Age- and Site-Specific Variation in the Dermoscopic Patterns of Congenital Melanocytic Nevi: An Aid to Accurate Classification and Assessment of Melanocytic Nevi

Objectives  To describe the dermoscopic features of congenital melanocytic nevi (CMN) and assess whether predominant dermoscopic patterns present in CMN are related to an individual's age (<12 years vs ≥12 years), sex, or lesional site (head, neck, and trunk vs extremities).

Design  Nonrandomized observational study.

Patients  A total of 77 consecutive patients, each with 1 CMN (n = 77 lesions), from an outpatient dermatology clinic. A diagnosis of CMN was established by (1) documentation of a melanocytic nevus during the first year of life or (2) by clinical examination and either clinical history or biopsy findings.

Main Outcome Measures  Images of CMN were evaluated for specific dermoscopic structures and patterns. The distribution of patterns was assessed by age, sex, and lesional site.

Results  Most of the 77 lesions exhibited 1 of the following predominant dermoscopic patterns: reticular (18 lesions [23%]), globular (14 [18%]), or reticuloglobular (12 [16%]). Globular CMN were present in 5 of the 19 individuals who were younger than 12 years (26%) but in only 9 of the 58 individuals 12 years or older (16%). Reticular CMN were seen exclusively in the individuals who were 12 years or older. Congenital melanocytic nevi exhibiting no predominant pattern were more commonly present in the individuals younger than 12 years. Globular CMN were present in 11 head, neck, and trunk lesions (30%) compared with 3 extremity lesions (8%). Conversely, reticular CMN were present in 16 extremity lesions (40%) compared with 2 head, neck, and trunk lesions (5%). The predominant dermoscopic pattern did not vary based on sex. The most commonly observed dermoscopic structures were globules (in 64 lesions [83%]), hypertrichosis (in 61 [79%]), and reticular networks (in 55 [71%]).

Conclusions  Our results suggest that the predominant dermoscopic patterns of CMN vary according to age and lesional site. These differences may inform future studies on the pathogenesis of CMN.

Congenital melanocytic nevi (CMN) are melanocytic nevi that are present at birth. In some cases, CMN may not be evident at birth owing to an initial lack of visible pigment and may become apparent only after the development of pigment months to years after birth.¹ Melanocytic nevi with clinical features consistent with CMN, but without a clinical history to conclusively establish their presence since early life, are termed congenital nevus–like nevi (CNLN). It is estimated that 1% to 6% of infants are born with 1 or more CMN,^2,3 whereas 2% to 6% of the general population have 1 or more CNLN.^4-6

The identification of distinct dermoscopic features of CMN may provide an important diagnostic tool for distinguishing between CMN and other pigmented lesions. Previous studies^7-9 indicate that CMN exhibit distinct dermoscopic features that may be helpful in distinguishing CMN from acquired melanocytic nevi and Becker nevi. There is also evidence that suggests that dermoscopy may be effective in detecting early malignant changes.¹⁰

The recognition of age-, sex-, and lesional site–related differences in the dermoscopic appearance of CMN may facilitate the accurate diagnosis and assessment of subsequent changes observed in follow-up dermoscopy. To date, little is known about the relationship of the dermoscopic appearance of CMN with factors such as age, sex, and lesional site. However, there is recent evidence suggesting that the predominant dermoscopic pattern present in CMN may vary according to an individual's age and lesional site.⁹ Accordingly, the purpose of this study was to provide a comprehensive description of the common dermoscopic features of CMN and to assess whether the predominant dermoscopic patterns present in CMN relate to an individual's age, sex, or lesional site.

Seventy-seven consecutive patients, each diagnosed with 1 CMN or CNLN (n = 77 lesions), who were seen in the outpatient clinic of one of the authors (A.A.M.) during the period of May 1996 to April 2001 were recruited for our study. With the exception of patients with large CMN, almost all other patients presented to the clinic for a primary purpose unrelated to their CMN. A total-body skin examination of each patient was performed, and the presence of any lesions with clinical features suggestive of CMN was noted. A diagnosis of CMN was established by documentation of the presence of a melanocytic nevus in the first year of life by birth record, baby photographs, or the examining physician. In cases in which such documentation was unavailable, a diagnosis of CMN or CNLN was established by the presence of (1) clinical features consistent with CMN, such as homogeneity, mamillated topography, well-defined borders, and hypertrichosis, and (2) either histologic findings consistent with CMN or a clinical history of the lesion having been present since early life. Lesions located on the palms, soles, mucosal, or subungual sites were excluded from our study. Hereinafter, unless otherwise noted, “CMN” will refer to both CMN and CNLN.

After oral consent was obtained from each patient, lesions were photographed with a conventional 35-mm camera to document their clinical appearance. Representative dermoscopic images of each lesion were then obtained using a Dermaphot lens (Heine Optotechnik, Herrsching, Germany) mounted on a conventional 35-mm camera, at 10-fold magnification. If a lesion was larger than the field of view, images were obtained of (1) the border, (2) representative architecture from the center of the lesion, and (3) any other areas of special interest.¹¹ All images were subsequently converted into digital format, and each digital image was evaluated by 2 examiners (L.C. and A.L.C.A.). Clinical images of the lesions were assessed for color, topography, shape, symmetry, borders, hypertrichosis, and homogeneity. Dermoscopic images were evaluated for color, symmetry, homogeneity, and the presence and quality of the structures listed in Table 1. Many of these structures have been previously described as common features of CMN.^7-9,12,15

If the dermoscopic appearance of the lesion revealed that a single pattern was present in greater than one-third of the total area of the lesion, that lesion was classified as having 1 of the following 5 predominant dermoscopic patterns: reticular, globular, reticuloglobular, diffuse background pigment with or without remnant structures, or other. A reticular pattern was defined as a primarily reticular pattern, a globular pattern was defined as a primarily globular pattern, and a reticuloglobular pattern was defined as peripheral reticulation with central globules.¹² A diffuse background pigment with or with remnant structures was defined as a diffuse structureless pattern with or without reticular network fragments and/or remnant globules.¹² A classification of “other” was assigned to lesions exhibiting a single predominant pattern that did not fall within 1 of the other 4 categories. If no single dermoscopic pattern was present in greater than one-third of the total area of any given lesion, the most prevalent patterns present (up to 3) in that lesion were noted. The absolute number and frequency of each clinical feature, dermoscopic feature, and predominant dermoscopic pattern were calculated for the entire study population. The absolute number and frequency of each predominant dermoscopic pattern were then determined for each of the following subgroups of the study population: age (dichotomized at age 12 years), sex, and lesional site (head, neck, and trunk vs extremities). Age was dichotomized at 12 years to coincide with the onset of adolescence. In addition, owing to the small number of head and neck lesions in our study population and the similarity in the timing of melanoblast migration in the head, neck, and dorsal trunk during embryogenesis, head, neck, and trunk lesions were combined into a single category. Differences in frequency between subgroups were evaluated using the χ² test of independence (the Fisher exact test was applied if any expected cell value in the contingency table was <5). In addition, the absolute number and frequency of each predominant dermoscopic pattern were determined for small lesions vs medium and large lesions. A formal multivariate analysis of the predominant dermoscopic pattern was not attempted owing to limitations imposed by the size of our data set. Given that the predominant dermoscopic pattern is a categorical response variable, a multivariate analysis of a cohort of 77 lesions would be likely to result in unreliable parameter estimates.

This retrospective review was approved by the Memorial Sloan-Kettering Cancer Center (New York, New York) institutional review board.

We examined 77 CMN in a study population of 77 individuals, consisting of 36 females (47%) and 41 males (53%) aged 1 month to 70 years (median age, 26 years). Of these 77 individuals, 19 were younger than 12 years (mean age, 3.4 years) and 58 individuals were 12 years or older (mean age, 35.2 years). Of the 77 lesions included in our study, 48 lesions were small (<1.5 cm in diameter) (62%), 25 lesions were medium (1.5-19.9 cm) (33%), and 4 lesions were large (≥ 20.0 cm) (5%). The diameter of lesions ranged from 0.4 to 51.0 cm (median diameter, 1.3 cm). Lesions were located on the head and neck in 6 cases (8%), the trunk in 31 cases (40%), and the extremities in 40 cases (52%).

As summarized in Table 2, most of the lesions in our study appeared clinically brown (73 lesions [95%]) or black (19 [25%]) in color, with macular (51 [66%]) or mamillated (28 [36%]) topography. Most lesions were symmetric (53 [69%]), round or oval in shape (61 [79%]), homogeneous (50 [65%]), with regular borders (57 [74%]), and with hypertrichosis (41 [53%]).

Table 3 presents the frequencies of various dermoscopic features in our study. Nearly all lesions contained dark brown or light brown pigment (75 [97%] and 74 [96%], respectively), whereas blue or white pigment was rarely observed. Globules, present in 64 lesions (83%), were the most common dermoscopic structure observed in our study. Most frequently, a mixed or uniformly small population of diffusely distributed globules was observed. Another common dermoscopic structure was the reticular network, which was present in 55 cases (71%) and most often was of fine quality (in 22 cases [29%]) and distributed throughout the lesion (in 40 cases [52%]). The predominant dermoscopic patterns of the lesions are presented in Table 4.

Overall, most of the lesions exhibited 1 of the following predominant dermoscopic patterns (Table 4): reticular (18 lesions [23%]), globular (14 [18%]), reticuloglobular (12 [16%]), diffuse background pigmentation (3 [4%]), and other (5 [7%]). No single predominant dermoscopic pattern was present in 25 lesions (33%). Of such lesions, 6 contained a reticular component (24%), 7 contained a globular component (28%), and 3 contained a diffuse background pigment component (12%).

The dominant dermoscopic patterns according to age are graphically shown in Figure 1. A single predominant dermoscopic pattern was present in 12 individuals younger than 12 years (63%), and 40 individuals 12 years or older (69%). Predominant dermoscopic pattern varied by the age of the individual (Figure 1) (P = .02): a globular pattern was present in 5 individuals younger than 12 years (26%) but was present in only 9 individuals 12 years or older (16%). A reticular pattern was seen exclusively in individuals 12 years or older. In addition, individuals 12 years or older were equally likely to demonstrate either a reticular pattern (18 [31%]) or no predominant pattern (18 [31%]), whereas those younger than 12 years were more likely to demonstrate no predominant pattern (7 [37%]) rather than a globular pattern (5 [26%]). The predominant dermoscopic pattern also varied by lesional site (Figure 2) (P = .006): a globular pattern was present in 11 head, neck, and axial lesions (30%) compared with only 3 extremity lesions (8%). Conversely, a reticular pattern was present in 16 extremity lesions (40%), compared with only 2 head, neck, and axial lesions (5%). The predominant dermoscopic pattern was not found to vary significantly based on sex (P  = .50) or on size (P  = .20). Small CMN were more likely to demonstrate a reticular pattern (40 lesions [83%]) compared with larger lesions (5 [17%]). However, small CMN were also more likely to demonstrate a globular pattern (31 [64%]) compared with larger lesions (10 [36%]).

Although the differential diagnosis of CMN is relatively limited, the lack of sensitive, well-defined clinical and histologic criteria may impede accurate diagnosis. This is particularly the case for adults, who may have multiple acquired melanocytic nevi and, remote from early childhood, are less likely to accurately report when a given pigmented lesion developed. Furthermore, although a reliable clinical history or birth record indicating that a particular lesion was present at birth may be sufficient to establish the presence of CMN, it does not effectively distinguish an acquired nevus from a tardive nevus, which is not apparent at birth owing to the initial absence of visible pigment. Distinct clinical and histologic features may assist in differentiating CMN from acquired nevi; however, these features are not reliably present in all cases. In fact, some of the histologic and clinical features typically attributed to CMN may on occasion be seen in acquired nevi, particularly in dysplastic nevi.^17-19 This histologic and clinical overlap, most commonly seen in small nevi,²⁰ may therefore preclude accurate classification of melanocytic nevi. The identification of dermoscopic features common to CMN provides an additional diagnostic tool that may help to differentiate CMN, particularly small CMN, from other pigmented lesions. Precise classification of melanocytic lesions is important not only for determining appropriate clinical treatment of such lesions but also for ensuring the integrity of epidemiological studies of CMN and acquired nevi. Although it is possible that some acquired nevi may have been classified as CMN in the study described herein, this is highly unlikely given our stringent inclusion criteria.

Previous studies^7-9 describing the dermoscopic features of CMN have identified globules, reticular networks, dots, focal network thickening, perifollicular hypopigmentation, target networks with globules or dots, and target globules as common features of CMN. Overall, the results of our assessment of dermoscopic features are consistent with those of previous studies (Table 5). However, certain structures, such as target networks with blood vessels, were observed more frequently than in previous studies. Other structures, such as blotches, dots, large globules, and perifollicular hypopigmentation, were less frequently observed. These discrepancies are likely due to differences in magnification, variation in the distribution of lesional size among study populations, minor differences in the criteria used to identify the presence of various structures, and the difficulty of distinguishing among certain dermoscopic features. For example, at times, it was extremely difficult to determine whether the structures within the spaces of a target network were blood vessels, small globules, or dots. In such cases, any structure within the spaces of a target network that appeared red was deemed to be a blood vessel, and any structure within the spaces of a network that appeared tan to brown was deemed to be a globule or dot.

Our results demonstrate dermoscopic structures that, to our knowledge, have not been identified as common features of CMN in previous studies. First, we note the frequent occurrence of milialike cysts, a characteristic dermoscopic finding of seborrheic keratoses.²¹ Second, we define the haloed globule, a dermoscopic structure that has not been previously described (Figure 3). We believe that the haloed globule, which was observed in 13 lesions (17%), is structurally distinct from the target globule. The target globule, which has been described as a globule containing a central dot,^8,9,15 essentially appears as a dot encircled by a halo that is hyperpigmented relative to surrounding background pigment (Figure 4). In contrast, the haloed globule appears as a globule surrounded by a hypopigmented halo.

The most common predominant dermoscopic patterns in our study population were reticular (18 [23%]), globular (14 [18%]), and reticuloglobular (12 [16%]). In addition, we observed an association between predominant dermoscopic pattern and age. A globular pattern was seen commonly in individuals younger than 12 years but far less frequently in individuals 12 years or older. A predominantly reticular pattern was observed exclusively in individuals 12 years or older. Our results are consistent with those reported by Seidenari et al.⁹ In an examination of 384 small and medium CMN, Seidenari et al⁹ similarly found that a globular pattern prevailed in children younger than 11 years, whereas reticular and homogeneous patterns were more common in individuals 11 years or older. Analogous findings have also recently been reported in acquired melanocytic nevi.²² Some researchers may speculate that the age-related differences in dermoscopic patterns observed in our study may reflect dynamic changes in lesions over time; that is, CMN may initially show a globular pattern, which may with age eventually evolve into a reticular or homogeneous pattern. Such a change would correlate histologically with the presence of intradermal nevus cells initially, followed by the presence of intraepidermal and junctional nevus cells later in life.²³ This is consistent with the concept of Hochsteigerung, the upward migration of melanocytes from the dermis to the epidermis.²⁴ It may be plausible to speculate that factors such as intermittent exposure to UV light resulting in the release of growth factors and chemokines from epidermal cells may lead to Hochsteigerung. However, such dynamic changes have not yet been observed, and longitudinal studies of nevus patterns over time would be required to adequately address this hypothesis. In addition, it should be noted that any true dynamic changes in CMN should be independent of lesion size and anatomic distribution.

To our knowledge, to date, no longitudinal studies examining the evolution of the dermoscopic appearance of CMN have been undertaken, and few studies have specifically examined age-related histologic changes in CMN. One study of a series of 38 CMN found no noteworthy correlation between histologic pattern and age.²⁵ But because most of the study participants were in the first decade of life and included only 2 individuals older than 20 years, these findings do not necessarily preclude the possibility that major histologic changes in CMN may become apparent after the first decade of life. In the study by Nickoloff et al,²⁶ biopsy samples of 10 small CMN from individuals younger than 1 year were obtained and compared with findings from repeated biopsies obtained after a mean follow-up period of 10.25 years. The histologic appearance of the findings from the repeated biopsies revealed no noteworthy change compared with the appearance of the original biopsy findings.²⁶ This suggests that major events in the evolution of CMN do not occur in the first and second decades of life but does not exclude the possibility that those events may occur later in life.

In addition, our results suggest that the dermoscopic pattern varies according to the lesional site. A reticular pattern was seen more frequently in lesions located on the extremities, whereas a globular pattern was more commonly seen in head, neck, and trunk lesions. Seidenari et al⁹ similarly observed that lesions of the trunk were more likely to be of the globular type, whereas those of the extremities were more likely to be of the reticular type. Differences in the timing of the migration of melanocytic precursor cells during embryogenesis may be one possible explanation. It has been theorized that during embryogenesis, melanoblasts migrate from the neural crest to the dermis via the paraspinal ganglia and the sheaths of their peripheral nerves.²⁴ Guided by chemotactic factors expressed within the developing epidermis, melanoblasts are thought to be present in the dermis only transiently and soon migrate to the basal layer of the epidermis.²⁷ In general, the sequence of melanoblast migration follows a cephalocaudal and proximal-to-distal sequence: melanoblasts have been identified by light microscopy in the dermis of the scalp, the nape of neck, and the sacrum by the 10th week of gestational life, whereas dermal melanoblasts are not apparent in the extremities and the ventral skin until the 12th week of development.²⁸

Although the pathogenesis of CMN is not well understood, it has been suggested that the perturbation in melanoblast migration and differentiation may result in the formation of CMN.¹ There is evidence that epidermal expression of growth factors, such as stem cell factor, may be important in directing the migration of melanoblasts from the dermis to the epidermis.^27,29,30 For example, in one recent study,²⁹ mice were injected with melanocytes genetically altered to express an endogenous constitutively active mutant Kit receptor. The altered melanocytes were observed to migrate through the dermis, ultimately settling near the dermal-epidermal junction and epidermis, whereas wild-type melanocytes remained within the lower dermis.²⁹ A delay in epidermal development or disruption of cellular pathways resulting in global delay of epidermal expression of growth factors required for dermal migration may have a differential effect on the ultimate location of nevomelanoblasts at different lesional sites. A transient delay in the expression of epidermal migratory factors is more likely to affect nevomelanoblasts destined for the head, neck, and dorsal trunk, which are programmed to reach the dermis at an earlier time. Such a delay may result in the arrest of nevomelanoblasts in the dermis, leading to the formation of nevi of the globular type. In contrast, nevomelanoblasts destined for the extremities, which are programmed to arrive at the dermis at a later time, may be relatively unaffected by a slight delay in the expression of epidermal migratory factors. In such cases, nevomelanoblast proliferation may occur more superficially, leading to formation of junctional nevi.

In our study, we use the term target network to refer generally to a reticular network that contains dots, globules, or blood vessels within its spaces. We note that this network has been previously described as a reticular network centered by a dot.^7-9 However, a variety of different structures, such as dots, globules, and vessels, may occupy the spaces of a reticular network (Figure 5). Because the spaces of the network have a targetlike appearance in each case, strictly speaking, a network may be accurately described as a target network regardless of the exact type of structure present. As a result, we propose redefining the term target network more generally to avoid confusion and to better conform with the definition of target globule (in which “target” similarly refers to the structure with a targetlike appearance).

A limitation of our study is the small size of our sample, which was in part a function of the application of relatively stringent inclusion criteria. The accurate diagnosis of CMN is considerably less certain in the absence of a reliable clinical history. For this reason, we required documentation of CMN by photograph, birth record, or medical record. If such documentation was unavailable, inclusion of a lesion required not only demonstration of clinical morphologic characteristics consistent with CMN but also a clinical history or biopsy finding suggestive of CMN.

In conclusion, our study describes the normal dermoscopic features of CMN, including the haloed globule, a new structure that to our knowledge has not yet been described. Our findings suggest that the predominant dermoscopic patterns of CMN vary according to age and lesional site but not sex. Globular CMN were more commonly observed in individuals younger than 12 years and on the head, neck, and trunk. In contrast, a reticular pattern was seen more frequently on the extremities and in individuals 12 years or older. Familiarity with the dermoscopic patterns commonly present in CMN may facilitate accurate classification of nevi and early detection of malignant change and therefore assist in the determination of whether prophylactic surgical excision of a given pigmented lesion is truly warranted. In addition, the identification of age- and site-related differences in dermoscopic patterns may inform future longitudinal studies examining the pathogenesis and subsequent evolution of CMN.

Correspondence: Ashfaq A. Marghoob, MD, Dermatology Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, 160 E 53rd St, Second Floor, New York, NY 10022 (marghooa@mskcc.org).

Financial Disclosure: None reported.

Accepted for Publication: February 7, 2007.

Author Contributions: Dr Marghoob had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Dusza, Korzenko, Halpern, and Marghoob. Acquisition of data: Changchien, Dusza, Agero, Korzenko, Sachs, Usman, and Marghoob. Analysis and interpretation of data: Changchien, Dusza, Agero, Braun, Halpern, and Marghoob. Drafting of the manuscript: Changchien, Dusza, Agero, Korzenko, Usman, and Marghoob. Critical revision of the manuscript for important intellectual content: Changchien, Dusza, Braun, Sachs, Halpern, and Marghoob. Statistical analysis: Dusza. Administrative, technical, or material support: Dusza, Agero, and Halpern. Study supervision: Dusza and Marghoob.