Wobm

C

a

c

d

a

K

l

S

v

©

0

d

J Oral Maxillofac Surg67:1069-1073, 2009

Predictors of Mandibular Involvement inCancers of the Oromandibular Region

Manoj Pandey, MS,* Latha P. Rao, MDS,† and

Shaima R. Das, MDS‡

Purpose: Invasion of the mandible by oral squamous carcinoma is not only a relative contraindicationto mandible conservation but also an indicator of poor prognosis. This study looks at clinical, radiologic,and operative variables that may help in predicting mandibular bone involvement.

Patients and Methods: A prospective study was carried out to evaluate the mandibular involvementand its predictors in 51 cases of oral squamous carcinoma located in the mandibular region. All patientsunderwent segmental- or hemimandibulectomy. A detailed clinical examination was followed by radio-logic assessment and operative assessment. Statistic analysis was carried out by �2 test (odds ratio [OR]with a significance level of 5%). Multivariate analysis was carried out by logistic regression analysis.

Results: Univariate analysis identified location of tumor on lower alveolus (OR � 8.5), sensorydisturbances of inferior alveolar nerve (OR � 16.2), location of tumor within 1 cm of mandible (OR �1.4), presence of findings on periosteal striping (OR � 2.0) like subperiosteal reaction (OR � 3.5),cortical expansion (OR � 8.8) and presence of pathologic fracture (OR � 2.3) as predictor of boneinvasion. Grade of tumor (P � .05) and radiologic bone involvement (P � .02) were found to besignificant independent predictors of pathologic bone involvement on multivariate analysis.

Conclusions: It is possible to identify mandibular invasion in almost all cases of oral squamouscarcinoma by combining clinical examination, radiologic findings, and findings on periosteal stripping.This helps surgeons to make an informed preoperative and intraoperative decision about mandibularconservation. However, one should be careful when evaluating bone involvement on periosteal strippingas this cannot be recommended as a method of choice due to fear of tumor dissemination and violationof oncologic principles.© 2009 American Association of Oral and Maxillofacial Surgeons

J Oral Maxillofac Surg 67:1069-1073, 2009

ddtpsbgaPreirautscmo

ith the increasing realization that the involvementf the mandible in cancers of the mandibular region isy direct invasion of tumor and not by lymphatics,1

ore mandibles are being saved.

*Formerly, Department of Surgical Oncology, Regional Cancer

enter, Medical College, Trivandrum, India; and Currently, Associ-

te Professor, Department of Surgical Oncology, Institute of Medi-

al Sciences, Banaras Hindu University, Varanasi, India.

†Formerly, Government Dental College, Medical College, Trivan-

rum, India; and Currently, Assistant Professor, Department of Oral

nd Maxillofacial surgery, Amrita Institute of Medical Sciences,

ochi, India.

‡Resident, Department of Pathology, Government Medical Col-

ege, Trivandrum, India.

Address correspondence and reprint requests to Dr Pandey: Head,

urgical Oncology, Institute of Medical Sciences, Banaras Hindu Uni-

ersity, Varanasi 221005, India; e-mail: manojpandey@vsnl.com

2009 American Association of Oral and Maxillofacial Surgeons

278-2391/09/6705-0022$36.00/0

oi:10.1016/j.joms.2008.06.059

a

1069

Carcinoma of the mandibular region is considered aistinct entity due to frequent involvement of the man-ible by direct extension.2 The survival of patients withumors in this area is low.3 This is one group of neo-lasms that has exclusively been treated by a minimumegmental resection of mandible, however, attempts areeing made at conservation of the mandible in thisroup of neoplasms if the depth of invasion is shallownd a clear margin of the bone can be achieved.4-6

reoperative evaluation of mandible is not always accu-ate. The sensitivity and specificity of the radiologicvaluation is low. For orthopantomogram, the sensitivitys 92% with a low specificity of 58%.7 Computed tomog-aphy (CT) scan, magnetic resonance imaging (MRI),nd bone scintigraphy have been found to be moreseful with sensitivity and specificity ranging from 85%o 95%.1,8-12 Clinical examination has been found to beensitive,7 however, the best results are obtained byombining a good clinical examination with one orore radiologic investigations.7 With the increasing use

f marginal resection of the mandible, several factors

dversely affecting the survival have been found, of

tiHmtt

P

2csmsdwEd

dsfsitv

sptm

wbcnlObtfiricm(mcoa

tlc

G

A

A

T

S

P

R

P

1070 PREDICTOR OF MANDIBULAR INVOLVEMENT

hese, stage of the disease, status of surgical margin, andnvasion of bone are of paramount importance.13-17

owever, literature on predictors of mandibular involve-ent is sparse. This study was carried out to evaluate

he factors predicting mandibular involvement in pa-ients with cancer of the mandibular region.

atients and Methods

In an open prospective study between September000 and May 2001, mandible invasion of squamousell carcinoma of the oromandibular region was as-essed in 51 consecutive patients. Patients with tu-ors located within 2 cm of the mandible that were

cheduled to undergo either segmental- or hemiman-ibulectomy were included in the study. The studyas approved by the Institutional Review Board and

thics Committee (Regional Cancer Center, Trivan-rum, India).A comprehensive history was taken including the

emography, past medical history, and dental andocial history (tobacco and alcohol habits). This wasollowed by a detailed examination that recorded theite of lesion, size of lesion, type of lesion (ulcerated,nfiltrative, proliferative, or combination) and a de-ailed examination of the neck for lymph node in-olvement. Fixation of the tumor to the mandible,

Table 1. RELATION OF DEMOGRAPHIC CHARACTERISTI

Invasion Present Invasion

enderMale 17 13Female 8 13

ge�50 5 12�50 20 14

ny Habit*Present 18 17Absent 7 9

obacco chewingPresent 15 16Absent 10 10

mokingPresent 7 4Absent 18 22

revious RTYes 9 15No 14 11

eligionHindu 16 20Non-Hindu 0.9 6

Abbreviations: CI, confidence interval; OR, odds ratio; RT*Includes patients who were chewers, smokers, or alcoh

andey, Rao, and Das. Predictor of Mandibular Involvement. J Oral M

ymptoms of inferior alveolar nerve or mental nervearesthesia, and bony irregularities on palpation wereaken as a clinical sign and symptom of bone involve-ent.After resection, the mandible was decalcified andas sectioned serially at 0.5-cm interval to determineone invasion. Pathologic parameters recorded in-luded grade of tumor, pathologic tumor stage andodal stage, perineural invasion, lympho-vascular and

ympho-plasmocytic invasion, and extranodal spread.ther considerations included invasion of the mandi-le, type of invasion (erosive, tumor extended intohe bone in a broad tumor front, with a layer ofbrous tissue separating the tumor cells from theeceding bone; or infiltrative, tumor invaded the bonen irregular cords and fingerlike projections of tumorells), pattern of spread (subperiosteal, subcortical,arrow, nerve-related spread), depth of invasion

shallow—extending up to upper part of marrow,oderate—extending up to but not involving the

anal, and deep involving or beyond canal), relationf tumor to neurovascular canal, host cell reactions,nd involvement of soft tissue.

Statistical analysis was carried out using cross tabula-ion and �2 test. Univariate odds ratio with a significanceevel of 5% were calculated. Multivariate analysis wasarried out by multiple logistic regression.

D HABITS TO MANDIBULAR INVASION

t �2 P Value OR 95% CI

1.7 .19 2.12 0.6-6.6

3.9 .48 0.29 0.84-1.01

0.25 .6 1.36 0.4-4.4

0.01 .9 0.9 0.3-2.8

1.9 .2 2.1 0.5-8.4

1.6 1.90.678 0.37-1.21.4 0.82-2.5

1.02 .3 0.53 0.15-1.8

otherapy.rs or used more than 1 substance.

CS AN

Absen

, radiol use

axillofac Surg 2009.

R

4gwsia(hcl(wwse4ctt

wctti

Trgwt(i

ttfibles(2m

iw(

api

S

C

P

F

S

D

S

S

P ral M

PANDEY, RAO, AND DAS 1071

esults

The mean age of the patients was 53.4 years and1% of the patients were female. The detailed demo-raphic and tumor characteristics are described else-here7 and in Tables 1 and 2. The mandibular inva-

ion was identified in 25 of 51 patients (49%). Thenvolvement was seen primarily on tumors of lowerlveolus (89%) followed by tumors of buccal mucosa35%). Thirty-five patients had tobacco or alcohol useabits of which majority (n � 31) were tobaccohewers a total of 19 (37.3%) had ulcero-proliferativeesion followed by ulcero-infiltrative lesion in 1019.6%). Clinically, most of the lesions were locatedithin 1 cm of the mandible (41; 80.4%) whereas 10ere between 1 to 2 cm, involvement of overlaying

kin was seen in 15 (29.4%) and 22 (43.1%) weredentulous. Most of the lesions were clinically T2 (23;5.1%) followed by T4 (19; 37.3%). Similarly, 51% hadlinically N1 disease at presentation. Twenty-nine ofhe lesions (58%) were well differentiated on histopa-hology.

Mandibular resections were carried out in all andere segmental in 16 (31.4%). Neck was addressed in all

ases whereas 32 (62.8%) underwent flap reconstruc-ion. Intraoperatively on periosteal stripping subperios-eal reaction was seen in 31 (60.8%), cortical expansion

Table 2. CLINICAL CHARACTERISTICS OF THE TUMOR A

Invasion Present I

iteLower alveolus 17Others 8

linical typeUlceroproliferative 10Other 15

roximity to mandible�1 cm 24�1 cm 1

ixation to mandibleYes 24No 1

kin invasionYes 6No 19

ental statusEdentulous 11Dentulous 14

ensory disturbances of inf alv NPresent 15Absent 9

tatus of tumorPrimary 17Radio residual/recurrent 8

Abbreviations: CI, confidence interval; inf alv N, inferior*Significant.

andey, Rao, and Das. Predictor of Mandibular Involvement. J O

n 19 (37.3%), and pathologic fracture in 5 (9.8%). o

wenty-four of the cases have received preoperativeadiotherapy whereas the rest underwent primary sur-ery. Neck dissection was carried out in all the cases, itas modified neck dissection (MND) type II in 18 pa-

ients, MND type III in 8 patients, radical neck dissectionRND) in 12, supraomohyoid neck dissection (SOHND)n 4, and MND type I in 9 patients.

Univariate odds ratio analysis identified location ofhe tumor on lower alveolus (OR � 8.5), location ofhe tumor within 1 cm of mandible (OR � 1.4),xation to mandible (OR � 2.77), sensory distur-ances of inferior alveolar nerve (OR � 16.2), radio-

ogic evidence of bone invasion (OR � 7.97), pres-nce of findings on periosteal stripping (OR � 2.0),ubperiosteal reaction (OR � 3.5), cortical expansionOR � 8.5), and absence of pathologic fracture (OR �.3) as the predictors of mandibular bone involve-ent (Table 3).The tumor size, nodal status, presence of perineural

nvasion, gender, age, or clinical type of the lesionas not found to be a predictor on univariate analysis

Table 3).Radiologic evidence of bone involvement (P � .02)

nd grade of the tumor (P �.05) were both independentredictors of pathologic bone involvement in multivar-

ate analysis. However, when the analysis was run using

EIR RELATION TO MANDIBULAR INVASION

n Absent �2 P Value OR 95% CI

19.1 .0002 8.5 2.1-32.9*3 0.34 0.19-0.62

0.15 .699 1.1 0.56-2.37 0.9 0.6-1.4

7.5 .0067 1.4 1.1-1.9*9 0.116 0.16-0.84

21.0 .0009 2.77 1.6-4.7*7 0.06 0.009-0.42

0.69 .49 0.6 0.28-1.67 1.1 0.8-1.6

0.01 .91 1.0 0.3-1.95 0.9 0.6-1.5

19.7 .0001 16.2 2.3-113.8*5 0.39 0.23-0.65

3.3 .061 1.6 0.92-9.15 0.55 0.28-1.07

ar nerve; OR, odds ratio.

axillofac Surg 2009.

ND TH

nvasio

2

1

1

1

1

11

2

11

alveol

nly the variables found to be significant on univariate

aw

D

tnsOevwdeirtt

towmths

tmtbethcsr

T

N

R

G

T

O

P

S

C

P

P

P ral M

1072 PREDICTOR OF MANDIBULAR INVOLVEMENT

nalysis, only the radiologic evidence of bone invasionas found to be a significant independent predictor.

iscussion

Involvement of mandible by tumor not only altershe management strategies but also affects the prog-osis6,13,14,18 and quality of life of patients with oralquamous cell carcinoma of the mandibular region.n the other hand, other studies failed to find anffect of degree of mandibular resection on sur-ival.15-17 A combination of clinical examinationith radiologic assessment is often accurate in pre-icting mandibular involvement, however, knowl-dge of additional factors that may influence bonenvolvement will help in interpreting the clinical andadiologic tests better. The present study reports onhe predictors of bone invasion and have identified

Table 3. TUMOR CHARACTERISTICS AND ITS RELATIONUNIVARIATE ANALYSIS

Invasion Present Invas

umor stage (clinical)�T2 13�T2 12

ode stage (clinical)�N1 16�N1 7

adiologic bone invasion� 23� 2

radeWell 15Other 10

ype of mandibulectomySM 5HM 20perative fixationPresent 23Absent 2

eriosteal strippingFinding present 23None 2

ubperiosteal reaction� 24� 1

ortical expansion� 17� 8

athologic fracture� 5� 20

erineural invasion� 5� 20

Abbreviations: CI, confidence interval; HM, hemimandibu*Significant.

andey, Rao, and Das. Predictor of Mandibular Involvement. J O

he grade of the tumor as the most important predic- s

or on multivariate analysis. The grade of the tumor isften associated with aggressiveness of the tumorith high grade (poorly differentiated) tumors beingore aggressive than low grade (well differentiated)

umors or from verrucous tumors. We also observedigher involvement in poorly differentiated tumors;imilar findings have been reported earlier.1

The parameters on clinical examination like loca-ion of the tumor on lower alveolus or within 1 cm ofandible, clinical fixation of the tumor, sensory dis-

urbance along inferior alveolar nerve were found toe significant predictors on univariate analysis, how-ver, they are not independent predictors as they losehe significance in multivariate analysis. On the otherand, gender, age, habits (tobacco chewing, alcohol)linical type of lesion, skin fixation of tumor, dentaltatus, and patients with recurrent cancers who haveeceived prior radiotherapy were not found to be

ANDIBULAR INVASION: RESULTS OF

sent �2 P Value OR 95% CI

1.3 .20.76 0.48-1.21.5 0.74-3.0

0.2 .60.9 0.6-1.21.2 0.49-3.2

33.0 .0007.97 2.7-23.2*0.09 0.02-0.34

0.08 .771.07 0.66-1.70.9 0.47-1.7

3.9 .0481.48 0.99-2.20.43 0.17-1.05

27.25 .0004.7 2.1-10.6*0.09 0.02-0.37

14.6 .0002.0 1.3-3.1*0.07 0.01-0.54

25.5 .0003.5 1.8-6.7*0.05 0.008-0.379

19.8 .0008.8 2.2-34.30.34 0.19-0.62

5.7 .01– –

2.3 1.6-3.1*1.6 .2

2.6 0.5-12.10.8 0.6-1.08

y; OR, odds ratio; SM, segmental mandibulectomy.

axillofac Surg 2009.

TO M

ion Ab

178

196

323

1411

1214

521

1214

719

224

026

224

lectom

ignificant predictors. A number of earlier re-

pan

pbsctfiewbfiwd

ademmtcimitpateonpimafimdmvmv

R

1

1

1

1

1

1

1

1

1

1

2

2

PANDEY, RAO, AND DAS 1073

orts14,15,17 found higher mandibular involvement onlveolar (gingival) primaries, however, the exact sig-ificance of these findings is not clear.An earlier study evaluated findings on periosteal strip-

ing and found them to be an important predictor ofone invasion.19 We looked at the findings on periostealtripping separately and found subperiosteal reaction,ortical expansion, and presence of pathologic fractureo be a significant predictor of a bone invasion. Thesendings lost the significance of multivariate analysis. Asxpected if shown, radiologic evidence of bone invasionas found to be a significant predictor of pathologicone invasion. This is due to high sensitivity and speci-city of radiologic investigations and is in accordanceith earlier studies,2,7,8,20,21 and is an independent pre-ictor of bone invasion.None of the other variables studied was found to

ffect the risk of bone invasion including neitherental status nor preoperative treatment. No differ-nce was observed among primary or recurrent tu-ors. Although a smaller sample size is a limitation toultivariate analysis, however, prospective nature of

he study and completeness of data helped in over-oming this problem. Appearance of grade of tumorn multivariate analysis using all variables and enter

ethod was a surprise and shows importance of us-ng all the variables for multivariate analysis ratherhen choosing variables. The results suggest that allatients with cancer of the oral cavity should undergothorough clinical examination and radiologic inves-

igations including CT scan and scintigraphy when-ver possible. The findings on preoperative strippingf periosteum are most sensitive, however, its use canot be recommended as this may violate oncologicrinciple and may lead to tumor dissemination. A high

ndex of suspicion is necessary to avoid unnecessaryandibular resections. Further prospective studies

re needed on large cohort to confirm the presentndings and their clinical relevance. In conclusion, itay be suggested that patients with radiologic evi-

ence of bone involvement and higher grade of tu-ors may be poor candidates for mandibular conser-

ation and findings on periosteal stripping althoughost useful can not be recommended as they may

iolate the oncologic safety of resection.

eferences

1. Lam KH, Lam LK, Ho CM, et al: Mandibular invasion in carci-

noma of the lower alveolus. Am J Otolaryngol 20:262, 1999

2. Soderholm AL, Lindqvist C, Heitanen J, et al: Bone scanning forevaluating mandibular bone extension of oral squamous cellcarcinoma. J Oral Maxillofac Surg 48:252, 1990

3. Langdon JD, Harvey PW, Rapides AD, et al: Oral cancer: Thebehavior and response to treatment of 194 cases. J MaxillofacSurg 5:221, 1977

4. Hong SX, Cha IH, Lee EW, et al: Mandibular invasion of lowergingival carcinoma in the molar region: Its clinical implicationson the surgical management. Int J Oral Maxillofac Surg 30:130,2001

5. Barttelbort SW, Ariyan S: Mandibular preservation with oralcavity carcinoma: Rim mandibulectomy versus sagittal mandi-bulectomy. Am J Surg 166:411, 1993

6. Ord RA, Sarmadi M, Papadittirou J: A comparison of segmentaland marginal bone resection for squamous cell carcinoma in-volving the mandible. J Oral Maxillofac Surg 55:470, 1997

7. Rao LP, Das SR, Mathews A, et al: Mandibular invasion in oralsquamous cell carcinoma. Investigation by clinical examinationand orthopantomogram. Int J Oral Maxillofac Surg 33:454,2004

8. Kalavrezos ND, Gratz KW, Sailer HF, et al: Correlation ofimaging and clinical features in assessment of mandibular in-vasion of oral carcinomas. Int J Oral Maxillofac Surg 25:439,1996

9. Noyek AM: Bone scintigraphy in otolaryngology. Laryngoscope89:1, 1979

0. Brown JS, Lewis-Jones H: Evidence for imaging the mandible inthe management of oral squamous cell carcinoma: A review.Br J Oral Maxillofac Surg 39:411, 2001

1. Higashi K, Wakao H, Ikuta H, et al: Bone scintigraphy indetection of bone invasion by oral carcinoma. Ann Nucl Med14:93, 1996

2. Lewis-Jones HG, Rogers SN, Beirne JC, et al: Radionuclide boneimaging for detection of mandibular invasion by squamous cellcarcinoma. Br J Radiol 73:488, 2000

3. Munoz Guerra MF, Naval Gias L, Campo FR, et al: Marginal andsegmental mandibulectomy in patients with oral cancer: Astatistical analysis of 106 cases. J Oral Maxillofac Surg 61:1289,2003

4. Tankere F, Golmard J, Barry B, et al: Prognostic value of man-dibular involvement in oral cancers. Rev Laryngol Otol Rhinol123:7, 2002

5. de Vicente JC, Recio OR, Pendas SL, et al: Oral squamous cellcarcinoma of the mandibular region: A survival study. HeadNeck 23:536, 2001

6. Dubner S, Heller KS: Local control of squamous cell carcinomafollowing marginal and segmental mandibulectomy. HeadNeck 15:29, 1993

7. Wald RM, Calcaterra TC: Lower alveolar carcinoma. SegmentalVS marginal resection. Arch Otolaryngol 109:578, 1983

8. Brown JS, Kalavrezos N, D’Souza L, et al: Factors that influencethe method of mandibular resection in the management of oralsquamous cell carcinoma. Br J Oral Maxillofac Surg 40:475,2002

9. Politi M, Costa F, Robiony A, et al: Review of segmental andmarginal resection of the mandible in patients with oral cancer.Acta Otolaryngol 120:569, 2000

0. Acton CH, Layt C, Gwynne R, et al: Investigative modalities ofmandibular invasion by squamous cell carcinoma. Laryngo-scope 110:2050, 2000

1. Shaha AR: Preoperative evaluation of the mandible in patients

with carcinoma of the floor of the mouth. Head Neck 13:398,1991