Sentinel Lymph Node Biopsy in Patients WithDiagnostically Controversial SpitzoidMelanocytic Tumors

Christina M. Lohmann, M.D., Daniel G. Coit, M.D., Mary S. Brady, M.D.,Marianne Berwick, Ph.D., and Klaus J. Busam, M.D.

Melanomas can be difficult to diagnose histologically if theydeviate in their growth pattern or cytology only minimally froma nevus. On occasion, even experts on melanocytic lesions maynot reach a consensus on whether a lesion is a benign butunusual nevus or a malignant melanoma mimicking a nevus.This diagnostic dilemma is particularly well known for thedistinction of Spitz nevus from melanoma. Diagnostic uncer-tainty and disagreement among consultant pathologists lead toconfusion about the prognosis and clinical management of pa-tients. In this study we present the clinical and pathologic find-ings of 10 patients with diagnostically controversial melano-cytic tumors, who underwent sentinel lymph node biopsy. In allof these cases, the diagnostic controversy among experts wasbetween Spitz nevus and melanoma. Seven patients were fe-male, and three were male, ranging in age from 7 to 46 years(mean 21 years). Histologic examination of the sentinel lymphnodes revealed tumor deposits in the lymph node parenchymain 5 of 10 patients. Among patients with positive sentinellymph nodes, two had satellite nodules and one showed addi-tional tumor deposits in three nonsentinel regional lymphnodes. All patients are alive and free of disease with a follow-up of 10–54 months (mean 34 months). Our study illustrates therole of a sentinel lymph node biopsy in the evaluation of pa-tients with diagnostically controversial melanocytic tumors.Although the presence of metastatic tumor deposits in thesentinel lymph node supports the diagnosis of malignantmelanoma, further studies are needed to determine the prog-nostic significance of the sentinel lymph node findings in suchpatients.Key Words: Sentinel lymph node—Spitz nevus—Melanoma.

Am J Surg Pathol 26(1): 47–55, 2002.

Since Sophie Spitz reported a series of “melanomas inchildhood” and discussed the difficulty of distinguishing

these tumors from conventional adult-type melanomas,45

the diagnosis of Spitz nevi and melanomas mimickingSpitz nevi has remained one of the more contentiousissues in dermatopathology.13,22,43 Spitz nevi are definedby a number of histologic features, which have evolvedover the past decades.1–3,10,13,22,35,47,48 The typical com-pound Spitz nevus is by silhouette a symmetric dome-shaped lesion composed of large spindle and epithelioidmelanocytes with eosinophilic cytoplasm. If the dermalmelanocytes show features of maturation and lackmarked pleomorphism and mitotic figures, the lesion isreadily recognized as a benign nevus.1–3,10,13,22,35,47,48

Clefts between melanocytes and neighboring cells,and/or large dull pink globules at the dermoepidermaljunction provide additional supportive evidence for a di-agnosis of Spitz nevus.

However, there are atypical melanocytic tumors withfeatures of a Spitz nevus, which deviate in growth patternand/or cytology from a benign nevus.5 Such lesions maycontain many mitoses, expansile dermal nests, minimalmaturation, marked pleomorphism, and other features,which are of concern for melanoma. Because they dif-fer in appearance from “conventional” adult-typemelanomas and tend to occur in young patients, patholo-gists usually hesitate to render a diagnosis of malignantmelanoma.

Various terms have been proposed to separate suchdiagnostically difficult melanocytic lesions from conven-tional melanomas, including “minimal deviation mela-noma, Spitz nevus-type,” “borderline nevomelanocyticneoplasm,” “nevoid melanoma,” and “atypical Spitznevus/tumor,”5,27,28,30,35–38,43,46,51,52 with no consensuson the most suitable term.37 In this study we refer to thelesions in question pragmatically as “diagnostically con-troversial Spitzoid melanocytic tumors” (DCSMT). Theyrepresent a morphologically heterogeneous group of me-lanocytic tumors but have in common the fact that expertpathologists reviewing their histology discuss the dis-

From the Departments of Pathology (C.M.L., K.J.B.), Surgery(D.G.C., M.S.B.), and Epidemiology and Biostatistics (M.B.),Memorial Sloan-Kettering Cancer Center, New York, NY, U.S.A.

Address correspondence and reprint requests to Klaus J. Busam,MD, Department of Pathology, Memorial Sloan-Kettering CancerCenter, 1275 York Ave., New York, NY 10021, U.S.A.; e-mail:busamk@mskcc.org

The American Journal of Surgical Pathology 26(1): 47–55, 2002 © 2002 Lippincott Williams & Wilkins, Inc., Philadelphia

47

tinction between Spitz nevus and melanoma and disagreeon the final diagnosis.

A recent study by the North America MelanomaPathology Study Group revealed considerable lack ofconsensus among pathologists in the diagnosis of so-called “atypical Spitz tumors.”5 Some of the tumors,judged as benign by many experts, proved to be malig-nant by clinical outcome. Likewise, when patients orclinicians obtain multiple professional opinions onDCSMT from various pathologists, there is often a lackof consensus as well, leading to confusion of patients andtheir physicians. As long as some pathologists, however,express concern about a lesion’s potential malignant be-havior and suggest the possibility of melanoma, mostpatients and their physicians decide to have a diagnosti-cally difficult lesion completely excised.

For many solid tumor types, including melanoma,lymph node status has become the most important prog-nostic indicator for patients despite the fact that there isprincipal limitation on its prognostic accuracy. Tumorscapable of early hematogenous spread may bypasslymph nodes en route to distant sites, thereby escapingdetection by lymph node analysis. Until a few years ago,pathologic staging of regional lymph nodes was not per-formed routinely in an attempt to avoid lymph node dis-section with its risk for complications, such as lymph-edema and paresthesias.26,29 With the advent of lym-phatic mapping and sentinel lymph node (SLN) biopsies,early metastases through the lymphatic route can now bedetected with minimal morbidity.12,16,26,29

Given the low morbidity of SLN biopsy, it has beenproposed that SLN biopsies may aid in the evaluation ofpatients with diagnostically difficult melanocytic tu-mors.20 We had previously reported a positive SLN in achild with a melanoma,9 which had been interpreted byseveral dermatopathologists as Spitz nevus or atypicalSpitz nevus. In this series we present our experience withnine additional patients with DCSMT or melanomas ofuncertain malignant potential.

MATERIALS AND METHODS

Cases were retrieved from the files of the Departmentof Pathology at Memorial Sloan-Kettering Cancer Center(MSKCC). Only those with outside reports by a board-certified dermatopathologist designating a given pig-mented lesion of the skin as “atypical Spitz nevus,”“Spitz nevus with severe atypia,” “atypical Spitz tumor,”or “Spitzoid melanocytic tumor of uncertain malignantpotential” were included. Upon review at MSKCC, eightcases were judged by at least one pathologist to be vari-ants of malignant melanomas. Two cases were thought tobe uncertain in their malignant potential, and melanomacould not be excluded. Only patients whose tumors mea-

sured at least 1 mm in thickness and/or extended into thereticular dermis underwent SLN biopsies.

All patients were treated with wide local excision.Seven patients had their SLN biopsy done at MSKCC inthe manner previously described.9,12 Three patients un-derwent SLN biopsies at outside hospitals. Their pathol-ogy material (hematoxylin and eosin-stained slides aswell as immunohistochemical studies and tissue blocksin selected cases) was reviewed. Follow-up was obtainedfrom the medical records at MSKCC, outside medicalrecords, or by contacting the physicians of respectivepatients.

Immunohistochemical studies were performed on theSLNs of all seven patients, who were biopsied atMSKCC. The antibodies used in this study includedS-100 protein (1:10,000; Dako, Carpinteria, CA, USA),HMB-45 (anti-gp100; 1:200; Dako), or A103 (anti-Melan-A; 1:50; Dako). Detection of the primary anti-body was performed with a biotinylated horse anti-mouse secondary reagent (Vector, Burlingame, CA,USA) followed by an avidin-biotin complex system(Vector) using diaminobenzidine tetrahydrochloride(DAB, Biogenix, San Ramon, CA, USA) as a chromogenas previously described.11

The statistical analysis was done using SAS(Statistical Analysis System) with t test for continuousvariables (Breslow thickness and number of mitoses) and�2 analysis for categorical variables (presence or absenceof a histologic parameter, such as Kamino bodies or sat-ellite nodule).

RESULTS

Clinical Findings

The clinical findings are summarized in Table 1.Seven patients were female, and three were male. Theirages at initial biopsy of the primary tumor ranged from 7to 46 years (mean 21 years). The anatomic sites varied.Four tumors occurred in the head and neck region, twoon the back, and the remainder on the lower or upperextremities. Two of the patients had a family history ofmelanoma. None had a history multiple atypical/dys-plastic nevi. Before their referral to MSKCC, two pa-tients developed satellite nodules after excisions of theprimary tumor. All patients with a positive SLN under-went regional lymph node dissection. One patient wasfound to have three additional lymph nodes involved bymetastatic tumor.

Pathologic Findings

Thickness and mitotic index are listed in Table 1. Allprimary tumors involved the epidermis and papillary andreticular dermis. Their thickness ranged from 0.9 to 12

C. M. LOHMANN ET AL.48

Am J Surg Pathol, Vol. 26, No. 1, 2002

mm. None of the tumors was ulcerated. Four were heav-ily pigmented. Four lacked readily detectable melaninpigment (pauci- or “amelanotic” tumors). All lesionswere composed of a mixed population of spindle andepithelioid melanocytes. A single neoplasm was com-posed predominantly of spindle cells and demonstratedareas of stromal desmoplasia (desmoplastic tumor). Raresmall dull pink globules (Kamino bodies) were seen atthe dermoepidermal junction of all but one tumor. Noneof the tumors demonstrated large well-formed and/or nu-merous Kamino bodies. All consisted of at least a minorpopulation of pleomorphic epithelioid melanocytes withlarge irregular nuclei, which were either hyperchromaticor showed prominent eosinophilic nucleoli. Features ofzonation in depth (“maturation”) were seen in eight le-sions. All but one had mitotic figures in dermal melano-cytes with the mitotic index ranging from 0 to 6mitoses/mm2. In six tumors mitotic figures were found inthe lower third of the dermal component.

Half of the patients had an SLN with melanocytictumor deposits in the lymph node parenchyma. In twocases the metastasis was recognized on routine hema-toxylin and eosin-stained sections. In one of them meta-static tumor occupied approximately 30% of the lymphnode, whereas in the other case only 5% of the node wasinvolved by tumor. In three cases microscopic tumordeposits became apparent only after multistep sectionsand immunohistochemical studies for melanocytic dif-ferentiation markers were examined.

Three examples of melanomas are shown in Figures1–3 to illustrate the spectrum of histopathologic changes.Figure 1 shows a desmoplastic melanocytic tumor. Ir-regular nests of epithelioid melanocytes were seen at thedermoepidermal junction (Fig. 1A). The dermal compo-nent was amelanotic and composed of a dense populationof epithelioid cells infiltrating the dermis in a single cellpattern (Fig. 1A). In the mid and deep dermis spindlecells predominated and were associated with fibrosis(Fig. 1B). Recurrent tumor involved the deep dermis and

subcutis in a sarcomatoid pattern (Fig. 1C). In the sub-sequent excision a satellite nodule was found (notshown). The SLN contained desmoplastic tumor recog-nizable on a hematoxylin and eosin-stained section(Fig. 1D), which was strongly immunopositive for S-100protein.

Figure 2 illustrates an epithelioid melanocytic tumor.The lesion was relatively small and fairly symmetric andhad sharp lateral borders (Fig. 2A). The V-shaped dermalcomponent was infiltrative at its base. At the periphery ofthe lesion, lymphocytes were seen around dermal ves-sels. The epidermis was hyperplastic. Melanocytes werediscohesive in junctional nests of varying size and shape(Fig. 2B). Focally, melanocytes in solitary units wereseen at all layers of the epidermis (pagetoid spread, Fig.2D). The underlying dermal component showed poormaturation. Dull pink globules were seen in the junc-tional zone (Fig. 2B, C). Mitoses were present in dermalmelanocytes (Fig. 2D). Clusters of atypical epithelioidmelanocytes were found in the SLN parenchyma. Thesewere best seen in a section immunostained for S-100protein (Fig. 2E).

An amelanotic tumor with a prominent spindle cellnodule is featured in Figure 3. Stromal desmoplasia wasnot observed. Clusters of small epithelioid and multinu-cleated melanocytes were adjacent to the spindle cellnodule (Fig. 3B). Mitoses were readily identified andpresent in the deep dermal component of the tumor (Fig.3C). Immunohistochemical studies for S-100 protein re-vealed small deposits of tumor cells in the parenchymaof the SLN (Fig. 3D).

A statistical analysis was performed to examine theassociation of variable histologic parameters with thestatus of the SLN. The parameters included tumor thick-ness, mitotic index, and the presence or absence of sat-ellite nodules and Kamino bodies. None of these param-eters was associated with a strong p value (tumor thick-ness: p � 0.46; number of mitoses: p � 0.74; Kaminobody: p � 0.32; and satellite nodule(s): p � 0.13).

TABLE 1.

Patientno.

Age(yr) Sex

Tumorsite

Tumorthickness

Mitosesper mm2

SLNstatus

Satellitenodules

OtherLNs Follow-up

1 14 F Back 7 mm 3 − No NED after 36 mo2 14 F Back 3.7 mm 2 − No NED after 34 mo3 25 F Right jaw 2.5 mm 6 − No NED after 13 mo4 46 F Right knee 0.35 mm 2 − No NED after 51 mo5 22 F Left thigh 1.5 mm 0 − No NED after 13 mo6 7 M Right ear 6.5 mm 2 + Yes (1) 3NL+ NED after 34 mo7 13 M Right ear 12 mm 2 + Yes (2) − NEd after 54 mo8 28 F Neck 2.3 mm 4 + No − NED after 10 mo9 20 M Right arm 2.4 mm 5 + No − NED after 53 mo

10 26 F Buttock 0.8 mm 2 + No − NED after 39 mo

+, positive for melanocytic tumor deposits in lymph node parenchyma.−, negative for melanocytic tumor deposits.NED, no evidence of disease.

DIAGNOSTICALLY CONTROVERSIAL SPITZOID MELANOCYTIC TUMORS 49

Am J Surg Pathol, Vol. 26, No. 1, 2002

FIG. 1. Desmoplastic Spitz tumor located on the ear of a 7-year-old boy. (A) Junctional nests of atypical epithelioidmelanocytes are seen. Melanocytes infiltrate the dermis in a dense and diffuse single cell pattern. (B) Spindle cells withdesmoplasia are present. (C) Local recurrence after excision of the primary tumor shows a dermal scar and infiltration ofthe dermis and subcutis by a dense population of melanocytes. (D) SLN with metastatic melanocytic tumor deposit (*).

FIG. 2. Epithelioid Spitz tumor located on the buttock of a 13-year-old girl. (A) Silhouette of the lesion. Sharp lateraldemarcation, associated epidermal hyperplasia, and a V-shaped dermal component are seen. (B) Epithelioid melanocytesare discohesive in junctional nests. Maturation is poor. The base of the lesion shows an infiltrative pattern. Dull pinkglobules are seen (arrows). (C) Several pink globules are seen at the dermoepidermal junction (arrows). (D) Melanocytesare present as solitary units in the spinous and granular cell layer of the epidermis (*). Mitoses (arrow) are identified indermal melanocytes. (E) Metastatic tumor deposits are seen in the parenchyma of the SLN (immunopositive for S-100protein).

DISCUSSION

Most melanocytic nevi, including those composed oflarge spindle and/or epithelioid cells (Spitz nevi), can bereliably distinguished from malignant melanoma by his-topathologic analysis. However, atypical melanocytic tu-mors, which show features intermediate between ob-vious nevus and clear-cut melanoma, exist. At times itcan be difficult or even impossible to determine withcertainty, on morphologic grounds alone, whether suchtumors are benign, albeit unusual nevi, nevi undergoingchanges toward melanoma, or outright malignant mela-noma in disguise (malignant melanoma mimicking anevus).5,27,28,32,43,47,51–53

Melanomas closely mimicking nevi have also beentermed “nevoid melanoma.”27,28,39,51–53 These tumorsare said to represent variants of nodular melanoma. Theytypically have a deceptively nevoid silhouette, lackprominent intraepidermal pagetoid spread, but show cy-tologic atypia, lack of maturation, and dermal mitoses.

Pathologists relying on architectural features at scanningmagnification alone risk misinterpreting these melano-mas as nevi. Although the term “nevoid melanoma” hasdidactic value in reminding pathologists that some mela-nomas are more difficult to distinguish from nevi thanothers, there is continued controversy over whether toaccept nevoid melanoma as a biologically distinct sub-type of melanoma or not. Some have suggested a betterthan expected prognosis for nevoid melanomas.28 Othershave argued that nevoid melanoma and “conventional”melanomas have a similar biology.39 Variants of nevoidmelanoma have been described simulating Spitz nevus,similar to the tumors described in this series.51,52

The difficulty in distinguishing Spitz nevi from Spitznevus-like malignant melanoma is mirrored by the factthat the pathology literature on the subject is confusing.Most authors agree on the importance of symmetry andsharp lateral demarcation for the diagnosis of a nevus,including a Spitz nevus.3,10,13,22,44,48 However, there isconflicting information in several articles and text-

FIG. 3. Nondesmoplastic spindle cell tumor on the neck of a 28-year-old woman. (A) Silhouette of the primary tumor. (B)A spindle cell nodule (*) is juxtaposed to small clusters of spindle and epithelioid melanocytes with multinucleated cells(arrow). (C) Mitoses are seen (arrow) near the base of the lesion. (D) Metastatic tumor deposits are present in theparenchyma of the SLN (immunopositive for S-100 protein).

DIAGNOSTICALLY CONTROVERSIAL SPITZOID MELANOCYTIC TUMORS 51

Am J Surg Pathol, Vol. 26, No. 1, 2002

books on various other salient histologic criteria saidto be useful for discriminating a Spitz nevus frommelanoma.13,21–24,33,43,48

Whereas Sophie Spitz suggested that giant cells wereimportant for diagnosis,45 others attribute to them atmost a minor diagnostic role.48 Many authors believethat evidence of maturation of dermal melanocytes ishelpful for the distinction of Spitz’s nevus from mela-noma.4,10,13,22 Kernen and Ackerman, however, thoughtthat maturation is uncommon in Spitz nevi and thereforeof little diagnostic value.21 Whereas Helwig stated thatSpitz nevi characteristically displayed an infiltrativegrowth pattern at their base,17 Kernen and Ackermansuggested the opposite,21 i.e., that Spitz nevi tended topush rather than infiltrate the stroma. There is alsocontroversy with regard to the significance of cyto-logic atypia and mitotic figures in dermal melano-cytes.1–3,22,33,44,46 Some authors have suggested thatthe presence of pleomorphism and mitotic figures is aclue to the recognition of malignant melanoma.14,35,44

Others have countered that too much attention to cytologicfeatures and mitoses may lead to misdiagnosing nevi asmelanoma.2

There is also uncertainty about the diagnostic signifi-cance of Kamino bodies. Kamino et al. reported in 1979that dull pink globules were commonly found in Spitznevi but rarely seen in association with melanomas.18

Although it was initially thought that the globules re-sulted from apoptosis of keratinocytes and/or melano-cytes,18 they have since been shown to represent aggre-gates of basement membrane material40 and not apop-totic cells.49 Kamino bodies are in our experience a veryhelpful clue for the diagnosis of Spitz nevi. When largeand numerous, their presence strongly favors a nevus.However, rare small Kamino bodies are not specific forSpitz nevi and can be seen in other nevi as well as inmalignant melanomas, which limits their diagnosticvalue.3,4,18,48 Kamino et al. noted eosinophilic globulesin 2% of melanomas.18 Others have found a higher in-cidence and reported that Kamino bodies can be numer-ous in some melanomas.4 In the series reported herein,Kamino bodies were seen, albeit only rarely so, in all butone case. Only hematoxylin and eosin-stained sectionswere examined. Additional sections were not available toperform trichrome or periodic acid–Schiff stains withdiastase to confirm the authenticity of Kamino bodiesbeyond dispute. As assessed in this study, Kamino bod-ies had no predictive value for the status of the SLN.

LeBoit had recently pointed out that the dermatopa-thology literature lacks a well-documented case of amelanoma, proven to be malignant by follow-up, withnumerous well-formed Kamino bodies in the primarytumor.22,23 Because Kamino bodies were small and rarein our cases, we could not evaluate the value of numer-

ous dull pink globules in predicting SLN status. Al-though we agree with LeBoit that well-formed Kaminobodies are exceedingly rare in melanoma, we have seenrare metastasizing melanomas with bona fide Kaminobodies (K.J.B., personal observations).

A major drawback of several studies on Spitz nevi isthe lack of long-term clinical follow-up. In the series byPaniago-Pereira et al., for example, clinical follow-upwas “three years at most” and no information was pro-vided on mean or median follow-up.33 Thus, proposedcriteria for the distinction of Spitz nevus from melanomafrom such studies cannot be considered as fully validatedand need to be applied with caution. Barnhill et al. re-cently pointed out that many reports on Spitz nevi arevirtual tautologies, in that many pathologists re-examined their own case collections retrospectively forthe purpose of publication.5

Given the limitations of morphologic analysis, severalattempts have been made to explore the use of ancillarytechniques for the diagnosis of Spitz tumors.8,19,24,25,34,50

A number of investigators have suggested a role forMIB-1 immunoreactivity as an adjunct to the histopath-ologic distinction of Spitz nevus from melanoma.8,19,24

Bastian et al. have explored the use of comparative ge-nomic hybridization in the analysis of melanocytic tu-mors.6,7 They have found distinct genetic alterations forsome Spitz tumors, such as an 11p copy number increasein desmoplastic Spitz nevi.6 This is in contrast to malig-nant melanomas, which typically have more complexchromosomal alterations. Although this approach showspromise in providing useful information on the biologyof melanocytic tumors, it is currently premature to applythese findings for diagnosis.

Currently, diagnostic judgments still need to rely onhistomorphologic analysis, and a number of criteria use-ful for the distinction between Spitz nevus and mela-noma have been reported. These include architecturaland cytologic parameters. In our experience of patientswith metastatic melanomas, whose primary tumor hadbeen misdiagnosed as Spitz nevus, failure to recognizemelanoma typically occurred by reliance on architecturalcriteria alone (K.J.B., personal observations). The tumorsappeared nevoid by silhouette (symmetric, sharp lateraldemarcation, minimal or no pagetoid spread, predomi-nant nested arrangement of cells) but showed atypicalcytologic features (presence of mitoses and pleomor-phism, lack of maturation and Kamino bodies). Recentstudies with >3 years’ follow-up of several patients havesuggested that a number of histologic parameters canhelp gauge the malignant potential of a Spitz tumor.44,47

Large size (>1 cm), tumor extension into the subcutis,the presence of ulceration, and a high mitotic index seemto correlate with malignant behavior, at least in patientsafter the onset of puberty. Because some Spitzoid mela-nocytic neoplasms are too difficult to classify reliably as

C. M. LOHMANN ET AL.52

Am J Surg Pathol, Vol. 26, No. 1, 2002

atypical nevus or melanoma, Spatz and Barnhill haveproposed replacing the dichotomous distinction of be-nign versus malignant with a grading system that allowsstratification of atypical tumors into low, intermediate,and high risk groups for recurrence.43 When applied toour series, all the tumors fell into the low risk categoryand the score did not correlate with SLN status.

When pathologists cannot agree on a diagnosis or ad-mit that a tumor is of uncertain malignant potential, cli-nicians and patients face a dilemma with regard to guide-lines for management and prognosis. As long as the pos-sibility of melanoma cannot be excluded, it is areasonable and common practice for surgeons to excisean atypical Spitz tumor as if it were a melanoma. Adifferent issue is the role of a SLN biopsy.

SLN biopsy has become the standard of care for theworkup of patients with primary cutaneous melanomawith a tumor thickness �1 mm.26 Recent evidence indi-cates that as with other solid tumors, the status of theSLN is a powerful predictor of clinical outcome for pa-tients with stage I and stage II melanoma.12,16,26,29

Given the low morbidity of the procedure, it has beensuggested that SLN biopsy may also serve as an adjunctprocedure in the evaluation of diagnostically difficultmelanocytic tumors by increasing the sensitivity of thediagnosis and providing potentially useful prognostic in-formation.20 Although a histologically negative SLN isnot a guarantee against future recurrences and spread ofa tumor, a positive lymph node (i.e., metastasis, notnodal nevus) can be assumed as evidence for the malig-nant potential of a tumor. We agree with Ackerman et al.that “the notion of metastasizing Spitz’s nevus is illogi-cal and without foundation.”2

In this report we document melanocytic tumor depos-its in the SLNs of half of our patients, whose primarytumor was reported at outside institutions as an atypicalSpitz nevus. The pattern of lymph node involvement wastypical of metastases, with multiple clusters of melano-cytes in the lymph node parenchyma. Furthermore, thecytology of the melanocytes in the SLNs was atypical(enlarged nuclei, presence of nucleoli) and resembled theappearance of the cells in the primary tumor. No involve-ment of the lymph node capsule by melanocytes wasseen in any of the cases. Thus, we believe that the evi-dence argues against nevic deposits. Although we con-sidered the possibility of a nodal nevus component as-sociated with Spitz nevus, there is insufficient knowl-edge about this phenomenon and the patterns of lymphnode involvement one might observe.23

The high incidence of metastases in our series is mostlikely a reflection of referral bias and the fact that manytumors were thick (mean thickness 3.1 mm). In a recentseries by Nguyen et al., <20% of SLNs were positive inpatients with melanomas thinner than 3 mm, whereas>50% were positive if the tumor thickness exceeded 3

mm.31 Other series have also found an association be-tween the incidence of metastasis and tumor thickness,however, with an overall lower frequency of positiveSLNs.12,16,29 When we performed a statistical analysis ofseveral histologic variables of the primary tumors for acorrelation with SLN status, no parameter had a strong pvalue, a result that is most likely due to the small samplesize of our series.

Although melanoma had already been considered orfavored by some pathologists at the time of review atMSKCC or elsewhere, the finding of a positive SLNsupported the diagnosis of malignant melanoma. Newquestions, however, have emerged. What is the prognos-tic significance of a positive SLN in a patient with aSpitzoid melanoma and does it differ from the prognosisof a positive SLN in a patient with conventional mela-noma? What are the diagnostic and prognostic implica-tions of node-negative DCSMT? These are questions thatrequire multi-institutional studies with long-term follow-up in parallel with conventional melanomas. They cannotbe answered at the current time. Many issues related toSLN biopsy findings are still unsettled even for conven-tional adult-type melanomas, including the prognosticsignificance of small metastatic tumor deposits, whichare recognized best or only by immunohistochemicalstudies. Preliminary data from patients with conventionalmelanoma indicate that patients with only small micro-scopic metastases in their SLN have a low risk for re-currence (D.G.C., M.S.B., K.J.B., unpublished observa-tions). Such small metastatic deposits were seen in threeof our five cases with positive SLN. The presence of onlymicrometastatic disease in three patients likely played arole in the fact that the overall clinical outcome of ourpatients with positive SLNs has so far been good (at lastfollow-up all were alive and free of disease).

Some have proposed that melanomas sharing featuresof Spitz nevi or closely mimicking nevi differ in theirbiologic potential from conventional epithelioid melano-mas. Although it is possible that melanocytic neoplasmsexist with genetic alterations different from both typicalmelanomas and benign nevi, the concept of a low-grademelanocytic tumor category remains at the current timean unproven albeit valid hypothesis. The notion thatmelanomas with features of Spitz nevus could carry abetter prognosis than melanomas without such featureseven after regional lymph node metastasis is more con-troversial.15 Contrary to the suggestion of some patholo-gists that metastases from spindle cell and epithelioidcell nevi with atypia may arrest at the regional lymphnode,41,42 we and others have seen several cases of me-tastases from Spitz nevus-like melanomas with lethaloutcome. The melanomas with positive SLN presented inthis series differ from the metastasizing melanocytic tu-mors previously reported as “malignant Spitz ne-vus.”41,42 The primary tumors in this series were smaller

DIAGNOSTICALLY CONTROVERSIAL SPITZOID MELANOCYTIC TUMORS 53

Am J Surg Pathol, Vol. 26, No. 1, 2002

(<1 cm in diameter), did not extend into the subcutis, andlacked ulceration. Furthermore, the tumor deposits in theSLN also tended to be smaller than the metastases fromthe tumors reported by Smith et al.42 and Skelton et al.41

Because the term “malignant Spitz nevus” is contradic-tory in itself, we advise against its use. Melanocytic tu-mors found to have metastatic tumor deposits in lymphnode(s) should be designated as malignant melanoma.

In conclusion, we report herein the first series of pa-tients with DCSMT who underwent SLN biopsy. Giventhe low procedural morbidity, it seems reasonable to con-sider this technique in the care of patients with suchtumors. An SLN biopsy, however, should not be used asa substitute for professional consultation. In our view, itis in the patient’s best interest for any pathologist uncer-tain about an unusual melanocytic tumor representingnevus or melanoma to first consult other experiencedcolleagues. If several pathologists suspect melanoma, anSLN biopsy seems justifiable. If an SLN is positive, thisfinding helps to support the diagnosis of malignant mela-noma. Conclusions about the prognostic significance of apositive SLN in a patient with a Spitz nevus-like mela-noma, however, are currently premature. Further studiesare needed to assess whether the features of the primarytumor, which deviated from both nevus and conventionalmelanoma, justify a separate prognostic category forthese patients. �

Acknowledgments

The authors thank Jennifer Nobrega and Kin Kong for theirassistance with the prints.

REFERENCES

1. Ackerman AB. Spitz nevus. Am J Dermatopathol 1997;19:419–21.2. Ackerman AB, Mendonca AMN, Guo Y. Spitz’s nevus, compound

type vs malignant melanoma. In: Differential Diagnosis in Der-matopathology I. Philadelphia: Lea & Febiger, 1992:146–9.

3. Ackerman AB, Miyauchi Y, Takeuchi A, et al. Melanomas thatsimulate Spitz’s nevi histopathologically (and vice versa): an ex-ercise in differentiation based on dependable criteria. Dermatopa-thology 1999;5:9–13.

4. Ackerman AB, Takeuchi A, Ohata C. Kamino bodies may benumerous in melanomas. Dermatopathology 1998;4:127–30.

5. Barnhill RL, Argenyi ZB, From L, et al. Atypical Spitznevi/tumors: lack of consensus for diagnosis, discrimination frommelanoma, and prediction of outcome. Hum Pathol 1999;30:513–20.

6. Bastian BC, LeBoit PE, Pinkel D. Mutations and copy numberincrease of HRAS in Spitz nevi with distinctive histopathologicalfeatures. Am J Pathol 2000;157:967–72.

7. Bastian BC, Wesselmann U, Pinkel D, et al. Molecular cytogeneticanalysis of Spitz nevi shows clear differences to melanoma. JInvest Dermatol 1999;113:1065–9.

8. Bergman R, Malkin L, Sabo E, et al. MIB-1 monoclonal antibodyto determine proliferative activity of Ki-67 antigen as an adjunct tothe histopathologic diagnosis of Spitz nevi. J Am Acad Dermatol2001;44:500–4.

9. Brady MS, Weinberg H, Kraus D, et al. Lymphatic mapping in the

management of melanoma in children. Pediatr Dermatol 1998;15:421–5.

10. Busam KJ, Barnhill RL. Spitz nevi and related lesions. In: BarnhillRL, ed. Pathology of Melanocytic Nevi and Malignant Melanoma.Newton, MA: Butterworth-Heineman, 1995:97–130.

11. Busam KJ, Jungbluth AA. Melan-A, a new melanocytic differen-tiation marker. Adv Anat Pathol 1999;6:12–8.

12. Clary BM, Brady MS, Lewis JJ, et al. Sentinel lymph node biopsyin the management of patients with primary cutaneous melanoma:review of a large single institution experience with emphasis onrecurrence. Ann Surg 2001;233:250–8.

13. Cochran AJ, Bailly C, Paul E, et al. Spitz and related nevi. In:Cochran AJ, Bailly C, Paul E, et al., eds. Melanocytic Tumors: AGuide to Diagnosis. Philadelphia: Lippincott-Raven, 1997:125–55.

14. Crotty KA, McCarthy SW, Palmer AA, et al. Malignant melanomaof childhood: a clinicopathologic study of 13 cases and comparisonwith Spitz nevi. World J Surg 1992;16:179–85.

15. Echevarria R, Ackerman LV. Spindle and epithelioid cell nevi inthe adult: clinicopathologic report of 26 cases. Cancer 1967;20:175–89.

16. Gershenwald JE, Thompson W, Mansfield PF, et al. Multi-institutional melanoma lymphatic mapping experience: the prog-nostic value of sentinel lymph node status in 612 stage I or IImelanoma patients. J Clin Oncol 1999;17:976–83.

17. Helwig EB. Malignant melanoma in children. From the TwentiethAnnual Clinical Conference on Cancer. Neoplasms of the Skin andMalignant Melanoma. Chicago: Year Book, 1975:11–26.

18. Kamino H, Flotte TJ, Misheloff E, et al. Eosinophilic globules inSpitz’s nevi. Am J Dermatopathol 1979;1:319–24.

19. Kanter-Lewensohn L, Hedblad M-A, Wejde J, et al. Immunohis-tochemical markers distinguishing Spitz nevi from malignantmelanomas. Mod Pathol 1997;10:917–20.

20. Kelley SW, Cockerell CJ. Sentinel lymph node biopsy as an ad-junct to management of histologically difficult to diagnose mela-nocytic lesions: a proposal. J Am Acad Dermatol 2000;42:527–30.

21. Kernen JA, Ackerman LV. Spindle cell nevi and epithelioid cellnevi (so-called juvenile melanomas) in children and adults: a clin-icopathologic study of 27 cases. Cancer 1960;13:612–25.

22. LeBoit PE. Spitz nevus: a look back and a look ahead. Adv Der-matol 2000;16:81–109.

23. LeBoit PE. Kamino bodies: what they may mean. Am J Dermato-pathol 2001;23:374–7.

24. Li LX, Crotty KA, McCarthy SW, et al. A zonal comparison ofMIB1-Ki-67 immunoreactivity in benign and malignant melano-cytic lesions. Am J Dermatopathol 2000;22:489–95.

25. Mackie RM, White SI, Seywright MM, et al. An assessment of thevalue of AgNOR staining in the identification of dysplastic andother borderline melanocytic nevi. Br J Dermatol 1989;120:511–6.

26. McMasters KM, Reintgen DS, Ross MI, et al. Sentinel lymph nodebiopsy for melanoma: controversy despite widespread agreement.J Clin Oncol 2001;19:2851–5.

27. McNutt NS. ‘Triggered trap’: nevoid melanoma. Semin DiagnPathol 1998;15:203–9.

28. McNutt NS, Urmacher C, Hakimian J, et al. Nevoid malignantmelanoma: morphologic patterns and immunohistochemical reac-tivity. J Cutan Pathol 1995;22:502–17.

29. Morton DL, Thompson JF, Essner R, et al. Validation of the ac-curacy of intraoperative lymphatic mapping and sentinel lymph-adenectomy for early-stage melanoma; a multicenter trial. Multi-center Selective Lymphadenectomy Trial Group. Ann Surg 1999;230:453–63.

30. Muhlbauer J, Randall J, Mihm MJ, et al. Minimal deviation mela-noma: a histologic variant of cutaneous malignant melanoma in itsvertical growth phase. J Invest Dermatol 1983;80(suppl):63–5.

31. Nguyen CL, McClay EF, Cole DJ, et al. Melanoma thickness andhistology predict sentinel lymph node status. Am J Surg 2001;181:8–11.

32. Okun MR. Melanoma resembling spindle and epithelioid cell ne-vus. Arch Dermatol 1979;115:1416–20.

33. Paniago-Pereira C, Maize JC, Ackerman AB. Nevus of largespindle and/or epithelioid cells (Spitz’s nevus). Arch Dermatol1978;114:1811–23.

34. Penneys N, Seigfried A, Nahaas G, et al. Expression of prolifer-

C. M. LOHMANN ET AL.54

Am J Surg Pathol, Vol. 26, No. 1, 2002

ating cell nuclear antigen in Spitz nevus. J Am Acad Dermatol1995;32:964–7.

35. Peters MS, Goellner JR. Spitz naevi and malignant melanomas ofchildhood and adolescence. Histopathology 1986;10:1289–302.

36. Reed RJ. Dimensionalities: borderline and intermediate melano-cytic neoplasia. Hum Pathol 1999;30:521–4.

37. Reed RJ. Atypical Spitz nevus/tumor. Hum Pathol 1999;30:1523–6.

38. Reed R, Ichinose H, Clark W, et al. Common and uncommonmelanocytic nevi and borderline melanomas. Semin Oncol 1975;2:119–47.

39. Schmoeckel C, Castro CM, Braun-Falco O. Nevoid malignantmelanoma. Arch Dermatol Res 1985;277:362–9.

40. Schmoeckel C, Stolz W, Burgeson R, et al. Identification of base-ment membrane components in eosinophilic globules in a case ofSpitz’s nevus. Am J Dermatopathol 1990;12:272–4.

41. Skelton HG III, Smith KJ, Holland TT, et al. Malignant Spitznevus. Int J Dermatol 1992;31:217–22.

42. Smith KJ, Barrett TL, Skelton HG, et al. Spindle and epithelioidcell nevi with atypia and metastasis (malignant Spitz nevus). Am JSurg Pathol 1989;13:931–9.

43. Spatz A, Barnhill RL. The Spitz tumor 50 years later: revisiting alandmark contribution and unresolved controversy. J Am AcadDermatol 1999;40:223–8.

44. Spatz A, Calonje E, Handfield-Jones S, et al. Spitz tumors inchildren: a grading system for risk stratification. Arch Dermatol1999;135:282–5.

45. Spitz S. Melanomas of childhood. Am J Pathol 1948;24:591–609.46. Stas M, van den Oord JJ, Garmyn M, et al. Minimal deviation

and/or naevoid melanoma: is recognition worthwhile? A clinico-pathological study of nine cases. Melanoma Res 2000;10:371–80.

47. Walsh N, Crotty K, Palmer A, et al. Spitz nevus versus spitzoidmalignant melanoma: an evaluation of the current distinguishinghistopathologic criteria. Hum Pathol 1998;29:1105–12.

48. Weedon D, Little JH. Spindle and epithelioid cell nevi in childrenand adults: a review of 211 cases of Spitz nevus. Cancer 1977;40:217–25.

49. Wesselmann U, Becker LR, Broecker EB, et al. Eosinophilic glob-ules in Spitz nevi: no evidence for apoptosis. Am J Dermatopathol1998;20:551–4.

50. Winokur TS, Palazzo JP, Johnson WC, et al. Evaluation of DNAploidy in dysplastic and Spitz nevi by flow cytometry. J CutanPathol 1990;17:342–7.

51. Wong T-Y, Duncan LM, Mihm MC Jr. Melanoma mimickingdermal and Spitz’s nevus (‘nevoid’ melanoma). Semin Surg Oncol1993;9:188–93.

52. Wong T-Y, Suster S, Duncan LM, et al. Nevoid melanoma: aclinicopathologic study of seven cases of malignant melanomamimicking spindle and epithelioid cell nevus and verrucous dermalnevus. Hum Pathol 1995;26:171–9.

53. Zembowicz A, McCusker M, Ciarelli C, et al. Morphologic analy-sis of nevoid melanoma: a study of 20 cases with review of theliterature. Am J Dermatopathol 2001;23:167–75.

DIAGNOSTICALLY CONTROVERSIAL SPITZOID MELANOCYTIC TUMORS 55

Am J Surg Pathol, Vol. 26, No. 1, 2002