Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

Jeremy Nikfarjam, MD1, and Earle Chambers, MPH, PhD2 1Department of Surgery, Division of Plastic & Reconstructive Surgery, Montefiore Medical Center, Bronx, NY; 2Department of Family and Social Medicine, Department of Epidemiology & Population Health, Albert Einstein College of Medicine, Bronx, NY

Objective: The objective of this review is to determine what formation from small CMN was determined to be 20.9 by his- size congenital melanocytic nevi (CMN) increases the risk of tory and 10.5 by histology in 238 patients in the case-control malignant melanoma in affected patients. study selected. No malignant transformation was found in a prospective study of 230 individuals with medium-sized mela- Background: Congenital melanocytic nevi are benign prolif- nocytic nevi. Finally, a 5% risk of malignant transformation erations of cutaneous melanocytes apparent at birth or in the was reported in a prospective study of patients with large con- first postnatal weeks. The Kopf system classifies nevi based genital nevi. on size: small, <1.5 cm in diameter; medium, 1.5–19.9 cm in diameter, and large, ≥20 cm in diameter. Great variability Conclusion: All patients should receive total body skin and exists in quantifying the risk of malignant transformation from mucosal surface exams. Patients with small CMN (<1.5 cm in congenital nevi of different sizes. Evidence-based standard diameter) and medium CMN (≥1.5 cm–19.9 cm in diameter) guidelines for clinical investigation need to be established. should be closely observed over their lifetimes and given the option of specialist referral. Finally, patients with large CMN Methods: Literature search included studies on medium, (≥20 cm in diameter) should be referred to specialists upon ini- large, and giant congenital nevi in association with melanoma. tial presentation. More original data and follow-up are needed for maturation of evidence-based clinical recommendations. Results: Three studies pertaining to small, medium, and large congenital nevi are defined. The odds ratio of malignant trans-

INTRODUCTION lesion may become more uniform in color with darker pigmentation, and acquire long, coarse, dark hairs. CMN Congenital melanocytic nevi (CMN) are benign prolif- can appear anywhere on the body, with a slight predi- erations of cutaneous melanocytes clinically apparent lection for giant CMN to appear on the posterior trunk at birth or within the first postnatal weeks (Walton et (Egan et al., 1998). Anatomical location also appears al., 1976). At least 2.5% of white infants are noted to to influence the progression in size of CMN with age. have some type of pigmented lesion at birth, but only According to Rhodes et al. (1996), the growth rate of 1% have biopsy-confirmed nevocellular nevi (Walton et CMN is usually correlated with anatomic site after early al., 1976). Prevalence of CMN ranges from 0.2% to 2.7% infancy, with lesions on the head increasing in size by in lesions less than 1.5 cm in greatest diameter (Pratt, a factor of 1.5, and lesions on other anatomic areas 1953; Walton et al., 1976; Alper et al., 1979) to 0.005% in increasing in size by a factor of 3. lesions greater than 10 cm in greatest diameter (Castilla et al., 1981). The embryological origin of melanocytes found in CMN is the neural crest. As the progression from formation to Clinical appearance of congenital melanocytic nevi var- closure of the neural tube takes place, the cells of the ies with respect to size, morphology, texture, age, and neural crest are delineated as a separate functional entity location. Although considered controversial in the past, from those of the neural tube. The neural crest cells ulti- the Kopf system is now the most widely accepted size- mately give rise to the ectoderm, most notably melano- based standard for the classification of CMN: small, cytes, the leptomeninges, and the peripheral nervous <1.5 cm in greatest diameter; medium, 1.5–19.9 cm in system. Cramer (1998) proposed a theory of the melano- greatest diameter; and large or giant, ≥20 cm in great- cytic differentiation pathway, describing the migration, est diameter (Kopf et al., 1979). In addition to size, CMN plasticity, and differentiation of pluripotent stem cells of can be distinguished on the basis of having a round or the neural crest. The consequences of this theory account oval shape, with a regular, smooth, well-demarcated for aberrations in the melanocyte differentiation path- border. The surface texture can be papular, rugose, peb- way, namely the capability of melanoma to arise in a bly, verrucous, or cerebriform, with combinations of nevus and leptomeninges as observed in neurocutaneous textures in a single lesion or in different lesions. During melanocytosis. Furthermore, nevi may also contain plu- the neonatal period, CMN may appear light in color, ripotent cells capable of differentiation and proliferation hairless and flat, or raised. As the child grows, the same throughout the lifetime of the nevi (Soyet et al., 2007).

The Einstein Journal of Biology and Medicine 59 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

The inception of CMN in utero has not been defini- taken in order to empower clinicians with practical and tively evidenced by imaging modalities or in utero stud- simple measures for evaluating malignancy risk in their ies. Nevertheless, Hamming (1983) reported a case of a patients. Among patients with giant CMN, a lifetime risk “kissing nevus” in which a congenital melanocytic nevus of malignancy of 5% to 60% has been reported in the developed on adjacent parts of the upper and lower literature (Kaplan, 1974; Kang et al., 1992; Sober and eyelids and appeared as a single, contiguous lesion Burstein, 1995). Marghoob et al. (1996) concluded a when the eyelids were closed. Based on this observation, lifetime increased risk of 3.3% for melanoma in large CMN were estimated to develop between the 9th and CMN measuring 20 cm or more in largest diameter, or 20th weeks of gestation. CMN predicted to attain this size by adulthood. Quaba and Wallace (1986) calculated an 8.52% lifetime risk for Histological analysis can aid in differentiating CMN from melanoma in a retrospective analysis of 39 patients with acquired nevi. Although there are no pathognomonic giant CMN. features of CMN, histological analysis can be employed when CMN are suspected in the absence of documen- Among patients with medium and small-sized CMN, the tation of nevus presence at birth (Cribier et al., 1999). risk of transformation to malignant melanoma seems to In contrast to acquired nevi, nevomelanocytes of CMN be controversial. Scalzo et al. (1997) of the Massachusetts proliferate in the subjacent dermal tissue and can General Hospital Pigmented Lesion Clinic reported that spread to the deep dermis and even subcutis (Zaal et in a 30-year period, no child (<20 years old) developed al., 2004). These nevomelanocytes can extend between melanoma in association with a congenital nevus less collagen bundles of the reticular dermis, within and than 5 cm in diameter. Similarly, Swerdlow et al. (1995) around hair follicles, sebaceous glands, vessel walls, and reported that none of their 239 patients with CMN from nerves (Mark et al., 1973; Rhodes et al., 1985; Barnhill 1 to 19 cm in diameter followed over a mean of 25 years and Fleischli, 1995). The depth and the pattern of nevus developed melanoma. However, in a study by Illig et al. cell migration are directly related to the size of the CMN (1985), 52 patients with melanoma that arose in CMN in infants (Barnhill and Fleischli, 1995). Nevertheless, were observed to occur in patients age 18 or older. Based nevus cell migration may continuously occur even after on these results, the authors concluded that melanoma birth. Walton et al. (1976) studied biopsies of small CMN can, in fact, arise from congenital nevi less than 10 cm taken within 72 hours after birth, and concluded that in diameter. the spread of congenital nevi into deeper layers of the dermis and subcutis occurs largely after birth. In addi- These disparate and sometimes conflicting study tion, Mark et al. (1973) demonstrated that the epider- results make it challenging for primary-care physicians mal-dermal junction aspect of a nevus may disappear to develop the appropriate standard of treatment for with time, rendering a dermal nevus without epidermal patients with CMN. Although these congenital lesions involvement. On the other hand, malignant melanoma may not be common in the population, primary-care arising within CMN usually originates at the epidermal- physicians may well be the first to detect such nevi, and dermal junction (Reed et al., 1965). the ones ultimately to decide whether such skin lesions warrant referral to a specialist. More importantly, the Thus, the determination of malignancy is of far greater role of the primary-care physician is critical in under- importance than the question of whether the nevus served areas, where the patient population has poor is congenital or acquired (Zaal et al., 2004). Studies access to healthcare, is largely uninsured, and may have attempting to quantify the percentage of primary mela- recently emigrated from another country with poor noma originating from a CMN have shown inconsistent healthcare. These underserved patients may present results. In a study conducted by Schmid-Wendtner et al. with CMN for the first time well into adulthood, when (2002), 9 of 6,931 (0.13%) patients under the age of 18 risk of melanoma from CMN has been reported. Such years developed cutaneous melanoma associated with delayed presentation may endanger the life of a patient a congenital nevus. Previous to this analysis, Illig et al. with a preventable disease. In addition, the African (1985) reported that at least 2.8% of melanoma cases American and Caribbean American population may be arose from congenital nevi, mostly those that were sin- protected from acquired melanoma due to dark skin gle and less than 10 cm. In a study assessing the risk of pigmentation, but CMN are equally present in these melanoma from CMN in a lower-risk black population, populations, as is the risk of melanoma from CMN. Not Shpall et al. (1994) calculated an actual risk in blacks up only may these factors contribute to physicians’ over- to age 75 to be approximately 1 in 2,000 (0.05%). The looking the importance of CMN in these populations, level of risk reported among these studies results var- but detection of CMN may be more difficult due to dark ies from tenfold to a hundredfold and thus is of lim- skin pigmentation. ited aid to clinicians in gauging the risk of malignancy in patients with CMN. In order for primary-care physicians to provide the best and most efficient treatment for patients with CMN, Several studies assessing the lifetime risk of malignant evidence-based standard guidelines for clinical inves- melanoma from CMN based on size have been under- tigation need to be established. The objective of this

60 EJBM, Copyright © 2012 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians review is to answer the following question: What size patients were diagnosed at the time of interview, 40 CMN increases the risk of malignant melanoma in the patients were interviewed within one year of diagno- affected patients? sis, 23 patients were interviewed between one and two years of diagnosis, and 38 patients were interviewed later than two years after diagnosis. Interviews included METHODS questions to assess the first appearance and morphologi- cal characteristics of the PPLs, along with the source of A literature search was performed using the biomedi- the information obtained (parents versus patient). cal bibliographical database Medline from 1966 to December 2008. Initial searches used the keywords “con- The second group was composed of 234 patients who genital nevus,” “congenital melanocytic nevus,” and did not report a detailed history of a PPL at the tumor “congenital nevocellular nevus.” We included studies site, but whose tumors were histologically examined for on medium, large, and giant congenital nevi in associa- the presence of melanoma-associated nevocellular nevi. tion with melanoma. We excluded all articles that were The 234 patients included 119 males and 115 females; not prospective analyses and/or case-control studies, 1 patient was black and 233 were white, with an age articles with five or fewer CMN patients, articles that range of 12 to 90 years (mean = 47.9 years). Histology dealt only with aspects of treatment, articles that clas- was determined in 234 unselected melanoma specimens sified size by methods other than the Kopf model, and entered into the Harvard Melanoma Registry within 30 data presented in letters to the editor and abstracts. The days of diagnosis during the period from September remaining studies were individually placed into Google 1, 1972, to May 30, 1977. Tumor-associated nevi were Scholar, and were selected based on the number of examined to determine whether they were nevocellular citations received. Manual cross-referencing was per- nevi with congenital features based on the presence of formed. Finally, one representative article was selected nevus cells in the dermis. with respect to CMN classification by size: small CMN, medium CMN, and large CMN. Data on the third group of patients were obtained from a previous study by Walton et al. (1976) in order to determine the prevalence of congenital nevocellular RESULTS nevi by histology in 841 white newborn neonates examined within 72 hours of birth. No prospective studies met our search criteria for small CMN, so a case-control study was included with the Finally, the data from all the groups were used to cal- caveat that a prospective study is preferred over a case- culate an odds ratio for primary cutaneous melanoma control study for our particular assessment. associated with SCN. This was done by dividing the exposure odds of SCN among patients with melanoma by the Three studies pertaining to small, medium, and large exposure odds of SCN among neonates. CMN met our inclusion criteria. The results are shown in Table 1. Results of the Rhodes and Melski (1982) study were determined based on the sample groups. Of the 134 The first article reviewed is by Rhodes and Melski (1982): patients with melanoma in the first group, 20 (14.9%) “Small congenital nevocellular nevi and the risk of cuta- gave a detailed history of the existence of a PPL at the neous melanoma.” This case-control study was created tumor site present since birth. Out of these 20 cases, 5 to determine the odds ratio of melanoma for persons were ascertained by direct interview with parents (2 of with small congenital nevi (SCN). This was done by which cases had photographs during the first 6 months comparing the frequency of SCN in unselected cases of of life), whereas the other 15 were ascertained indi- melanoma to the frequency of SCN in newborn infants. rectly by parental statements in the past. Among these Postpubertal patients with all histological types of cuta- 20 patients, the estimated greatest diameter for the PPL neous melanoma invasive to at least the papillary dermis ranged from 3 to 15 mm (mean = 9 mm +/- 4 mm). Out of were included and divided into three groups. the 234 melanoma specimens in the first group who did not report a detailed history of a PPL at the tumor site, The first group was composed of 134 patients who 19 (8.1%) were designated tumor-associated nevocellu- reported a detailed history of preexisting pigmented lar nevi with congenital features. Furthermore, out of lesions (PPLs) at the tumor site, but whose tumors were these 19 cases, a total of 6 (2.6%) had nevus cells in the not histologically examined for melanoma-associated lower two thirds of the reticular dermis (a highly sensi- nevocellular nevi. The 134 patients included 67 males tive indicator of congenital nevi). Of the 30 patients who and 67 females with an age range of 16–78 years (mean were included in both groups, 2 patients were among = 47.9 years) with histological documentation of primary the 20 PPLs in the first group, and 2 different patients cutaneous melanoma diagnosed between April 1960 were among the 19 tumor-associated nevocellular nevi and September 1979 at the Pigmented Lesion Clinic of with congenital features specimens in the second group. Massachusetts General Hospital. In this study group, 33 Of the Walton et al. (1976) third study group, 7 patients

The Einstein Journal of Biology and Medicine 61 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

TABLE 1. AN OVERVIEW OF LITERATURE ON MELANOMA AND CONGENITAL MELANOCYTIC NEVI (CMN Type according to Kopf et al. [1979]. CMN[n] – Number of subjects with CMN in study. MM – Malignant melanoma)

(0.83%) had nevocellular nevi confirmed by biopsy, all of from the patients in the first subject group was in some which were ≤3.0 cm in greatest diameter. Based on these cases not obtained directly from the patients or families. results, the study determined an odds ratio for cutane- Among those cases in which information was directly ous melanoma in persons with SCN to be 20.9 based obtained, only two had photographic evidence of the on history (95% confidence interval 11.0 to 39.9) and PPL. Thus, great subjectivity in the form of recall bias 10.5 based on histology (95% confidence interval 5.1 to during patient interviews is likely to confound the esti- 21.6). The odds ratio was 3.1 for the 6 specimens taken mation of PPL size. In the second subject group, nearly from the 19 nevocellular nevi with congenital features. two times more patients were present than in the first The authors concluded that congenital nevi of any size group. Though more subjects contribute to the power of may be precursors to cutaneous melanoma. the study results, it is optimal to have close to an equal number of subjects in both groups in order to minimize The Rhodes and Melski (1982) study was among the first the effects of statistical outliers. Among the 30 patients studies to address the risk of malignant melanoma in in the third subject group, 2 reported a history of PPL patients with SCN. The study methodology of distin- (6.66%) and 2 others had positive histology (6.66%). guishing three different subject groups further aided in These results were substantially different from those in minimizing confounding effects of various other param- the first group (14.9%) and the second group (2.6%), eters that were not controlled. In addition, by separat- raising the issue of the accuracy of the study results. ing patients based on history versus tissue diagnosis, the Finally, the authors chose to use the study by Walton et authors were able to note a difference in the odds ratio, al. (1976) to estimate the prevalence rate of SCN. This possibly raising the issue that PPL history alone may is also represents a flaw in study design because the overestimate the odds ratio of melanoma. authors make the assumption that prevalence of SCN in newborn infants is constant and that the prevalence of Although the authors’ conclusion may ultimately be SCN in adults is comparable to that in neonates. supported by their data, the methodology in calculat- ing the odds ratio is inherently flawed. Information The second report considered here, “Risk of melanoma

62 EJBM, Copyright © 2012 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians in medium-sized congenital melanocytic nevi: A follow- the power of the study results. However, certain flaws in up study,” by Sahin et al. (1998), is a prospective, clinical the study design did exist. First, the authors determined cohort study created to assess the risk for melanoma in whether the nevus was congenital by clinical charac- individuals with medium-sized melanocytic nevi clini- teristics. These are inherently subjective inclusion crite- cally judged to be congenital. Methodology consisted ria, as nevi are known to evolve with respect to clinical of a chart review of patients first seen from November appearance. Second, among the 227 patients ultimately 1955 to September 1996 in a private dermatology clinic enrolled in the study, only 174 received follow-up; while in order to identify cases of medium-sized CMN. The fol- those patients lost to follow-up were not known to have lowing data were taken from the chart: each patient’s developed melanoma, the loss of this potentially valu- age at initial visit and last follow-up and the anatomic able data from the study is considerable. Furthermore, location, shape, size, elevation, surface characteristics the average length of follow-up was merely 6.7 years (presence of hair), and color of the patient’s lesion. from an average patient age of 19 years, an interval The only inclusion criterion established was a Polaroid that may not be adequate to observe transformation of photograph taken of the CMN for clinical diagnosis. medium CMN to melanoma. Finally, among the patients Lesions smaller than 1.5 cm in greatest diameter and with follow-up, the authors mentioned that they were those greater than 19.9 cm in greatest diameter were asked “whether an MM [malignant melanoma] had excluded, without an attempt to predict adult sizes of developed within the CMN or elsewhere.” Once again, CMN smaller than 1.5 cm in greatest diameter. In total, without tissue diagnosis, this introduced subjectivity and 16,871 charts were reviewed, yielding 329 patients with bias into the study design because patients may have Polaroid photographs of lesions that met inclusion cri- had undiagnosed melanoma, or, conversely, may mistak- teria for CMN. These patients or parents of the patients enly have stated that they had melanoma without an were contacted and asked whether there had been a actual diagnosis. change in the CMN, whether a malignant melanoma had developed within the CMN or elsewhere, and when and The third article reviewed, “Giant pigmented nevi: how the CMN had been treated. Out of the 329 patients, Clinical, histopathologic, and therapeutic consider- 102 patients with less than one year of follow-up were ations,” by Ruiz-Maldonado et al. (1992), is a prospec- excluded. The remaining 227 patients (230 CMN) were tive, clinical cohort study created to assess the clinical enrolled in the study. An additional follow-up via tele- and histopathologic features, as well as the treatment of phone was carried out with 174 of these patients. giant CMN. This was done by entering all patients clinically diagnosed with giant pigmented nevus (defined Results of the Sahin et al. (1998) study were based on therein as cutaneous hamartomas of nevomelanocytes demographic data and clinical characteristics of the CMN measuring ≥20 cm in diameter) during the period from entered into the study. CMN were first seen at an aver- January 1971 to December 1990 at the Department of age age of 19 years (median, 12 years) with an average Dermatology at the National Institute of Pediatrics in age at last follow-up of 25.5 years (median, 19 years). Mexico City. The sole inclusion criterion was the pres- The average follow-up period was 6.7 years (median, 5.8 ence of a giant pigmented nevus. Each of the 80 patients years). The largest diameters of CMN ranged from 15 to enrolled in the study received a physical examination, 190 mm, with an average of 40 mm. No patients devel- routine laboratory tests (blood cell counts, chemistry oped malignant melanoma in medium-sized congenital studies, urinalysis), and nevus biopsy. Patients with nevi nevi during follow-up. However, three patients who met involving the skin overlying the head or upper portion the inclusion criteria developed melanoma at anatomic of the vertebral column received an electroencephalo- sites other than those were the CMN were located. One gram. Patients with extracutaneous abnormalities were of these patients developed melanoma during follow- examined by appropriate specialists. up, another patient had melanoma before the initial visit, and a third patient had been diagnosed with three Results of the Ruiz-Maldonado et al. (1992) study melanomas, two before his first visit and one during included data for various parameters not mentioned in follow-up. Based on these results, the authors concluded the methods section. The incidence of giant pigmented that medium-sized congenital nevi are not at increased nevi was determined to be 1 in 4,150 among general risk of developing into melanoma. pediatric outpatients, and 1 in 405 among pediatric dermatology outpatients (with 21% of patients lost to Several methodological attributes of the study by Sahin follow-up). The age of patients at the time of consulta- et al. (1998) contributed to the soundness of the authors’ tion ranged from 2 days to 16 years (mean = 20 months). conclusion and limited the confounding factors. First, The prevalence of giant pigmented nevi in girls (65% of the patient data taken from the chart review allowed sample size) was found to be significantly greater than the authors to define the characteristics of medium CMN in boys (35% of sample size) (p <0.001), which was not and melanoma. Second, each nevus was documented by a due to disparities in normal ratios in the population or Polaroid photograph, providing evidence of a skin lesion in the study subjects. In addition, the prevalence of large preceding a melanoma. Third, the relatively large initial nevi in second-degree relatives was found to be signifi- sample size of 329 patients with medium CMN increased cantly higher than the expected frequency in the popu-

The Einstein Journal of Biology and Medicine 63 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

TABLE 2. MALIGNANT TRANSFORMATION (Adapted with permission from Ruiz-Maldonado et al., 1992) Histopathologic Diagnosis Age of Patient Follow-Up

Malignant Melanoma 8 months Deceased

Malignant Melanoma 2 years, 4 months Deceased

Neuroblastoma 7 months Deceased

Minimal 14 years Alive & Well after 1 Year Deviation Melanoma

lation (p <0.001). About 80% of nevi were observed to Tsao et al., 2004). Among these deaths, melanomas aris- be black in color, and 86% of nevi were pigmented and ing from CMN compose a portion, but the rate of this hairy. Predominant locations of nevi included the trunk risk remains a matter of controversy. (47%), limbs (30%), and head (22%). Histopathological features of the 62 of 80 (78%) patients thus examined Several studies have suggested a vast array of risks asso- with giant pigmented nevi were: 64% predominantly ciated with CMN, but the level and nature of the risks intradermal, 12% predominantly compound, and 22% uncovered often varies among individual studies. These cellular blue nevi. Finally, four malignancies were found disparities may well be due to a lack of standardization among the 80 patients. Two of the patients diagnosed in defining CMN, as several differing criteria have been with malignant melanoma at the ages of 8 months and employed, such as greatest size in diameter (Kopf et 28 months died of metastasis despite chemotherapy and al., 1979), the amount of body surface area covered by surgery. One patient diagnosed with neuroblastoma the nevus (Lanier et al., 1976), and ease of excision and at the age of 7 months died as well. Another patient primary closure without deformity (Kaplan, 1974). The was diagnosed with minimal deviation melanoma at adoption of a standard definition is particularly impor- the same position; the nevus was previously treated by tant for primary-care physicians, as they are most likely chemical peel, and the patient was alive and well one to be the first medical professionals to interact with year after the diagnosis (Table 2). The authors concluded patients with CMN. Therefore, we chose the Kopf model that there is a “high risk of malignant transformation of to define CMN in our literature review due to its wide GPN [giant pigmented nevi].” acceptance among clinicians and its relative ease of use in practice. Subsequently, our literature review focused The Ruiz-Maldonado et al. (1992) study has several primarily on ascertaining what size of CMN increases the strong points. The inclusion criteria were clearly defined risk of malignant melanoma in the affected patients. and relatively simple, thus improving generalization of results. The mean age of patients was 20 months, which The study conducted by Rhodes and Melski (1982) deter- is closer to the ideal age of birth in assessing lifetime risk mined an odds ratio for cutaneous melanoma in persons of malignant melanoma. Additionally, this study made with SCN of 20.9 based on history and 10.5 based on his- an attempt to further quantify the prevalence of giant tology, leading the authors to conclude that congenital pigmented nevi according to patient gender and family nevi of any size may be precursors to cutaneous mela- history. Nonetheless, imperfections in the study design noma. The calculation of the odds ratio was inherently did exist. The authors reported the average length flawed due to faulty assumptions made by the authors, of follow-up time to be 4.7 years, leaving readers to along with poor evidence in determining a history of question whether long-term risk is properly addressed PPL in patients with cutaneous melanoma. Sahin et al. in such a short period of time. Moreover, the authors (1998) reported no incidence of malignant melanoma in briefly mention that 21% of patients enlisted were lost 227 patients arising from medium CMN, and concluded to follow-up. Not only does this limit the power of the that medium-sized CMN did not reveal an increased risk study results, but it also lends greater weight to outlier for melanoma. Their study, however, is limited by sub- effects and miscalculation of melanoma risk since those jective inclusion data, loss of many patients to follow- lost to follow-up may have developed malignancy. up, and an inadequate follow-up interval to observe transformation of medium CMN to melanoma were it to occur. Finally, Ruiz-Maldonado et al. (1992) studied DISCUSSION giant CMN in 80 patients and concluded that there exists a high risk of malignant transformation for this type of Both worldwide and in the United States, the incidence CMN. Although the statistical power of the study was of melanoma is increasing more rapidly than that of any weakened by a substantial loss of patients to follow-up, other cancer (Lens and Dawes, 2004; Tsao et al., 2004). the diagnosis of melanoma in four patients is remark- On average, one person dies every hour from metastatic able in view of the diminished sample size. Moreover, melanoma in the United States (Lens and Dawes, 2004; the death of three patients younger than four years of

FIGURE 1. DIAGRAM SUMMARIZING RECOMMENDATIONS FOR PATIENTS WITH CONGENITAL MELANOCYTIC NEVI (CMN) BASED ON SIZE.

age in this study highlights the potential catastrophic risk of malignant transformation (Figure 1). This plan of harm that CMN possess in the event of malignant trans- treatment does not take into account resectability, cos- formation. mesis, or age, since these factors are beyond the scope of this article.

CONCLUSIONS Further research is needed to enhance our understand- ing of this critical issue. Most importantly, efforts should Considering the results of this literature review, we have be made to establish evidence-based standard guide- come to the following conclusions. First, the odds ratio lines for primary-care physicians to confidently address calculated by Rhodes and Melski (1982) constitutes an the issue of CMN in their patients. The adoption of such important initial attempt to quantify the risk of small guidelines is imperative in underserved areas because CMN transforming into melanoma, but more evidence the consequences of CMN are preventable and ulti- is needed to properly evaluate the risk. Second, several mately lifesaving. Thus, we have attempted to address shortcomings present in the Sahin et al. (1998) study this issue and provide clinical recommendations in an likely limit the authors’ claim that medium-sized CMN effort to establish standard guidelines for primary-care do not impart an increased risk for melanoma. Third, physicians when dealing with CMN. In the future, more- despite the suboptimal methodology present in the specific evidence-based clinical recommendations should Ruiz-Maldonado et al. (1992) study, the evidence pre- be possible pending more original data and epidemio- sented therein is substantial and corroborates the risk of logical studies. melanoma from giant CMN.

Based on this analysis, clinicians may consider several REFERENCES points in treating a patient with CMN. The consensus Alper, J., Holmes, L.B., and Mihm, M.C., Jr. (1979). Birthmarks with serious medi- remains that each patient should receive a total body cal significance: Nevocellular nevi, sebaceous nevi, and multiple café au lait skin and mucosal surface exam for skin lesions upon spots. J Pediatr 95(5 pt. 1):696–700. initial presentation. Those patients with small CMN Barnhill, R.L. and Fleischli, M. (1995). Histologic features of congenital mela- measuring <1.5 cm in largest diameter and medium nocytic nevi in infants 1 year of age or younger. J Am Acad Dermatol 33(5 pt. 1):780–785. CMN measuring ≥1.5 cm to 19.9 cm in largest diameter should be photographed and observed for any changes Castilla, E.E., da Graça Dutra, M., and Orioli-Parreiras, M. (1981). Epidemiology of congenital pigmented naevi: I. Incidence rates and relative frequencies. Br J throughout the patient’s lifetime, or given the option Dermatol 104(3):307–315. for specialist referral upon initial presentation. This flex- Cramer, S.F. (1988). The melanocytic differentiation pathway in congenital ibility may be beneficial in the absence of unequivocal melanocytic nevi: Theoretical considerations. Pediatr Pathol 8(3):253–265. indications of melanoma, particularly because patients Cribier, B.J., Santinelli, F., and Grosshans, E. (1999). Lack of clinical-pathological should be given the autonomy to make their own deci- correlation in the diagnosis of congenital naevi. Br J Dermatol 141(6):1004– sions in the context of possible melanoma transforma- 1009. tion. Lastly, patients with large CMN measuring ≥20 cm Egan, C.L., Oliveria, S.A., Elenitsas, R., Hanson, J., and Halpern, A.C. (1998). should be referred to specialists upon initial presentation Cutaneous melanoma risk and phenotypic changes in large congenital nevi: A in view of the strong evidence supporting the increased follow-up study of 46 patients. J Am Acad Dermatol 39(6):923–932.

The Einstein Journal of Biology and Medicine 65 4MEDICAL REVIEWS Congenital Melanocytic Nevi and the Risk of Malignant Melanoma: Establishing a Guideline for Primary-Care Physicians

Hamming, N. (1983). Anatomy and embryology of the eyelids: A review Ruiz-Maldonado, R., Tamayo, L., Laterza, A.M., and Durán, C. (1992). Giant with special reference to the development of divided nevi. Pediatr Dermatol pigmented nevi: Clinical, histopathologic, and therapeutic considerations. J 1(1):51–58. Pediatr 120(6):906–911.

Illig, L., Weidner, F., Hundeiker, M., Gartmann, H., Biess, B., Leyh, F., and Paul, E. Sahin, S., Levin, L., Kopf, A.W., Rao, B.K., Triola, M., Koenig, K., Huang, C., (1985). Congenital nevi less than or equal to 10 cm as precursors to melanoma: and Bart, R. (1998). Risk of melanoma in medium-sized congenital melanocytic Fifty-two cases, a review, and a new conception. Arch Dermatol 121(10):1274– nevi: A follow-up study. J Am Acad Dermatol 39(3):428–433. 1281. Scalzo, D.A., Hida, C.A., Toth, G., Sober, A.J., and Mihm, M.C., Jr. (1997). Kang, S., Milton, G.W., and Sober, A.J. (1992). Childhood melanoma. In: Childhood melanoma: A clinicopathological study of 22 cases. Melanoma Res Cutaneous Melanoma. Balch, C.M., Houghton, A.N, Milton, G.W., Sober, A.J., 7(1):63–68. and Soong, S-J. (eds.), J.B. Lippincott, Philadelphia. pp. 312–318. Schmid-Wendtner, M.H., Berking, C., Baumert, J., Schmidt, M., Sander, C.A., Kaplan, E.N. (1974). The risk of malignancy in large congenital nevi. Plast Plewig, G., and Volkenandt, M. (2002). Cutaneous melanoma in childhood and Reconstr Surg 53(4):421–428. adolescence: An analysis of 36 patients. J Am Acad Dermatol 46(6):874–879.

Kopf, A.W., Bart, R.S., and Hennessey, P. (1979). Congenital nevocytic nevi and Soyer, H.P., Argenzian, G., Hofmann-Wellenhof, R., and Johr, R. (2007). Color malignant melanomas. J Am Acad Dermatol 1(2):123–130. Atlas of Melanocytic Lesions of the Skin. Springer, Berlin and Heidelberg.

Lanier, V.C., Jr., Pickrell, K.L., and Georgiade, N.G. (1976). Congenital giant Shpall, S., Frieden, I., Chesney, M., and Newman, T. (1994). Risk of malignant nevi: Clinical and pathological considerations. Plast Reconstr Surg 58(1):48–54. transformation of congenital melanocytic nevi in blacks. Pediatr Dermatol 11(3):204–208. Lens, M.B. and Dawes, M. (2004). Global perspectives of contemporary epi- demiological trends of cutaneous malignant melanoma. Br J Dermatol Sober, A.J. and Burstein, J.M. (1995). Precursors to skin cancer. Cancer 75(2 150(2):179–185. suppl.):645–650.

Marghoob, A.A., Schoenbach, S.P., Kopf, A.W., Orlow, S.J., Nossa, R., and Bart, Swerdlow, A.J., English, J.S., and Qiao, Z. (1995). The risk of melanoma in R.S. (1996). Large congenital melanocytic nevi and the risk for the development patients with congenital nevi: A cohort study. J Am Acad Dermatol 32(4):595– of malignant melanoma: A prospective study. Arch Dermatol 132(2):170–175. 599.

Mark, G.J., Mihm, M.C., Liteplo, M.G., Reed, R.J., and Clark, W.H. (1973). Tsao, H., Atkins, M.B., and Sober, A.J. (2004). Management of cutaneous mela- Congenital melanocytic nevi of the small and garment type: Clinical, histo- noma. N Engl J Med 351(10):998–1012. logic, and ultrastructural studies. Hum Pathol 4(3):395–418. Walton, R.G., Jacobs, A.H, and Cox, A.J. (1976). Pigmented lesions in newborn Pratt, A.G. (1953). Birthmarks in infants. AMA Arch Dermatol Syphilol infants. Br J Dermatol 95(4):389–396. 67(3):302–305. Zaal, L.H., Mooi, W.J., Sillevis Smitt, J.H., and van der Horst, C.M. (2004). Quaba, A.A. and Wallace, A.F. (1986). The incidence of malignant melanoma Classification of congenital (0 to 15 years of age) arising in “large” congenital nevocellular nevi. Plast melanocytic naevi and malignant transformation: A review of the literature. Reconstr Surg 78(2):174–181. Br J Plast Surg 57(8):707–719.

Reed, W.B., Becker, S.W., Sr., Becker, S.W., Jr., and Nickel, W.R. (1965). Giant pigmented nevi, melanoma, and leptomeningeal melanocytosis: A clinical and histopathological study. Arch Dermatol 91:100–119. Corresponding Author: Address correspondence to Jeremy Nikfarjam, MD Rhodes, A.R., Albert, L.S., and Weinstock, M.A. (1996). Congenital nevome- ([email protected]). lanocytic nevi: Proportionate area expansion during infancy and early childhood. J Am Acad Dermatol 34(1):51–62. Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. No conflicts Rhodes, A.R. and Melski, J.W. (1982). Small congenital nevocellular nevi and were noted for either author. Dr. Nikfarjam, a member of the EJBM Staff, was the risk of cutaneous melanoma. J Pediatr 100(2):219–224. not involved in the review of this manuscript.

Rhodes, A.R., Silverman, R.A., Harrist, T.J., and Melski, J.W. (1985). A histologic Acknowledgements: The authors would like to thank Lorenzo Agoni, MD, MS comparison of congenital and acquired nevomelanocytic nevi. Arch Dermatol for designing the figures in this manuscript. 121(10):1266–1273.