IOLII~AL o f t h e
AIlleRICaN A C a D e m Y OF
AAD' %'2J
DerMaTOLOGY VOLUME 27
NUMBER 6
PART 1
DECEMBER 1992
IIII
s
medical e d u c a t i o n Spitz nevi Elen M. Casso, MD, Caron M. Grin-Jorgensen, MD, and Jane M. Grant-Kels, M D
Farmington, Connecticut The Spitz nevus has long been defined as a benign melanocytic lesion that shares many histologic features with malignant melanoma. Despite the diagnostic criteria established for these two entities, their histologic similarities continue to make their distinction somewhat difficult. Uncertainties also exist with regard to the natural history of the Spitz nevus; the true pattern of this lesion's biologic behavior remains elusive. As a result, controversies exist with respect to appropriate therapy. To examine these controversies, the epidemiology, clinical features, and histopathology of Spitz nevi, as well as the role of recent molecular and immunohistochemical diagnostic studies, are discussed. However, the primary focus of this article is the natural course, prognosis, and treatment of the Spitz nevus. A review of 716 cases of Spitz nevi, compiled from 13 papers published from 1948 to 1990, is presented. After analyzing this and other available data, we propose that at this time Spitz nevi and malignant melanoma cannot easily be categorized as distinct entities and that perhaps they actually exist along one continuum of disease. Because of this uncertainty and the difficulties in differentiating these two lesions, we recommend that treatment include complete excision of all Spitz nevi followed by reexcision of positive margins if present. (J AM ACAD DERMATOL 1992;27:901-13.) The Spitz nevus was first described in 1948 by Spitz,1 who referred to it as a "juvenile melanoma." She established the initial histologic criteria for differentiating this lesion from malignant melanoma. Since then, the term benign juvenile melanoma has been criticized as a misnomer for two reasons: the lesion is encountered not only in children but in adults and the term melanoma carries a connotation of malignancy, whereas Spitz ned are considered benign. More recently, the designation spindle and~or epithelioid cell nevus has been used to reflect its histologic features. The term Spitz nevus has
The CME articles are made possible through an educational grant from tile Dermatological Division, Ortho Pharmaceutical Corporation. From the Division of Dermatology, The University of Connecticut, Farmington, CT 06030. Reprint requests: Caron M. Grin-Jorgensen, MD, Division of Dermatology, The University of Connecticut, Farmington, CT 06030. ORTHO
16/2/42525
been adopted and retained in acknowledgement of the lesion's historical identification and description by Spitz. Many authors have contributed to the understanding of Spitz nevi by defining several criteria that helps the pathologist to distinguish between Spitz nevi and malignant melanomas.l2i However, few studies have focused on the natural history and prognosis after the diagnosis of a Spitz nevus. The reason for this is twofold. First, the majority of Spitz nevi have been surgically excised. As a result, the natural course of these lesions has been disrupted. Second, few data are available on long-term follow-up of patients either with intact Spitz nevi or after surgical removal. The purpose of this article is to discuss the epidemiology and clinical and histopathologic features of Spitz nevi, while focusing primarily on the course and prognosis to determine appropriate treatment options. In addition, various controversies about Spitz nevi will be addressed. 901
902
Casso et al.
EPIDEMIOLOGY
Spitz nevi occur most often during the first two decades of life. However, they can appear at any age from birth 1, 22 to the eighth decade. 6 In a study of 202 cases of Spitz nevi, 79% of the patients were younger than 20 years of age. 6 In another report of 33 cases, 60% occurred in children younger than 14 years of age. 15 The incidence in adults has been shown to decrease with increasing age. A study of 200 cases of Spitz nevi found that 28.5% of the patients were older than 30 years of age, and only 8.5% were older than 45.14 All studies have shown that Spitz nevi occur predominantly in the white population. Only seven cases of Spitz nevi in blacks have been reported.14' 23, 24 Bovenmyer23 concluded that this simply reflects a lower incidence of this nevus in blacks. However, Carr et al. 24 hypothesized that the small number of cases m a y be related to underdiagnosis of these nevi in the black population. A slight female preponderance of Spitz nevi has been documented in m a n y series. 6' 7,10, 12,14,21 This is especially true in women between the ages of 15 and 29 years. 14 It has been theorized that this difference most likely results from increased cosmetic concern that leads to a higher rate of removal and subsequent reporting. 6, 10, 14, 21 The actual incidence of Spitz nevi in the population is not known. Various authors have estimated that between 1% and 10% of all melanocytic nevi in children are, in fact, Spitz nevi.11, 25 Paniago-Pereira e t a l . 14 proposed that these are overestimates and that Spitz nevi constitute less than 1% of all nevi in children. CLINICAL FEATURES
Clinically, Spitz nevi most commonly appear as asymptomatic, dome-shaped, round to oval, firm nodules. They have a smooth surface and are usually pink, red, or tan. 6' 10,12, 14, 15, 22 Variations of these clinical characteristics are often encountered. For instance, the lesion may also be macular, polypoid, or verrucous. 6, 12,14,22 In a study of 211 cases of Spitz nevi, 57% were described as dome-shaped, 19% were classified as fiat, and 24% were called polypoid. 21 Spitz nevi also appear in a wide spectrum of colors, ranging from translucent to purple 22 to brown or black. 7, 10, 14 One clinical feature that is notably absent is ulceration. 7' 12, 14,22 For simplicity, the clinical appearance can be divided into four basic types 14, 24, 2s, 26. 1. Light-colored and soft. These are usually smooth, pink, or slightly pigmented. They may also
Journal of the American Academy of Dermatology
resemble a pyogenic granuloma or h e m a n gioma. 2. Light-colored and hard, resembling a dermatofibroma or small keloid. 3. Dark-colored. This less common type varies in degree of pigmentation; more darkly pigmented lesions occur less frequently. Several studies have suggested that the degree of pigmentation varies with the age of the patient. This leads to more darkly pigmented lesions in older persons. For example, Peters and Goellner 15 noted that in 33 patients, all of whom were younger than 20 years of age, 4 h a d Spitz nevi that were black and 3 of those 4 patients were older than 16 years of age. Another study of 27 patients cited that 4 of 6 adults (67%) had brown or black nevi, whereas only 6 of the 21 children (29%) had a dark lesion, l~ In contrast to those results, the study of Coskey and Mehregan 6 of 202 cases found no significant clinical differences in the pigmentation of Spitz nevi in 53 adults and 149 children. 4. Multiple or agminated lesions that can further be subdivided into the following four types27: (1) disseminated, (2) grouped and arising on normal skin, (3) grouped and arising on a hypopigmented macule, or (4) grouped and arising on a congenital pigmented macule. Although Spitz nevi can occur anywhere, 14 the head, neck, and especially the face appear to be the most common locations. 6, 7, 10,12,15, 2~ One study of 27 patients revealed that 56% of the Spitz nevi occurred on the face. 10 Several reports have f o u n d t h e legs to be another common site, with less frequent involvement of the trunk and arms. 6, 7, 10,12. 14, 15 In contrast, a study of 211 cases found that most lesions (39%) were located on the trunk; the legs were next in frequency. 21 In women, there may be a slight predilection for these nevi to occur on the legs. 14, 21 Coskey and Mehregan 6 reported that the distribution of Spitz nevi differs in adults and children. In a study of 202 cases, the most frequent location in children was the face, head, and neck (48%), whereas in adults, 45% were located on the legs. A n o t h e r study of 33 cases failed to find a correlation between age and anatomic location. 15 Spitz nevi can range in size from 1 m m to 3 ClTI.6,7, 10,21, 22 However, most are tess than 1 c m in diameter 7, 10; one study of 211 cases determined that 73% of the nevi were smaller than 6 ram. 21 Several studies have shown that before diagnosis, the duration of Spitz nevi may range from 1 m o n t h to 20 years or more. 7' 10,22 However, the majority of patients have had their lesion for only a short time, usually less than 6 months, before seeking medical
Volume 27 Number 6, Part 1 December 1992 advice. In several studies, the changes most often noted by patients have been slow enlargement (in 35% to 46% of cases) 7, 10, i4, 15; changes in color (in 8% to 27% of cases) 7, 15; bleeding (in 8% to 18% of cases)7'l~ and localized pruritus (in 9% of cases). 15 HISTOPATHOLOGIC FEATURES Histologically, Spitz nevi are composed of spindle cells, epithelioid cells, or a mixture of both; they may be classified as junctional, intradermal, or, most frequently, compound, i2, 14 Spitz nevi share several histologic features with malignant melanomas and therefore make the distinction between the two entities difficult at times. Both are characterized by nests of atypical melanocytes within the epidermis and the dermis. In some cases, Spitz nevi may even possess more extensive nuclear atypia than do malignant melanomas. 2, 17 Since Spitz's initial description in 1948, many authors have attempted to establish histologic criteria to distinguish between these two entities. These criteria are summarized in Table I.* One of the more recent additions to these criteria has been the description of eosinophilic globules, at the dermoepidermal junction in Spitz nevi. These globules were first described by Kamino in 1979 who reported the presence of these eosinophilic bodies within the epidermis in 60% of 293 cases of Spitz nevi and in only 2% of 293 cases of malignant melanoma. In addition, the staining of these epidermal globules was positive with both periodic acid-Schiff and trichrome in Spitz nevi and was negative for both in malignant melanoma. 9 These globules were n a m e d "Kamino bodies" and were confirmed as morphologic markers of Spitz nevi by Arbuckle and Weedon 4 who described their presence in 86% of 50 cases of Spitz nevi and in only 12% of 50 cases of malignant melanoma. 4 Peters and Goeliner] 5 in a smaller study, found Kamino bodies in 52% of Spitz nevi and in 26% of malignant melanomas. It was proposed that the frequent occurrence of Kamino bodies in Spitz nevi may be caused by increased proliferation of keratinocytes and melanocytes in these tumors and subsequent death by apoptosis of some of these cells. 4 Awareness of this morphologic marker has aided in the ability to distinguish between Spitz nevi and malignant melanoma. However, despite weU-established criteria, there are still cases in which the diagnosis is unclear because of an overlap in histologic features between *References 2, 5, 10, 13-15, 17, 21.
Spitz nevi 903 Table I. Histologic criteria to distinguish Spitz nevi from malignant melanoma Spitz nevi
Symmetric lesion Diameter usually 6ram Poor circumscription of intraepidermal melanocytic component Wide variation in size and shape of melanocytic nests Nests of melanocytes within epidermis are neither elongated nor perpendicular to skin surface Usual absence of Kamino bodies within epidermis Confluence of nests of melanocytes within epidermis Few clefts, if any, between nests of melanocytes and keratinocytes Nests of dermal melanocytes tend to become confluent and form sheets of cells Usually no maturation of melanocytes with descent into the dermis Mitotic figures near base of lesion are common Inflammatory cell infiltrate, when present, may be bandlike, beneath, intermingling with, or often extending beyond the melanocytic component
Adapted from Ackerman et al. Differential diagnosis in dermatopathology. Philadelphia: Lea & Febiger, 1982.
Spitz nevi and malignant melanomas. 15 There are many papers that describe various atypical forms ot~ Spitz nevi that seem to fall into a "gray area" between the diagnoses of Spitz nevus and malignant
904
Casso el al.
melanoma. The following features have been identified by several authors as important in favoring the diagnosis of an atypical Spitz nevus over a malignant melanoma: ( 1) low mitotic activity, 10,11,1s especially in the deeper dermal component2~ (2) absence of atypical mitoses 1~ l 1, 2l; (3) slight pleomorphisml0, 11, 15,20; (4) rarity of large nucleolil~ (5) absence of hyperchromasia 11, 15,21; ( 6 ) marked spindle cell component 14' 15; (7) minimal or no significant upward epidermal spread of melanocytes2~ 21; (8) absence of intense melanization, TM15 especially deep within the lesion13; (9) absence of marked ulcerationlS; and (10) absence of deep penetration of the lesion into the deeper reticular dermis and/or subcutaneous fat.15 Histologie differences in adults versus children
Another controversial topic is the histologic differences between Spitz nevi in adults and children. Because the incidence &malignant melanoma is low in children and rises sharply after puberty, the chances of making an erroneous diagnosis of Spitz nevus increase proportionately in adults. 13 Therefore any histologic distinction found between Spitz nevi in adults and children might prove to be a valuable diagnostic tool. In general, the overall histologic appearance of Spitz ncvi is similar in adults and children.6, 7, 10, la-~4,21 However, several differences have been observed. For example, a study of 200 Spitz nevi revealed histologic characteristics that were absent in adults and present in young children. 14 These were papiltomatosis, edema of the papillary dermis, telangiectases, a high concentration of epithelioid nevus cells, and acantholytic nevus cells. In another study of 89 cases, it was noted that involvement of the suprabasal epidermis of Spitz nevi occurs much more frequently in children than in adults.t2 The authors stated that this difference in degree of intraepidermal involvement is a result of the gradual regressive changes of the Spitz nevus that normally occur with age. On the basis of this theory of involution and the observation that this histologic feature is rare in adults, these authors suggested strong consideration of the diagnosis of melanoma when melanocytes involve the suprabasal epidermis within a spitzoid nevus in an "adult." In analyzing their data, however, it can be seen that there exists much overlap regarding the patients' ages in the study groups defined as "with" and "without" intraepidermaI involvement. As a result,
Journal of the American Academy of Dermatology it is difficult to determine the age of onset of the nevus in their "adult" population; therefore it is unclear at what age the presence of suprabasal epidermal involvement in a Spitz nevus should be considered worrisome. Additional studies have found that Spitz nevi in adults contain greater amounts of melanin pigment than do those in children.7, a0,14 However, some authors classify these lesions not as Spitz nevi but as pigmented spindle cell nevi (PSCN) 16' 18, 19,28-3o that occur most commonly in women during the third decade of life. 30 It remains a point of controversy as to whether the PSCN is a variant of the Spitz nevus, or a distinct entity. It is generally agreed that the most prominent features that differentiate PSCN from Spitz nevi are the following16, 18,19,28-30:(1) abundant melanin pigment; (2) nests of spindle cells that are confined to the epidermis and papillary dermis, with absence of reticular dermal involvement; (3) virtual absence of epithelioid cells composing the neoplasm; the cells are spindled with delicate nuclei and eventy dispersed chromatin; (4) an expansile growth pattern that compresses and displaces the reticular dermis, with absence of an infiltrative pattern; (5) absence of stromal edema and telangiectasia; and (6) rare upward intraepidermal spread of melanocytes. Despite these differences, there are significantly more similarities between the two entities. In fact, one author admits that junctional Spitz nevi, in particular, can be distinguished from PSCN only by the relative absence of pigment both clinically and histologically.lS This appears to favor the classification of PSCN as a special variant of Spitz nevi. However, some authors maintain that the most important reason to classify PSCN as a distinct entity is to prevent its misdiagnosis as malignant melanoma, especially in young adults. 16,18,19,28 In examination of the possible differences between Spitz nevi in adults and children, it is necessary to include the issue of the desmoplastic Spitz nevus, a histologic variant of the Spitz nevus that one author suggested was found only in adults older than 30 years of age. 14It has been postulated that this lesion, described as a Spitz nevus with excessive fibrosis, results from tumor regression or a slowly progressive process of involution that can be correlated with increasing age of the patient. 3' z2, 14, 16,22 However, in a study of 75 Spitz nevi, 14 of which were desmoplastic, there was no significant correlation between desmoplasia and either patient age or
Volume 27 Number 6, Part I December 1992
duration of the lesion. 31 These findings were supported by Kernen and Ackerman,10 who reported no histologic evidence of transformation with age, and by Coskey and Mehregan, 6 who noted no signs of regression, even in lesions of long duration. Several large studies, including a study of 2l 1 cases 21 and another of 202 cases,6 failed to substantiate the findings of all of the aforementioned studies regarding the histologic variations of the Spitz nevus in children and adults. Both studies concluded that no histologic differences exist between the Spitz nevi of children and adults. Although the histopathology of Spitz nevi has been well studied, it is evident that many controversies still exist, such as the question of whether or not histologic distinctions can be made in children versus adults. In addition, its differentiation from malignant melanoma cannot always be unequivocally established, especially because the two lesions may have more features in common than they have differences.17 Furthermore, as stated by Peters and Goellner,15 "no single criterion is reliable in distinguishing Spitz nevi from malignant melanomas." (p. 1301)
Potentially erroneous histopathologic diagnoses It is possible that erroneous diagnoses might be responsible for unexpected prognoses. For example, in the original series described by Spitz, it is probable that the 12-year-old girl diagnosed with juvenile melanoma (Spitz nevus), who died of metastatic disease, actually had a malignant melanoma. This is especially likely in light of the author's acknowledgment that at the time, in 1948, the histologic distinction between Spitz nevus and melanoma "could not be made with certainty in most cases. ''I In addition, the author admitted that the lesion in question had both clinical and histologic characteristics that were distinctly different from those observed in the rest of the study group. The problem of melanoma simulating a Spitz nevus histologically has also been noted in more recent studies. Okun 13 reported three cases of melanoma that histologically resembled spindle and epithelioid cell nevi. Two of these cases were erroneously diagnosed as Spitz nevi. Six of seven pathologists involved in the first case, and two of four pathologists in the second case, confirmed the incorrect diagnoses. The third case was accurately diagnosed despite its striking resemblance to a Spitz nevus. All three patients died of malignant melanoma. A similar case of a 10-year-old boy with a mela-
Spitz nevi 905 noma masquerading as a Spitz nevus was described by Goldes et al. 32The case was further complicated because 4 years earlier, the boy had acute lymphoblastic leukemia that had been treated and was in complete remission. The initial shave biopsy specimen of the patient's lesion, as well as the specimens of his two subsequent recurrences, were histologically misinterpreted by several pathologists as Spitz nevi. Two years after the discovery of the first lesion, the boy died of metastatic melanoma. A study by Peters and Goellner 15revealed several cases in which Spitz nevi and melanomas were correctly diagnosed initially and then misdiagnosed retrospectively. The histologic material of 33 cases of Spitz nevi and 19 cases of malignant melanoma was reviewed without knowledge of the clinical appearance or behaivor of the lesions. Of the 19 cases known to be melanomas, two were erroneously categorized as Spitz nevi by criteria used in the study, and of the 33 cases known to be Spitz nevi, two were erroneously diagnosed as melanoma by histologic criteria alone. Finally, there have been descriptions of "malignant Spitz nevi.''2~ 3o,33, 34 In one study of 32 cases (six of which were originally diagnosed as melanoma), the lesions were categorized as Spitz nevi despite the existence of several atypical histologic features including: (1) high mitotic rate, (2) mitoses deep in the lesion, (3) lack of maturation of melanocytes, (4) increased cellularity, (5) increased cellular pleomorphism, (6) extension into the subcutaneous fat, (7) ulceration, (8) loss of cellular cohesion, and (9) large lesional diameter of more than 1 cm.34 Furthermore, 6 of the 32 patients in the study were found to have spindle or epithelioid cells in regional lymph nodes. Smith et al. 34 stated that despite the greater degree of atypia, all the 32 lesions studied "fit cytologically within the spectrum of over 100 additional spindle and/or epithelioid cell nevi reviewed" (p. 9 33), except for the features of large size, ulceration, and lymph node involvement. Therefore the authors designated these lesions as "malignant Spitz nevi." However, it appears that not only are those characteristics unusual for Spitz nevi, but of the remaining seven atypical features listed, at least three (mitoses deep in the lesion, lack of maturation of melanocytes, and extension into the subcutaneous fat) do not meet the criteria for Spitz nevi in Table I, and four of them (high mitotic rate, mitoses deep in the lesion, increased cellular pleomorphism, and extension into the subcutaneous fat) fail to meet
906
Casso et aL
even the criteria of atypical Spitz nevi described previously. Despite these data, the authors maintain that these lesions lack many features of typical malignant melanoma, such as poor circumscription and an infiltrative pattern, and do in fact fit within the spectrum of Spitz nevi. 34 These studies illustrate that histologic differentiation between a Spitz nevus and malignant melanoma is often difficult and can lead to the over diagnosis or underdiagnosis of one or the other, sometimes with serious consequences. The study by Peters and Goellner 15 emphasized that histologic data alone are not always sufficient to establish a correct diagnosis; they must be correlated with the clinical appearance and behavior of the lesion to ensure a more accurate assessment. In addition, adequate biopsy specimens are necessary for accurate diagnosis. 14Specimens that lack the deep and lateral margins of the lesion may give misleading impressions and may precipitate erroneous diagnoses. Even an ideal specimen, however, cannot guarantee an accurate diagnosis. The overlap in histologic criteria make each individual case a diagnostic challenge. POTENTIAL ROLE OF MOLECULAR AND IMMUNOHISTOCHEMICAL STUDIES IN DIAGNOSIS OF SPITZ NEVI Because the clinical and histologic differentiation of Spitz nevi from malignant melanoma can often be difficult, other methods are being explored that might aid in the differential diagnosis. These include flow cytometry, silver staining techniques, and immunohistochemical studies. One method under investigation is flow cytometry. Flow cytometry analyzes the D N A content within cells. It has been suggested that the finding of aneuploidy, abnormal quantities of DNA within cells, might be associated with various malignancies, including melanoma. Several investigators have attempted to determine whether DNA aneuploidy can be used as a marker of malignant melanoma. 35-39It is hypothesized that cells with an abnormal DNA content may have lost control mechanisms important in limiting tumor growth and therefore might result in an increased propensity for tumor progression.39 Kheir et al.,36 in a study of 162 paraffin-embedded stage I cutaneous melanomas, determined the aneuploid rate to be 22%. Another study used the same technique on 53 primary melanomas and found 25% to be aneuploid. 3s A third series used fresh-frozen tissue to determine that 74% of their 3 5
Journal of the American Academy of Dermatology melanomas were aneuploid. 4~ One theory proposes that the discrepancies in aneuploid rates are most likely secondary to the sampling of different populations with lesions of different thicknesses. It has been noted by several authors that aneuploidy is found more commonly, although not exclusively, i in association with thick or high Clark level le-/ sions.36, 38,39A secondtheory, proposed by Herzberg et a1.,35suggests that the discrepancies in aneuploidy rates may be partially caused by differing methods. In a study of 14 patients with a primary melanoma, these authors used image rather than flow cytometry, thereby improving their ability to measure aneuploidy, especially in thin lesions. They were unable to confirm the aforementioned correlations between aneuploidy and thickness. However, their findings that aneuploidy could be detected in 78% (7 of 9) of fresh-frozen samples and in only 44% (4 of 9) of the paraffin-embedded samples supports the hypothesis that differing methods of investigation may be responsible for the discrepant results found in this and the three previously cited studies. 36, 38,40 A study by Winokur et al.39investigated DNA content in 14 Spitz nevi. All 14 lesions were found to be diploid. This finding suggests that the use of DNA ploidy might help in differentiation. However, other studies of benign melanocytic nevi have discovered aneuploidy rates ranging from 3%38 to 7%37 to 25%.40 Therefore, it appears that DNA aneuploidy occurs more frequently in malignant melanomas than in benign melanocytic lesions such as Spitz nevi. However, because benign lesions can be aneuploid in approximately 3% to 25% of cases37-40 and because melanomas can be diploid,36,38-40 DNA ploidy cannot be reliably used to distinguish definitively benign from malignant melanocytic lesions, including Spitz nevi and melanomas. Perhaps future studies comparing the DNA ploidy of these two lesions with the use of image cytometry and fresh-frozen tissue techniques will be able to define better the role of aneuploidy as a diagnostic and prognostic tool. Another technique that has been investigated in hopes of finding a good means to differentiate Spitz nevi from melanomas involves the use of silver staining nucleolar organizer regions (AgNOR). 41"43 AgNORs are loops of D N A that transcribe ribosomal RNA. They can be identified when fixed with a silver (Ag) stain. It is hypothesized that if the number of AgNORs per cell correlates with cellular activity, then increased numbers may indicate
Volume27 Number 6, Part 1 December 1992 malignancy.43 A study by Crocker and Stillbeck41 demonstrated a difference in AgNOR counts between benign and malignant melanocytic lesions. Melanomas were found to have a mean value of 7.9 AgNORs per cell, whereas benign melanocytic lesions had only 1.2 AgNORs per cell. Unfortunately, a later study of 29 Spitz nevi and 39 invasive melanomas was unable to confirm these findings.43 As in the previous study, these investigators calculated the same mean of 1.2 AgNORs per cellin Spitz nevi, with a range of 1.0 to 1.6. However, they found the mean count in melanomas to be only 2.0 AgNORs per cell, with a range of 1.2 to 4.2. Therefore, because of the significant overlap in the AgNOR counts found by these authors and confirmed by others,42it cart be concluded that AgNOR counts are unable to differentiate reliably Spitz nevi from melanomas. However, Howat et al.43proposed that a count greater than 2.0 would certainly favor a diagnosis of malignant melanoma. Further investigation of this technique is essential to establish whether in fact it will be useful clinically. Immunohistochemical studies have also been investigated. This technique uses antibodies to either S- 100 protein or the monoclonal antibody HMB-45. S-100 protein staining is sensitive in detecting melanocytes, but it lacks specificity. 3~This protein has been found not only in melanocytes, but also in many other cells including Langerhans cells, Schwann cells, and histiocytes. 3~ 44, 45 In addition, S-100 protein has been found in many tumors associated with these cells, including benign melanocytic nevi and malignant melanomas. 3~ 44,46 In studies by Kahn et al. 44 and by Skelton et al.,45 positive staining for the S-100 protein was detected in melanomas and in all types of melanocytic nevi except in monophasic cellular blue nevi. Therefore it can be concluded that the S-100 protein can be a useful marker in determining the histogenesis of various skin tumors but cannot be used to distinguish between Spitz nevi and melanomas because both display a similar staining pattern. HMB-45 is a monoclonal antibody, originally described as a marker specific for malignant melanoma and junctional melanocytic nevi.47 However, subsequent studies have been unable to confirm this. One study of 12 Spitz nevi showed an HMB-45 staining pattern similar to that of malignant melanoma. 46Skelton et al., 4s in a series 0f225 benign and malignant melanocytic lesions, determined that all lesions with a junctional component, whether benign
Spitz nevi 907 or malignant, showed apositive reaction with HMB45. In addition, 94% (28 of 30 patients) of the mafignant melanomas studied showed positive staining in the dermal components, primarily in a diffuse pattern. The Spitz nevi also demonstrated positive HMB-45 staining in 94% (28 of 30) of cases. However, the majority of these benign lesions had positive reactivity in the nevus cells at or near the dermoepidermal junction, with staining often focal in pattern; only a few Spitz nevi showed positive reactivity deep within the dermis. HMB-45 has proved to be most valuable as a highly specific and sensitive marker for the differentiation of malignant melanoma from nonmelanocytic malignant tumors. HMB-45 correlates with active early melanosome production, melanocytic differentiation, and therefore a melanocytic origin of positive-staining cells. However, it is not a melanoma marker and cannot be used to differentiate between Spitz nevi and malignant melanomas. As stated by Skelton etal., 45 "HMB-45 is a melanocytic antigen, not a melanoma antigen." (p. 548) It is evident from these investigations that currently there exists no known sensitive and specific marker that can reliably differentiate a Spitz nevus from a malignant melanoma. The histologic, molecular, and immunohistochemical analyses all remain somewhat flawed. Even when used in combination, these tools are not definitive, making diagnosis an imperfect science, and giving rise to controversies regarding prognosis and treatment. PROGNOSIS: RECURRENCES AND
"METASTATIC" POTENTIAL A review of 13 articles published from 1948 to 1990 was undertaken to address the controversial issues related to the prognosis of Spitz nevi, the incidence of local recurrence, and the development of metastatic disease. Unfortunately, the majority of these studies involve a small number of cases and short follow-up periods. Therefore, drawing conclusions from these data is difficult and they should be recognized as crude estimates. In this section, we attempt to present a general overview of the studies available (summarized in Table II); several of them are discussed in detail. Of the 13 articles analyzed, 11 discussed Spitz nevi, one "malignant" Spitz nevus, and one recurrent Spitz nevus. The number of cases in each series ranged from a single patient 49 to a review of 211 patients. 21The combined total number of cases is 716.
908
Journal of the American Academy of Dermatology
Casso et al.
Table II. S u m m a r y of available studies on Spitz nevi
Author(s>
No. of No. with INo. oflocall No. of IAveragefollow-] leases [positivemargins[recurrenceslme~t~s~ I .p
AIlleRICaN A C a D e m Y OF
AAD' %'2J
DerMaTOLOGY VOLUME 27
NUMBER 6
PART 1
DECEMBER 1992
IIII
s
medical e d u c a t i o n Spitz nevi Elen M. Casso, MD, Caron M. Grin-Jorgensen, MD, and Jane M. Grant-Kels, M D
Farmington, Connecticut The Spitz nevus has long been defined as a benign melanocytic lesion that shares many histologic features with malignant melanoma. Despite the diagnostic criteria established for these two entities, their histologic similarities continue to make their distinction somewhat difficult. Uncertainties also exist with regard to the natural history of the Spitz nevus; the true pattern of this lesion's biologic behavior remains elusive. As a result, controversies exist with respect to appropriate therapy. To examine these controversies, the epidemiology, clinical features, and histopathology of Spitz nevi, as well as the role of recent molecular and immunohistochemical diagnostic studies, are discussed. However, the primary focus of this article is the natural course, prognosis, and treatment of the Spitz nevus. A review of 716 cases of Spitz nevi, compiled from 13 papers published from 1948 to 1990, is presented. After analyzing this and other available data, we propose that at this time Spitz nevi and malignant melanoma cannot easily be categorized as distinct entities and that perhaps they actually exist along one continuum of disease. Because of this uncertainty and the difficulties in differentiating these two lesions, we recommend that treatment include complete excision of all Spitz nevi followed by reexcision of positive margins if present. (J AM ACAD DERMATOL 1992;27:901-13.) The Spitz nevus was first described in 1948 by Spitz,1 who referred to it as a "juvenile melanoma." She established the initial histologic criteria for differentiating this lesion from malignant melanoma. Since then, the term benign juvenile melanoma has been criticized as a misnomer for two reasons: the lesion is encountered not only in children but in adults and the term melanoma carries a connotation of malignancy, whereas Spitz ned are considered benign. More recently, the designation spindle and~or epithelioid cell nevus has been used to reflect its histologic features. The term Spitz nevus has
The CME articles are made possible through an educational grant from tile Dermatological Division, Ortho Pharmaceutical Corporation. From the Division of Dermatology, The University of Connecticut, Farmington, CT 06030. Reprint requests: Caron M. Grin-Jorgensen, MD, Division of Dermatology, The University of Connecticut, Farmington, CT 06030. ORTHO
16/2/42525
been adopted and retained in acknowledgement of the lesion's historical identification and description by Spitz. Many authors have contributed to the understanding of Spitz nevi by defining several criteria that helps the pathologist to distinguish between Spitz nevi and malignant melanomas.l2i However, few studies have focused on the natural history and prognosis after the diagnosis of a Spitz nevus. The reason for this is twofold. First, the majority of Spitz nevi have been surgically excised. As a result, the natural course of these lesions has been disrupted. Second, few data are available on long-term follow-up of patients either with intact Spitz nevi or after surgical removal. The purpose of this article is to discuss the epidemiology and clinical and histopathologic features of Spitz nevi, while focusing primarily on the course and prognosis to determine appropriate treatment options. In addition, various controversies about Spitz nevi will be addressed. 901
902
Casso et al.
EPIDEMIOLOGY
Spitz nevi occur most often during the first two decades of life. However, they can appear at any age from birth 1, 22 to the eighth decade. 6 In a study of 202 cases of Spitz nevi, 79% of the patients were younger than 20 years of age. 6 In another report of 33 cases, 60% occurred in children younger than 14 years of age. 15 The incidence in adults has been shown to decrease with increasing age. A study of 200 cases of Spitz nevi found that 28.5% of the patients were older than 30 years of age, and only 8.5% were older than 45.14 All studies have shown that Spitz nevi occur predominantly in the white population. Only seven cases of Spitz nevi in blacks have been reported.14' 23, 24 Bovenmyer23 concluded that this simply reflects a lower incidence of this nevus in blacks. However, Carr et al. 24 hypothesized that the small number of cases m a y be related to underdiagnosis of these nevi in the black population. A slight female preponderance of Spitz nevi has been documented in m a n y series. 6' 7,10, 12,14,21 This is especially true in women between the ages of 15 and 29 years. 14 It has been theorized that this difference most likely results from increased cosmetic concern that leads to a higher rate of removal and subsequent reporting. 6, 10, 14, 21 The actual incidence of Spitz nevi in the population is not known. Various authors have estimated that between 1% and 10% of all melanocytic nevi in children are, in fact, Spitz nevi.11, 25 Paniago-Pereira e t a l . 14 proposed that these are overestimates and that Spitz nevi constitute less than 1% of all nevi in children. CLINICAL FEATURES
Clinically, Spitz nevi most commonly appear as asymptomatic, dome-shaped, round to oval, firm nodules. They have a smooth surface and are usually pink, red, or tan. 6' 10,12, 14, 15, 22 Variations of these clinical characteristics are often encountered. For instance, the lesion may also be macular, polypoid, or verrucous. 6, 12,14,22 In a study of 211 cases of Spitz nevi, 57% were described as dome-shaped, 19% were classified as fiat, and 24% were called polypoid. 21 Spitz nevi also appear in a wide spectrum of colors, ranging from translucent to purple 22 to brown or black. 7, 10, 14 One clinical feature that is notably absent is ulceration. 7' 12, 14,22 For simplicity, the clinical appearance can be divided into four basic types 14, 24, 2s, 26. 1. Light-colored and soft. These are usually smooth, pink, or slightly pigmented. They may also
Journal of the American Academy of Dermatology
resemble a pyogenic granuloma or h e m a n gioma. 2. Light-colored and hard, resembling a dermatofibroma or small keloid. 3. Dark-colored. This less common type varies in degree of pigmentation; more darkly pigmented lesions occur less frequently. Several studies have suggested that the degree of pigmentation varies with the age of the patient. This leads to more darkly pigmented lesions in older persons. For example, Peters and Goellner 15 noted that in 33 patients, all of whom were younger than 20 years of age, 4 h a d Spitz nevi that were black and 3 of those 4 patients were older than 16 years of age. Another study of 27 patients cited that 4 of 6 adults (67%) had brown or black nevi, whereas only 6 of the 21 children (29%) had a dark lesion, l~ In contrast to those results, the study of Coskey and Mehregan 6 of 202 cases found no significant clinical differences in the pigmentation of Spitz nevi in 53 adults and 149 children. 4. Multiple or agminated lesions that can further be subdivided into the following four types27: (1) disseminated, (2) grouped and arising on normal skin, (3) grouped and arising on a hypopigmented macule, or (4) grouped and arising on a congenital pigmented macule. Although Spitz nevi can occur anywhere, 14 the head, neck, and especially the face appear to be the most common locations. 6, 7, 10,12,15, 2~ One study of 27 patients revealed that 56% of the Spitz nevi occurred on the face. 10 Several reports have f o u n d t h e legs to be another common site, with less frequent involvement of the trunk and arms. 6, 7, 10,12. 14, 15 In contrast, a study of 211 cases found that most lesions (39%) were located on the trunk; the legs were next in frequency. 21 In women, there may be a slight predilection for these nevi to occur on the legs. 14, 21 Coskey and Mehregan 6 reported that the distribution of Spitz nevi differs in adults and children. In a study of 202 cases, the most frequent location in children was the face, head, and neck (48%), whereas in adults, 45% were located on the legs. A n o t h e r study of 33 cases failed to find a correlation between age and anatomic location. 15 Spitz nevi can range in size from 1 m m to 3 ClTI.6,7, 10,21, 22 However, most are tess than 1 c m in diameter 7, 10; one study of 211 cases determined that 73% of the nevi were smaller than 6 ram. 21 Several studies have shown that before diagnosis, the duration of Spitz nevi may range from 1 m o n t h to 20 years or more. 7' 10,22 However, the majority of patients have had their lesion for only a short time, usually less than 6 months, before seeking medical
Volume 27 Number 6, Part 1 December 1992 advice. In several studies, the changes most often noted by patients have been slow enlargement (in 35% to 46% of cases) 7, 10, i4, 15; changes in color (in 8% to 27% of cases) 7, 15; bleeding (in 8% to 18% of cases)7'l~ and localized pruritus (in 9% of cases). 15 HISTOPATHOLOGIC FEATURES Histologically, Spitz nevi are composed of spindle cells, epithelioid cells, or a mixture of both; they may be classified as junctional, intradermal, or, most frequently, compound, i2, 14 Spitz nevi share several histologic features with malignant melanomas and therefore make the distinction between the two entities difficult at times. Both are characterized by nests of atypical melanocytes within the epidermis and the dermis. In some cases, Spitz nevi may even possess more extensive nuclear atypia than do malignant melanomas. 2, 17 Since Spitz's initial description in 1948, many authors have attempted to establish histologic criteria to distinguish between these two entities. These criteria are summarized in Table I.* One of the more recent additions to these criteria has been the description of eosinophilic globules, at the dermoepidermal junction in Spitz nevi. These globules were first described by Kamino in 1979 who reported the presence of these eosinophilic bodies within the epidermis in 60% of 293 cases of Spitz nevi and in only 2% of 293 cases of malignant melanoma. In addition, the staining of these epidermal globules was positive with both periodic acid-Schiff and trichrome in Spitz nevi and was negative for both in malignant melanoma. 9 These globules were n a m e d "Kamino bodies" and were confirmed as morphologic markers of Spitz nevi by Arbuckle and Weedon 4 who described their presence in 86% of 50 cases of Spitz nevi and in only 12% of 50 cases of malignant melanoma. 4 Peters and Goeliner] 5 in a smaller study, found Kamino bodies in 52% of Spitz nevi and in 26% of malignant melanomas. It was proposed that the frequent occurrence of Kamino bodies in Spitz nevi may be caused by increased proliferation of keratinocytes and melanocytes in these tumors and subsequent death by apoptosis of some of these cells. 4 Awareness of this morphologic marker has aided in the ability to distinguish between Spitz nevi and malignant melanoma. However, despite weU-established criteria, there are still cases in which the diagnosis is unclear because of an overlap in histologic features between *References 2, 5, 10, 13-15, 17, 21.
Spitz nevi 903 Table I. Histologic criteria to distinguish Spitz nevi from malignant melanoma Spitz nevi
Symmetric lesion Diameter usually 6ram Poor circumscription of intraepidermal melanocytic component Wide variation in size and shape of melanocytic nests Nests of melanocytes within epidermis are neither elongated nor perpendicular to skin surface Usual absence of Kamino bodies within epidermis Confluence of nests of melanocytes within epidermis Few clefts, if any, between nests of melanocytes and keratinocytes Nests of dermal melanocytes tend to become confluent and form sheets of cells Usually no maturation of melanocytes with descent into the dermis Mitotic figures near base of lesion are common Inflammatory cell infiltrate, when present, may be bandlike, beneath, intermingling with, or often extending beyond the melanocytic component
Adapted from Ackerman et al. Differential diagnosis in dermatopathology. Philadelphia: Lea & Febiger, 1982.
Spitz nevi and malignant melanomas. 15 There are many papers that describe various atypical forms ot~ Spitz nevi that seem to fall into a "gray area" between the diagnoses of Spitz nevus and malignant
904
Casso el al.
melanoma. The following features have been identified by several authors as important in favoring the diagnosis of an atypical Spitz nevus over a malignant melanoma: ( 1) low mitotic activity, 10,11,1s especially in the deeper dermal component2~ (2) absence of atypical mitoses 1~ l 1, 2l; (3) slight pleomorphisml0, 11, 15,20; (4) rarity of large nucleolil~ (5) absence of hyperchromasia 11, 15,21; ( 6 ) marked spindle cell component 14' 15; (7) minimal or no significant upward epidermal spread of melanocytes2~ 21; (8) absence of intense melanization, TM15 especially deep within the lesion13; (9) absence of marked ulcerationlS; and (10) absence of deep penetration of the lesion into the deeper reticular dermis and/or subcutaneous fat.15 Histologie differences in adults versus children
Another controversial topic is the histologic differences between Spitz nevi in adults and children. Because the incidence &malignant melanoma is low in children and rises sharply after puberty, the chances of making an erroneous diagnosis of Spitz nevus increase proportionately in adults. 13 Therefore any histologic distinction found between Spitz nevi in adults and children might prove to be a valuable diagnostic tool. In general, the overall histologic appearance of Spitz ncvi is similar in adults and children.6, 7, 10, la-~4,21 However, several differences have been observed. For example, a study of 200 Spitz nevi revealed histologic characteristics that were absent in adults and present in young children. 14 These were papiltomatosis, edema of the papillary dermis, telangiectases, a high concentration of epithelioid nevus cells, and acantholytic nevus cells. In another study of 89 cases, it was noted that involvement of the suprabasal epidermis of Spitz nevi occurs much more frequently in children than in adults.t2 The authors stated that this difference in degree of intraepidermal involvement is a result of the gradual regressive changes of the Spitz nevus that normally occur with age. On the basis of this theory of involution and the observation that this histologic feature is rare in adults, these authors suggested strong consideration of the diagnosis of melanoma when melanocytes involve the suprabasal epidermis within a spitzoid nevus in an "adult." In analyzing their data, however, it can be seen that there exists much overlap regarding the patients' ages in the study groups defined as "with" and "without" intraepidermaI involvement. As a result,
Journal of the American Academy of Dermatology it is difficult to determine the age of onset of the nevus in their "adult" population; therefore it is unclear at what age the presence of suprabasal epidermal involvement in a Spitz nevus should be considered worrisome. Additional studies have found that Spitz nevi in adults contain greater amounts of melanin pigment than do those in children.7, a0,14 However, some authors classify these lesions not as Spitz nevi but as pigmented spindle cell nevi (PSCN) 16' 18, 19,28-3o that occur most commonly in women during the third decade of life. 30 It remains a point of controversy as to whether the PSCN is a variant of the Spitz nevus, or a distinct entity. It is generally agreed that the most prominent features that differentiate PSCN from Spitz nevi are the following16, 18,19,28-30:(1) abundant melanin pigment; (2) nests of spindle cells that are confined to the epidermis and papillary dermis, with absence of reticular dermal involvement; (3) virtual absence of epithelioid cells composing the neoplasm; the cells are spindled with delicate nuclei and eventy dispersed chromatin; (4) an expansile growth pattern that compresses and displaces the reticular dermis, with absence of an infiltrative pattern; (5) absence of stromal edema and telangiectasia; and (6) rare upward intraepidermal spread of melanocytes. Despite these differences, there are significantly more similarities between the two entities. In fact, one author admits that junctional Spitz nevi, in particular, can be distinguished from PSCN only by the relative absence of pigment both clinically and histologically.lS This appears to favor the classification of PSCN as a special variant of Spitz nevi. However, some authors maintain that the most important reason to classify PSCN as a distinct entity is to prevent its misdiagnosis as malignant melanoma, especially in young adults. 16,18,19,28 In examination of the possible differences between Spitz nevi in adults and children, it is necessary to include the issue of the desmoplastic Spitz nevus, a histologic variant of the Spitz nevus that one author suggested was found only in adults older than 30 years of age. 14It has been postulated that this lesion, described as a Spitz nevus with excessive fibrosis, results from tumor regression or a slowly progressive process of involution that can be correlated with increasing age of the patient. 3' z2, 14, 16,22 However, in a study of 75 Spitz nevi, 14 of which were desmoplastic, there was no significant correlation between desmoplasia and either patient age or
Volume 27 Number 6, Part I December 1992
duration of the lesion. 31 These findings were supported by Kernen and Ackerman,10 who reported no histologic evidence of transformation with age, and by Coskey and Mehregan, 6 who noted no signs of regression, even in lesions of long duration. Several large studies, including a study of 2l 1 cases 21 and another of 202 cases,6 failed to substantiate the findings of all of the aforementioned studies regarding the histologic variations of the Spitz nevus in children and adults. Both studies concluded that no histologic differences exist between the Spitz nevi of children and adults. Although the histopathology of Spitz nevi has been well studied, it is evident that many controversies still exist, such as the question of whether or not histologic distinctions can be made in children versus adults. In addition, its differentiation from malignant melanoma cannot always be unequivocally established, especially because the two lesions may have more features in common than they have differences.17 Furthermore, as stated by Peters and Goellner,15 "no single criterion is reliable in distinguishing Spitz nevi from malignant melanomas." (p. 1301)
Potentially erroneous histopathologic diagnoses It is possible that erroneous diagnoses might be responsible for unexpected prognoses. For example, in the original series described by Spitz, it is probable that the 12-year-old girl diagnosed with juvenile melanoma (Spitz nevus), who died of metastatic disease, actually had a malignant melanoma. This is especially likely in light of the author's acknowledgment that at the time, in 1948, the histologic distinction between Spitz nevus and melanoma "could not be made with certainty in most cases. ''I In addition, the author admitted that the lesion in question had both clinical and histologic characteristics that were distinctly different from those observed in the rest of the study group. The problem of melanoma simulating a Spitz nevus histologically has also been noted in more recent studies. Okun 13 reported three cases of melanoma that histologically resembled spindle and epithelioid cell nevi. Two of these cases were erroneously diagnosed as Spitz nevi. Six of seven pathologists involved in the first case, and two of four pathologists in the second case, confirmed the incorrect diagnoses. The third case was accurately diagnosed despite its striking resemblance to a Spitz nevus. All three patients died of malignant melanoma. A similar case of a 10-year-old boy with a mela-
Spitz nevi 905 noma masquerading as a Spitz nevus was described by Goldes et al. 32The case was further complicated because 4 years earlier, the boy had acute lymphoblastic leukemia that had been treated and was in complete remission. The initial shave biopsy specimen of the patient's lesion, as well as the specimens of his two subsequent recurrences, were histologically misinterpreted by several pathologists as Spitz nevi. Two years after the discovery of the first lesion, the boy died of metastatic melanoma. A study by Peters and Goellner 15revealed several cases in which Spitz nevi and melanomas were correctly diagnosed initially and then misdiagnosed retrospectively. The histologic material of 33 cases of Spitz nevi and 19 cases of malignant melanoma was reviewed without knowledge of the clinical appearance or behaivor of the lesions. Of the 19 cases known to be melanomas, two were erroneously categorized as Spitz nevi by criteria used in the study, and of the 33 cases known to be Spitz nevi, two were erroneously diagnosed as melanoma by histologic criteria alone. Finally, there have been descriptions of "malignant Spitz nevi.''2~ 3o,33, 34 In one study of 32 cases (six of which were originally diagnosed as melanoma), the lesions were categorized as Spitz nevi despite the existence of several atypical histologic features including: (1) high mitotic rate, (2) mitoses deep in the lesion, (3) lack of maturation of melanocytes, (4) increased cellularity, (5) increased cellular pleomorphism, (6) extension into the subcutaneous fat, (7) ulceration, (8) loss of cellular cohesion, and (9) large lesional diameter of more than 1 cm.34 Furthermore, 6 of the 32 patients in the study were found to have spindle or epithelioid cells in regional lymph nodes. Smith et al. 34 stated that despite the greater degree of atypia, all the 32 lesions studied "fit cytologically within the spectrum of over 100 additional spindle and/or epithelioid cell nevi reviewed" (p. 9 33), except for the features of large size, ulceration, and lymph node involvement. Therefore the authors designated these lesions as "malignant Spitz nevi." However, it appears that not only are those characteristics unusual for Spitz nevi, but of the remaining seven atypical features listed, at least three (mitoses deep in the lesion, lack of maturation of melanocytes, and extension into the subcutaneous fat) do not meet the criteria for Spitz nevi in Table I, and four of them (high mitotic rate, mitoses deep in the lesion, increased cellular pleomorphism, and extension into the subcutaneous fat) fail to meet
906
Casso et aL
even the criteria of atypical Spitz nevi described previously. Despite these data, the authors maintain that these lesions lack many features of typical malignant melanoma, such as poor circumscription and an infiltrative pattern, and do in fact fit within the spectrum of Spitz nevi. 34 These studies illustrate that histologic differentiation between a Spitz nevus and malignant melanoma is often difficult and can lead to the over diagnosis or underdiagnosis of one or the other, sometimes with serious consequences. The study by Peters and Goellner 15 emphasized that histologic data alone are not always sufficient to establish a correct diagnosis; they must be correlated with the clinical appearance and behavior of the lesion to ensure a more accurate assessment. In addition, adequate biopsy specimens are necessary for accurate diagnosis. 14Specimens that lack the deep and lateral margins of the lesion may give misleading impressions and may precipitate erroneous diagnoses. Even an ideal specimen, however, cannot guarantee an accurate diagnosis. The overlap in histologic criteria make each individual case a diagnostic challenge. POTENTIAL ROLE OF MOLECULAR AND IMMUNOHISTOCHEMICAL STUDIES IN DIAGNOSIS OF SPITZ NEVI Because the clinical and histologic differentiation of Spitz nevi from malignant melanoma can often be difficult, other methods are being explored that might aid in the differential diagnosis. These include flow cytometry, silver staining techniques, and immunohistochemical studies. One method under investigation is flow cytometry. Flow cytometry analyzes the D N A content within cells. It has been suggested that the finding of aneuploidy, abnormal quantities of DNA within cells, might be associated with various malignancies, including melanoma. Several investigators have attempted to determine whether DNA aneuploidy can be used as a marker of malignant melanoma. 35-39It is hypothesized that cells with an abnormal DNA content may have lost control mechanisms important in limiting tumor growth and therefore might result in an increased propensity for tumor progression.39 Kheir et al.,36 in a study of 162 paraffin-embedded stage I cutaneous melanomas, determined the aneuploid rate to be 22%. Another study used the same technique on 53 primary melanomas and found 25% to be aneuploid. 3s A third series used fresh-frozen tissue to determine that 74% of their 3 5
Journal of the American Academy of Dermatology melanomas were aneuploid. 4~ One theory proposes that the discrepancies in aneuploid rates are most likely secondary to the sampling of different populations with lesions of different thicknesses. It has been noted by several authors that aneuploidy is found more commonly, although not exclusively, i in association with thick or high Clark level le-/ sions.36, 38,39A secondtheory, proposed by Herzberg et a1.,35suggests that the discrepancies in aneuploidy rates may be partially caused by differing methods. In a study of 14 patients with a primary melanoma, these authors used image rather than flow cytometry, thereby improving their ability to measure aneuploidy, especially in thin lesions. They were unable to confirm the aforementioned correlations between aneuploidy and thickness. However, their findings that aneuploidy could be detected in 78% (7 of 9) of fresh-frozen samples and in only 44% (4 of 9) of the paraffin-embedded samples supports the hypothesis that differing methods of investigation may be responsible for the discrepant results found in this and the three previously cited studies. 36, 38,40 A study by Winokur et al.39investigated DNA content in 14 Spitz nevi. All 14 lesions were found to be diploid. This finding suggests that the use of DNA ploidy might help in differentiation. However, other studies of benign melanocytic nevi have discovered aneuploidy rates ranging from 3%38 to 7%37 to 25%.40 Therefore, it appears that DNA aneuploidy occurs more frequently in malignant melanomas than in benign melanocytic lesions such as Spitz nevi. However, because benign lesions can be aneuploid in approximately 3% to 25% of cases37-40 and because melanomas can be diploid,36,38-40 DNA ploidy cannot be reliably used to distinguish definitively benign from malignant melanocytic lesions, including Spitz nevi and melanomas. Perhaps future studies comparing the DNA ploidy of these two lesions with the use of image cytometry and fresh-frozen tissue techniques will be able to define better the role of aneuploidy as a diagnostic and prognostic tool. Another technique that has been investigated in hopes of finding a good means to differentiate Spitz nevi from melanomas involves the use of silver staining nucleolar organizer regions (AgNOR). 41"43 AgNORs are loops of D N A that transcribe ribosomal RNA. They can be identified when fixed with a silver (Ag) stain. It is hypothesized that if the number of AgNORs per cell correlates with cellular activity, then increased numbers may indicate
Volume27 Number 6, Part 1 December 1992 malignancy.43 A study by Crocker and Stillbeck41 demonstrated a difference in AgNOR counts between benign and malignant melanocytic lesions. Melanomas were found to have a mean value of 7.9 AgNORs per cell, whereas benign melanocytic lesions had only 1.2 AgNORs per cell. Unfortunately, a later study of 29 Spitz nevi and 39 invasive melanomas was unable to confirm these findings.43 As in the previous study, these investigators calculated the same mean of 1.2 AgNORs per cellin Spitz nevi, with a range of 1.0 to 1.6. However, they found the mean count in melanomas to be only 2.0 AgNORs per cell, with a range of 1.2 to 4.2. Therefore, because of the significant overlap in the AgNOR counts found by these authors and confirmed by others,42it cart be concluded that AgNOR counts are unable to differentiate reliably Spitz nevi from melanomas. However, Howat et al.43proposed that a count greater than 2.0 would certainly favor a diagnosis of malignant melanoma. Further investigation of this technique is essential to establish whether in fact it will be useful clinically. Immunohistochemical studies have also been investigated. This technique uses antibodies to either S- 100 protein or the monoclonal antibody HMB-45. S-100 protein staining is sensitive in detecting melanocytes, but it lacks specificity. 3~This protein has been found not only in melanocytes, but also in many other cells including Langerhans cells, Schwann cells, and histiocytes. 3~ 44, 45 In addition, S-100 protein has been found in many tumors associated with these cells, including benign melanocytic nevi and malignant melanomas. 3~ 44,46 In studies by Kahn et al. 44 and by Skelton et al.,45 positive staining for the S-100 protein was detected in melanomas and in all types of melanocytic nevi except in monophasic cellular blue nevi. Therefore it can be concluded that the S-100 protein can be a useful marker in determining the histogenesis of various skin tumors but cannot be used to distinguish between Spitz nevi and melanomas because both display a similar staining pattern. HMB-45 is a monoclonal antibody, originally described as a marker specific for malignant melanoma and junctional melanocytic nevi.47 However, subsequent studies have been unable to confirm this. One study of 12 Spitz nevi showed an HMB-45 staining pattern similar to that of malignant melanoma. 46Skelton et al., 4s in a series 0f225 benign and malignant melanocytic lesions, determined that all lesions with a junctional component, whether benign
Spitz nevi 907 or malignant, showed apositive reaction with HMB45. In addition, 94% (28 of 30 patients) of the mafignant melanomas studied showed positive staining in the dermal components, primarily in a diffuse pattern. The Spitz nevi also demonstrated positive HMB-45 staining in 94% (28 of 30) of cases. However, the majority of these benign lesions had positive reactivity in the nevus cells at or near the dermoepidermal junction, with staining often focal in pattern; only a few Spitz nevi showed positive reactivity deep within the dermis. HMB-45 has proved to be most valuable as a highly specific and sensitive marker for the differentiation of malignant melanoma from nonmelanocytic malignant tumors. HMB-45 correlates with active early melanosome production, melanocytic differentiation, and therefore a melanocytic origin of positive-staining cells. However, it is not a melanoma marker and cannot be used to differentiate between Spitz nevi and malignant melanomas. As stated by Skelton etal., 45 "HMB-45 is a melanocytic antigen, not a melanoma antigen." (p. 548) It is evident from these investigations that currently there exists no known sensitive and specific marker that can reliably differentiate a Spitz nevus from a malignant melanoma. The histologic, molecular, and immunohistochemical analyses all remain somewhat flawed. Even when used in combination, these tools are not definitive, making diagnosis an imperfect science, and giving rise to controversies regarding prognosis and treatment. PROGNOSIS: RECURRENCES AND
"METASTATIC" POTENTIAL A review of 13 articles published from 1948 to 1990 was undertaken to address the controversial issues related to the prognosis of Spitz nevi, the incidence of local recurrence, and the development of metastatic disease. Unfortunately, the majority of these studies involve a small number of cases and short follow-up periods. Therefore, drawing conclusions from these data is difficult and they should be recognized as crude estimates. In this section, we attempt to present a general overview of the studies available (summarized in Table II); several of them are discussed in detail. Of the 13 articles analyzed, 11 discussed Spitz nevi, one "malignant" Spitz nevus, and one recurrent Spitz nevus. The number of cases in each series ranged from a single patient 49 to a review of 211 patients. 21The combined total number of cases is 716.
908
Journal of the American Academy of Dermatology
Casso et al.
Table II. S u m m a r y of available studies on Spitz nevi
Author(s>
No. of No. with INo. oflocall No. of IAveragefollow-] leases [positivemargins[recurrenceslme~t~s~ I .p
