Melanocytic Nevi and Melanoma

Melanocytic Nevi and Melanoma

Community Perceptions About the Important Signs of Early Melanoma

Background.—The early detection of melanoma often relies on patients becoming concerned about a particular pigmented lesion. It is unclear what specific features the public views as being important. The importance that individuals place on various characteristics of skin lesions when looking for early signs of melanoma was explored.

Methods.—Residents of 60 rural communities in Queensland, Australia, were telephoned and asked to participate in an interview. A total of 1,148 individuals agreed to the interview, for a participation rate of 78%.

Findings.—A change in a lesion was considered the most important feature. Other factors considered important were the presence of more than 1 color, uneven edges, elevation, large size, and hairiness. Perceived levels of importance were influenced by age, sex, education, self-efficacy, perceived knowledge, and recent self-examination, but not by a recent skin examination by a family physician.

Melanocytic Nevi and MelanomaConclusions.—Guidelines specifically for the public may be necessary to enhance the self-examination skills of the community. In addition, family physicians should be encouraged to discuss important features of potentially cancerous lesions with their patients when examining their skin.

► Early detection is the best way to reduce melanoma mortality. However, there is concern that self-examination may not be as effective as it could be because lay persons are unclear as to what signs are most important in detecting early melanoma. In a telephone survey, the authors found that most respondents considered a change in a lesion or a mixture of colors to be the most important warning signs. Less importance was given to large size, elevation, and unevenness of the edges. The rank order of importance of these features, however, varied with different segments of the population interviewed. Because of the variability of responses, it is apparent that more specific guidelines for self-examination are necessary and that all opportunities to educate the public need to be fully utilized.

Family Physicians’ Knowledge of Malignant Melanoma

Objective.—The incidence and mortality of malignant melanoma have been increasing rapidly. Melanoma is detectable and curable in its early phases. Physicians must know not only the signs and symptoms of malignant melanoma but also the host factors associated with melanoma risk. Previous studies suggest that physicians need education regarding the assessment of pigmented skin lesions, particularly in the need for skin examination. Family practitioners were surveyed to determine their baseline knowledge of melanoma. Toward the development of a continuing education program, the physicians were also asked how they preferred to learn new information about the skin.

Methods.—A questionnaire was mailed to a random sample of 750 Canadian family physicians. The 4-section survey included questions about experience with various skin lesions, knowledge specific to malignant melanoma, opinions about various methods of continuing education regarding skin lesions, and data on medical practice.

Results.—The response rate was 47%. Fifty-six percent of respondents expressed a lack of confidence in their ability to recognize melanoma, compared with 30% for basal cell carcinoma. Fifty-eight percent doubted their ability to recognize squamous cell carcinoma or dysplastic nevi, and 78% said they would have trouble identifying early or subtle melanomas. When referring patients with suspicious pigmented lesions, the physicians were most likely to refer to a dermatologist. However, 61 % of respondents rarely if ever referred such patients. Two thirds said they would excise such lesions themselves, whereas nearly one third said they had never excised a suspect lesion. Ninety-seven percent of physicians knew that the pathology report of lesion thickness is a very important prognostic factor. However, most did not recognize that color variation within a mole is an important sign of melanoma. The respondents preferred a hands-on demonstration approach to learn new information about the skin.

Conclusions.—Although they lack confidence in their ability to recognize melanoma, most family physicians do not commonly refer patients with pigmented lesions to a dermatologist. These physicians may take on themselves the responsibility for excising or not treating pigmented lesions. Family physicians are well informed about the prognostic factors of melanoma, although they may lack key information about the history, physical examination, and risk factors. Family practitioners need education in the recognition of melanoma and patients at risk.

► Because many patients with significant risk factors for the development of melanoma may never see a dermatologist, it is important that primary care physicians learn to (1) distinguish between the various pigmented lesions; (2) evaluate the patient with suspected melanoma; (3) appropriately refer patients with suspicious pigmented lesions; (4) be knowledgeable regarding the risk factors for melanoma; and (5) educate and evaluate their patients accordingly. This will be even more important with the advent of managed care and the gatekeeper concept. In this questionnaire survey of family physicians in Canada, which had a 47% response rate, more than one half did not feel confident in diagnosing a melanoma or atypical mole. Interestingly, however, more than 60% of those responding never or infrequently referred their patients with pigmented lesions to a dermatologist. Clearly, primary care physicians need to be better educated regarding the recognition of melanoma and the importance of appropriately referring patients with pigmented lesions to a dermatologist.

The Independent Pathology Laboratory as a Reporting Source for Cutaneous Melanoma Incidence in Iowa, 1977-1994

Background.—Cutaneous melanoma is being diagnosed and treated more often in nonhospital settings because of changes in medical practice, including the emergence of diagnosis-related groups, health care reform, and a greater awareness of cutaneous melanoma. It is, therefore, possible that some cases of cutaneous melanoma may be missed by population-based registries. This could result in significant underestimations of the incidence of cutaneous melanoma. Studies in California, Connecticut, Massachusetts, and Washington have found that cases of cutaneous melanoma have been underreported by 12% to 21%. Causes include diagnosis or treatment in private physicians’ offices, out-of-state pathology laboratories, and nonhospital laboratories.

Methods.—Changing trends in the incidence of cutaneous melanoma were examined and case-finding sources—such as hospitals, clinics, hospital pathology laboratories, and independent pathology laboratories— were compared. A survey of dermatologists in Iowa was also conducted.

Results.—From 1977 to 1994, the incidence of invasive cutaneous melanoma increased 82% and the incidence of in situ cutaneous melanoma increased 900%. The proportion of cutaneous melanomas diagnosed in independent pathology laboratories increased to 25% of all cases. Underreporting of cutaneous melanoma was estimated at 10.4% to 17.1%, based on the survey of dermatologists.

Discussion.—The estimated underreporting of 10.4% to 17.1% is comparable to that of studies in other states. Population-based cancer registries must make greater attempts to identify and obtain pathology reports from nonhospital settings to improve the accuracy of surveillance. This may be achieved by direct reporting by physicians or through a nationwide reporting system from pathology laboratories to central cancer registries.

► The proportion of melanomas reported from hospitals/clinics dropped from 96.3% in 1977 to 81.4% in 1994, whereas the proportion of cases from independent laboratories increased from 1.3% to 15.2%. This reflects the change in practice habits over this period. There is an increasing tendency toward diagnosis of thin melanomas and, concomitantly, there is increased interest among dermatologists in surgical treatment of skin cancer. Dermatology practice is also unique in that many dermatologists send their specimens to dermatopathologists, including many large dermatopathology laboratories that process tens of thousands of specimens per year. Population-based cancer registries will underreport melanoma unless the ascertainment process includes independent laboratories. This study found an underreporting estimate of 10.4% to 17.1%, which is similar to the underreporting estimates obtained in other states.

Declining Effect of Latitude on Melanoma Mortality Rates in the United States: A Preliminary Study

Objective.—Recent reports suggest a slowing of increases in melanoma incidence and mortality, particularly in younger age groups. The more recent the age of birth, the smaller the increase in melanoma mortality and in the effect of latitude, suggesting a possible common factor. Data from the United States are consistent with a declining latitude gradient for melanoma. Mortality from melanoma was analyzed by geographic location over time.

Methods and Findings.—Semilogarithmic models fitted to state-specific mortality from melanoma and the latitudes of the states’ capital cities were used to analyze the gradient of melanoma mortality with latitude among U.S. whites. From the 1950s through 1992, the upward gradient of melanoma, from north to south, has been on the decline. The data suggested that, by the early 21st century, latitude will have no effect on melanoma mortality in the continental United States. Though melanoma mortality has risen over the years, the rate of increase has slowed. The pattern suggested that nationwide melanoma mortality may stabilize in the near future.

Conclusions.—The effects of latitude on U.S. melanoma mortality appear to be decreasing. Many social factors may affect the incidence and mortality of melanoma; the geographic mobility of the population has probably contributed to the decline in latitude gradients. However, it may also be that melanoma mortality has reached the point at which increased sunlight exposure has little additional effect.

► In recent years the lay public and medical profession have been bombarded with data on the rapid increase in the incidence of melanoma. However, this article suggests that in the United States the rate of increase may be slowing, and that by the second decade of the next century an actual decline may be seen. Moreover, because of the population’s mobility, the once-sharp contrast in melanoma mortality between northern and southern latitudes also appears to be declining.

Melanoma and Sun Exposure: An Overview of Published Studies

Background.—It is generally accepted that sun exposure is the most important cause of cutaneous melanoma in humans. However, the evidence for such an association is not very strong. There are some data to suggest that melanoma may be related specifically to intermittent sun exposure, rather than continued or chronic exposure. Data from case-control studies of the relationship between melanoma, sun exposure, and sunburn were reviewed.

Methods.—The analysis included 35 published case-control studies, including 9,121 incident cases of melanoma. Twenty-nine studies included information on sun exposure and 21 on sunburn. The data were analyzed to determine the association between melanoma; intermittent, occupational, and total sun exposure; and history of sunburn at various ages.

Results.—.Intermittent sun exposure was significantly and positively associated with melanoma, with an odds ratio (OR) of 1.71. Melanoma risk was significantly reduced in individuals with heavy occupational exposure (OR 0.86), and nonsignificantly increased for total exposure (OR 1.18). Melanoma was significantly associated with sunburn at all ages or in adulthood (OR 1.91). Comparable elevations in risk were noted for sunburn in adolescence (OR 1.73) or in childhood (OR 1.95). However, except for risk associated with sunburn at all ages or in adulthood, these risk estimates were variable.

Conclusions.—The available case-control data suggest that melanoma risk is specifically and positively associated with intermittent sun exposure. Intermittent exposure probably also accounts for the noted association with sunburn. Melanoma risk does not appear to be associated with any particular age at sunburn. The effects of sun exposure on melanoma risk are similar to those for basal cell skin cancer but different from those for squamous cell cancer. It is unclear how intermittent exposure increases melanoma risk whereas other types of exposure do not.

► Parallels continue to be drawn between melanoma and basal cell carcinoma. Both tumors appear to be associated with intermittent intense sun exposure, whereas cutaneous squamous cell carcinoma appears to be more often related to continuous (e.g., occupational) sun exposure.

Ultraviolet Radiation A-induced Precursors of Cutaneous Melanoma in Monodelphis domestica

Introduction.—The increased incidence of malignant melanoma in the white population cannot be attributed to increased exposure to ultraviolet radiation (UV)-B resulting from stratospheric ozone depletion. In populated areas, the amount of UV-B actually decreased from 1974 to 1985. One recent study suggests that sunscreen use, which may allow prolonged exposure without sunburn, allows increased exposure to UV-A. An animal study was conducted to examine the capacity of UV-A to induce melanocytic hyperplasia (MH), a melanoma precursor.

Methods.—Experiments were performed on 2 groups of 30 dorsally shaved opossums (Monodelphis domestica). The animals were exposed 3 times/week for 81 weeks to 250 J/m2 of UV radiation from FS40 sunlamps (a UV radiation spectrum rich in UV-B) or to 2.5 X 104 J/m2 of UV-A from filtered F40BLB fluorescent lamps. During the exposure period, animals were monitored for the appearance of nonmelanoma skin tumors (NMSTs) and MH.

Results.—Compared with animals exposed to UV-A, those exposed to UV-B had NMSTs significantly earlier in the experimental period. The prevalence of NMSTs was 71% in the UV-B group and 4% in the UV-A group (only 1 animal exposed to UV-A was scored as having a NMST). There was no significant difference between groups, however, in the prevalence of MH (31% and 22% for the UV-B and UV-A groups, respectively). Thus, UV-A was as effective as UV-B in the induction of MH, despite its considerably reduced ability to induce NMSTs.

Conclusion.—In this opossum model, UV-B radiation proved equally effective in the induction of both NMSTs and MH. At a dose 100 times greater than the UV-B dose, UV-A appears to be more effective at inducing MH than at inducing NMSTs.

► The author resurrects the argument that sunscreen use, by extending the time an individual can stay outdoors without getting sunburned, could result in enhanced exposure to UV-A. However, the applicability of the results of this study to human melanoma is uncertain.

The Intermediate Filament Peripherin is Expressed in Cutaneous Melanocytic Lesions

Introduction.—Peripherin is an intermediate filament that is present in cells derived from the neural tube and neural crest, including motor, sensory, and autonomic neurons. It has a molecular weight of 56,000 d and is encoded on chromosome 12. The function of peripherin is not known. The possible expression of peripherin in melanocytes and a variety of cutaneous melanocytic lesions was assessed.

Methods.—Immunohistochemistry was used to assess 74 cutaneous melanocytic lesions for expression of peripherin, including the following: primary invasive malignant melanoma (IMM); melanoma in situ (MIS); atypical nevus (AN); nevus with architectural disorder and cytologic atypia of melanocytes, spindle and epithelioid cell nevus, or Spitz nevus, (SN); blue nevus (BN); and common intradermal benign melanocytic nevus (BMN).

Results.—Peripherin was observed in cytoplasmic distribution within tumor cells in 14 of 14 IMM and in 8 of 10 MIS. Peripherin was localized to the intraepidermal and invasive dermal components in IMM. Labeling was throughout the lesions of 2 desmoplastic melanomas. Most tumor cells were labeled (+++) in one and only scattered cells (+) were labeled in the other. Peripherin was seen in all 10 AN and in 9 SN. Immunoreactivity was localized predominantly to the intraepidermal component, with only focal labeling in the upper portion of the dermal component of these 2 lesions. For all 15 BN, strong immunoreactivity was observed through-out the lesions. Almost all cells were labeled. The expression of peripherin was absent or limited to rare, scattered cells located in the superficial portion of the lesions in most BMN. Elongated cell processes in cutaneous nerves also were immunolabeled in the skin nearest the lesions. Immunoreactive cells were not seen in the epidermis or adnexa, so keratinocytes, normal resting adult melanocytes, Merkel cells, and Langerhans’ cells did not express peripherin.

Conclusion.—Expression of peripherin is frequently seen in benign and malignant melanocytic lesions, but it is absent in normal resting adult melanocytes. Only axons were labeled for peripherin in a limited sample of 3 schwannomas, suggesting that Schwann cells are not immunoreactive. Peripherin was expressed throughout the dermal component, including the deep portion in IMM and BN.

► It may be difficult to distinguish small melanoma cells, which may be present in the intradermal component of invasive melanoma, from remnants of a pre-existing melanocytic nevus. Peripherin is consistently expressed by melanoma cells, including those in the deepest portion of the lesion, but it is either absent or restricted to rare, scattered cells in the superficial portion of benign melanocytic nevi. Therefore, it may be useful in ascertaining the presence of a pre-existing nevus in doubtful cases and in accurately determining melanoma thickness.

Proto-Oncogene c-kit Expression in Malignant Melanoma: Protein Loss With Tumor Progression

Background.—The proto-oncogene c-kit—sometimes called stem cell factor receptor—is a transmembrane receptor tyrosine kinase that is related to the platelet-derived growth factor receptor family and is involved in hematopoiesis, gametogenesis, and melanogenesis. Previous studies have suggested that c-kit expression may be lost in melanoma. This immunohistochemical study evaluated the expression of c-kit in a series of melanocytic lesions.

Methods.—The study included 85 melanocytic lesions, including banal nevi, junctional and compound nevi with melanocytic dysplasia, nontumorigenic radial growth phase melanoma, tumorigenic vertical growth phase melanoma, and metastatic melanoma. Expression of c-kit in each lesion was evaluated by immunohistochemistry using a monoclonal antibody.

Results.—Normal melanocytes and mast cells showed intense membrane staining for c-kit. The junctional and superficial dermal components of compound nevi showed the strongest staining, although only weak expression was seen in dermal nevi. Strong staining was noted in dysplastic nevi, especially in junctional cells. Melanomas in radial growth phase showed the strongest reactivity, compared with weak-to-absent staining in vertical growth phase melanomas or metastatic lesions.

Conclusions.—Expression of the c-kit protein is noted in normal melanocytes, benign nevi, dysplastic nevi, and nontumorigenic melanomas. No c-kit expression is observed in tumorigenic primary melanomas or metastatic melanomas. More study is needed to evaluate the significance of loss of c-kit in advanced melanomas.

► The decrease (but not always absence) of the c-kit protein with tumor progression suggests that this proto-oncogene might act as a tumor suppressor gene. High concentrations might prevent development of vertical growth phase lesions and inhibit melanoma tumor progression, whereas low concentrations might act as a regulator of cellular adhesion in dermal stroma and promote vertical growth phase lesion development. The observation of a decrease in protein expression with tumor progression may shed light on the biology of invasive melanoma.

Expression of lnterleukin-8 by Human Melanoma Cells Up-regulates MMP-2 Activity and Increases Tumor Growth and Metastasis

Purpose.—Previous studies have shown that the level of interleukin-8 (IL-8) expression by melanoma cells is related to their metastatic potential. When primary cutaneous melanoma cells are subjected to ultraviolet B radiation, the result is increased tumorigenicity and metastatic potential via induction of IL-8. Interleukin-8 is a known angiogenic factor; however the effects of increased IL-8 production by melanoma cells and its role in the metastatic process are unclear. The role of IL-8 in the growth and metastasis of human melanoma cells was studied.

Methods and Results.—Experiments were performed with transfected nonmetastatic SB-2 melanoma cells, which expressed very low levels of IL-8. In response to transfection with IL-8 cDNA, the cells showed greatly enhanced tumorigenic and metastatic potential, compared with parental and control-transfected cells. After IL-8 transfection, the cells showed upregulation of Mr 72,000 collagenase type IV (MMP-2) mRNA and collagenase activity. They also demonstrated increased invasiveness in experiments using Matrigel-coated filters. Linking the MMP-2 promoter upstream of the chloramphenicol acetyltransferase (CAT) reporter gene led to upregulation of CAT activity in IL-8 cells, but not in control-transfected cells.

Conclusions.—Enforced expression of IL-8 in nonmetastatic melanoma cells leads to increased tumorigenic and metastatic potential. The results suggest that IL-8 is directly involved in melanoma metastasis. This cytokine may activate type IV collagenase, leading to enhanced invasion of the host stroma, increased angiogenesis, and metastasis. Interleukin-8 also seems to be involved in MMP-2 gene transcription.

► Proper vascularization and an ability to degrade type IV collagen enhance the metastatic potential of tumor cells. Thus, the collagenase enzyme may play a critical role in the local spread and metastasis of several malignant tumors, including melanoma. Luca et al. demonstrate that IL-8 increases the tumorigenic and metastatic potential of melanoma cells in nude mice, providing direct evidence for the involvement of IL-8 in metastasis. They suggest that IL-8 may exert its angiogenic effect through the induction of type IV collagenase in tumor cells.

Increase of Melanocytic Nevus Counts in Children During 5 Years of Follow-up and Analysis of Associated Factors

Background.—Given the significance of the number of acquired melanocytic nevi as a risk factor for cutaneous melanoma, it is important to identify factors associated with the development of such nevi. There are questions about the apparent causative factors, such as pigment phenotype and sun exposure. These factors were evaluated in a longitudinal study of the increase of melanocytic nevi in children.

Methods.—Nevus counts were performed in 866 children younger than 7 years, and repeated 5 years later in 377 children. The analysis included data on 357 children who underwent 2 examinations. The examinations were performed in standardized fashion, and revealed the numbers and locations of melanocytic nevi. A parent questionnaire was used to gather information on potential risk factors, including phenotypic traits and sun exposure.

Results.—At the first examination, the mean number of moles measuring 1 mm or larger was 9, while the number of moles measuring 2 mm or larger was 4. At repeat examination, the children had a mean of 40 moles measuring 1 mm or larger and 16 moles measuring 2 mm or larger. Large increases in the number of nevi were significantly related to fair skin color, freckles, sun exposure, and sunscreen use. On multivariate analysis, the number of nevi measuring 1 mm or larger was independently related to freckles, skin type, and days of intensive sun exposure. The freckling tendency doubled the relative risk of increased nevi. For children with more than 21 days of intensive sun exposure per year, the relative risk of melanocytic nevi was 1.6. The same factors were independently related to the number of nevi measuring 2 mm or larger.

Conclusions.—This study documents the increasing number of melanocytic nevi during the early school-age years. The childhood development of nevi is significantly related to pigmentation factors associated with poor sun tolerance, including freckling and the tendency to sunburn. The amount of recreational sun exposure also appears to play a significant role.

► Controversy exists regarding the relationships among Fitzpatrick skin type, sun exposure, and the number of moles a person develops. Because the number of moles strongly correlates with the tendency to develop melanoma, further investigation might indicate a link. In this 5-year longitudinal study of children, the authors found more moles in fair-complected children who were freckled, had suffered more than 3 sunburns, and who regularly experienced intense sun exposure. However, there were also some puzzling and thought-provoking findings. For example, regular use of a sunscreen correlated with an increased number of nevi, and in children with nevi more than 2 mm in size, intense sun exposure correlated with a decrease in the number of nevi. Although the authors explain these latter findings by citing the development of a tan with chronic sun exposure and noting the inadequacies of sunscreens, are other explanations possible or was there a deficiency in the methodology of the study?

A Study of Large Congenital Melanocytic Nevi and Associated Malignant Melanomas: Review of Cases in the New York University Registry and the World Literature

Introduction.—Earlier reports indicate that patients with large congenital melanocytic nevi (LCMN) are at increased risk for having malignant melanoma (MM). Data from the New York University (NYU)-LCMN Registry were reviewed to identify the clinical features of patients with LCMN in whom MM actually developed.

Methods.—Records from 117 patients with LCMN from the NYU-LCMN registry were compared with those of 177 patients with LCMN in the world literature. Clinical features and patient characteristics were analyzed for patients from the NYU-LCMN Registry and for the case reports from the world literature.

Results.—Ninety-one percent of LCMN were located on the trunk. Of 289 patients with LCMN evaluated, 3 from the NYU-LCMN Registry and 31 from the world literature had primary MMs that arose from within the LCMN. Malignant melanomas developed at cutaneous sites other than within the nevi in 2 additional patients. Nevi were in axial locations in all patients with LCMN in whom MM developed. No MM arose from any of the 26 LCMN confined to the extremities. No MM arose from thousands of satellite nevi.

Conclusion.—These findings are in agreement with others regarding increased risk of MM in patients with LCMN. No patients with LCMN of the extremity had MM, which suggests that patients in this cohort are at decreased risk. The absence of MM arising in satellite nevi was an extraordinary finding.

► The take-home message of this retrospective study lies in the observation that all patients who had melanoma within LCMN had nevi in axial locations. Melanoma was not found in any such lesions confined to the extremities. Nevertheless, melanoma in medium-sized extremity congenital melanocytic nevi has been reported; thus, the risk, although small, is not nil. Also noteworthy in this report is the observation that no melanomas were found in the thousands of satellite nevi.

MR Imaging of Symptomatic Neurocutaneous Melanosis in Children

Introduction.—Neurocutaneous melanosis is a rare and lethal disorder in children characterized by multiple or diffuse congenital pigmented nevi of the skin and melanocytosis of die leptomeninges of the CNS. The CNS MR images of 5 children with neurocutaneous melanosis were reviewed and findings were compared with those in the literature.

Methods.—Age range of 5 children with neurocutaneous melanosis was 7-10 years. All children had multiple pigmented cutaneous nevi, seizures, signs of raised intracranial pressure, cranial nerve palsies, or myelopathy. Patients underwent MR studies with T1-weighted, T2-weighted, and T1- weighted-postgadolinium images of the brain and T1-weighted post-gadolinium images of the entire spine.

Results.—In all children, MR imaging showed marked, diffuse enhancement of thickened leptomeninges surrounding the brain and spinal cord. This was observed only on the postgadolinium T1-weighted images. Mild to moderate hydrocephalus was also observed.

Conclusion.—A review of literature confirmed the findings in this patient series of leptomeningeal involvement, demonstrating either marked enhancement or hyperintense nonenhancement. Leptomeningeal enhancement is common in patients with neurocutaneous melanosis and should be considered characteristic of this disorder particularly in significantly symptomatic children.

► When neurocutaneous melanosis presents with CNS signs, the prognosis is dismal. In all 5 symptomatic children in this report, the MR study demonstrated dilatation of ventricles; however, only on postgadolinium studies were the thickened leptomeninges demonstrated. Each of these children died within 1 month to 2 years of diagnosis. The authors stress that these MR findings do not differentiate between benign and malignant leptomeningeal melanosis. Enhancement of the leptomeninges may be seen in chronic meningitis and metastatic spread from both primary brain tumors and from carcinomas located outside the CNS. Whether performance of this costly examination produces significant clinical benefit in this unfortunate group of children remains to be clarified.

Epiluminescence Microscopy of Small Pigmented Skin Lesions: Short-term Formal Training Improves the Diagnostic Performance of Dermatologists

Introduction.—Early recognition and surgical excision of thin lesions is the most effective management of malignant melanoma. There are a number of small pigmented skin lesions in which it is difficult to distinguish between the benign and the malignant with the naked eye, and in specialized centers, the diagnostic accuracy is only 60%. A noninvasive clinical technique for diagnosing pigmented skin lesions is epiluminescence microscopy. Interested clinicians can now use small hand-held epiluminescence devices, which are available for routine use. It was determined whether a short epiluminescence microscopy training course would have any effect on the diagnostic performance of untrained dermatologists.

Methods.—One hundred pigmented skin lesions were photographed with (epiluminescence microscopy) and without oil immersion (surface microscopy). None of the lesions was larger than 1 cm. Before and 24 hours after a 9-hour teaching course, tests were performed. The test consisted of showing 100 images, in random order, of the surface microscopy or epiluminescence microscopy slides. The 11 participating interpreters determined whether the results were definitely or almost definitely not melanoma, probably not melanoma, possibly not melanoma, probably melanoma, or definitely or almost definitely melanoma.

Results.—In 8 of 11 testees, the diagnostic accuracy was decreased rather than increased in those who used epiluminescence microscopy without training. The diagnostic performances of the testees were significantly enhanced with training, with an average gain of 8.4% after 9 hours of formal training in epiluminescence microscopy.

Conclusion.—For the useful application of epiluminescence microscopy, formal training is required. With formal training, the diagnostic accuracy with epiluminescence microscopy was significantly better than with surface microscopy.

► Epiluminescence microscopy has been said to increase diagnostic accuracy when examining pigmented skin lesions. However, as suspected and confirmed by this article, considerable experience and training in this technique are required to make it a worthwhile adjunct.

Persistent Nevus: An Exception to the ABCD Rule of Dermoscopy

Introduction.—The ABCD rule of dermoscopy is helpful in differentiating benign melanocyte neoplasms from melanoma. False positive scores for benign melanocytic lesions have been reported. This false positive score may be observed in melanocytic nevi with a globular pattern; melanocytic nevi with a papillary, lentiginous, or congenital component; and in pigmented spindle-cell or Spitz nevi. Another melanocytic neoplasm that may have a false positive ABCD dermoscopy score was identified.

Case Report.—Female, 18, with a histologically diagnosed compound melanocytic nevus reported her “mole was coming back” 5 months after saucerization. Recurrent nevus or pseudomelanoma was diagnosed. The original nevus had no congenital melanocytic nevus features. Two examiners diagnosed the lesion as melanoma on dermoscopy. The total scores assigned by the 2 examiners were 7.9 and 7.0, respectively (4.8-5.45 is suspect for melanoma). The lesion was excised to fat with a 2-mm margin of normal skin. The specimen was step-sectioned every 2 millimeters. Pathologic examination showed nested melanocytes within the epidermis above a scar and nested melanocytes in the dermis. The histopathologic diagnosis was recurrent nevus.

Conclusion.—Persistent melanocytic nevus may be added to the list of melanocytic lesions that can give a false positive ABCD score on dermoscopy.

► Epiluminescence microscopy of pigmented skin lesions has proved to increase the diagnostic accuracy of clinicians experienced in the use of the dermatoscope. Every rule has its exceptions as does the ABCD rule of dermoscopy. In this article, Marghoob and Kopf independently came to the same conclusion that a persistent melanocytic nevus (pseudomelanoma or recurrent nevus) can give a false positive score. Other pigmented lesions that may give a false positive score include melanocytic nevi with a papillary, lentiginous, or congenital component and pigmented spindle-cell or Spitz nevi. As collective experience with dermoscopy grows, we can be sure that more conditions will be added to this list.

Nonpigmented Dysplastic Melanocytic Nevi

Introduction.—Early recognition and treatment of dysplastic melanocytic nevi (DMN) can help to reduce the risk of transformation to invasive melanoma. Clinical criteria for DMN include irregular perimeter, size greater than 5 mm, background erythema, and variegated color, while the histologic criteria include basilar melanocytic proliferation with nuclear atypia, a patchy lymphocytic infiltrate with concentric eosinophilic fibroplasia, and lamellar fibroplasia. There have been no reported cases of nonpigmented DMN. A patient with multiple, nonpigmented papules that proved to have the histologic features of dysplasia was described.

Case.—Woman, 31, was evaluated for increasing number of small, pigmented lesions. The patient was white, with brown hair and brown eyes; she had no history or family history of DMN or melanoma. Examination showed about 100 discrete, nonindurated, nonpigmented macules and slightly elevated papules, associated with slightly accentuated skin markings. The lesions, which ranged in size from 2 to 5 mm, were found in a symmetric, primarily truncal distribution. Biopsy specimens revealed proliferating melanocytic cells with enlarged or clustered hyperchromatic nuclei within a lentiginous epidermis. Although concentrated along the basilar layer, the cells were observed up to the midepidermal level. The papillary dermis showed mild fibroplasia with a round cell- lymphohistiocytic infiltrate. Strong positivity for S100 protein was noted for cells within the epidermis, along with focal positivity for HMB-45 and negativity for leukocyte common antigen. The basilar keratinocytes of lesional epidermis showed a few black pigment granules, whereas the uninvolved dermis showed many granules. The surprising findings were confirmed in biopsy specimens from 3 clinically similar lesions.

Discussion.—The first case report of nonpigmented DMN was presented. This patient had many small, nonpigmented, slightly elevated papules that had the microscopic appearance of nevi with histologic dysplasia. The risk of transformation to malignant melanoma in this case is difficult to assess; however, the guidelines for melanoma risk in pigmented DMN should probably apply. Nonpigmented DMN may be a clinical variant of pigmented DMN, and should be considered in the differential diagnosis of hypopigmented lesions.

► It is well established that atypical or dysplastic melanocytic nevi serve as markers for a predisposition to the development of melanoma. Our visual concept of a DMN is a large, irregularly shaped lesion with variegation of pigmentation that at times is clinically indistinguishable from a melanoma. In this case report of a young woman with multiple hypopigmented lesions, the authors found on biopsy specimens histopathologic findings consistent with DMN of the junctional type with poor pigment production. The patient also had scattered pigmented nevi, but apparently biopsies were not performed on these lesions. The authors could not explain their observations. More importantly, the prognostic significance of these lesions was unclear.

Histopathology of Solar Lentigines of the Face: A Quantitative Study

Background.—Solar lentigines, pigmented lesions that occur on sun-damaged skin, generally have both keratinocytic and melanocytic hyperplasia but not melanocytic atypia or nests of melanocytes. Most histopathologic studies of these lesions have reported on sites other than the face and have not assessed the diagnostic features systematically and quantitatively.

Methods.—Fifty-one solar lentigines from 51 patients were assessed retrospectively. The hematoxylin-eosin, Mel-5 immunoperoxidase, and Fontana-Masson stains were reviewed to quantify melanocytes, the melanin content of the epidermis, and the epidermal area using a computer-assisted image analysis program.

Findings.—Frequently, lentigines from the face had a flattened epidermis. Compared with normal photoexposed facial skin, these lesions had an average 2.1-fold increase in melanocytes, a 2.1-fold increase in epidermal region, and a 2.2-fold increase in epidermal melanin content.

Conclusion.—The epidermal area and the number of melanocytes in facial solar lentigines were significantly increased compared with facial skin with a similar degree of photodamage. However, facial lentigines often lacked the rete-ridge hyperplasia that is typically associated with lentigines from other anatomical sites. These findings may be useful in the assessment of biopsy specimens from pigmented lesions on photodamaged facial skin.

The Starburst Giant Cell Is useful for Distinguishing Lentigo Maligna From photodamaged Skin

Introduction.—It may be difficult to differentiate atypical melanocytes in lentigo maligna (LM) from the pleomorphic atypical melanocytes in skin subjected to long-term sun exposure. The likelihood of finding multinucleated melanocytes in LM and determining their usefulness in distinguishing chronically sun-exposed skin from LM were evaluated.

Methods.—The tips of the ellipses of biopsy specimens from 89 patients with LM and 107 elliptical excisions of basal cell carcinoma or squamous cell carcinoma were reviewed for the presence of multinucleated melanocytes. When present, the maximum number of nuclei were recorded.

Results.—The most common location of lesions in patients with LM and controls was the face (74% for patients with LM and 73% for controls). Patients whose specimens contained multinucleated melanocytes were significantly younger than patients without them. Multinucleated melanocytes were observed in 85% of LM specimens and in 21% of control specimens. The multinucleated melanocytes had prominent dendritic processes that gave them the appearance of starburst giant cells (SGCs). There were numerous SGCs along the basal cell layer in some LM specimens and, occasionally, within the follicular epithelium and the dermis. In LM specimens, the maximum number of nuclei per SGC ranged from 2 to 30, whereas the number ranged from 2 to 6, in control specimens. When SGCs with more than 2 nuclei were considered, 76% of LM specimens and 8% of control specimens had SGCs. These figures were 64% for LM specimens and 3% for control specimens when there were more than 3 nuclei.

Conclusion.—Starburst giant cells may be considered a sensitive and specific marker for LM. As the number of nuclei in the SGC rises, the diagnosis of LM is more likely.

► Andersen et al. make the important observation that it can be difficult to differentiate solar lentigines arising on severely photodamaged skin from the diffuse melanocytic hyperplasia of evolving lentigo maligna. They ask, “What degree of melanocytic hyperplasia is required for a diagnosis of lentigo maligna?” Pleomorphic atypical melanocytes can be seen both in severely photodamaged skin and in lentigo maligna. This may cause problems in the evaluation of the margins of excision of such lesions, as the lentigo maligna and the surrounding sun-damaged skin may share similar histopathologic characteristics. Cohen suggests that multinucleated melanocytes with a “starburst” appearance may help in this distinction. Unfortunately, her study was not a blind one; a blind study would help confirm the reliability of this criterion. Furthermore, as stated by the author, there is no absolute standard for diagnosing lentigo maligna—different dermatopathologists require various criteria to make the diagnosis.

From Melanotic to Amelanotic Lentigo Maligna: An Aggressive Variant Presenting as an Inflammatory Lesion

Introduction.—The usual picture of lentigo maligna is a slowly growing lesion of irregular pigmentation on sun-exposed skin in an elderly patient. Very few amelanotic lesions are seen. In only 2 cases has a nonpigmented tumor developed from a primarily pigmented lesion. Nonpigmented lesions can be misdiagnosed as benign inflammatory disorders. A case of aggressive amelanotic lentigo maligna developing from pigmented lentigo maligna was reported.

Case.—Woman, 79, had a 4-cm erythematous macule of the left cheek, which had been slowly growing over the past year. The lesion had irregular borders with changing discoloration and slight scaling. The patient had undergone repeated excisions for pigmented lentigo maligna, the first 17 years ago. The new lesion was located beside and within the area of a previous skin transplant. A biopsy specimen was obtained, confirming the histologic diagnosis of amelanotic lentigo maligna. Immunohistochemical studies showed expression of S-100 protein, HMB45, and NK/C3 antibodies, confirming that the tumor cells were melanocytic. However, Fontana stain was negative for melanin pigment. The initial lentigo maligna specimen from 17 years ago showed deposits of melanin pigment in melanophages and pigmented tumor cells, with a lichenoid and perifollicular lymphocytic infiltrate and some macrophages. The new biopsy specimen showed no melanin pigment. The tumor cells had a pagetoid infiltrating pattern in the epidermal layers. Both specimens showed a similar inflammatory infiltrate, with no regression. Upper dermal solar elastosis had greatly progressed over the years.

Discussion.—This is only the third known case of amelanotic lentigo maligna developing from an earlier pigmented lesion. However it arises, the diagnosis of amelanotic lentigo maligna may be difficult to make. This case underscores the need for histologic examination of persistent areas of erythema in sun-damaged skin of elderly patients.

► Amelanotic lentigo maligna can develop de novo or may evolve from a pigmented lentigo maligna. Amelanotic lentigo maligna can also be seen after inadequate cryosurgical removal of a pigmented lentigo maligna. Such lesions are extremely difficult to evaluate and, as in the case reported, often are misdiagnosed as an inflammatory process.

Usefulness of the Staged Excision for Lentigo Maligna and Lentigo Maligna Melanoma: The “Square” Procedure

Objective.—Lentigo maligna (LM) can progress to invasive lentigo maligna melanoma (LMM). Reported LM recurrence rates of 20% to 100% 5 years after superficial destruction result from failure to destroy the atypical junctional melanocytes that extend into the adnexal epithelium. Removal of LM and LMM from the head and neck area presents problems because of the atypical junctional melanocyte hyperplasia (AJMH) that extends beyond the visible pigmented and Wood’s lamp margins. Therefore, a variety of surgical techniques have been proposed for tissue sparing and to prevent local recurrences. A simple alternative method that makes use of staged permanent peripheral vertical section margin control is discussed.

Technique.—The excision site is geometrically outlined, leaving a surgical-margin of 0.5-1.0 cm that is excised with a 2-bladed knife, having 2-4 1-mm spacers, to the underlying adipose tissue. The strip of tissue corresponding to the margin is labeled for orientation and sectioned for pathologic examination. The procedure is repeated until all margins are free of AJMH; the lesion is then removed. Primary closure or flap reconstruction is performed at the time of excision. For LMM, excision of the invasive component is performed simultaneously with the first margin excision to lessen the risk of metastasis.

Discussion.—Whereas melanoma represents a small number of cancers, it is responsible for 75% of deaths from skin cancer with the LMM type requiring wide margin excisions. Wide surgical excision provides the highest rate of local control. The staged excision is a simple way to increase local control.

Conclusion.—No local recurrence has been observed in the 35 patients treated with the simple and quick stage technique, but a 5- to 10-year follow-up period will be required before long-term local control rates can be determined.

► The significant recurrence rate of lentigo maligna and lentigo maligna melanoma after routine surgical excision may result from unsuspected sub-clinical extension and the inadequacy of the routine histopathologic process, which allows for examination of only a limited portion of the edges (margins) of the excised specimen. Recent reports have suggested that Mohs micro-graphic surgery, with its superior histologic control, might offer a superior cure rate. However, Mohs surgery is not always readily available. In an effort to offer better histologic control but at the same time not necessitate the help of a Mohs surgeon, the authors have developed a combined surgical-pathologic technique that uses permanent sections and a double-bladed scalpel. With this technique the peripheral margins are rendered tumor free before the remaining central island of tissue is removed and pathologically examined. The authors suggest that excising the bulk of the tissue last facilitates the delayed repair of the final wound. Although this innovative approach to the surgical removal of lentigo maligna and lentigo maligna melanoma may improve histologic control and circumvent the lack of availability of Mohs surgery, it has several potential drawbacks. Bread-loafing the central island of tissue, which is excised last, does not offer the same 100% margin control as would taking horizontal sections from the undersurface of the island of tissue, as is done with Mohs surgery. Moreover, excising the central island of tissue only after the lateral margins are tumor free results in additional surgery and a further delay in reconstruction.

Malignant Melanoma Masquerading as Malignant Fibrous Histiocytoma

Introduction.—There are numerous histologic variants of malignant melanoma (MM), the most common of which are the superficial spreading, lentigo maligna, nodular and acral lentiginous types. The 2 cases presented here illustrate a less common variant of MM, with multiple giant cells resembling superficial malignant fibrous histiocytoma (MFH).

Case Report 1.—White man, 69, was seen for an ulcerated tumor on the sole of his right foot. A spindle cell tumor reported to be S-100 negative was identified at shave biopsy. At a subsequent biopsy the tumor was found to have atypical pleomorphic multinucleated giant cells, similar to the appearance of MFH. Lentiginous spread of melanoma was apparent along the dermal-epidermal junction lateral to the tumor nodule. The tumor was interpreted as MFH because of negative S-100 staining and the presence of giant cells. But after a review of the slides, together with immunohistochemical studies, the diagnosis of MM was made. The tumor’s epithelioid, spindle, and multinucleated cells were S-100 positive, CD68 negative, and weakly vimentin positive.

Case Report 2.—The second patient had a histologically similar tumor that was originally interpreted as a severely dyspiastic junctional nevus. Subsequent review of the initial biopsy and a repeated biopsy confirmed the diagnosis of MM.

Discussion.—The histologic pattern in some MM can resemble MFH, but few MFH are S-100 positive. Common ultrastructural features in this type of MM are lysosomal-dense bodies, intertwining cell processes, and intermediate cell junctions.

► Malignant melanoma may present with numerous histologic variants, including patterns that resemble malignant fibrous histiocytomas. Both tumors may be composed of spindle cells containing atypical and pleomorphic multinucleated giant cells. Helm reviews the assertion in the literature that the presence of lentiginous and pagetoid spread of melanocytes within the epidermis is the most reliable distinguishing factor. Immunohistochemical and electromicroscopic studies demonstrate that the majority of melanomas are S-100 or homatropine methobromide (HMB)-45 positive as well as vimentin positive. Rare melanomas, however, may also be α-1-antitrypsin positive (a presumptive marker of malignant fibrous histiocytomas.) Conversely, rare malignant fibrous histiocytomas may be S-100 positive but not homatropine methobromide (HMB)-45 positive. Electromicroscopically, only 28% of desmoplastic melanomas demonstrate the presence of premelanosomes or melanosomes, thus leading to further difficulty in interpreting these cases. Lysosomal-dense bodies, intertwining cellular processes, and intermediate cell junctions appear to be better ultrastructural markers of melanoma.

Malignant Mucosal Melanoma of the Head and Neck: Review of the Literature and Report of 14 Patients

Introduction.—Mucosal melanoma of the head and neck is a rare but highly malignant tumor with poor long-term survival rates. The most lethal forms of this disease are those that occur in the sinuses, followed by those that occur in the oral cavity, pharyngeal, and intranasal areas. Local failure is common in treated patients. The adequacy of margins is hard to evaluate in this anatomic area. The total number of reported cases of malignant mucosal melanoma of the head and neck is about 1,000 since it was first described in 1856. Fourteen patients with characteristics similar to those described in the literature are presented.

Findings.—Age range of 14 patients with a diagnosis of primary mucosal melanoma of the head and neck was 52-85 years (mean, 70.7 years). Thirteen melanomas were of sinonasal origin and one was a lip melanoma. Twelve of 14 patients were initially evaluated for symptoms related to the nose. Three patients with obstruction, epistaxis, and proptosis on initial evaluation were dead of disease at an average of 6 months. Of 7 patients who died of disease, 6 died of distant metastases, and 1 died of advanced local disease. Five-year survival was 14.3%. Of 3 patients who underwent positron emission tomography (PET), distant osseous metastases were detected in 1, no distant metastases were detected in another who remained disease-free at final follow-up, and 1 patient had a false-positive PET scan (confirmed by computed tomography and biopsy). The mean follow-up was 31.4 months (range, 4 months to 13 years). Six patients (42.9%) had local recurrence at a range of 7 months to 7 years. The only 2 patients with 5-year survival had local recurrences. Ten patients (71%) underwent full surgical treatment with full macroscopic resection. Three patients underwent radiotherapy.

Conclusion.—It is not known whether modern surgical techniques have impacted survival or local control. It is likely that whole body PET and radiotherapy will grow in importance in diagnosing and treating malignant mucosal melanoma of the head and neck.

► The screening examination for malignant melanoma should include not only examination of the entire skin surface but also an inspection of the oral and nasal mucosa and external genitalia. As demonstrated by this study, malignant mucosal melanoma of the head and neck has a poor prognosis, with an overall 5-year survival rate of 17.1%. Over 50% of patients experience local treatment failure. These results can most likely be attributed to lack of an early diagnosis. Most oral melanomas develop on either the maxillary gingiva or the hard palate. These areas should be routinely examined whenever a patient is screened for melanoma.

Multiple Primary Melanomas: Implications for Screening and Follow-up Programs for Melanoma

Purpose.—The incidence of melanoma is increasing, suggesting the need to identify high-risk patients as candidates for screening programs. The risk factors, including sun exposure, skin type, melanocytic nevi, freckling, and family history, are readily indentifiable by the public. For patients with 1 primary site, the risk of a second primary site is as much as 12.5 times higher than in subjects with no previous history of melanoma. Personal history of melanoma was evaluated for use as a single risk factor in screening for melanoma.

Methods.—The investigators performed an 8-year review of 2,600 consecutive patients with melanoma. One hundred one patients (4.3%) were identified as having a second primary melanoma: 67 synchronous (at presentation or within 2 months) and 44 metachronous (more than 2 months after diagnosis). Seventy-nine percent of the patients were men. The study examined the effects of the Moffitt/University of South Florida Cutaneous Oncology Program follow-up protocol, an intensive though relatively routine protocol, on the stage at diagnosis of second primary melanomas.

Results.—Mean tumor thickness at diagnosis was 1.72 mm for the initial melanomas and 0.58 mm for the second melanomas. For metachronous melanomas, mean thickness was 2.27 mm for the first tumors vs. 0.90 mm for the second tumors. The initial melanomas were thicker by 3.8 mm, on average. Nine percent of the initial melanomas were Clark level I, compared with 36% of the second melanomas. The percentage of potentially curable Clark level I and II melanomas increased from 34% to 59%, and the percentage of ulcerated lesions decreased from 23% to 2%. Ninety-three percent of the second primary melanomas were recognized by the attending physician.

Conclusions.—Intensive screening for melanoma appears to be highly beneficial in patients with a personal history of the disease. The follow-up protocol used in this study significantly reduced tumor thickness, the key prognostic factor for stage 1 and 2 melanomas, in subsequent melanomas. Because patients with previous melanoma have a high relative risk of second primary tumors, this group should be screened initially at the time of diagnosis, then followed up frequently to detect curable second primary tumors. Most second primary melanomas are detected by the attending physician.

► In order to further reduce the mortality due to malignant melanoma, it will be important to detect these lesions early while they are still thin. This can be accomplished in part through public education and the education of primary care physicians. Although mass screenings will yield a few patients with melanoma, for the effort expended the return is small. The screening of patients at high risk for developing melanoma is more cost-effective and yields greater returns. In this study the investigators found that 4.3% of their patients with melanoma developed a second melanoma. Because of close follow-up, these second lesions were detected at an early stage. Despite patient education and emphasis on self-examination, only 7% of these second primaries were detected by the patients themselves. Although recent studies have questioned the necessity of close follow-up of melanoma patients to detect recurrence,  this article emphasizes a different aspect of the importance of follow-up: early detection of a second primary melanoma.

20 MHz Sonometric Determination of Tumor Thickness and Invasion Index Correlates With Breslow Thickness and Clark Level in Malignant Melanoma

Introduction.—The preoperative evaluation of primary malignant melanoma using 20 MHz US systems is useful in planning an operative approach. Preoperative US does not provide information on the level of tumor invasion. The accuracy of high-resolution US examination of 155 primary melanomas of the skin was assessed in relation to Breslow thickness.

Methods.—A 20 MHz scanner was used to assess invasive, malignant melanoma tumors located within the head-neck (15.5%), on the trunk (40%), and on the extremities (45.5%). The histologic tumor types were as follows: 131 (84.5%) superficial spreading melanomas, 12 (7.7%) nodular melanomas, 11 (7.1%) lentigo maligna melanomas, and 1 (0.6%) acral-lentiginous melanoma. For 80, 40, 27, and 8 patients, respectively, the Breslow thickness was 0.75 mm or less, 0.76-1.5 mm, 1.51-4 mm, and greater than 4 mm.

Results.—Histologic tumor thickness and preoperative sonometric thickness were highly correlated (mean deviation of sonometric thickness, 0.25 mm). The correlation was even better in thick melanomas (0.76-1.50 mm; more than 1.50 mm), compared with thin melanomas (up to 0.75 mm). The sonometric invasion index and the histologic level of invasion had good correlation. There was a significant difference between Clark levels III and IV, as indicated by sonographic invasion index.

Conclusion.—The sonometric tumor thickness and invasion index should both be considered in the preoperative treatment planning of patients with malignant melanoma. Correlation was poor with thin melanomas. Typically, sonometric tumor thickness is overstated. There was a particularly high predictive value for the differentiation of Clark levels III and IV.

► The standard of care in the United States is for excision margins for melanoma to be determined by Breslow thickness based on histologic examination of a representative biopsy specimen. High-resolution US is useful for determining melanoma thickness. Although it cannot currently substitute for direct measurement of tumor thickness by micrometer measurement, it could be very useful in determining the appropriate portion of the lesion to biopsy. Somach et al. previously have demonstrated that invasive melanoma is missed in approximately 20% of biopsy specimens of melanomas on actinically damaged skin. Ultrasound may help prevent a sampling error by selecting out the thickest portion of the lesion for biopsy.

Surgical Margins for Excision of Primary Cutaneous Melanoma

Introduction.—The width of surgical margin necessary for complete excision of melanoma is still a major controversy. Margins of 5 cm or more were standard until recently. Current consensus evidence supports 1-cm margins with an additional 2 cm of subcutaneous fat for melanomas less than 1-mm thick on the trunk and proximal extremities; and 2-cm margins for melanomas less than 4-mm thick on the trunk and proximal extremities. There are no guidelines for melanomas more than 4 mm, in situ melanomas, or those on the head, neck, hands, and feet. To develop guidelines for predetermined surgical margins for the excision of cutaneous melanoma, the extent of subclinical melanoma extensions was measured with prospective data.

Methods.—A total of 535 patients with 553 primary cutaneous melanomas were prospectively studied. The fresh tissue technique of Mohs micrographic surgery was used to excise all melanomas. There was frozen-section examination of the margin. By measuring the invisible extensions of tumor around the melanoma, the surgical margin needed for excision of melanoma was determined. Six millimeters was the minimal surgical margin, and an additional 3 mm was added for any melanoma that required a subsequent stage to remove the tumor completely.

Results.—With a 6-mm margin, 83% of melanomas were successfully excised. With a 9-mm margin, 95% were removed. With a 1.2-cm margin, 97% were removed. There were wider margins for removal of melanomas on the head, neck, hands, and feet than on the extremities and trunk. There were wider margins for removal of melanomas that were more than 2-3 cm in diameter than for smaller melanomas. The new guidelines were compared with the current guidelines and it was found that if all melanomas in this study were excised according to the current recommendations, 8% of the melanomas would have been inadequately excised. There would have been an excess margin of 5 mm for 59% of the melanomas, and 23% with an excess of 1 cm. If the new guidelines were used, only 1% of the melanomas would have been inadequately excised, and there would have been unnecessary margins in excess of 1 cm for 6% of melanomas.

Conclusions.—For melanoma in situ or melanomas on the trunk and proximal extremities that are smaller than 2 cm in diameter 1 cm of normal-appearing skin should be the predetermined surgical margin for excision by standard surgical technique. For tumors larger than 2 cm in diameter, there should be a 1.5-cm margin. A minimal surgical margin of 1.5 cm is recommended for melanomas on the head, neck, hands, and feet. For melanomas larger than 3 cm in diameter the recommended margin is 2.5 cm. When more narrow margins are desired, Mohs micrographic surgery is a useful alternative to standard surgery, particularly for melanomas larger than 2.5 cm in diameter melanomas without distinct clinical margins, or for melanomas on the head, neck, hands, and feet.

► One of the most controversial issues in the management of cutaneous melanoma has been the margin of resection. For years dogma was based on the conclusions of a single physician who dissected out a metastatic lesion in a single patient and, based on his observations, made recommendations on the surgical management of melanoma. It was years before anyone realized the fallacy of these recommendations. However, even today there have been very few objective studies that have addressed this issue. Although the role of Mohs surgery (MS) in the management of melanoma has been debated, no one can argue that it is the most precise means by which to determine the subclinical spread of a variety of cutaneous neoplasms.

In this study of 553 primary lesions, contrary to dogma, it was the diameter and location of the lesion and not its thickness that determined the margin of resection necessary to achieve clear margins. Larger lesions and those on the head, neck, hands, and feet demonstrated greater subclinical spread. The incidence of metastases after MS was less than that seen in historical controls. The local recurrence rate was low (0.5% or less), and 2 of the 3 recurrences may have actually represented satellites. Overall, 83% of the lesions were completely excised with a 6-mm margin, and 97% were excised with a 1.2-cm margin.

This article is important for several reasons. First it establishes a role for MS in the management of melanoma, and gives some guidance as to which lesions should be managed by MS. Second, for the first time it provides hard data on margins necessary to completely excise most melanomas. This article should serve as an impetus for the reevaluation of surgical margins recommended for melanoma. Hopefully, the day of arbitrarily wide margins to encompass satellite lesions will soon be a thing of the past. Even if one is fortunate enough to encompass these, there is no benefit in terms of survival because they represent metastatic disease and are a harbinger of systemic spread.

Mohs Micrographic Surgery for the Treatment of Primary Cutaneous Melanoma

Background.—Mohs micrographic surgery should be successful in treating melanomas, if frozen-section analysis is accurate. The survival, metastatic, and local recurrence rates after Mohs micrographic surgical treatment of primary cutaneous melanoma were assessed.

Methods.—Over 15 years, 353 patients with primary cutaneous melanoma were treated with Mohs micrographic surgery. Tumors were staged by history, physical examination, and formalin-fixed, paraffin-embedded sections; for tumors greater than 1 mm thick, chest roentgenography and serum lactate dehydrogenase levels were also used. Mohs micrographic surgery involved excising the biopsy site or visible tumor with a 3-mm surgical margin into the subcutaneous tissue plane (i.e., debulking). After debulking, another 3-mm margin was excised for frozen-section analysis, and excision was repeated until a tumor-free margin was identified. Follow-up consisted of semiannual visits or more frequent examinations for patients whose melanomas were greater than 1.5 mm thick. Results after Mohs micrographic surgery were compared with results from 15,798 historical controls treated by standard wide margin surgical techniques. The 2 patient groups were stratified and compared on the basis of tumor thickness.

Findings.—There were 553 melanomas treated in these 535 patients, and all but 3 patients were available at 5-year follow-up. Results with Mohs micrographic surgery were overall as good as those using standard techniques, and were in fact significantly better for tumors 0.76-1.49 mm and greater than 4.0 mm thick. Similarly, at each thickness compared, there were fewer metastases with the micrographic technique. Local recurrences developed in 0.5% of patients who underwent the microsurgical technique, compared with local recurrence rates between 3% and 15% for standard surgical excision. Surgical margins of 6 mm sufficed to excise completely 83% of the melanomas, whereas margins of 1.2 cm completely cleared 97% of melanomas. Tumors that required more extensive excision typically had in situ melanoma at the lateral margins, rather than invasive melanoma in the deeper part of the skin.

Conclusions.—Mohs micrographic surgery successfully treated primary melanomas and was associated with low recurrence and metastasis rates. Thus, particularly in cases in which margins are ill defined, this technique can help identify microscopic disease and preserve normal tissue.

► The author presents a large series of patients with primary cutaneous melanoma treated with Mohs micrographic surgery. His data show that the 5-year survival rate after Mohs micrographic surgery is equivalent or greater than the survival rate after traditional surgical excision. The patients treated with Mohs micrographic surgery also have a lower metastatic rate than historical controls. I agree with Dr. Zitelli that Mohs micrographic surgery is an effective therapy for primary cutaneous melanoma. This is especially true in areas in which tissue conservation is important. However, one must be experienced in the interpretation of melanoma using frozen sections, and have an excellent frozen-section laboratory to ensure the quality of the slides being examined.

Does Surgical Removal of Primary Melanoma Trigger Growth of Occult Metastases? An Analytical Epidemiological Approach

Purpose.—It has been suggested that surgery to remove a primary tumor might provoke metastatic dissemination or growth of micrometastatic deposits. Supportive data have been reported for breast and prostate carcinoma. The epidemiologic evidence that surgical removal of a primary melanoma enhances the growth of metastases was assessed.

Methods.—The data set included 1,224 consecutive patients with primary cutaneous melanomas. The risk that patients with “thin” melanomas would have metastases within 1 year was compared with the risk that those with “thick” melanomas would have metastases at diagnosis. The cases were matched in pairwise fashion, based on assumptions regarding tumor volume and tumor doubling time.

Results.—In the presence of a long tumor doubling time—200-800 days—surgically removed tumors appeared to be associated with a higher risk of clinically apparent metastases within 1 year; compared with matched tumors seen with metastases at diagnosis. For tumors with short doubling times—50 or 100 days—there was no significant difference. However, neither was their any apparent benefit of surgical removal.

Conclusions.—Surgical removal of primary melanoma may enhance the growth of micrometastases, thus leading to clinical metastases earlier than if the primary tumor had been left in place. The study has many limitations, including its crude method of determining tumor volume and its failure to consider variations in tumor doubling time. For patients in whom micrometastases are already present, removing the primary tumor may lead to metastatic growth. For those who do not yet have micrometastases, removing the primary tumor will produce a cure. Rather than suggesting that melanomas should be left in situ, the data outline a theoretical approach to questions about melanoma surgery and metastatic spread.

► This article represents a theoretical discussion on whether surgical removal of a melanoma from a patient with occult micrometastases can enhance the growth of the metastases. As suggested by the authors, their calculations are based on a number of assumptions that may or may not be correct; however, they certainly present food for thought, i.e., do we harm the patient with a high-risk melanoma by removing the primary lesion? Without well-controlled studies, which are unlikely to be done, it will be difficult to answer this question. Moreover, even if the study was done and the hypothesis was proven correct, it is well known that not all high-risk patients develop metastases. Thus, some patients with a theoretically poor prognosis could be denied curative surgery.

The Utility of Fine Needle Aspiration in the Diagnosis of Melanoma Metastatic to Lymph Nodes

Introduction.—The detection of nodal metastases is crucial in determining the prognosis and treatment of patients with melanoma. Fine-needle aspiration (FNA) is efficient and accurate in evaluating suspected nodes in patients with a variety of malignancies. The FNA technique is unfamiliar to many clinicians and dermatologists who are responsible for the care of patients with melanoma. The utility and diagnostic reliability of FNA in melanoma patients with suspect nodules in lymph node basins were evaluated.

Methods.—The records of 46 patients with melanoma who underwent 56 FNAs for suspect nodules in lymph node basins were reviewed retrospectively. The aspiration diagnosis was compared with an open biopsy diagnosis or clinical follow-up to predict the sensitivity and specificity of FNA.

Results.—Of 56 FNAs, 24 were positive for melanoma, 31 were negative, and 1 was suspect but inconclusive. Melanoma was confirmed by open biopsy in 22 patients and by clinical progression of disease in 2 patients. Negative FNAs were confirmed by open node biopsy in 12 patients and by the clinical course in 19 patients. Definitive diagnoses of melanoma, other cancer, or benign findings were determined in 50 of 56 FNAs (89%). When definitive diagnoses were reached, the sensitivity and specificity of FNA was 100%.

Conclusion.—Fine-needle aspiration biopsy of suspect nodules in lymph node basins may be considered an accurate, rapid, and cost-effective tool in the management of melanoma. An algorithmic approach to patients with melanoma who have palpable lymph nodes is offered by the authors.

► Fine-needle aspiration is useful for the evaluation of lymph node metastases in a variety of neoplasms. However, it has not been extensively studied in patients with melanomas. Basler et al. demonstrate 100% specificity and sensitivity of FNA in definitive cases. Although it is unlikely that any laboratory test consistently performs at that high level, FNA may prove helpful for evaluation of patients with melanoma who develop 1 or more enlarged lymph nodes during the course of their disease, whether it be in the initial staging period or later in the course. Confirmation of these results by others is awaited.

Sentinel Node Biopsy in Melanoma Patients: Dynamic Lymphoscintigraphy Followed by intraoperative Gamma Probe and Vital Dye Guidance

Introduction.—Experts do not agree about the utility of elective lymph node dissection. An alternative to this approach is biopsy of the first tumor-draining lymph node (sentinel node, SN). Since the SN is the first to receive metastatic tumor cells, a negative SN biopsy would theoretically preclude the presence of lymphatic malignant involvement and obviate the need for lymph node dissection. The advances of three known techniques—preoperative (dynamic) lymphoscintigraphy, gamma probe localization, and intraoperative blue dye mapping—were combined to improve the yield of successful SN biopsies.

Methods.—One hundred thirty-five consecutive patients with histologically proven clinical stage I cutaneous melanoma underwent the 3 diagnostic tests to evaluate 150 drainage areas. Patients initially underwent dynamic and static lymphoscintigraphy after injection of technetium 99m colloidal albumin. One to 3 focal accumulations concordant with SNs were observed in the lymphatic drainage areas in all patients. This occurred within 20 minutes of tracer injection in 97%. Within 2-24 hours after tracer injection, a handheld gamma probe was used for perioperative identification of SNs, to estimate the optimal site for incision and guide preparation. Vital dye was injected immediately preoperatively to facilitate final identification and biopsy of the SN. Biopsies were obtained from 216 SNs.

Results.—Micrometastases were detected in 39 SNs in 30 drainage areas. The only node-harboring tumor was the SN in 22 of 30 drainage areas. The SN did not contain blue dye and would not have been detected without the gamma probe in 5 of 30 drainage areas. At a range of 233-691 days, no recurrences were observed in any of the lymphatic drainage areas where the SN was tumor free.

Conclusion.—By combining preoperative (dynamic) lymphoscintigraphy, gamma probe localization, and intraoperative blue dye mapping, SNs were detected and excised in all 135 patients. The sentinel node concept has far-reaching consequences for surgical oncology and may provide new possibilities for intervention.

► SN biopsy is rapidly gaining acceptance as an alternative to elective lymph node dissection for determining occult lymph node involvement in melanoma patients with intermediate-thickness lesions. There is a learning curve in the surgeon’s ability to find the SN. Intraoperative lymphatic mapping with vital blue dye alone is not always easy. The authors of this study have demonstrated that preoperative dynamic lymphoscintigraphy is useful to detect the SN. Intraoperative localization of the SN with a handheld gamma probe facilitates the SN biopsy. The use of vital blue dye is helpful because the SN can then actually be visualized. Combining preoperative lymphoscintigraphy with intraoperative blue dye mapping and gamma probe guidance led to a 100% success rate in detecting and removing SNs with no false-negative biopsies.

Prospective Evaluation of Selective Lymph Node Biopsy for Cutaneous Malignant Melanoma

Introduction.—Intraoperative sentinel lymph node biopsy involves injecting blue dye around the primary melanoma site and tracing it to regional lymph node basins to identify the sentinel lymph node. The sentinel lymph node detection rate was improved to a rate of 90% with the addition of preoperative lymphoscintigraphy to blue dye lymphatic mapping. A high success rate of 98% has been found by combining technetium sulfur colloid lymphoscintigraphy with the use of the hand-held gamma probe to identify sentinel lymph nodes intraoperatively. In the treatment of stage I and II melanoma, a prospective clinical experience was reported using preoperative lymphoscintigraphy and sentinel lymph node biopsy.

Methods.—During a 3-year period, 63 patients with primary cutaneous malignant melanoma had lymphoscintigraphy with technetium sulfur colloid followed by gamma probe-guided lymph node biopsy. They had a mean age of 51.1 years and 51% were women. The melanomas had a mean Breslow thickness of 2.13 mm. They were located on the head and neck (13%), trunk (38%), upper extremity (21%), and lower extremity (29%). In 49 patients (78%), selective lymph node biopsy was performed on an outpatient basis, and in 14 patients (22%), it was performed in the operating room. Biopsy was performed on 1 lymph node site in 73% of patients, on 2 lymph node sites in 25%, and on 3 lymph node sites in 2%. Patients had a mean of 1.64 labeled lymph nodes removed per site.

Results.—In 9 selective lymph node biopsy sites in 8 patients, micrometastatic disease was identified. No additional pathologic lymph node involvement occurred in 5 of 8 patients who had lymph node dissection (63%). These patients had a mean Breslow thickness of the primary tumor of 6.01 mm. No evidence of disease was found in 59 patients (94%); they remained disease-free after a mean follow-up of 579 days from selective lymph node biopsy. Death occurred in 3 patients, 2 with systemic disease and 1 from myocardial infarction. Regional recurrence was not noted in these patients. Six patients had seroma or infection (10%).

Conclusions.—Patients with cutaneous malignant melanoma can be managed with gamma probe-directed selective lymph node biopsy, which is a straightforward procedure that can be done in the outpatient setting.

► Data and experience continue to accumulate on the role of the selective (sentinel) lymph node biopsy (SLNB) in the management of melanoma. This procedure has at long last brought some rationale to the management of melanoma and has spared many patients from undergoing unnecessary lymph node dissections. As this article demonstrates, the technique continues to evolve. Loggie et al., like many surgeons, no longer use the blue dye but rely solely on a gamma probe in combination with the technetium sulfur colloid that is used for the lymphoscintigraphy. Lymphoscintigraphy is usually done 4 hours before the node biopsy to permit a greater concentration of tracer. The radiologist can mark the sentinel node(s) for the surgeon to facilitate its (their) detection and removal. In 78% of cases the procedure was done under local anesthesia. The incidence of adverse effects was low (10%), with infection and seroma formation being the 2 most common complications. The authors found the procedure to be extremely reliable. Problems encountered included access to axillary nodes, which can be deeply located, and sometimes inaccessibility of nodes of the head and neck, biopsy of which can be associated with significant morbidity (e.g., the parotid nodes). Interestingly, the authors found that prior wide excision of the melanoma did not affect their ability to do a SLNB. Finally, unlike other studies, patients with high-risk (e.g., regressing) thin lesions (smaller than 0.76 mm) were also included in their study. Whether SLNB improves survival is still unanswered; however, as better adjuvant therapy becomes available, it should allow physicians to select those patients who would most benefit from the procedure.

Classification of Localized Melanoma by the Exponential Survival Trees Method

Introduction.—Many clinical and pathologic factors are associated with survival of patients with cutaneous melanoma. Dominant factors that affect the outcome of melanoma have been identified, including tumor thickness, tumor ulceration, level of invasion, primary lesion site, and gender. The current staging system is based on tumor thickness and level of invasion, but the cutoff points for tumor thickness are still controversial, and there is still room for improving the staging system so that it can be applied more uniformly. The most important prognostic factors of localized melanoma were identified. The possible interactions among these prognostic factors were examined. A localized melanoma risk grouping system was defined.

Methods.—To analyze 4,568 patients with melanoma, the exponential survival method was used. This was created according to prognostic factors that classified patients into homogeneous subgroups by survival. The analysis included 6 clinical and pathologic factors. A superior means of prognostic classification was provided with this tree-based method. Interactions among the variables were detected to a greater degree than with regression models.

Results.—The most important prognostic factor was tumor thickness, followed by tumor ulceration and primary lesion site. Among these prognostic factors, some important interactions were identified. Tumor thickness, ulceration, and primary lesion site were used to create 5 distinct risk groups. The best prognosis was found in patients with tumor thickness of 0.75 mm or less and lesions on their arms or legs. The poorest prognosis was found in patients who had ulcerated tumors with thickness greater than 4.50 mm. The first risk group included patients with a tumor thickness of 0.75 mm or less and lesions on arms or legs. This group accounted for 17.3% of the total number of patients and had a 10-year survival rate of 96%. Patients in the second risk group had a tumor thickness of 0.75 mm or less and lesions located on sites other than arms or legs. They accounted for 28.6% of the total patients and had a 10-year survival rate of 88.3%. The third risk group consisted of patients with a tumor thickness between 0.76 and 1.50 mm with or without ulceration, and patients with a tumor thickness greater than 1.50 mm and lesions on arms or legs without ulcerations. This risk group comprised 25.9% of the total patients and had a 10-year survival rate of 74.6%. The fourth risk group consisted of patients with a tumor thickness of more than 1.50 mm and lesions on sites other than arms or legs without ulceration, and patients with tumor thickness between 1-51 mm and 4.50 mm with ulceration. They accounted for 24.1% of the total patients and had a 10-year survival rate of 55.1%. The fifth risk group (4.2% of the total) consisted of patients with a tumor thickness greater than 4.50 mm and ulceration; their 10-year survival rate was 32.7%.

Conclusions.—A comprehensive, easy-to-use risk group system for classifying patients with localized melanoma was created and was based on exponential survival trees. In the clinical management of patients with melanoma, and in designing and analyzing clinical trials, this grouping system would be useful.

► To inform patients with melanoma about what they can expect in terms of survival, and to evaluate adjuvant therapy for melanoma, it is important to determine what variables are most important in determining prognosis, and based on these variables develop a reliable prognostic model. In this study, Huang et al. used an experimental survival tree method to predict prognosis. Based on the thickness of the lesion, the presence or absence of ulceration, and the location of the lesion (arm and leg vs. other), the authors were able to construct 5 risk groups whose 10-year survival varied from 96.0% to 32.7%. Although thickness is the single most important variable in determining prognosis, by combining it with lesion location and the presence or absence of ulceration, the authors were able to further refine their prognostication.

Predicting Ten-Year Survival of Patients With Primary Cutaneous Melanoma: Corroboration of a Prognostic Model

Objective.—Several approaches have been tried to improve prediction of survival in patients with primary cutaneous melanomas. In a model recently reported by the Pigmented Lesion Group of the University of Pennsylvania, 4 variables—tumor thickness, anatomic site, age, and sex—were used to predict 10-year survival. The advantage of this model is that information on all 4 variables is readily available. The University of Pennsylvania predictive model was tested using a large melanoma data base.

Methods.—The New York University Melanoma Cooperative Group data base included 1,142 patients with 1,170 primary cutaneous melanomas. The patients were entered into the data base between November 1972 and March 1982, most often before any definitive surgery was performed. The current analysis included 780 patients with 1 primary melanoma who were followed up prospectively for at least 10 years or until death from melanoma. Seventy-seven percent of the patients had superficial spreading melanomas; the subtype distribution was similar to that reported in previous studies. The 4 variables were age 60 years or younger vs. older than 60, male vs. female sex, anatomic site (i.e., extremity sites excluding palms, soles, and subungual regions [group 1] vs. palms, soles and subungual regions plus all axial sites [group 2]), and Breslow tumor thickness. Predicted probabilities of 10-year survival were assessed by multivariate logistic regression.

Findings.—Ten-year survival was 78%. Median age was 50 years at baseline, and 54% of patients were women. Fifty-six percent of lesions were in group 2 sites. Median tumor thickness was 1.2 mm. On multivariate analysis, age, site, and thickness were independent predictors of survival; sex was not. A fitted logistic regression model was used to create a table for use in estimating probability of 10-year survival. Tumor thickness was the most powerful prognostic factor. Age and site had a greater effect on prognosis in patients with intermediate or thick melanomas than in those with thin melanomas. Anatomic site was the second most powerful prognostic factor. Older age continued to be an adverse prognostic factor after controlling for lesion thickness and other risk factors.

Conclusions.—This study, using a large, independent data base, supports the prognostic multivariate model of the Pigmented Lesions Group of the University of Pennsylvania. However, sex is not an independent predictor in the current analysis. Though tumor thickness is the most important prognostic factor, survival estimates vary widely according to age and anatomic site.

► Although numerous studies have demonstrated the prognostic importance of tumor thickness for patients with malignant melanoma, it is obvious that other variables also have an impact on survival. Over the years a number of models have been developed to better predict survival. Unfortunately, many of these models included variables not recorded by all pathologists. Recently an attempt was made to construct a simple model that included only parameters readily accessible to the clinician. In this study the New York University Melanoma Cooperative Group applied their, data to this model and found it to be useful and applicable to most patients with melanoma.

Life-long Screening of Patients With Intermediate-thickness Cutaneous Melanoma for Asymptomatic Pulmonary Recurrences: A Cost-effectiveness Analysis

Introduction.—The incidence of cutaneous melanoma continues to rise in this era of increased emphasis on containment of health care expenditures. It has been recommended that blood tests and chest radiographs be eliminated from follow-up evaluations of patients with cutaneous melanoma because they are never the sole indicator of disease recurrence. The costs and potential benefits of an intensive chest radiograph screening program to detect asymptomatic pulmonary metastases in patients with intermediate-thickness, local, cutaneous melanoma were analyzed.

Methods.—A cost-effectiveness analysis was performed to create a model based on data from the following sources: an historic cohort at Roswell Park Cancer Institute and other published studies; estimates of new patients with melanoma in 1996 from the National Cancer Institute’s Surveillance, Epidemiology, and End Results program; and estimates of cost and treatment benefits from published articles retrieved through MEDLINE. The net costs were estimated as the added cost of chest radiographic screening to follow-up and costs of surgical treatment of lung recurrences. The net benefits were considered as potential savings in non-quality-adjusted life years (NQALY) and quality-adjusted life years (QALY) consequent to surgical treatment. The cost-effectiveness ratios were assessed as the current value of net costs divided by net benefits (discounted and undiscounted).

Results.—The cost of screening per NQALY and QALY was $150,000 and $165,000, respectively, using undiscounted health benefits. Screening and treatment costs were responsible for 80% and 20% of program costs, respectively. The annual cost-effectiveness ratios were the least expensive in years 3-10 of screening. Estimates of the total cost of a 20-year screening program for patients diagnosed in 1996 were between $27 and $32 million.

Conclusion.—Significant cost-savings may be realized if screening frequency could be decreased in the first 2 years and limited to the first 5-10 years after diagnosis. At some point, the risk of recurrent disease becomes so slight that greater premiums are paid to detect 1 recurrence. Patients with stage II tumors (more than 1.5 mm thick) probably benefit the most from screening.

► Researchers at the Mayo Clinic recently have demonstrated that medical history and physical examination are much more cost-effective than chest x-ray examinations and blood chemistry panels in detecting recurrent disease in patients with melanoma. Weiss et al. found that 94% of recurrences in 145 patients with melanoma were identified by history or physical examination, and the remaining 6% had abnormal chest x-ray results. Laboratory results were never a sole indicator of recurrence. The National Institutes of Health Consensus Development Conference statement on Diagnosis and Treatment of Early Melanoma did not recommend routine screening for occult visceral lesions in patients with melanomas less than 1 mm thick. The authors of the current study conclude that significant cost-savings may be possible by decreasing the frequency of chest x-ray examinations during the first 2 years after surgery and limiting screening to the first 5-10 years after diagnosis. Even more cost-savings may be achieved by limiting chest x-ray examinations to patients with melanomas more than 1 mm thick.

Computed Tomography in Evaluation of Patients With Stage III Melanoma

Introduction.—In asymptomatic patients with primary cutaneous melanoma, routine CT scanning or nuclear scintigraphy has a low metastasis detection rate. Because of their high rate of relapse and distant disease involvement, patients with stage III disease are frequently evaluated with brain, chest, abdominal, and pelvic CT. However, the diagnostic yield of these studies and their effect on patient care remain unknown. The ability of CT to identify metastatic melanoma in patients with clinical stage III disease was evaluated, including its influence on patient management.

Methods.—The retrospective study included 347 patients with asymptomatic stage III melanoma seen over a 6-year period. Of these, 289 patients had at least 1 CT scan of the chest, abdomen, or pelvis, and 136 had all 3 scans. Data on history and physical findings, radiologic studies, and liver function tests were collected by chart review. Histopathologic, cytologic, or other firm evidence was required to confirm the radiologic finding of metastatic melanoma or a second primary neoplasm.

Results.—The analysis included a total of 788 cranial, thoracic, abdominal, or pelvic CT scans. The presence of metastatic melanoma was confirmed by 4.2% of individual scans, while the false-positive rate was 8.4%. Head CT scanning did not identify metastases in any patient. Chest CT was positive in 20% of patients with cervical adenopathy, 5% of patients with axillary adenopathy, and 2% of patients with groin adenopathy. Abdominal CT was positive in 8% of patients with cervical adenopathy and 4% of patients with axillary or groin adenopathy. Pelvic CT scans were negative in patients with neck or axillary adenopathy, but did identify metastatic melanoma in 7.4% of patients with groin adenopathy. In patients undergoing total body imaging, metastases were detected in 8.1% and second primaries in 2.2%, for an overall true positive yield of 10.3%. The finding of metastatic melanoma led to a change in treatment plan for 6.6% of patients. Two hundred eighty-nine patients underwent a total of 40 surgeries or other invasive procedures; 3% had complete resection of metastatic disease.

Conclusions.—Though CT scanning is widely used in the diagnostic workup of patients with stage III melanoma, the yield appears too low to justify its routine use. The findings suggest that head CT is not indicated in asymptomatic patients, chest CT is not indicated in patients with groin adenopathy, and pelvic CT is not indicated in patients with axillary or cervical adenopathy. There may be a role for selective chest CT in patients with cervical adenopathy or pelvic CT in patients with groin disease. A prospective study is needed to define the true yield and benefit of CT scanning in the evaluation of patients with stage III melanoma.

► The use of CT to examine asymptomatic patients with primary melanomas has been shown to be of little value. Kuvshinoff et al. studied the use of CT in patients with stage III disease (in-transit lesions and/or regional nodal disease). Overall, 5% of patients were found to have “occult” disease; the false-positive rate was 8.4%. Moreover, the few patients with abnormal CT scans often had abnormal chest x-ray films or elevated liver function studies. (Interestingly, liver function studies accurately identified metastatic melanomas in only 1.4% of patients who had abdominal CT scans to evaluate the liver.) A better yield was seen in patients with cervical adenopathy who had chest CT scans and in patients with inguinal adenopathy who had pelvic CT scans, which suggests a possible role in certain clinical settings. Indiscriminate use of CT scans, even in patients with stage III disease, appears to be unjustified. Moreover, of those with positive scans, only 7% had their management altered.

Outcomes of Patients With Local Recurrence of Cutaneous Malignant Melanoma: A Population-based Study

Introduction.—Local recurrence of malignant melanoma has typically been associated with poor prognosis, with reports of 50% to 82% of patients dying of disease progression. The clinical consequences and prognostic impact of local recurrence from cutaneous malignant melanoma as the first event in relation to putative prognostic factors was evaluated in a population-based investigation.

Methods.—The definition of local recurrence of cutaneous malignant melanoma differs in the literature. Local recurrence is defined here as recurrence strictly within the scar or transplant. Data were collected from a population-based regional cancer registry in Stockholm, Sweden, of 3,706 patients with cutaneous malignant melanoma registered in the period 1976-1997. The prognostic value of putative prognostic factors was assessed. Melanoma specific survival (MS) and disease-free survival (DFS) were calculated.

Results.—Local recurrence as a first event occurred in 48 of 3,706 patients (1.3%). In 11 of 39 patients (28%) with local recurrence of invasive primary melanoma distant metastases developed, and they died. Ulceration had prognostic significance in univariate analysis. With melanoma as the end point and local recurrence as a time-dependent covariate, there was a nonsignificant relative risk of 1.3 associated with local recurrence. The 5-year DFS and MS were 69% and 83%, respectively. When considering only the patients with primary invasion, these rates were 61% and 69%, respectively.

Conclusion.—Local recurrence was rare in this population-based cohort of patients with cutaneous malignant melanoma. Patient outcomes were relatively favorable after diagnosis. The putative excess risk of having a local recurrence develop was considered moderate.

► Previous studies have shown dismal survival rates after local recurrence of melanoma, with mortality rates in excess of 50%. However, many earlier studies have included patients with satellite or in transit metastases in addition to patients who had recurrence within the scar. In this population-based study, inclusion was restricted to patients with recurrence in the scar (or split-thickness skin graft in wounds that were not closed primarily). The authors found a low rate (1.3%) of true local recurrence over a 15-year period. The 5-year disease-free survival rate for patients with primary invasive melanoma in this study was 61 %, and the melanoma-specific survival rate was 69%. For patients with local recurrences, the 5-year disease free survival rate was 69%, and the melanoma-specific survival rate was 83%. Therefore, there was no major deleterious effect associated with true local recurrence. The difference with previous studies undoubtedly is the strict definition of local recurrence used here. Satellite metastases are a harbinger of distant metastases, and they indicate a prognosis that is markedly different from that for patients with true local recurrence.

Serum S100—A Marker for Disease Monitoring in Metastatic Melanoma

Introduction.—Serum S100 is an acidic calcium-binding protein with a molecular weight of 21 kD. The usefulness of S100 protein as the marker of choice for immunohistochemical identification of malignant melanoma has been confirmed by several reports since it was detected in cultured melanoma cells in 1980. It has been suggested that serum S100 may be a useful marker for staging disease. Serum S100 levels of patients with different stages of malignant melanoma were analyzed to determine the value of serum S100 as an indicator of response to treatment.

Methods.—Serum S100 concentrations were measured, in 73 patients with malignant melanoma and 130 healthy controls, by means of an immunoradiometric assay based on the two-site sandwich method. Four patients with metastatic melanoma underwent serial serum S100 determinations.

Results.—Detectable serum S100 levels were observed in 1 of 25 stage I/III patients. Serum S100 levels were elevated in 3 of 14 patients with lymph node metastases (stage III, 21.4%). In patients with metastatic disease, disease progression was indicated by increasing serum S100 levels. A complete decline in serum S100 levels was observed in 2 patients in disease remission. Serum S100 correlated with clinical stage. Sensitivity for patients with metastatic melanoma was 62.5%.

Conclusion.—Serial measurements of serum S100 were useful in monitoring efficacy of treatment. Serum S100 was not helpful in screening and early diagnosis of malignant melanoma. Serum S100 may be valuable as a clinical marker for disease progression and for serologic monitoring during systemic therapies in patients with malignant melanoma.

► A serologic test for detecting metastatic melanoma that is sensitive and specific would be a great advance in the management of melanoma patients. An immunoradiometric assay for epitopes of the S100B subunit appears to be quite sensitive for monitoring patients with metastatic malignant melanoma. Remarkably, none of the 130 control subjects had elevation of serum S100 levels. Now that improved therapies are available, such as interferon-α 2B for patients with advanced disease, the availability of a serologic marker for progression of melanoma and serologic monitoring during systemic therapy could prove very valuable. Confirmation of these data by other investigators is awaited.

Serum S-100 Protein: A Potentially Useful Prognostic Marker in Cutaneous Melanoma

Introduction.—Neuroendocrine cells contain the acidic calcium-binding protein S-100, which is of known value in diagnosing tumors of melanocytic origin. It has been suggested that serum protein S-100 concentration could be a useful marker of disease activity in patients with metastatic melanoma. Serum S-100 protein concentration was evaluated as a prognostic indicator in patients with malignant melanoma.

Methods.—Participants included 97 patients receiving care for histologically confirmed malignant melanoma and 48 controls without this disease. Both groups underwent measurement of serum S-100 protein concentration. The results were compared with clinical disease stage, as classified by the criteria of the American Joint Committee on Cancer, and Breslow tumor thickness.

Results.—Median serum S-100 protein concentration was 0.11 pg/L in patients with stage I and II melanoma, 0.24 pg/L in those with stage III disease, and 0.39 pg/L in those with stage IV disease. These were all significantly higher than the median 0.1 pg/L value in controls. Using a value of 0.2 pg/L as the cutoff point, serum S-100 protein was 82% sensitive and 91% specific for advanced disease. Serum S-100 protein was also significantly correlated with Breslow tumor thickness. The combination of a serum S-100 protein concentration greater than 0.22 pg/L and a Breslow thickness greater than 4 mm was 91% sensitive and 95% specific for the detection of secondary spread of melanoma.

Conclusions.—Measurements of S-100 protein level in serum could be a useful prognostic indicator for patients with malignant melanoma, and may be useful in monitoring their therapy. Combined with Breslow thickness, this serum measure might improve the ability to predict prognosis at the time of diagnosis. More research is needed to determine whether the increased S-100 protein level in patients with advanced disease precedes the clinical signs, and whether S-100 protein is a useful marker of disease regression during treatment for advanced disease.

► The true usefulness of this test can only be determined by comparison with existing methodologies to detect spread of melanoma; it is perhaps wishful thinking to hope that a single blood test could replace the incredibly expensive imaging studies now in widespread use. It is interesting that patients with stage I and stage II melanoma (primary localized disease), many of whom were presumably cured of their disease, had median serum S-100 protein levels significantly higher than the control group. As indicated by the authors, there seemed to be a correlation between Breslow thickness and serum S-100 protein concentration; hence, their suggestion that Breslow thickness and serum S-100 protein concentration together may provide a better indication of prognosis than either factor alone.

DNA-ploidy Abnormalities Are a Reflection of the Metastatic Potential of Malignant Melanoma: Microfluorometric DNA Analysis

Introduction.—Studying DNA-ploidy in primary and metastatic melanomas may offer useful insights into tumor progression. However studies using flow cytometry have not shown significant differences between primary and metastatic melanomas. The authors have found 4′,6-diamidino-2-phenylindole(DAPI)-DNA microfluorometry to be a sensitive technique for studying the biological behavior or malignant potential of pigmentary neoplasms. This technique was used to compare DNA-ploidy in primary and metastatic melanomas.

Methods.—The study included 15 primary malignant melanomas and 20 regional or distant metastases from the same patients. Abnormalities of DNA-ploidy were assessed by DAPI-DNA microfluorometry. Another 15 metastasis-free patients were studied to assess the relationship between metastases and various prognostic factors.

Results.—Mean DNA index values were 1.76 in primary melanomas vs. 2.07 in metastatic melanomas. Eighty percent of the metastatic melanomas had higher DNA index values than their primary tumors. Eleven of the 30 primary melanomas studied metastasized; the DNA index in this group was 1.97, compared with 1.49 in the nonmetastasizing melanomas. Average tumor thickness was 6.25 and 3.82 mm, respectively.

Conclusions.—As measured by DAPI-DNA microfluorometry, metastatic melanomas have greater DNA-ploidy abnormalities than their corresponding primary tumors. In primary melanomas, DNA-ploidy abnormalities are a significant risk factor for later metastasis. The higher the DNA index of a primary tumor, the more likely it is to have metastatic potential.

► The authors report that DNA-ploidy abnormalities, measured using DAPI- DNA microfluorometry, are a reliable risk factor for melanoma metastasis. Could this increased tendency to metastases also reflect the increased thickness of tumors in the metastasizing group? This technique also has been used to differentiate Spitz nevus from melanoma. Another technique to examine DNA-ploidy abnormalities, flow cytometry, does not appear to differentiate primary from metastatic melanoma.

Evaluation of the Clinical Usefulness of Measuring Urinary Excretion of 5-S-cysteinyldopa in Melanoma: Ten Years’ Experience of 50 Patients

Purpose.—Both normal melanocytes and melanoma cells produce the melanin pigments black eumelanin and reddish brown pheomelanin. Patients with melanoma show increased urine, serum, and tumor levels of 5-S-cysteinyldopa (5-S-CD), which is a precursor of pheomelanin. Thus patients with malignant melanoma may be monitored with urine and serum 5-S-CD measurements. The diagnostic value of urinary 5-S-CD measurements over the long-term was analyzed, including comparison with radiographic and physical examination.

Patients and Findings.—Of 116 Japanese patients with melanoma of the skin, eye, or mucosa or melanoma that had metastasized from an unknown primary, 50 underwent measurement of urinary 5-S-CD levels on at least 3 occasions. Twenty-six patients were free of metastases, 10 had regional metastases, and 2 had amelanotic melanoma; no patient in any of these groups had a significant increase in 5-S-CD levels. However, significant increases occurred in 12 of 12 patients with distant metastases. In 7 patients who did not have distant metastases at presentation, the increase in urinary 5-S-CD occurred before metastases were detected by radiography. Mean survival was 8 months for patients with 5-S-CD levels greater than 1,000 µg/day, compared with 3.5 months for those with levels of greater than 10,000 µg/day. Four patients had a peak 5-S-CD excretion greater than 40,000 µg/day; all had multiple liver metastases.

Conclusions.—In patients with malignant melanoma, serial urinary 5-S- CD measurements provide useful information regarding the presence of distant metastases and prognosis. These measurements are not useful for detecting early regional lymph node metastases or for monitoring amelanotic melanoma.

► Although measurement of urinary excretion of 5-S-CD was not useful in the detection of early regional lymph node metastases, increased levels did indicate the presence of distant metastases and also provided prognostic information. The role of this test in patient management remains to be established.

Cutaneous Malignant Melanoma: Scintigraphic Detection of Recurrent Disease With Tc-99m MIBI

Objective.—In addition to a high risk of postoperative recurrence, many patients with malignant melanoma (MM) already have nodal or systemic metastases when their cancer is diagnosed. Postoperative follow-up would be greatly aided by the availability of some technique for early detection of local recurrences or subclinical nodal and distant metastases. Previous reports have described uptake of the radiopharmaceutical agent technetium-99m methoxyisobutylisonitrile (Tc-99m MIBI) by primary MM lesions The findings of Tc-99m MIBI scanning in 3 patients with recurrent MM were reported.

Patients.—The patients were 2 men and 1 woman, 57-68 years of age. All patients were treated by surgery, which achieved complete resection of the primary lesion. All patients were referred with evidence of possible recurrent MM. After Tc-99m MIBI imaging studies were completed, biopsies of the lesions were performed. Other laboratory and imaging tests were performed as well. One patient who had had primary MM of the forearm had a circumscribed, raised, nonpigmented lesion on the surgical scar. The Tc-99m MIBI scan revealed increased tracer uptake in the area of the new lesion. A few nests of melanoma cells were found on histologic examination of the papillary dermis.

The second patient had multiple, subcutaneous, pigmented nodules along his leg 16 months after resection of MM on his right heel. On whole-body MIBI scanning, focal areas of uptake were found in association with the nodules, as well as in both inguinal regions. Metastatic cells were found in the nodular lesions and atypical cells in both inguinal regions, although none in the surgical scar. The third patient had respiratory symptoms and radiographic evidence of metastatic nodules in the lungs. He also reported headaches. He had previously undergone resection of a pigmented lesion of his left arm. Whole-body MIBI scanning showed abnormal uptake in the mediastinum and lungs, corresponding to the radiographic findings. Uptake was also noted in the brain and at the site of the surgical scar, confirmed by biopsy.

Discussion.—Scanning with Tc-99m MIBI, a radiotracer commonly used for myocardial perfusion imaging, can sensitively demonstrate areas of locally recurrent or metastatic MM. Tracer uptake is noted in areas of massive tumor involvement, as well as in lesions with only a few nests of melanoma cells. This imaging technique could aid in the early diagnosis of recurrent MM.

Pharmacokinetics of lodine-123-IMBA for Melanoma Imaging

Introduction.—The early detection of metastasis is an important diagnostic goal for the successful treatment of melanoma. For some time, the development of an effective radiopharmaceutical with melanoma affinity has been pursued. The synthesis, characterization, and biological evaluation of radioiodinated N-(2-diethylaminoethyl)-3-iodo-4-methoxybenz- amide (IMBA) was described and compared with 7 other benzamides. There was also an evaluation of the scintigraphic efficacy of this compound.

Methods.—Mice that were subcutaneously transplanted with B16 melanomas were investigated for the organ distribution of the 8 benzamides. Measurements were taken of the whole-body retention and the erythrocyte- and serum protein-bound fractions of radioiodinated benzamides.

Results.—The organ distribution was not any more improved as compared with benzylamine (BZA) when structural changes were made in the amide substituents of N-(2-dialkylaminoalkyl)-4-iodobenzamides 2-7. There was a high melanoma uptake with the 3-iodo-4-methoxyphenyl form of IMBA and significantly higher melanoma/nontarget tissue ratios. The average ratio improved after 1 hour by a factor of 8 compared with BZA, and it was still 4 times better after 6 hours. Almost identical n-octanol-phosphate buffer partition coefficients are exhibited by BZA and IMBA; however, faster urinary excretion is facilitated by a lower affinity to erythrocytes and serum proteins with IMBA. This could explain the improved tissue partitioning.

Conclusion.—At 4 hours after injection, IMBA showed high tumor contrast because of rapid background clearance and high melanoma affinity, which makes this a promising new radiopharmaceutical for the scintigraphic detection of melanoma metastases.

► The usefulness of the imaging techniques described in Abstracts 19-39 and 19-40 in patients with melanoma requires further investigation. Questions that need to be answered are (1) What is the diagnostic accuracy of the tests? (2) Is their sensitivity and specificity superior to existing technology? and (3) Does earlier detection of metastatic spread translate into improved survival and quality-of-life parameters?

Mortality, Major Amputation Rates, and Leukopenia After Isolated Limb Perfusion With Phenylalanine Mustard for the Treatment of Melanoma

Background.—This review examined the risks of isolated limb perfusion chemotherapy for advanced primary, regionally recurrent, or metastatic melanoma of the extremity.

Study Design.—A MEDLINE search of the English language medical literature from 1980 to 1995 was conducted to obtain all reports of patients treated with phenylalanine mustard isolated limb perfusion therapy for cutaneous melanoma of the extremities. This search identified 13 articles—11 retrospective case studies and 2 prospective trials—which were analyzed for mortality rates, major amputations and leukopenia incidence. There were 1,315 women and 667 men included in the study.

Findings.—The overall 30-day mortality rate was 0.6%. The incidence of major amputation was 0.8%. Leukopenia occurred in 0.7% of the patients.

Conclusion.—A review of 35 years of reported experience with phenylalanine mustard isolated limb perfusion therapy for patients with cutaneous malignant melanoma of the extremity indicated that the risk of mortality, limb amputation, and leukopenia after this therapy was low.

► Isolated limb perfusion with phenylalanine mustard has proven useful for the treatment of regionally confined recurrent melanoma. This study by Taber and Polk on the risk of serious complications from this therapy is timely, as the final data analysis of the World Health Organization (WHO) and European Organization for Research and Treatment of Cancer in the use of isolated limb perfusion as adjuvant treatment for cutaneous melanoma is pending. These investigators, found a low rate of serious complications. Moreover, most of the serious complications occurred in the early days of the procedure.

This is a complex and technically demanding procedure that should probably be offered only in those centers that have had experience and developed expertise in performing it. If the use of isolated limb perfusion as an adjuvant therapy for high risk stage I melanoma is proven effective in the WHO study, its application may become more widespread.

The authors of the study abstracted here have demonstrated that there is increased risk for amputation in persons with significant underlying atherosclerotic peripheral vascular disease. Therefore, this factor must be taken into account in determining which patients with advanced melanomas of the lower extremities may be acceptable candidates for this procedure.

Chemoembolization of a Solitary Liver Metastasis From Malignant Melanoma

Introduction.—With chemoembolization of the liver, embolizing and cytostatic drugs are administered into the hepatic artery to produce selective necrosis of tumor tissue, while leaving healthy liver tissue with sufficient oxygen and nutrients delivered via the portal vein. Two patients underwent chemoembolization of a solitary liver metastasis from malignant melanoma.

Case Report 1.—Male, 40, had a malignant melanoma excised from his right thigh in March 1992. Lymph node metastasis was detected in the right groin in October 1992. A solitary liver metastasis was detected in the right lobe of the liver in July 1993 on routine examination. The patient refused surgery, but consented to chemoembolization. In August 1993, cisplatin solution and the peripheral occluding substance Lipiodol were injected into the right hepatic artery. Gelatin sponge was used to enhance embolization. The patient agreed to a hemihepatectomy in September 1993. Histologic examination revealed a centrally necrotic metastasis of malignant melanoma with intact tumor cells at the edges of the lesion. Metastases to the brain, lung, and abdomen were detected in January 1993. The patient died in April 1994, 9 months after liver metastases.

Case Report 2.—Male, 50, had a malignant melanoma of the midthoracic spine excised in June 1991. A metastasis was detected in the right lobe of the liver in April 1992. It was unresectable because of its size and proximity to the vena cava. Polychemotherapy was initiated, but the patient refused this treatment after a drug rash and polyneuropathy developed. Chemoembolization was performed via injection of cisplatin and Lipiodol into the right hepatic artery. The solitary tumor reduced in size from 12 cm at baseline to 7 cm in 2 months, to 5 cm at 4 months. The course of the disease was stable until March of 1993, when liver metastases increased. Chemoembolization could not be repeated because the supplying arteries were displaced by the tumor. He died in May 1993, 13 months after diagnosis of liver metastasis.

Conclusion.—Chemoembolization selectively administers chemotherapeutic and embolizing agents into tumor-feeding arteries. This is a suitable palliative approach for patients with unresectable disease and offers acceptable quality of life.

► Both patients died shortly after the procedure was undertaken. Certainly, chemoembolization can be considered no more than a palliative form of therapy; early diagnosis with complete excision of the primary tumor offers the patient the best chance for survival.

Possible Role of the Multidrug Resistance-associated Protein (MRP) in Chemoresistance of Human Melanoma Cells

Background.—Conventional chemotherapy is not effective in patients with malignant melanoma. Melanoma-derived cell lines are often very chemoresistant, indicating that cellular mechanisms may mediate the multidrug resistance phenotype. The multidrug resistance-associated protein (MRP), a drug transporter, is associated with resistance to a wide variety of lipophilic drugs. Whether MRP is involved in intrinsic drug resistance in human melanoma was studied.

Methods.—The expression and functional activity of MRP and its effect on chemoresistance were determined in 40 melanoma cell lines, 1 dysplastic nevus-derived cell line, and normal melanocytes.

Findings.—Various levels of MRP mRNA were detected in all melanoma cell lines by reverse transcriptase-polymerase chain reaction. Immunoblot identified the corresponding protein in a high percentage of them. Probenecid, N-ethylmaleimide, and benzbromarone moderately but significantly increased intracellular accumulation of MRP substrate probes corresponding to MRP expression. In addition, MRP gene expression was inversely correlated with the sensitivity of melanoma cell lines to daunomycin and doxorubicin but not to vinblastine, etoposide, and cisplatin.

Conclusions.—Multidrug resistance-associated protein may be a mechanism that causes the multidrug resistance phenotype of melanoma cell lines. Prospective clinical studies are needed to clarify the contribution of MRP to chemoresistance in vivo and its possible value as a prognostic factor in human malignant melanoma.

► Data reported to date regarding expression of the MRP gene in melanoma are contradictory. Until now, no functional MRP assays or correlation studies with chemoresistance had been performed. Thus, the current study by Berger et al., in which they assess MRP expression and function as well as impact on chemosensitivity in a panel of established and standard cell lines, is especially noteworthy. In addition to the observation that the MRP gene is frequently expressed intrinsically in melanoma cell lines, there was marked MRP gene expression in a dysplastic nevus-derived cell line and also in melanocytes. Further studies are necessary to clarify the role of the MRP in normal pigmented cells.