CLINICAL ARTICLESJ Oral Maxillofac Surg60:748-753, 2002

Recurrent Neck Disease in Oral CancerD.R.P. Godden, FDSRCS, FRCS(OMFS),* N.F.F. Ribeiro, FDSRCS,†

K. Hassanein, FRCS,‡ and S.G. Langton, FDSRCS, FRCS§

Purpose: The goals were to examine the clinical and pathologic features of patients who developedmetastatic squamous cell carcinoma in the cervical lymph nodes after initial treatment and to identify anycommon patterns.

Patients and Methods: A retrospective analysis of 35 patients of varying initial tumor stage wasperformed. There were 18 patients who had an initial neck dissection and 17 patients whose neck wasmanaged by a “watch and wait” policy.

Results: Recurrence frequently involved level II nodes, and extracapsular spread was invariablypresent. The time taken for recurrence to develop was the same in both groups of patients (15 months,P � .35), and the overall median survival time after recurrence was 18 months (12 to 25 months, 95%confidence interval). In 27 of 29 patients (93%) who had the primary tumor resected, the thickness oftumor was greater than 5 mm.

Conclusion: Neck recurrence may represent residual disease; it has histologically unfavorable featuresand consequently a poor prognosis. The frequency of recurrence at level II emphasizes the need formeticulous dissection in this region, and tumor thickness needs to be considered in planning treatmentof the clinically negative neck.© 2002 American Association of Oral and Maxillofacial SurgeonsJ Oral Maxillofac Surg 60:748-753, 2002

Many factors can influence survival in patients withoral cancer; probably the most significant is lymphnode metastasis. Advanced tumor stage, site, andthickness together with adverse histologic featuresfurther reduce survival. When metastatic carcinoma ispresent in the neck nodes, the presence of extracap-sular spread further diminishes survival.1-7

The initial management of the neck remains a con-troversial area in early-stage oropharyngeal tumors.Khafif et al8 and Vandenbrouck et al9 have both ar-gued that elective neck dissection is not necessaryand that it is safe to adopt a “watch and wait” policy,delaying neck dissection until cervical metastases be-

come clinically evident. Part of the problem with thisapproach is that it is known that 20% to 30% ofclinically negative necks in early-stage tumors actuallyhave metastases present5,10 and that 3% to 16% overallhave extracapsular spread present when the neck isexamined histologically.10-12 Recently, much evidencehas been produced indicating that a “watch and wait”policy may not be safe, and Shingaki et al13 andHaddadin et al14 have shown in their retrospectivestudies that patients who have undergone delayedneck dissection have an increased incidence of extra-capsular spread, involvement of multiple levels, andreduced survival.It has been proposed that tumor thickness, histo-

logic differentiation, and DNA content are useful pre-dictors of occult nodal metastasis. Unfortunately,preoperative investigation with ultrasound and com-puted tomography has limited value in predictingwho will have positive nodes.4 The type of neckdissection required has also been subject to scrutiny.It is evident from the work of Byers et al15 and Wool-gar7 that tongue tumors do not necessarily spread in asequential fashion from the upper to lower levels.Isolated level IV node involvement has been demon-strated in 16% of tongue cancers. These authors stressthe importance of the inclusion of the level IV lymphnodes in the dissection. It has been shown that mod-

*Specialist Registrar, Blackburn Royal Infirmary, Lancashire, En-

gland.

†Senior House Officer, North Manchester General Hospital,

Manchester, England.

‡PhD Student, Manchester Royal Infirmary, Manchester, En-

gland.

§Consultant, Blackburn Royal Infirmary, Lancashire, England.

Address correspondence and reprint requests to Dr Godden:

Pigeon Box Farm, Rodley, Westbury-on-Severn, Gloucestershire,

GL14 1QZ England; e-mail: daryl.godden@doctors.org.uk

© 2002 American Association of Oral and Maxillofacial Surgeons

0278-2391/02/6007-0005$35.00/0

doi:10.1053/joms.2002.33240

748

ified neck dissections are equally successful in con-trolling early-stage neck disease.16,17 Neck failure oc-curs in around 10% to 30% of dissections, and thefrequency increases with increasing N stage16-21 (Fig 1).Radiotherapy has been used as an alternative to

dissection for the clinically negative neck. There islittle difference in the survival of patients treated withelective neck irradiation or dissection.22 It does seem,however, that elective neck irradiation offers a lowerneck failure rate than a “watch and wait” policy,23,24

although clearly neck irradiation does not allow forfull histologic assessment of the neck nodes, which ispossible when elective neck dissection is performed.The aims of this study were to identify the common

clinical and pathologic features of patients with recur-rent neck disease and analyze the data in an attemptto identify any common themes or patterns, includingthe type of treatment, outcome, and survival.

Patients and Methods

All patients in whom recurrent neck disease haddeveloped after attempted initial curative treatmentwere identified from the cancer databases of 3 max-illofacial centers: Blackburn Royal Infirmary, NorthManchester General Hospital, and Manchester RoyalInfirmary. The databases contained the records of 450patients, from which 35 patients could be identified

with recurrent neck disease. The pathology was re-viewed and data were recorded. Statistical analysiswas performed with the SPSS statistical package (SPSSInc, Chicago, IL) on a personal computer. The Kaplan-Meier and log rank tests were used to compare sur-vival. Statistical significance was determined at the 5%level.

Results

Twenty-two men and 13 women were studied. Themean age at presentation was 61.5 years. All except 2patients were smokers, and 13 patients had an alcoholconsumption in excess of the recognized safe limit of28 units per week. Tumors of the tongue were mostcommon, the posterior tongue disproportionately so(Fig 2). The majority of tumors were either moder-ately or poorly differentiated.Twenty-nine patients had surgical excision of the

primary tumor; the remainder were treated with ra-diotherapy to the primary site. One patient presentedwith neck disease from an unknown primary andreceived radiotherapy to the pharynx after a neckdissection. Of the surgical resections, 4 patients (14%)had tumor at the margin, 7 (24%) were less than 5mm, and 18 patients (62%) had marginal clearance ofmore than 5 mm.Eighteen patients had a neck dissection at initial

presentation; 17 had a “watch and wait” policy. Thedecision to operate on the neck was determined bytumor stage, patient’s health, and the clinician’s andpatient’s preference. The lowest level of dissectionwas level V in 9 cases, level IV in 6 cases, and level IIIin the remaining 3 patients. Ten patients who had aninitial neck dissection received postoperative radio-therapy, and 6 patients from the “watch and wait”group received radiotherapy to the site of the primaryand levels I and II neck nodes only as part of theirinitial treatment.The majority of patients had advanced tumors at

presentation. The mean stage was 3.7 for the neck

FIGURE 2. Site of primary tumor. FOM, floor of mouth.

FIGURE 1. The anatomic lymph node levels and staging of cervicalmetastases: N1, single node less than 3 cm; N2a, single node 3 to 6cm; N2b, multiple nodes less than 6 cm; N2c, bilateral or contralat-eral nodes; N3, single or multiple nodes more than 6 cm. (Adaptedfrom Color Atlas of Head and Neck Surgery (ed 1) by Shah JP, p 225,1987, by permission of the publisher Mosby.)

GODDEN ET AL 749

dissection group and 2 for the “watch and wait”group (Fig 3). Only 2 patients had primary tumors thatwere less than 5-mm thick.At initial presentation, the distribution of involved

lymph nodes in the neck dissection group reveals thatlevel II nodes were involved most commonly; 6 of the18 patients (33%) who had an initial neck dissectionhad level II disease. Five patients (28%) had patholog-ically clear necks, 9 patients (50%) had extracapsularspread present, and in 4 cases the metastasis wasintranodal. Three patients (17%) had multilevel dis-ease (Fig 4).The median time for development of neck recur-

rence in the initial neck dissection group was 15.16months. In the “watch and wait” group, it was 15.72months. The difference is not statistically significant(log rank, P � .35)The recurrence was managed by neck dissection in

25 patients, 21 had a level V dissection, 3 had a levelIV dissection, and 1 patient had a level III dissection.Of the 25 patients who had the recurrence managedby a neck dissection, 18 patients (72%) had adjunctiveradiotherapy after treatment. Ten patients had inop-erable disease in the neck or primary site and receivedpalliative treatment only. The distribution of the re-current metastasis was focused mainly at level II,where 20 of the 35 patients (57%) had disease (Fig 5).

Of the 18 patients who had an initial neck dissection,9 patients (50%) had failure at level II. Six patients hadrecurrence on the contralateral side of the neck afteran initial neck dissection, of whom five had posteriortongue or oropharyngeal tumors. Extracapsular spreadwas noted in the lymph nodes of 24 patients (96%),and in only 1 patient was the metastatic disease in-tranodal. Eight patients (23%) had neck failure atmultiple levels. Overall from recurrence to death, themedian survival time was 18.68 months (95% confi-dence interval, 12.29 to 25.07 months). The mediansurvival time of the initial neck dissection group was8 months, and that of the “watch and wait” group was19.3 months. This apparent difference is not statisti-cally significant (log rank test, P � .11) (Fig 6). Theseresults have been condensed for ease of reference inTable 1.

Discussion

This study confirms that patients with recurrentneck disease often get metastases at multiple levelsand invariably have extracapsular spread. The timetaken to develop recurrence is the same regardless ofwhether a neck dissection was performed. This sug-gests that in some circumstances, recurrence repre-

FIGURE 3. Tumor stage at presentation.

FIGURE 4. Distribution of lymph node metastasis of initial surgerygroup. Pathologically clear, 5; intranodal metastasis, 4; extracapsularspread, 9.

FIGURE 5. Distribution of recurrent neck disease.

FIGURE 6. Survival after neck recurrence.

750 RECURRENT NECK DISEASE IN ORAL CANCER

sents disease not successfully removed by electiveneck dissection. The median time to develop metas-tasis was 15 months; this is the interval it takes for themetastasis to become palpable.Recurrent neck disease has histologically unfavor-

able features; most patients had disease that had be-come extracapsular and frequently involved multiplelevels. Ten patients had disease that was inoperable,and in 4 patients who underwent a salvage neckdissection, removal of the tumor was incomplete.Contralateral disease occurred in 6 patients, of whom5 had posterior tongue or oropharyngeal tumors; thisconfirms the erratic nature of lymphatic spread in thisregion of the tongue.28 In addition, this study con-firms previous reports that show cervical metastasisgenerally occurs with thick tumors,3 that almost all

cases occurred with primary tumors that were morethan 5-mm thick, and that only 2 patients had tumorsthat were smaller. Tumor thickness needs to be con-sidered when planning the treatment of the N0 neck.Most patients who underwent an initial neck dis-

section had metastasis at level II; recurrence was alsolocated at level II. What makes level II special? Dis-section in this area is difficult, dissection along theinternal jugular vein deep to the posterior belly ofdigastric and up to the skull base is hazardous, anddissection around the spinal accessory nerve beneaththe sternomastoid is awkward. There are few studieson the extent of the necessary dissection in this re-gion, and it is not clear how far back one should go toremove any likely metastases. Byers has divided thisarea into level IIa, which is anterior to the internal

Table 1. SUMMARY OF TUMOR LOCATION, DEPTH, AND OUTCOME

Patient No. Site of TumorPathologic

TNM

Depth ofTumor(mm)

Initial NeckDissection

InitialRadiotherapy

Time toRecurrence

(mo)

SurvivalAfter

Recurrence(mo)

CurrentStatus*

1 FOM T2 N0 11 � — 15.2 12.1 D with SCC2 Posterior tongue T2 N2b 15 � � 9.1 4.2 D with SCC3 Lateral tongue T4 N2b 18 � � 2.2 3.8 D with SCC4 Mandibular

alveolusT4 N1 7 � � 4.6 4.7 D with SCC

5 Tonsillar fossa T2 N1 5 � � 24.2 3.0 D with SCC6 Lateral tongue T3 N2b 8 � � 3.9 2.2 D with SCC7 Lateral tongue T4 N0 10 � — 6.4 18.0 A, free of SCC8 Posterior tongue T4 N2b 24 � � 5.1 2.0 D with SCC9 FOM T4 N3 20 � � 12.0 0.7 D with SCC10 Posterior tongue T4 N2b 18 � � 6.4 13.9 A with SCC11 Tonsillar fossa T4 N1 20 � — 8.5 9.6 A, free of SCC12 Posterior tongue T4 N1 20 � — 14.9 51.6 A, free of SCC13 FOM T4 N1 10 � — 40.5 22.4 D with SCC14 FOM T2 N0 41 � — 5.1 12.0 A with SCC15 Unknown 1° TX N2b NA � � 18.2 8.1 D with SCC16 Maxilla T4 N2b 5 � — 10.8 12.4 A with SCC17 Posterior tongue T4 N0 14.5 � — 14.3 4.8 A, free of SCC18 Lateral tongue T2 N0 11 � � 8.1 5 D with SCC19 Maxilla T4 NX 60 — � 6.2 1.2 D with SCC20 Lateral tongue T1 NX 6 — — 11.5 15.8 D with SCC21 Lip T4 NX 10 — — 5.4 18.1 A, free of SCC22 Lateral tongue T2 NX 2 — — 15.9 42.3 D with SCC23 Lateral tongue T4 NX 10 — — 22.8 22.0 D with SCC24 Posterior tongue T4 NX Incisional — � 7.0 0.3 A with SCC25 Buccal mucosa T1 NX 3 — — 14.5 4.4 A, free of SCC26 Lateral tongue T4 NX 9 — — 3.4 19.6 A with SCC27 Posterior tongue T2 NX Incisional — � 49.5 46.6 A, free of SCC28 Maxilla T2 NX 8 — — 4.1 27.9 A with SCC29 Posterior tongue T1 NX 8 — — 79.1 13.1 D with SCC30 Tonsillar fossa T4 NX Incisional — � 11.5 19.3 D with SCC31 Lateral tongue T4 NX Incisional — � 15.7 15.7 D with SCC32 Lateral tongue T2 NX 11.4 — — 7.8 6.0 D with SCC33 Lateral tongue T4 NX Incisional — � 6.8 4.9 A with SCC34 Buccal mucosa T1 NX 8 — — 14.8 27.2 D with SCC35 Maxilla T2 NX 7 — — 3.2 18.7 D with SCC

Abbreviations: FOM, floor of mouth, A, alive; D, dead; SCC, squamous cell carcinoma.

GODDEN ET AL 751

jugular vein, and level IIb, which is distal to it.15 It hasbeen reported that levels IIb to IVb are at risk ofoccult metastases, although this is rare in the N0neck.29

Recurrent neck disease has an appalling prognosis.Although the “watch and wait” group survives longer,it should be remembered that they had early-stagedisease and comparison is inappropriate. This studycannot answer the question of whether it is safe toleave the N0 neck and adopt a “watch and wait”policy; such a question can only be answered by alarge randomized multicenter trial. Until such infor-mation is available, the authors argue that electiveneck dissection offers the ability to stage the neckaccurately and to plan subsequent treatment; in addi-tion, it allows vascular access for free tissue transfer.We suggest that the prevention of recurrent neck

disease lies with adequate primary treatment. Selec-tive neck dissection is as good as radical neck dissec-tion for the N0 or N1 neck.16,17 Adjuvant radiotherapyreduces the recurrence rate and improves survival inadvanced neck disease.21,25 Dissection must be metic-ulous, especially in the region of the skull base andaccessory nerve.Earlier detection of recurrence may be beneficial. It

has been shown that 20% of 1-cm nodes may containtumor. For a carcinoma to reach this size, it mustcontain 109 tumor cells, and the tumor will havedoubled in size 27 times. Detection of occult tumor isnotoriously difficult. The radiologic criteria for nodalinvolvement have been defined and include a mini-mum node length of 10 mm.26 Even so, 42% to56%10,11 of positive nodes in the clinically negativeneck have been shown to be less than 10 mm. Con-sequently, preoperative imaging with computed to-mography (CT) or magnetic resonance imaging (MRI)and clinical examination is only about 70% sensi-tive.10,27 This has 2 implications. First, is a “watch andwait” policy safe if you do not know whether there istumor present in the neck? Second, is it safe to per-form a selective neck dissection in a neck that maylater be shown on pathologic examination to containmultiple positive nodes, in which case a more radicalprocedure would have been indicated? Better detec-tion of occult tumour with sentinel node biopsy andpositron emission tomography scanning may help de-lineate in which patients a “watch and wait” policycan be adopted.The problem with the postoperative or postradio-

therapy patient lies in the fact that the routine detec-tion of recurrence relies solely on clinical examina-tion. This is made especially difficult by the presenceof scar tissue and edema, which can give the neck awooden feel and make interpretation difficult. Thisfailure to detect early recurrence allows continuedtumor growth, spread into adjacent tissue, and diffi-

culty with subsequent resection. There are no studieson postoperative screening in such patients, but withthe acceptance that CT and MR have their limitations,because the median time to detect recurrence is 15months, it is possible that a screening CT or MRI at 12months may detect a significant number of tumorsbefore they present. We are currently investigatingthis possibility.

References1. Dreyfuss AI, Clark JR: Analysis of prognostic factors in squa-

mous cell carcinomas of the head and neck. Haematol OncolClin North Am 5:701, 1991

2. Noguchi M, Kido Y, Kubota H, et al: Prognostic factors andrelative risk for survival in N1-3 oral squamous cell carcinoma:A multivariate analysis using Cox’s hazard model. Br J OralMaxillofac Surg 37:433, 1999

3. Spiro RH, Huvos AG, Wong G, et al: Predictive value of tumorthickness in squamous carcinoma confined to the tongue andfloor of mouth. Am Med J 152:345, 1986

4. Byers RM, El-Naggar AK, Lee YY, et al: Can we detect or predictthe presence of occult nodal metastases in patients with squa-mous carcinoma of the oral tongue? Head Neck 20:138, 1998

5. Woolgar JA, Scott J, Vaughan ED, et al: Survival, metastasis andrecurrence of oral caner in relation to pathological features.Ann R Coll Engl 77:325, 1995

6. Woolgar JA, Rogers S, West CR, et al: Survival and patterns ofrecurrence in 200 oral cancer patients treated by radical sur-gery and neck dissection. Oral Oncol 35:257, 1999

7. Woolgar JA: Histological distribution of cervical lymph nodemetastases from intraoral/oropharyngeal squamous cell carci-nomas. Br J Oral Maxillofac Surg 37:175, 1999

8. Khafif RA, Gelbfish GA, Tepper P, et al: Elective radical neckdissection in epidermoid cancer of the head and neck: Aretrospective analysis of 853 cases of mouth, pharynx, andlarynx cancer. Cancer 67:67, 1991

9. Vandenbrouck C, Sancho-Garnier H, Chassagne D, et al: Elec-tive versus therapeutic radical neck dissection in epidermoidcarcinoma of the oral cavity: Results of a randomized clinicaltrial. Cancer 46:386, 1980

10. Woolgar JA: Pathology of the N0 neck. Br J Oral Maxillofac Surg37:205, 1999

11. Woolgar JA, Vaughan ED, Scott J, et al: Pathological findings inclinically false-negative and false-positive neck dissections fororal carcinoma. Ann R Coll Surg Engl 76:237, 1994

12. Van den Brekel MW, van der Waal I, Meijer CCJ, et al: Theincidence of micrometastases in neck dissection specimensobtained from elective neck dissections. Laryngoscope 106:987, 1996

13. Shingaki S, Kobayashi T, Suzuki I, et al: Surgical treatment ofstage I and II oral squamous cell carcinomas: Analysis of causesof failure. Br J Oral Maxillofac Surg 33:304, 1995

14. Haddadin KJ, Soutar DS, Oliver RJ, et al: Improved survival forpatients with clinically T1/T2, N0 tongue tumors undergoing aprophylactic neck dissection. Head Neck 21:517, 1999

15. Byers RM,Weber RS, Andrews T, et al: Frequency and therapeuticimplications of “skip metastases” in the neck from squamouscarcinoma of the oral tongue. Head Neck 19:14, 1997

16. Anderson PE, Shah JP, Cambronero E, et al: The role of com-prehensive neck dissection with preservation of the spinalaccessory nerve in the clinically positive neck. Am J Surg168:499, 1994

17. Khafif RA, Gelbfish GA, Asase DK, et al: Modified radical neckdissection in cancer of the mouth, pharynx, and larynx. HeadNeck 12:476, 1990

18. Manni JJ, van den Hoogen FJA: Supraomohyoid neck dissectionwith frozen section biopsy as a staging procedure in the clin-ically node-negative neck in carcinoma of the oral cavity. Am JSurg 162:373, 1991

752 RECURRENT NECK DISEASE IN ORAL CANCER

19. Lingeman RE, Helmus C, Stephens R, et al: Neck dissection:Radical or conservative. Ann Otol 86:737, 1977

20. Spiro JD, Spiro RH, Shah JP, et al: Critical assessment of su-praomohyoid neck dissection Am J Surg 156:286, 1988

21. Byers RM: Modified neck dissection: A study of 967 cases from1970-1980. Am J Surg 150:414, 1985

22. Chow JM, Levin BC, Krivit JS, et al: Radiotherapy or surgery forsubclinical cervical node metastases. Arch Otolaryngol HeadNeck Surg 115:981, 1989

23. Leborgne F, Leborgne JH, Barlocci LA, et al: Elective neckirradiation in the treatment of cancer of the oral tongue. Int JRadiat Oncol Biol Phys 13:1149, 1987

24. Dearnaley DP, Dardoufas C, A’Hearn RP, et al: Interstitial irra-diation for carcinoma of the tongue and floor of mouth: RoyalMarsden Hospital experience 1970-1986. Radiother Oncol 21:183, 1991

25. Huang DT, Johnson CR, Schmidt-Ullrich R, et al: Postopera-tive radiotherapy in head and neck carcinoma with extra-capsular lymph node extension and/or positive resectionmargins: A comparative study. Int J Radiat Oncol Biol Phys23:737, 1992

26. Van den Brekel MW, Stel HV, Castelijns JA, et al: Cervival lymphnode metastasis: Assessment of radiologic criteria. Radiology177:379, 1990

27. Woolgar JA: T2 carcinoma of the tongue: The histopatholo-gist’s perspective. Br J Oral Maxillofac Surg 37:187, 1999

28. Prince S, Bailey BMW: Squamous carcinoma of the tongue:Review. Br J Oral Maxillofac Surg 37:164, 1999

29. Godden DRP, Cardozo C, Prescott R, et al: 2b or not 2b: Fatis the question. Br J Oral Maxillofac Surg 39:237, 2001(abstr)

753