MELPREDICT: a logistic regression model to estimate CDKN2A carrier probability.

BACKGROUND: Heritable alterations in CDKN2A account for a subset of familial melanoma cases although no robust method exists to identify those at risk of being a mutation carrier. METHODS: We set out to construct a model for estimating CDKN2A mutation carrier probability using a cohort of 116 consecutive familial cutaneous melanoma patients evaluated at Massachusetts General Hospital Pigmented Lesion Center between April 2001 and September 2004. Germline CDKN2A and CDK4 status on the familial melanoma cases and clinical features associated with mutational status were then used to build a multiple logistic regression model to predict carrier probability and performance of model on external validation. RESULTS: From the 116 kindreds prone to melanoma in the Boston area, 13 CDKN2A mutation carriers were identified and 12 were subsequently used in the modeling. Proband age at diagnosis, number of proband primaries, and number of additional family primaries were most closely associated with germline mutations. The estimated probability of the proband being a mutation carrier based on the logistic regression model (MELPREDICT) is given by e(L)/(1 + e(L) where L = 1.99+[0.92x(no. of proband primaries)]+[0.74x(no. of additional family primaries)]-[2.11xln(age)]. The mean estimated probabilities for subjects in the Boston dataset were 55.4% and 5.1% for the mutation carriers and non-carriers respectively. In a receiver operator characteristic analysis, the area under the curve was 0.881 (95% confidence interval 0.739 to 1.000) for the Boston model set (n = 116) and 0.803 (0.729 to 0.877) for an external Toronto hereditary melanoma cohort (n = 143). CONCLUSIONS: These results represent the first-iteration logistic regression model to approximate CDKN2A carrier probability. Validation of this model with an external dataset revealed relatively robust performance.

Influence of estrogen administration on tumor characteristics and survival in women with cutaneous melanoma.

A prospective study on 289 women with clinical stage I cutaneous melanoma was done to determine the relationship between estrogen administration, tumor characteristics, and survival. Eighty-two women took oral contraceptives (OC) and 44 took menopausal estrogens (MPE) prior to the diagnosis of melanoma. Users of OC presented with thinner primary tumors than nonusers of OC (P less than .01). A similar trend was observed in users of MPE. Women who used OC in the year prior to the diagnosis of melanoma had statistically thinner tumors than those who had discontinued use of OC more than 1 year prior to diagnosis (P less than .025). A statistically significant preponderance of truncal lesions was observed among users of OC (P less than .01). Other tumor characteristics were unaltered by estrogen administration. Duration of use and time in relation to diagnosis of melanoma did not affect survival. Women who took hormones had slightly better 5- and 9-year survival rates than nonusers. These results suggest that prior estrogen use and, particularly, use of OC in women developing melanoma have no deleterious effect.

Novel mutations in the p16/CDKN2A binding region of the cyclin-dependent kinase-4 gene.

Mutations in genes that lie in the retinoblastoma pathway have been implicated in the pathogenesis of many tumor types. Two critical components that determine progression from G1 to S include p16/CDKN2A and CDK4. Alterations in p16/CDKN2A have been well documented in multiple cancers, including melanoma. However, changes in CDK4 are apparently more rare. Only two alterations, both at codon 24, have been identified in CDK4: an activating arginine-to-cysteine transition and a germ-line arginine-to-histidine substitution in one French kindred. In a survey of 20 neuroblastomas, 17 uncultured metastatic melanomas, 33 uncultured primary uveal melanomas, 8 colon cancer cell lines, and 20 primary colon cancer samples, we found no evidence of mutations in exon 2 of CDK4. From our cell lines derived from metastatic melanomas, we detected two alterations in the functionally critical exon 2 of CDK4: a lysine-to-glutamine transition at codon 22 and the arginine-to-histidine mutation at codon 24. These findings document several novel changes in the p16-binding region of CDK4.

Melastatin expression and prognosis in cutaneous malignant melanoma.

PURPOSE: Melastatin (MLSN-1), a novel melanocyte-specific gene recently identified using a genomic approach, is expressed in murine and human melanoma cells at levels inversely proportional to metastatic rates in vivo. We studied the relationship between expression of melastatin mRNA in the primary cutaneous tumor and prognosis in patients with localized malignant melanoma. PATIENTS AND METHODS: Melastatin mRNA was evaluated by in situ hybridization in primary cutaneous melanoma from 150 patients with localized disease (American Joint Committee on Cancer [AJCC] stage I and II). Multivariate Cox proportional hazards regression analysis was performed to assess the prognostic utility of melastatin mRNA expression while adjusting for other prognostic indicators. RESULTS: Uniform melastatin mRNA expression in the primary tumor was correlated with prolonged disease-free survival (P < .0001). Multivariate analysis revealed that melastatin status, mitotic rate, and tumor thickness influence disease-free survival independently. The 8-year disease-free survival rate in AJCC stage I patients whose tumors diffusely expressed melastatin mRNA was 100%, whereas in stage I patients with melastatin loss, the disease-free survival rate was 77% +/- 15% (median +/- SE). In patients with stage II disease whose tumors diffusely expressed melastatin mRNA, the 8-year disease-free survival rate was 90% +/- 7%, whereas in patients with melastatin loss, the disease-free survival rate was 51% +/- 8%. CONCLUSION: Downregulation of melastatin mRNA in the primary cutaneous tumor is a prognostic marker for metastasis in patients with localized malignant melanoma and is independent of tumor thickness and other variables. Used in combination, melastatin status and tumor thickness allow for the identification of subgroups of patients at high and low risk of developing metastatic disease.

Patient preferences for adjuvant interferon alfa-2b treatment.

PURPOSE: Although trials of adjuvant interferon alfa-2b (IFN alpha-2b) in high-risk melanoma patients suggest improvement in disease-free survival, it is unclear whether treatment offers improvement in overall survival. Widespread use of adjuvant IFN alpha-2b has been tempered by its significant toxicity. To quantify the trade-offs between IFN alpha-2b toxicity and survival, we assessed patient utilities for health states associated with IFN therapy. Utilities are measures of preference for a particular health state on a scale of 0 (death) to 1 (perfect health). PATIENTS AND METHODS: We assessed utilities for health states associated with adjuvant IFN among 107 low-risk melanoma patients using the standard gamble technique. Health states described four IFN alpha-2b toxicity scenarios and the following three posttreatment outcomes: disease-free health and melanoma recurrence (with or without IFN alpha-2b) leading to cancer death. We also asked patients the improvement in 5-year disease-free survival they would require to tolerate IFN. RESULTS: Utilities for melanoma recurrence with or without IFN alpha-2b were significantly lower than utilities for all IFN alpha-2b toxicities but were not significantly different from each other. At least half of the patients were willing to tolerate mild-moderate and severe IFN alpha-2b toxicity for 4% and 10% improvements, respectively, in 5-year disease-free survival. CONCLUSION: On average, patients rate quality of life with melanoma recurrence much lower than even severe IFN alpha-2b toxicity. These results suggest that recurrence-free survival is highly valued by patients. The utilities measured in our study can be applied directly to quality-of-life determinations in clinical trials of adjuvant IFN alpha-2b to measure the net benefit of therapy.

Lentigo maligna melanoma has no better prognosis than other types of melanoma.

We studied 48 patients with lentigo maligna melanoma (LMM) and compared the clinical stage I patients with non-LMM melanoma patients (matched by site and thickness) to see if prognosis differed. There was no significant difference in mortality from melanoma between the two groups (P = .68) after a mean follow-up time of five years (67.5 months for LMM, 60.5 months for non-LMM). In addition, a Cox multivariate analysis of the entire matched group showed that only thickness was significantly associated with death from melanoma (P = .0007) while histology (LMM v non-LMM) did not make a significant contribution (P = .61). Our data suggest that after accounting for primary tumor thickness and site, LMM and non-LMM have the same prognosis and biologic behavior, in contrast to the widely held belief that LMM has a better prognosis than other forms of melanoma.

Prognosis of clinical stage I melanoma patients with positive elective regional node dissection.

We tested 12 clinical and histologic variables to see which ones best predicted death from melanoma in 66 patients with positive elective regional node dissections (clinical stage I, pathologic stage II [CSI, PSII]). Despite the presence of lymph node metastases, not all patients had poor prognoses. Patients with tumors less than or equal to 3.5 mm and a percentage of positive nodes less than or equal to 20% had a 7-year survival rate of 66%. Within this low-risk group the subset with primary lesions on the trunk or extremities (except hands and feet) had a 7-year survival rate of 76%. This compares with poor 7-year survivals of 29% and 30% observed in other defined high-risk groups. Our results confirm and extend earlier observations concerning the prognoses of CSI, PSII melanoma patients and are relevant to any ongoing and future studies concerning elective regional node dissection (ERND) or adjuvant therapy trials in melanoma.

Cutaneous melanoma in Japan and the United States: comparative prognostic factors.

The ability to predict outcome for an individual patient with melanoma is important not only for the prognostic interests of the particular patient involved, but also for the design and accurate interpretation of surgical and adjuvant therapy studies. Comparison of prognostic studies between centers in the United States and Japan may result in understanding biologic differences between the disease as it manifests in whites and mongoloids. Studies from the Melanoma Clinical Cooperative Group (U.S.A.) and from Tohoku University (Japan) show some significant similarities and some differences. Overall five-year survival is quite different for clinical state I patients, 85% and 65%, respectively. However, when five-year survival for lesions on the hands and feet in the U.S. series is compared with the mainly acral melanomas in the Japanese series, the results show five-year survivals of 63% and 73%, respectively. Comparison of five-year survivals for clinical stage I acral lentigenous melanoma of the soles subdivided by thickness would establish whether data from the large series collected in the United States can be extrapolated to melanoma behavior in Japan.

Case-control study of melanoma and dietary vitamin D: implications for advocacy of sun protection and sunscreen use.

The rapid increase in melanoma incidence and mortality has given rise to nationwide and international campaigns that encourage the public to protect themselves from solar radiation with clothing, sunscreens, and other measures. The basis of these campaigns has been challenged by proponents of the theory that vitamin D, which is generated in the skin by ultraviolet B radiation, inhibits the development of melanoma. The present investigation tests this theory by examining the relation between dietary vitamin D and melanoma risk in a case-control study. Vitamin D intake was assessed by a food-frequency questionnaire in 165 melanoma patients and 209 controls. After controlling for age, hair color, and family history of melanoma, there was no association of melanoma risk with total vitamin D intake, calorie-adjusted vitamin D intake, vitamin D intake from foods, or consumption of milk or vitamin D supplements. We find no evidence to suggest that vitamin D protects against melanoma, and therefore continue to support the ongoing public health campaigns aimed at reducing sun exposure for the prevention of melanoma.

Increased intraepidermal melanocyte frequency and size in dysplastic melanocytic nevi and cutaneous melanoma. A comparative quantitative study of dysplastic melanocytic nevi, superficial spreading melanoma, nevocellular nevi, and solar lentigines.

Dysplastic melanocytic nevi (DMN) are distinguished histologically by a hyperplasia of variably atypical intraepidermal melanocytes in a lentiginous epidermal pattern. In order to further characterize the intraepidermal melanocytes of DMN, 4 representative specimens each of DMN, acquired nevocellular nevi (NCN), solar lentigines (SL), and superficial spreading melanoma (SSM) were selected on the basis of predetermined criteria, confirmed in a blind histologic assessment, and compared in a quantitative morphologic study using 6 micron-thick hematoxylin and eosin stained sections of L-dihydroxyphenylalanine (dopa) preincubated vertical tissue slices of lesion and adjacent normal skin. The average melanocyte frequency, expressed as the percent of dopa-reactive perikarya among 600 consecutive basal unit cells, was significantly greater in DMN (60 +/- 23%) than in NCN (18 +/- 3%), SL (25 +/- 7%), and adjacent skin (14 +/- 3%), but similar to that in SSM (71 +/- 11%). The average mean diameter of 200 consecutive epidermal basal unit melanocytes was significantly larger in DMN (11 +/- 2 microns) than in NCN (7 +/- 0.4 microns), SL (6 +/- 0.1 microns), and adjacent skin (6 +/- 0.4 microns), but significantly smaller than in SSM (16 +/- 3 microns). The observed similarities of intraepidermal melanocytes in selected DMN and SSM, as well as distinct differences from melanocytes in selected NCN and SL, support the hypothesis that some varieties of DMN may represent potential precursors of cutaneous melanoma.

Early death from clinical stage I melanoma.

We studied 13 prognostic factors in 582 patients with clinical stage I melanoma to determine which factor or combination of factors was associated with death from melanoma within the first 24 months following diagnosis. Thirty-six patients died during this period. Only 2 deaths occurred in patients with primary tumors thinner than 1.70 mm, and only 2 patients of 189 died with tumors located on the non-BANS extremities, excluding the hands and feet. Individual factors associated with high risk for death within 2 years included level V tumors, acral location, thickness greater than or equal to 3.65 mm, histologic ulceration greater than 3 mm, nodular type, presence of microscopic satellites, greater than 6 mitoses/mm2, positive elective node dissection, absence of lymphocyte response at the tumor base, and absence of an associated nevus histologically. Many of the preceding individual factors are highly correlated. By the use of logistic regression analysis, only one very high risk group was found: 71 percent of patients with level V tumors greater than 1.70 mm thick with histologic ulceration width greater than 3 mm located in an area other than the extremities (excluding hands and feet) had died within 2 years of diagnosis. The ability to select high-risk groups should be useful to investigators involved with the design and evaluation of adjuvant therapy studies.

Factors associated with death from melanoma from 2 to 5 years following diagnosis in clinical stage I patients.

We studied 14 prognostic factors in 428 patients with clinical stage I melanoma to determine which factor or combination of factors was associated with death from melanoma from 24 to 60 months following diagnosis. Forty-eight patients (11 percent) died during this period. All 17 patients who had visceral metastases present at 24 months died during this period. All surviving patients were followed for at least 60 months. Individual high risk factors included ulceration width (as determined by histology), level IV or V tumor, recurrence other than visceral, 6 or more mitoses per square millimeter, presence of involved nodes on elective dissection, absent or slight lymphocyte response, tumor type other than superficial spreading, location other than extremities (excluding hands and feet), microscopic satellites, thickness, sex, and wide local excision. The presence of sex as a risk factor for patients dying from 2 to 5 years following diagnosis is noteworthy because no sex difference was noted in the early death (less than 24 months) group. Age, presence of a nevus, and histologic regression were not significant factors. A logistic regression analysis selected a combination of the following independent factors: (1) location on extremities excluding hands and feet (favorable), (2) thickness, (3) recurrence other than visceral, (4) positive elective nodal dissection, (5) 6 or more mitoses per square millimeter, and (6) moderate to marked lymphocyte response (favorable). Twenty-five percent of patients with level IV lesions died between 24 and 60 months compared with only a 6 percent death rate within the first 24 months.

An evaluation of the effectiveness of azelaic acid as a depigmenting and chemotherapeutic agent.

In the past five years, it has been reported that certain dicarboxylic acids (C8-C13) and azelaic acid (C9) (AZA), in particular, have a remarkable effect in the management of lentigo maligna, human malignant melanoma, and certain disorders of hyperpigmentation. Preclinical trials, therefore, were undertaken in order to evaluate the effectiveness of AZA as a depigmenting agent and as a chemotherapeutic agent. Twenty-seven uniformly black pigmented guinea pigs were given topical applications of various concentrations (3, 5, 10, 15, and 20%) of AZA preparations for 8 weeks, and their effects on the melanocytes of epilated skin of the backs and the nonepilated ears of guinea pigs were compared to the effects of well-known depigmenting agents. Whereas 4-isopropylcatechol, monobenzylether of hydroquinone, monoethylether of hydroquinone, hydroquinone, and 4-hydroxyanisole were found to be selectively cytotoxic to melanocytes in black-skinned guinea pigs, AZA has little or no visually recognizable effect on melanocytes in these animals. The therapeutic effect of local s.c. injections of various concentrations of AZA preparations on the development of s.c. implanted B-16 melanoma tumor was evaluated in 96 C57BL/6J mice. In addition, 31 BDF1 mice, implanted i.p. with B-16 melanoma tumor, were used to assess the effect of 100-500 mg/kg concentrations of AZA administered i.p. In both studies, AZA revealed no significant tumoristatic or tumoricidal effect on the size, color, and growth of melanoma. The effect of AZA was also evaluated on S-91A (melanotic or pigmented) and S-91B (amelanotic) human melanoma cells in culture. Low concentrations (10(-5) and 10(-3) M) of AZA had no inhibitory effect on the growth of these cells. Only at higher concentrations (greater than 10(-3) M) was a cytotoxic effect on cell viability observed. These observations indicate AZA is not selectively cytotoxic to normal and proliferative melanocytes and has no apparent inhibitory effect on the formative process of melanin pigmentation.

Early death from clinical stage I melanoma.

We studied 13 prognostic factors in 582 patients with clinical stage I melanoma to determine which factor or combination of factors was associated with death from melanoma within the first 24 months following diagnosis. Thirty-six patients died during this period. Only 2 deaths occurred in patients with primary tumors thinner than 1.70 mm, and only 2 patients of 189 died with tumors located on the non-BANS extremities, excluding the hands and feet. Individual factors associated with high risk for death within 2 years included level V tumors, acral location, thickness greater than or equal to 3.65 mm, histologic ulceration greater than 3 mm, nodular type, presence of microscopic satellites, greater than 6 mitoses/mm(2), positive elective node dissection, absence of lymphocyte response at the tumor base, and absence of an associated nevus histologically. Many of the preceding individual factors are highly correlated. By the use of logistic regression analysis, only one very high risk group was found: 71 percent of patients with level V tumors greater than 1.70 mm thick with histologic ulceration width greater than 3 mm located in an area other than the extremities (excluding hands and feet) had died within 2 years of diagnosis. The ability to select high-risk groups should be useful to investigators involved with the design and evaluation of adjuvant therapy studies.

Factors associated with death from melanoma from 2 to 5 years following diagnosis in clinical stage I patients.

We studied 14 prognostic factors in 428 patients with clinical stage I melanoma to determine which factor or combination of factors was associated with death from melanoma from 24 to 60 months following diagnosis. Forty-eight patients (11 percent) died during this period. All 17 patients who had visceral metastases present at 24 months died during this period. All surviving patients were followed for at least 60 months. Individual high risk factors included ulceration width (as determined by histology), level IV or V tumor, recurrence other than visceral, 6 or more mitoses per square millimeter, presence of involved nodes on elective dissection, absent or slight lymphocyte response, tumor type other than superficial spreading, location other than extremities (excluding hands and feet), microscopic satellites, thickness, sex, and wide local excision. The presence of sex as a risk factor for patients dying from 2 to 5 years following diagnosis is noteworthy because no sex difference was noted in the early death (<24 months) group. Age, presence of a nevus, and histologic regression were not significant factors. A logistic regression analysis selected a combination of the following independent factors: (1) location on extremities excluding hands and feet (favorable), (2) thickness, (3) recurrence other than visceral, (4) positive elective nodal dissection, (5) 6 or more mitoses per square millimeter, and (6) moderate to marked lymphocyte response (favorable). Twenty-five percent of patients with level IV lesions died between 24 and 60 months compared with only a 6 percent death rate within the first 24 months.

Dermatologic manifestations of infections in immunocompromised patients.

Thirty-one immunocompromised patients (22 renal allograft recipients, 5 patients receiving chronic corticosteroid therapy, and 4 patients undergoing chemotherapy for acute leukemia) with significant dermatologic infection, excluding typical cellulitis and herpesvirus infections, were retrospectively identified over a 12-year period. Of these 31 patients, 15 (48%) had infection restricted to their skin, 6 (19%) appeared to have primary cutaneous infection that spread hematogenously to other parts of the body, 2 (6%) had infections of adjoining nasal tissue that spread to contiguous skin, and 8 (26%) appeared to have disseminated systemic infection that spread to the skin. In six of the eight patients with apparent secondary skin involvement, the development of the cutaneous lesion was the first clinical indication of disseminated infection. Eleven immunocompromised patients (35%) with bacterial infection of the skin or subcutaneous tissue were identified. These patients could be divided into three categories: leukemic patients with bacteremic gram-negative infection metastasizing to the skin (3 cases), renal transplant recipients with recurrent staphylococcal infection on and around the elbow ("transplant elbow") or streptococcal sepsis from a site of cellulitis (5 cases), and immunocompromised patients with opportunistic bacterial infection due to Nocardia asteroides or atypical mycobacteria (3 cases). Seventeen immunocompromised patients (55%) with fungal infection of the skin or subcutaneous tissue were identified. These included 12 patients with opportunistic fungal infection (Cryptococcus neoformans, 4 cases; Aspergillus species, 3 cases; Paecilomyces, 2 cases; Rhizopus species, 2 cases; and Candida tropicalis, 1 case) and 5 patients with extensive, confluent cutaneous dermatophyte infections. One patient with protothecosis and two patients with extensive papillomavirus infection were identified. Of these latter two cases, one had his immunosuppression discontinued, with clearing of his extensive warts; the other had confluent warts of the face and neck that subsequently underwent malignant degeneration to squamous cell carcinoma while chronic immunosuppressive therapy was continued.(ABSTRACT TRUNCATED AT 400 WORDS)

Prognosis in Stage 1 malignant melanoma: seven-year follow-up study of splenic radiocolloid uptake as predictor of death.

In an earlier study we found that patients with clinical Stage 1 and 2 cutaneous malignant melanoma and increased splenic radiocolloid uptake had more frequent recurrence at 24 mo, compared with melanoma patients having normal liver-spleen scintigrams. This report, an 80-mo follow-up study, gives further information on 119 clinical Stage 1 patients. Fifteen of 35 patients with increased splenic uptake (42.9%) died from melanoma as opposed to only 16 of 84 (19.1%) with normal liver-spleen images (p less than 0.01). Multivariate analysis showed that augmented splenic uptake of technetium-99m sulfur colloid is a marker for adverse prognosis in patients with malignant melanoma but does not appear to be an independent variable in predicting death. In clinical Stage 1 patients, increased splenic uptake correlated significantly with pathologic stage (positive elective node biopsy) as well as thickness and mitotic rate in patients with thicker lesions. It may be that patients with thicker, pathologically aggressive tumors have an increased splenic blood flow and/or enhanced immune and reticuloendothelial response (as manifested by abnormal liver-spleen scintigram). If so, the enhanced immune response does not appear to contribute to overall survival.

The significance of augmented radiocolloid uptake by the spleen in patients with malignant melanoma.

Increase in splenic uptake of Tc-99m sulfur colloid was noted in 47 of 147 (32%) patients with cutaneous malignant melanoma early in the coure of disease. Patients with disseminated disease and/or clinical or laboratory evidence of hapatic dysfunction were excluded from study. Recurrence rate of 2 yr was higher for those patients with splenic scans demonstrating augmented uptake compared with patients having normal scans, 36% against 16% (p less than 0.02). These differences resulted from a much more favorable prognosis in women with normal scans contrasted with women with increased uptake, 6% against 26% (p less than 0.05). Women with increased splenic uptake, and all men regardless of scan status, seemed to have a higher rate of recurrence than women with normal spleen scans. Scan status may be an adjunctive prognostic marker in women.

Estrogen and progesterone receptor analysis in pregnancy-associated melanoma: absence of immunohistochemically detectable hormone receptors.

The role of estrogen in the initiation and progression of melanoma remains unclear. Some findings that suggest a hormonal role in melanoma initiation or progression include the following: (1) melanomas arising during pregnancy are thicker than those in nonpregnant women, (2) pregnant women with stage II (regional nodal metastases) melanoma have a worse prognosis than nonpregnant women of similar stage, and (3) melanoma is rare prior to puberty. Although biochemical assays have shown that estrogen-binding proteins are present in malignant melanoma, studies using a sensitive and more specific immunohistochemical technique have not found estrogen receptors (ERs) in melanoma. In our laboratory an immunohistochemical technique using monoclonal antibody H222 can detect ER in tumors with receptor levels lower than 9 fmol/mg protein and detects ER in a variety of tissues and species. In addition, monoclonal antibody KD68 is used to detect progesterone receptors immunohistochemically. We studied 14 cases of pregnancy-associated melanoma. None of our cases, ranging from melanoma in situ to metastatic melanoma, showed positive nuclear staining for ER, nor did any of these cases show positive immunohistochemical staining for progesterone receptor. Despite the wide tissue and species distribution of ER detected by the monoclonal antibody H222, this immunohistochemical technique does not appear to be useful in the study of possible hormonal effects on the progression of malignant melanoma. The estrogen-binding proteins in melanoma detected by biochemical techniques in previous studies probably are distinct from the well-defined human ER.

The spectrum of minimal deviation melanoma: a clinicopathologic study of 21 cases.

A retrospective study of 21 patients with the histopathologic diagnosis of minimal deviation melanoma (MDM; n = 18) and borderline melanoma (BM; n = 3) was undertaken to determine the prognosis for these patients compared with that for patients with other types of malignant melanoma. The findings indicate that the prognosis for these uncommon nevomelanocytic tumors is somewhat better than that for other malignant melanomas. Follow-up periods in this series ranged from 18 to 96 months (mean, 57 months). Primary lesions ranged in thickness from 1.6 to 10.4 mm. The histopathologic subtypes included the Spitz variant (nine patients), the spindle cell variant (six patients), the combined spindle and epithelioid cell type (three patients), and the small epithelioid cell type (three patients). Only two of the patients died of widespread metastatic disease. Comparison of the histologic and clinical prognostic indicators of mortality in patients who have malignant melanoma with the clinical and pathologic features seen in this series of 21 patients would appear to indicate a diminished tendency toward metastatic or recurrent disease in patients with MDM and BM.