Improved Staging of Cervical Metastases in ClinicallyNode-Negative Patients With Head and Neck Squamous

Cell Carcinoma

G. L. Ross, MRCSEd, D. S. Soutar, ChM, D. G. MacDonald, FRCPath, T. Shoaib, FRCSEd,I. G. Camilleri, FRCS (Plast), and A. G. Robertson, PhD

Background: The management of the N0 neck in oral and oropharyngeal cancer is oftendetermined by the risk of metastases related to features of the primary tumor. Where the risk ofmetastases is 20%, elective neck dissection (END) has been advocated. This study reviewedclinical staging, surgical staging, pathologic staging, and histopathologic parameters to determinethe prediction of nodal metastases and micrometastases in patients with head and neck squamouscell carcinoma.

Methods: A prospective series of 61 clinically neck nodenegative patients undergoing surgicalresection of a T1/2 intraoral or oropharyngeal invasive squamous cell carcinoma and surgicalstaging of the neck, with sentinel node biopsy (SNB) alone or SNB-assisted END, between June1998 and March 2002 were included in this study.

Results: Pathologic upstaging of the clinically N0 neck occurred in 27 (44%) of 61 patients.Routine pathology with hematoxylin and eosin upstaged disease in 22 of 27 patients (sensitivity of81%). Five patients with micrometastasis were staged pN1mi after stepped serial sectioning andimmunohistochemistry. Tumor thickness, a noncohesive invasive front, and perineural and boneinvasion were all histological predictors for cervical metastases. Five patients with micrometastaseswere staged pN1mi.

Conclusions: Both clinical staging and routine pathologic staging underestimate the presence ofnodal metastases. Staging with either SNB alone or SNB-assisted END shows promise in themanagement of the N0 neck by identifying patients with micrometastases (pN1mi).

Key Words: Head and neckNeoplasmsSentinel node biopsyElective neck dissectionCervical metastases.

Numerous retrospective studies indicate that patientswith head and neck squamous cell carcinoma (HNSCC)whose lymph nodes are treated on an elective basis havemuch better regional control and survival outcome thanpatients whose necks are salvaged after failure of treat-ment of the primary site alone.14 Because the prognosis

of delayed cervical metastasis is poor, most authorscautiously favor elective neck dissection (END) and be-lieve that the elective dissections are predominantly stag-ing procedures and that the information obtained bypathologic examination should be used to direct adjuvanttreatments.5,6 Where micrometastasis has been found inelective lymph node dissections, surgery has often beenconsidered curative.

Assessment of the status of the neck is imperative forstaging/treatment planning. However, 70% of HNSCCpatients with N0 necks will have no detectable metastaticdeposits within the neck specimen and may therefore notbenefit from an END. Assessment with clinical palpationalone has proven unreliable for the detection of lymphnode disease and is hence unreliable for staging purpos-es.7,8 Since the advent of sophisticated imaging technol-

Received January 2, 2003; accepted October 1, 2003.From the Plastic Surgery Unit (GLR, DSS, TS, IGC), Canniesburn

Hospital, Bearsden, Glasgow; Oral Pathology Unit (DGM), GlasgowDental Hospital and School, Glasgow; and Beatson Oncology Centre(AGR), Western Infirmary, Glasgow, United Kingdom.

Address correspondence and reprint requests to: Gary Ross, MRC-SEd, Plastic Surgery Unit, Canniesburn Hospital, Switchback Road,Bearsden, Glasgow, G61 1QL, UK; Fax: 00-44-141-211-5652; E-mail:gary.ross@canniesburn.org.Published by Lippincott Williams & Wilkins 2004 The Society of SurgicalOncology, Inc.

Annals of Surgical Oncology, 11(2):213218DOI: 10.1245/ASO.2004.03.057

213

ogy, clinicians have focused on diagnostic refinement inan attempt to define and hence stage the N0 neck. Com-pared with routine pathologic evaluation of nodal dis-ease, diagnostic imaging such as computed tomogra-phy,912 magnetic resonance imaging,10,12 and ultrasoundscan12,13 have shown a specificity of 75% to 92%. Ul-trasound fine-needle aspiration cytology12,14,15 has beenreported as showing 100% specificity, although the sen-sitivity is between 42% and 73%, which is similar to thesensitivities of the other imaging techniques alreadymentioned (40% to 80%).913 Recent reports of the useof positron emission tomography show a sensitivity of87%,16 a specificity of 90%,17 and a low false-positiverate. The $1 to $2 million cost of a dedicated positronemission tomography scanner and the small number offalse positives prevent this imaging option as a standardof care, and its use as an indicator for surgery is stillcontroversial.17

Because uniform criteria to identify the presence ofmetastatic disease have not been established diagnosti-cally, the gold standard to stage the neck remains theidentification of metastatic deposits by careful examina-tion of the surgical specimen obtained at the time ofEND. END, however, carries with it associated morbid-ity and an effect on quality of life1821 and does notbenefit most patients. In HNSCC, a technique of super-selecting lymph nodes that reflect the status of the lymphnode basin is now available in the form of sentinel nodebiopsy (SNB). This less-invasive staging technique hasbeen mainly used in other cancers, such as melanomaand breast cancer.22,23 SNB was used initially in HNSCCto assist END2426 and more recently has been describedalone as a staging tool in HNSCC.2628

The reliability of routine histopathologic examinationof lymph nodes in neck dissection specimens, thus es-tablishing the number and level of lymph node metasta-ses, is seldom questioned. Guidelines for such identifi-cation are available in the United Kingdom from theroyal college of pathologists.29 However, the sensitivityof current routine histopathologic methods in assessingregional lymph node involvement is imperfect in that it,like clinical staging, can rarely detect the presence ofmicrometastasis within those nodes.30 Also, greater sam-pling of lymph nodes will lead to a greater probability ofmetastasis detection.31 It is important to realize that theterm micrometastasis is not synonymous with occultdisease. The term occult describes nodal metastatic dis-ease that has not been detected by clinical staging inves-tigations. The accuracy of staging is influenced by therange of available investigations, as well as by multipletechnical and patient factors.32 The pathologic extent ofoccult disease varies from a single positive node to

multiple positive nodes. Hence, outcome with minimalmetastatic disease and the wide variation in the patho-logic stage of N0 neck dissections account for the diverseoutcome of patients with occult nodal disease.32 Thedefinition of micrometastatic disease is a single depositor multiple deposits of tumor within the lymph nodesinuses with minimal replacement of the nodal architec-ture and measuring in total no more than 2 mm at anylevel of sectioning.33 The nomenclature of N1mi hasbeen attributed to such patients within this article.

We looked prospectively at a group of patients inwhom SNB-assisted END or SNB alone was used tostage the clinically N0 neck for patients with T1/2 tu-mors of the oral cavity or oropharynx. We assessedclinical staging (cT), surgical staging, pathologic staging(pT), and histopathological parameters to determine theprediction of nodal metastasis (pN) and micrometastasis(pNmi).

METHODS

A prospective series of 61 clinically node-negativepatients undergoing surgical resection of a T1/2 intraoralor oropharyngeal invasive SCC and surgical staging ofthe neck between June 1998 and March 2002 were in-cluded in this study. None of the patients had receivedpreoperative radiotherapy, chemotherapy, or previoussurgery other than the primary procedure. Only patientsin whom tumor thickness was able to be accuratelyassessed and recorded were included.

Thirty-eight patients underwent SNB to stage the clin-ically N0 neck. Twenty-three patients underwent anSNB-assisted END in the form of a modified radicalneck dissection.34 This comprehensive neck dissectionallowed a complete and thorough investigation of theentire nodal basin for accurate comparison with the sen-tinel node pathology.

The SNB procedure has already been de-scribed24,25,27,28 and involved the triple diagnostic tech-nique of preoperative lymphoscintigraphy, intraoperativeuse of blue dye, and a gamma probe detector. All nodes,including sentinel nodes, were fixed in 10% neutralbuffered formalin. After fixation, they were bisectedthrough the hilum (if this was identifiable) or through thelong axis of the node. If the thickness of the halves wasmore than 2.5 mm, the nodes were further trimmed toprovide additional 2.5-mm blocks. One routine hematox-ylin and eosin (H&E)-stained section was taken fromeach block.

Twenty-two patients were initially staged positive onH&E. Out of the 39 remaining patients, full pathologicevaluation for micrometastasis was performed in 27 pa-

214 G. L. ROSS ET AL.

Ann Surg Oncol, Vol. 11, No. 2, 2004

tients. Within the full pathologic protocol, if H&E sec-tions were found to be free from tumor, all blocks werestep serial sectioned with a ribbon of six exactly serialsections cut at 150-m intervals. The numbered sectionswere individually mounted. Section 3 from each levelwas stained with H&E and examined. If these sectionsseemed to be negative, section 2 at each level wasreacted with cytokeratin antibody AE1/AE3. Immunocy-tochemically positive areas were compared with the cor-responding area in the H&E-stained serial section toconfirm that the positivity represented viable tumor cells.Patients pathologically upstaged with step serial section-ing and immunohistochemistry were staged as pN1mi.33

Histopathologic parameters of disease that were re-corded prospectively were evaluated. These included tu-mor thickness, maximum tumor diameter, invasive front,and presence of perineural, perivascular, and bone inva-sion. Statistical analysis was performed with SPSS 8.0software (SPSS Inc., Chicago, IL) by using the t-test tocompare primary parameters between groups.

RESULTS

The tumor staging for 61 patients was cT1, n 34;cT2, n 27; pT1, n 36; pT2, n 16; 6pT2, n 1;pT3, n 1; and pT4, n 8. Nine patients (15%) wereupstaged from early T1/2 to advanced pT3/4 disease.

Pathologic upstaging of the clinically N0 neck oc-curred in 27 (44%) of 61 patients. Patients nodal stageswere pN0, n 34; pN1mi, n 5; pN1, n 11; andpN2B, n 11. Pathologic upstaging with routine H&Eevaluation of sentinel nodes was 22 (36%) of 61. The useof immunocytochemistry and step serial sectioning up-staged disease in 5 (19%) of 27 patients.

Cervical metastases were present in 12 (29%) of 34 T1tumors and 17 (63%) of 27 T2 tumors (P .05). Clinicaland pathologic staging and rates of cervical metastasisare listed in Table 1.

Although tumor thickness was a significant predictorof metastasis (P .05), there were cervical metastasespresent in 3 (23%) of 13 tumors with a thickness of 2mm and 24 (50%) of 48 tumors with a thickness of 2

mm (P .063; Table 2). For tongue tumors, none of 3tumors with a thickness of 2 mm had cervical metas-tasis, compared with 7 of 18 tumors with a thickness of2 mm (not significant). Although this was not statisti-cally significant, there were five tongue tumors with atumor thickness of 3 mm, none of which developedcervical metastases.

Cervical metastases were present in 9 (45%) of 20floor-of-mouth tumors, 7 (33%) of 21 anterior tonguetumors, 3 (75%) of 4 posterior tongue tumors, 5 (71%) of7 retromolar trigone (RMT) tumors, and 3 (33%) of 9other sites (not significant; Table 3). The presence ofperineural invasion was recorded in 44 patients; 8 of 9tumors with perineural invasion and 8 of 35 withoutperineural invasion had cervical metastasis (P .05).The sensitivity of using perineural invasion as a predictorwas 8 (50%) of 16, and specificity was 27 (96%) of 28.

An invasive front was documented in 36 patients aseither cohesive or noncohesive. Seven of 11 tumors witha noncohesive front, compared with 5 of 25 tumors witha cohesive front, had cervical metastasis (P .05). Thesensitivity of using the presence of a noncohesive frontas a predictor of metastases was 58% (7 of 12), and thespecificity was 83% (20 of 24).

Both an invasive front and the presence of perineuralinvasion were recorded in 33 cases. All five tumors withboth a noncohesive front and the presence of perineuralinvasion developed cervical metastasis. There was noevidence of cervical metastasis in 86% (18 of 21) oftumors with a cohesive front and no evidence of peri-neural invasion.

The presence of vascular invasion was recorded in 46patients: 3 of 4 patients with tumors with vascular inva-sion and 15 of 42 without vascular invasion developedcervical metastases (P .214). Documented presence orabsence of bone invasion was available in 35 patients.Six of 8 with bone invasion compared with 8 of 27 withno bone invasion developed metastasis (P .05).

A total of 25 patients staged SNB negative have beenfollowed up for an average of 20 months (range, 936months). Twelve patients staged SNB positive under-went a therapeutic lymph node dissection (TLND). One

TABLE 1. Cervical metastases detected pathologically (pN) according to clinical (cT) andpathologic (pT) tumor staging

Variable pN0 pN pN1mi pN1 pN2B% Patients with

metastases

cT1 24 10 2 3 5 29cT2 10 17 3 8 6 63pT1 23 13 4 4 5 36pT2 9 7 1 3 3 44pT3/4 2 7 0 4 3 78

215SENTINEL NODE BIOPSY IN HEAD AND NECK CANCER

Ann Surg Oncol, Vol. 11, No. 2, 2004

patient staged SNB positive did not undergo a therapeu-tic neck dissection because of concomitant medical prob-lems. One patient developed nodal recurrence 26 monthsafter SNB and underwent a TLND. This patient devel-oped further nodal disease 3 months later after TLND.Eleven patients staged negative with SNB-assisted ENDhave been followed up on average for 32 months (range,951 months). None of these patients has developednodal recurrence.

Three patients have developed local recurrence, threehave developed a second primary tumor, and two havedeveloped distant metastases. Two patients are dead be-cause of disease, and three are dead because of unrelatedcauses during the follow up period. A total of 56 patientscontinued to be followed up in the clinic.

SNB sensitivity was 25 of 27 (sensitivity 93%); bothfailures were in tumors in the floor of mouth. A total of17 (71%) of 24 patients staged SNB positive had nofurther disease in the TLND.

DISCUSSION

A total of 27 (44%) of 61 patients in our series wereupstaged from cN0 to pN. Clinical palpation alone hashence proven unreliable in the detection of lymph nodedisease and is unreliable for staging purposes.35,36

Tumor thickness, the presence of a noncohesive inva-sive front, the presence of perineural invasion, and thepresence of bone invasion were statistically significanthistopathologic predictors of the presence of cervicalmetastasis. Tumor site and the presence of vascular in-

vasion were not able to predict the presence of cervicalmetastasis. As with clinical palpation, however, we can-not rely on histopathologic parameters alone to deter-mine the need for an END, with individual sensitivitiesall 60%. The only way, therefore, currently to stagepatients adequately is pathologic dissection of the lymphnode basin.

The use of SNB allows a more detailed search formetastasis in relatively few nodes. The use of extensivesampling of nodes in cancers such as melanoma andbreast cancer has shown the percentage of patients beingupstaged by serial sectioning as 3% to 33% and byimmunocytochemistry as 8% to 41%.35

The prognostic significance of micrometastasis in ma-lignant melanoma has been highlighted in a study byGershenwald et al.,36 in which a large number of recur-rences in melanoma could have been prevented hadextensive sampling been performed prospectively ratherthan in retrospect. In breast cancer, reports suggest thatthere is a worse prognosis associated with even a singlemicrometastasis.37 The Ludwig trial38 noted that the find-ing of a single micrometastasis in a single node by moreextensive pathologic evaluation was associated with adecreased survival of approximately 50%. Similar prog-nostic influences have been reported for othertumors.3943

In HNSCC, relying on serial sectioning and immuno-cytochemistry with an antibody to pan-cytokeratin, Am-brosch and Brinck44 uncovered eight micrometastases(7.9%) in six specimens from six patients. Woolgar32found that 8% of patients with oral or oropharyngeal

TABLE 3. Nodal staging according to tumor site

Site pN0 pN pN1mi pN1 pN2B

Buccal 1 0 0 0 0FOM 9 11 1 6 4Hard palate 2 1 1 0 0Lower alveolus 2 0 0 0 0RMT 2 5 1 0 4Soft palate 2 0 0 0 0Anterior tongue 14 7 2 4 1Tonsil 1 0 0 0 0Post tongue 1 3 0 1 2

FOM, floor of mouth; RMT, retromolar trigone.

TABLE 2. Nodal staging (pN) in comparison to tumor thickness

Thickness (mm) pN0 pN pN1mi pN1 pN2B% Patients with

metastases

2 10 3 0 0 3 232 to 5 15 7 4 1 2 415 to 10 5 10 1 7 2 6710 4 7 0 3 4 64

216 G. L. ROSS ET AL.

Ann Surg Oncol, Vol. 11, No. 2, 2004

squamous cell carcinoma had micrometastasis by usingstep serial sectioning at 100-m intervals.

For breast cancer and melanoma, extensive his-topathologic investigation, including step serial H&Esectioning and immunocytochemistry, is now consideredmandatory for reliable detection of metastases in sentinelnodes.4547 As yet, molecular techniques such as reversetranscription-polymerase chain reaction, which may de-tect a single cancer cell among 106 normal cells, is notpart of the recommended protocol to evaluate sentinelnodes.4547 It remains to be proven whether reverse tran-scription-polymerase chain reaction may identify thosefew patients not identified by step serial sectioning andimmunohistochemistry.4547

In breast cancer and melanoma, the resultant upstagingof disease by the presence of micrometastases withincurrent staging criteria remains controversial.45 Althoughthe micrometastases must be .2 mm in greatest dimen-sions to be classified as pN1mi33 rather than N1, it is theupstaging of micrometastasis from N0 to N1mi that isdebated. The College of American Pathologists has sug-gested that immunocytochemistry-positive cells shouldbe visualized and confirmed on the H&E-stained sectionto qualify as a micrometastasis before disease is upstagedfrom N0 to N1mi. With this classification, difficultiesstill exist in pathologic interpretation of small clusters ofepithelial cells as benign, degenerating malignant, orviable malignant. This has been addressed by the Inter-national Union Against Cancer (UICC). For staging aspNmi, a nodal metastasis must extravasate out of lym-phatic sinuses, thus demonstrating stromal reaction andtumor cell proliferation.33

Pathologic staging with routine H&E, like clinicalpalpation, proved unreliable in predicting metastasis inour series; 22 (36%) of 61 patients were upstaged fromcN0 to pN with routine H&E. The sensitivity of routinepathology was 22 (81%) of 27, and 5 (19%) of 27patients were upstaged from pN0 to pN1mi with addi-tional pathologic evaluation of sentinel nodes.

It may not be cost-effective to stage a lymph nodedissection specimen for the presence of micrometastasisin the absence of metastases on routine H&E.32,43,48 SNBor SNB-assisted END, however, is an effective and re-liable means of assessing nodal basins for micrometas-tasis with additional pathologic techniques being used onthe sentinel nodes only.

The identification of micrometastasis by additionalpathologic techniques, such as step serial sectioning andimmunocytochemistry, provides a dilemma for the cur-rent staging classifications of the UICC in HNSCC.Patients who are upstaged by additional pathologic meth-ods but show no further disease in the neck specimen

have only micrometastatic disease. These patients maywell have a different prognosis than patients with one2-mm metastasis, which is classified as N1 in the currentUICC classification.49 The nomenclature of N1mi hasbeen attributed to such patients within this article, and itremains to be seen whether this will be adopted by theUICC in the future for HNSCC, as it has already done forboth breast cancer and melanoma.33,50

There may be a prognostic difference between patientsstaged as N1mi and N1. Further adjuvant trials may bepossible for patients staged as N1mi who currently re-ceive no form of adjuvant therapy in our series.

SNB alone in combination with additional pathologictechniques was successful in upstaging the clinically N0neck in 25 (93%) of 27 patients. It remains unclearwhether SNB will replace END in treatment of the clin-ically N0 neck, and we require the results of multicentertrials to determine this.2628 However, it is clear that thebest staging option for patients remains a combination ofclinical staging, surgical staging, and pathologic staging.

ACKNOWLEDGMENTSThe acknowledgments are available online in the full-text version at www.annalssurgicaloncology.org. They

are not available in the PDF version.

REFERENCES

1. Haddadin KJ, Soutar DS, Oliver RJ, Webster MH, Robertson AG,MacDonald DG. Improved survival for patients with clinicallyT1/T2, N0 tongue tumors undergoing a prophylactic neck dissec-tion. Head Neck 1999;21:51725.

2. Lydiatt DD, Robbins KT, Byers RM, Wolf PF. Treatment of stageI and II oral tongue cancer. Head Neck 1993;15:30812.

3. Franceschi D, Gupta R, Spiro RH, Shah JP. Improved survival inthe treatment of squamous cell carcinoma of the oral tongue. Am JSurg 1993;166:3605.

4. Yuen APW, Wei WI, Wong YM, Tang KCT. Elective neck dis-section versus observation in the treatment of early oral tonguecarcinoma. Head Neck 1997;19:5838.

5. Kligerman J, Lima RA, Soares JR, et al. Supraomohyoid neckdissection in the treatment of T1/T2 squamous cell carcinoma oforal cavity. Am J Surg 1994;168:3914.

6. Forastiere AA, Ang K, Argiris A, et al. NCCN practice guidelinesfor cancer of the head and neck, version 1.2002. Rockville, PA.Available at: http://www.nccn.org/physician_gls/index.html. Ac-cessed April 11, 2003.

7. OBrien CJ, Traynor SJ, McNeil E, McMahon JD, Chaplin JM.The use of clinical criteria alone in the management of the clini-cally negative neck among patients with squamous cell carcinomaof the oral cavity and oropharynx. Arch Otolaryngol Head NeckSurg 2000;126:3605.

8. Woolgar JA, Beirne JC, Vaughan ED, Lewis-Jones HG, Scott J,Brown JS. Correlation of histopathologic findings with clinical andradiologic assessments of cervical lymph-node metastases in oralcancer. Int J Oral Maxillofac Surg 1995;24:307.

9. Stern WB, Silver CE, Zeifer BA, Persky MS, Heller KS. Com-puted tomography of the clinically negative neck. Head Neck1990;12:10913.

217SENTINEL NODE BIOPSY IN HEAD AND NECK CANCER

Ann Surg Oncol, Vol. 11, No. 2, 2004

10. Friedman M, Mafee MF, Pacella BL Jr, Strorigl TL, Dew LL,Toriumi DM. Rationale for elective neck dissection in 1990. La-ryngoscope 1990;100:549.

11. Moreau P, Goffart Y, Collignon J. Computed tomography ofmetastatic cervical lymph nodes. A clinical, computed tomo-graphic, pathologic correlative study. Arch Otolaryngol Head NeckSurg 1990;116:11903.

12. van den Brekel MW, Castelijns JA, Stel HV, Golding RP, MeyerCJ, Snow GB. Modern imaging techniques and ultrasound-guidedaspiration cytology for the assessment of neck node metastases: aprospective comparative study. Eur Arch Otorhinolaryngol 1993;250:117.

13. John DG, Anaes FC, Williams SR, et al. Palpation compared withultrasound in the assessment of malignant cervical lymph nodes. JLaryngol Otol 1993;107:8213.

14. Righi PD, Kopecky KK, Caldemeyer KS, Ball VA, WeisbergerEC, Radpour S. Comparison of ultrasound-fine needle aspirationand computed tomography in patients undergoing elective neckdissection. Head Neck 1997;19:60410.

15. Takes RP, Knegt P, Manni JJ, et al. Regional metastasis in headand neck squamous cell carcinoma: revised value of US withUS-guided FNAB. Radiology 1996;198:81923.

16. Kau RJ, Alexiou C, Laubenbacher C, Werner M, Schwaiger M,Arnold W. Lymph node detection of head and neck squamous cellcarcinomas by positron emission tomography with fluorodeoxy-glucose F 18 in a routine clinical setting. Arch Otolaryngol HeadNeck Surg 1999;125:13228.

17. Stokkel MP, ten Broek FW, Hordijk GJ, Koole R, van Rijk PP.Preoperative evaluation of patients with primary head and neckcancer using dual-head 18fluorodeoxyglucose positron emissiontomography. Ann Surg 2000;231:22934.

18. Ewing MR, Martin H. Disability following radical neck dissection.Cancer 1952;5:87383.

19. Shone GR, Yardley MP. An audit into the incidence of handicapafter unilateral radical neck dissection. J Laryngol Otol 1991;105:7602.

20. Kuntz AL, Weymuller EA. Impact of neck dissection on quality oflife. Laryngoscope 1999;109:13348.

21. Saffold SH, Wax MK, Nguyen A, et al. Sensory changes associ-ated with selective neck dissection. Arch Otolaryngol Head NeckSurg 2000;126:4258.

22. Morton DL, Wen DR, Wong JH, et al. Technical details of intra-operative lymphatic mapping for early stage melanoma. Arch Surg1992;127:3929.

23. Alex JC, Krag DN. Gamma-probe guided localization of lymphnodes. Surg Oncol 1993;2:13743.

24. Shoaib T, Soutar DS, Prosser JE, et al. A suggested method forsentinel node biopsy in squamous cell carcinoma of the head andneck. Head Neck 1999;21:72833.

25. Shoaib T, Soutar DS, MacDonald DG, et al. The accuracy of headand neck carcinoma sentinel lymph node biopsy in the clinicallyN0 neck. Cancer 2001;91:207783.

26. Ross GL, Shoaib T, Soutar DS, et al. The first internationalconference on sentinel node biopsy in mucosal head and neckcancer and adoption of a multicenter trial protocol. Ann Surg Oncol2002;9:40610.

27. Ross GL, Shoaib T, Soutar DS, et al. The use of sentinel nodebiopsy to upstage the clinically N0 neck in head and neck cancer.Arch Otolaryngol Head Neck Surg 2002;128:128791.

28. Ross GL, Soutar DS, Shoaib T, et al. The ability of lymphoscin-tigraphy to direct sentinel node biopsy in the clinically N0 neck forpatients with head and neck squamous cell carcinoma. Br J Radiol2002;75:9508.

29. The Royal College of Pathologists. Minimum dataset for head andneck carcinoma histopathology reports. London: The Royal Col-lege of Pathologists, 1998.

30. Ferlito A, Devaney KO, Rinaldo A, Devaney SL, Carbone A.Micrometastases: have they an impact on prognosis? Ann OtolRhinol Laryngol 1999;108:11859.

31. Wilkinson EJ, Hause L. Probability in lymph node sectioning.Cancer 1974;33:126974.

32. Woolgar JA. Micrometastasis in oral/oropharyngeal squamous cellcarcinoma: incidence, histopathological features and clinical im-plications. Br J Oral Maxillofac Surg 1999;37:1816.

33. Hermanek P, Hutter RVP, Sobin LH, Wittekind C. Classificationof isolated tumour cells and micrometastasis. Cancer 1999;86:266873.

34. Robbins KT. Classification of neck dissection: current conceptsand future considerations. Otolaryngol Clin North Am 1998;31:63955.

35. van Diest PJ, Peterse HL, Borgstein PJ, Hoekstra O, MeijerCJ. Pathological investigation of sentinel lymph nodes. Eur J NuclMed 1999;26:S439.

36. Gershenwald JE, Colome MI, Lee JE, et al. Patterns of recurrencefollowing a negative sentinel lymph node biopsy in 243 patientswith stage I or II melanoma. J Clin Oncol 1998;6:225360.

37. Dowlatshahi K, Fan M, Bloom KJ, Spitz DJ, Patel S, Snider HC Jr.Occult metastases in the sentinel lymph nodes of patients withearly stage breast carcinoma: a preliminary study. Cancer 1999;86:9906.

38. Cote RJ, Peterson HF, Chaiwun B, et al. Role of immunohisto-chemical detection of lymph node metastases in the managementof breast cancer. International Breast Cancer Study Group. Lancet1999;354:896900.

39. Ishida K, Katsuyama T, Sugiyama A, Kawasaki S. Immunohisto-chemical evaluation of lymph node micrometastases from gastriccarcinomas. Cancer 1997;79:106976.

40. OSullivan GC, Collins JK, Kelly J, Morgan J, Madden M, Shana-han F. Micrometastases: marker of metastatic potential or evidenceof residual disease? Gut 1997;40:5125.

41. Natsugoe S, Mueller J, Stein HJ, Feith M, Hofler H, Siewert JR.Micrometastasis and tumour cell microinvolvement of lymphnodes from esophageal squamous cell carcinoma. Frequency, as-sociated tumour characteristics, and impact on prognosis. Cancer1998;83:85866.

42. Glickman JN, Torres C, Wang HH, et al. The prognostic signifi-cance of lymph node micrometastasis in patients with esophagealcarcinoma. Cancer 1999;85:76978.

43. Yokoyama N, Shirai Y, Hatakeyama K. Immunohistochemicaldetection of lymph node micrometastases from gall bladder carci-noma using monoclonal anticytokeratin antibody. Cancer 1999;85:14659.

44. Ambrosch P, Brinck U. Detection of nodal micrometastasis in headand neck cancer by serial sectioning and immunostaining. Oncol-ogy (Huntingt) 1996;10:12216.

45. Turner RR. Histopathologic assessment of the sentinel lymph nodein breast cancer. Ann Surg Oncol 2001;8:56S59S.

46. Cochran AJ, Huang RR, Guo J, Wen DR. Current practice andfuture directions in pathology and laboratory evaluation of thesentinel node. Ann Surg Oncol 2001;8:13S17S.

47. van Diest PJ, Torrenga H, Meijer S, Meijer CJ. Pathologic analysisof sentinel lymph nodes. Semin Surg Oncol 2001;20:23845.

48. van den Brekel MW, Stel HV, van der Valk P, van der Waal I,Meyer CJ, Snow GB. Micrometastases from squamous cell carci-noma in neck dissection specimens. Eur Arch Otorhinolaryngol1992;249:34953.

49. Sobin LH, Wittekind C, eds. International Union Against Cancer:TNM Classification of Malignant Tumours. 6th ed. New York:Wiley, 2002.

50. Balch CM, Buzaid AC, Soong SJ, et al. Final version of theAmerican Joint Committee on Cancer staging system for cutaneousmelanoma. J Clin Oncol 2001;19:363548.

218 G. L. ROSS ET AL.

Ann Surg Oncol, Vol. 11, No. 2, 2004