodate the volume ofdrug injection. Similarly, the dose selected for vo-cal fold injection was based on the potential fluidexpansion capacity of the phonatory mucosal com-partment (Reinke's space).̂ ^^—^ Tbe volumes of in-jection proposed (0.3 to 0.5 mL) were well toleratedin all 20 subjects. Tbe variation in subepithelial in-fusion volume was related to tbe amount of scarringpresent from prior procedures in the subepithelialsuperficial lamina propria.

Tbe dose of 7.5 to 12.5 mg (0.3 to 0.5 mL) beva-cizumab delivered locally is much less than the 5 to10 mg/kg (300 to 600 mg for a 60-kg human) doserange approved for systemic administration for thetreatment of colorectal or lung cancer. Witb 7.5 to12.5 mg bevacizumab injected into a confined spacein a vocal fold, the systemic exposure to the drugwill be substantially less than that experienced bycancer patients treated witb higher intravenous (sys-temic) administered dosages. Unlike in reports onintravenous cancer treatment, there were no system-ic complications. Tbis mirrors the opbtbalmologicexperience,-'*-'' as well as the substantially higher-dose (approximately 100 mg) local intranasal treat-ment of hereditary bemorrhagic telangiectasia.'-''

The 20 patients studied herein demonstrated awide spectrum of disease presentation, which is notsurprising. We have not noted that effective surgicaltechniques that preserve the layered microstructureofthe glottis result in a biological change in the pap-illomatosis-host relationship and recurrence pattern.'This concept was reinforced by the current investiga-tion. Both vocal folds received similar surgical treat-ment, yet by the end of the 6-month study period,there continued to be discernible disease in the saline-treated vocal fold in 17 of the 20 patients (85%). Giv-en the frequency of the procedures, it is not surpris-ing that 15% ofthe patients did not have discernibledisease at this relatively short follow-up. However, itwas remarkable that only 13 of the 20 patients (65%)had discernible RRP in the bevacizumab-treated vo-cal folds, which .started with more disease. This is es-pecially noteworthy because it is well accepted thatlaryngeal RRP generally retains repetitive geograph-ic patterns of disease recurrence, often with consis-tent asymmetry of vocal fold involvement.

It was also promising to observe that of the 17subjects wbo had disease after the fourth injection,16 had less disease in the bevacizumab-treated vo-cal fold despite starting with more disease in thatfold. Only 1 ofthe 17 continued to have more RRPin the bevacizumab-treated vocal fold. These obser-vations reached statistical significance. There wasalso a trend toward observable resolution of tbe dis-ease in the bevacizumab-treated vocal folds (7/20)as compared to tbe saline-injected side (3/20), al-though this did not reach statistical significance.Tbis trend toward enhanced resolution migbt havereached statistical significance with a larger patientcohort. Alternatively stated, there is substantial evi-dence tbat the biological behavior of the RRP wasaltered in tbe localized region in which tbe bevaciz-umab was injected, because the surgical interven-tions were consistently applied to botb vocal folds

Zeiteis et al. Bevacizumab & KTP Laser Treatment of Papillomatosis 633

in the same subject.

We expect that there will be a spectrum of respons-es to bevacizumab with varied patterns of recurrencecommensurate with the spectrum of disease presen-tation. Remarkably, the opera singer whose imagesand history were reported in the prior publication'did not participate in this study because he has beenin complete remission since December 2008, whenhe received the last of his 5 office-based bevacizu-mab injections. It is likely that more-substantial an-giogenic activity associated with more-aggressiveneoplastic growth will be more susceptible to bev-acizumab.

With regard to voice outcome, the patients dis-played substantial improvement in posttreatmentvocal function. This was reflected most clearly bythe statistically significant increases in vocal effi-ciency (vocal sound pressure level divided by av-erage subglottic air pressure) and by patients' self-assessments on the V-ROOL, accompanied by a sig-nificant decrease in the level of noise in the voiceacoustic signal (noise-to-harmonics rario). In addi-tion, the male subjects displayed a significant post-treatment drop in vocal pitch (Fo) to normal lim-its. However, even though the vocal efficiency andacoustic noise level improved significantly, neitherof the group-based averages for these measures at-tained the normal range, and there was a decrease inthe average vocal pitch to a slightly abnormal lowvalue for the 3 female subjects. This is likely an ef-fect of chronic inflammation of the superficial lami-na propria similar to that seen in female smokers.^^

Although there was no evidence that the beva-cizumab negatively affected vocal function, as hasbeen reported with cidofovir,^'' it is challenging toestablish whether it contributed to the observed im-provements in vocal function. It is extraordinarilydifficult to decouple the influence of the bevaciz-umab and the KTP laser treatments on the observedvoice improvements, because the interventions oc-curred simultaneously.

Despite the fact that it is difficult to prove, our ob-servation is that a majority of the bevacizumab-treat-ed patients developed a better voice than we wouldhave predicted given their prior extensive surgery,associated scarring, and initial diminished pliabilityof the phonatory mucosal superficial lamina propria.It is likely that the papillomavirus remains in the mu-cosa, which is why there is a tendency for the neo-plastic lesions to geographically recur in the samelocations. Our impression is that the bevacizumabturns off the virally induced angiogenic switch thatproduces papillomatosis neoplasms but does not turnoff the virally induced production of extracellular

matrix proteins^^ responsible for regeneraring pho-natory mucosal pliability and vocal fold vibration.

It is clear to any clinician who manages substan-tial numbers of RRP patients who have had largenumbers of procedures (more than 50) with oldertechnologies that a majority of the patients have re-markably stable voices despite having undergoneextensive imprecise and ablative surgery. This vocalstability is not typically observed after repeated sur-gical treatment of other phonatory mucosal lesions.Most analogous to RRP would be patients with glot-tal dysplasia who have had many ablative mucosalprocedures and who typically have poor voice qual-ity. Alternatively, one can hypothetically envisionhow poor a patient's voice would be if he or shewere to undergo 50 carbon dioxide laser proceduresfor a limited surgical zone of vocal fold nodules.

Our observarions thus far suggest that localizedlaryngeal injection of an antiangiogenesis agentshows potential to improve angiolytic KTP lasertreatment of RRP. It serves as an adjuvant medicalmanagement to diminish the severity and frequencyof recurrence and consequently the number of ab-lative mucosal procedures, which poses the mostsignificant threat to permanent voice loss. Becausethere is less mucosal ablation and disruption, fluc-tuations in vocal function are mitigated.

Until there is a systemic medical treatment forRRP, localized injections of antiangiogenesis agentsprovide a reasonable bridge strategy to diminish themorbidity of the disease and mitigate the impact ofrepeated surgical removal. An injection of beva-cizumab into the superficial lamina propria of thephonatory mucosa leads to several hours of signifi-cant hoarseness due to the hydrostatic pressure inReinke's space and is mostly resolved by the nextday. Patients do not require voice rest. It is impor-tant to recognize that there will be a learning curvefor surgeons to perfect subepithelial office-basedReinke's space infusions, since it is not typically intheir skill-sets and deep infusion of the drug is un-likely to work. A sublesional subepithelial injectionmust be done, since the virus likely resides intramu-cosally and a typical injection laryngoplasty is donedeep to the vocal ligament.

The promising clinical findings above are note-worthy because there has been limited private-sectorpharmaceutical investment in solving laryngeal pap-illomatosis. Recurrent respiratory papillomatosis isa classic orphan disease that has not been adoptedby the pharmaceutical industry because it provideslimited financial incentive for the private sector todesign, produce, and market medications to treator prevent it. Gonsidering the enormous resources

634 Zeiteis et al, Bevacizumab & KTP Laser Treatment of Papillomatosis

already invested by the pharmaceutical industry todevelop antiangiogenesis agents in oncology, alongwith their mixed success thus far, this study suggeststhat a company with an effective VEGF inhibitormight consider reformulating it for specific use inRRP. Further evidence of efficacy would also pro-vide valuable support to investigating the use of in-travenous antiangiogenic agents^^ for patients withpotentially lethal airway obstruction.

CONCLUSIONS

Treating RRP by coupling the antiangiogenesisagent bevacizumab with KTP laser photoangiolysis

is conceptually and scientifically synergistic, sincethe mechanisms of action are complementary. Ob-servations from this second investigation have pro-vided further evidence that pharmacologie antian-giogenesis is efficacious in treating RRP and that itappears to be relatively safe locally and systemically.Larger patient cohorts and multi-institutional studiesare needed. This effort provides additional evidencethat clinicians should explore new avenues for inves-tigating local injections of antiangiogenesis pharma-cologie agents as a primary or adjuvant treatment forother mucosal diseases and neoplasms of the aerodi-gestive and genitourinary systems, as well as skin.

REFERENCES1. Zeitels SM, Lopez-Guerra G, Bums JA, Lutch M, Fried-

man AM, Hillman RE. Microlaryngoscopic and office-basedinjection of bevacizumab (Avastin) to enhance 532-nm pulsedKTP laser treatment of glottal papillomatosis. Ann Otol RhinolLaryngol Suppl 2009; 118(suppl 201 ): 1 -13.

2. Hooper FH. On the treatment and recurrence of papil-loma of the larynx: candidate's thesis to the American Laryngo-logical Association. Arch Laryngol 1882;3:334-43.

3. Jako GJ, Kleinsasser O. Endolaryngeal micro-diagno-sis and microsurgery. Reprint from the Annual Meeting of theAmerican Medical Association, 1966.

4. Folkman J. Tumor angiogenesis: therapeutic implica-tions. N EngI J Med 1971;285:1182-6.

5. Folkman J. Clinical applications of research on angio-genesis. N Engl J Med 1995;333:1757-63.

6. Zeitels SM. Atlas of phonomicrosurgery and other en-dolaryngeal procedures for benign and malignant disease. SanDiego, Calif: Singular, 2001:119-31.

7. Zeitels SM, Franco RA Jr, Dailey SH. Burns JA, HillmanRE, Anderson RR. Office-based treatment of glottal dysplasiaand papillomatosis with the 585-nm pulsed dye laser and localanesthesia. Ann Otol Rhinol Laryngol 2004; 113:265-76.

8. Zeitels SM, Akst LM, Bums JA, Hillman RE, BroadhurstMS, Anderson RR. Pulsed angiolytic laser treatment of ectasiasand varices in singers. Ann Otol Rhinol Laryngol 2006;l 15:571-80.

9. Zeitels SM, Akst LM, Burns JA, Hillman RE, Broad-hurst MS, Anderson RR. Office-based 532-nm pulsed KTP la-ser treatment of glottal papillomatosis and dysplasia. Ann OtolRhinol Laryngol 2006; 115:679-85.

10. Bums JA, Zeitels SM, Akst LM, Broadhurst MS, Hill-man RE, Anderson R. 532 nm Pulsed potassium-titanyl-phos-phate laser treatment of laryngeal papillomatosis under generalanesthesia. Laryngo.scope 2007;l 17:1500-4.

11. Zeitels SM, Anderson RR, Hillman RE, Bums JA. Letterto the Editor. Ann Otol Rhinol Laryngol 2007;! 16:317-8.

12. Zeitels SM, Burns JA. Office-based laryngeal laser sur-gery with the 532-nm pulsed-potassium-titanyl-phosphate laser.Curr Opin Otolaryngol Head Neck Surg 2007;15:394-400.

13. Rahbar R, Vargas SO, Folkman J, et al. Role of vascularendothelial growth factor-A in recurrent respiratory papilloma-tosis. Ann Otol Rhinol Laryngol 2005;l 14:289-95.

14. Lynch SS, Cheng CM. Bevacizumab for neovascular oc-ular diseases. Ann Pharmacother 2007;41:614-25.

15. Simonds J, Miller F, Mandel J, Davidson TM. The effectof bevacizumab (Avastin) treatment on epistaxis in hereditary

hemorrhagic telangiectasia. Laryngoscope 2009;l 19:988-92.

16. Hogikyan ND, Sethuraman G. Validation of an instru-ment to measure voice-related quality of life (V-RQOL). J Voice1999;13:557-69.

17. Zeitels SM, Bums JA, Lopez-Guerra G, Anderson RR,Hillman RH. Photoangiolytic laser treatment of early glotticcancer: a new management strategy. Ann Otol Rhinol LaryngolSuppl 2008; 117(suppl 199):3-24.

18. Folkman J. Tumor angiogenesis. Adv Cancer Res 1985;43:175-203.

19. Folkman J. Angiogenesis. In: Braunwald E, Häuser SL,Fauci AS, Kasper DL, Longo DL, Jameson L, eds. Harrison'sprinciples of intemal medicine. 15th ed. New York, NY: Mc-Graw-Hill, 2001:517-30.

20. Eskens FA, Sleijfer S. The use of bevacizumab in colo-rectal, lung, breast, renal and ovarian cancer: where does it fit?Eur J Cancer 2008;44:2350-6.

21. Incze JS, Lui PS, Strong MS, Vaughan CW. ClementeMP. The morphology of human papillomas of the upper respira-tory tract. Cancer 1977;39:1634-46.

22. Zeitels SM, Vaughan CW. A submucosal true vocal foldinfusion needle. Otolaryngol Head Neck Surg 1991 ;105:478-9.

23. Zeitels SM. Premalignant epithelium and microinvasivecancer of the vocal fold: the evolution of phonomicrosurgicalmanagement. Laryngoscope 1995;105(suppl 67): 1-51.

24. Fung AE, Rosenfeld PJ, Reichel E. The Intemational In-travitreal Bevacizumab Safety Survey: using the internet to as-sess drug safety worldwide. Br J Ophthalmol 2006;90:I344-9.

25. Wu L. Martínez-Castellanos MA, Ouiroz-Mercado H, etal. Twelve-month safety of intravitreal injections of bevacizu-mab (Avastin): results of the Pan-American Collaborative Reti-na Study Group (PACORES). Graefes Arch Clin Exp Ophthal-mol 2008 ;246:81-7.

26. Zeitels SM, Hillman RE, Bunting GW, Vaughn T. Rein-ke's edema: phonatory mechanisms and management strategies.Ann Otol Rhinol Laryngol 1997;IO6:533-43.

27. Lee AS, Rosen CA. Efficacy of cidofovir injection forthe treatment of recurrent respiratory papillomatosis. J Voice2004; 18:551-6.

28. Gray SD, Titze IR, Chan R, Hammond TH. Vocal foldproteoglycans and their influence on biomechanics. Laryngo-scope 1999; 109:845-54.

29. Nagel S, Busch C, Blankenburg T, Schütte W. Treatmentof respiratory papillomatosis — a case report on systemic treat-ment with bevacizumab [in German]. Pneumologie 2009;63:387-9.

Copyright of Annals of Otology, Rhinology & Laryngology is the property of Annals Publishing Company and

its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.