Nevus Type in Dermoscopy Is Related to Skin Type in White Persons FREE

Acquired melanocytic nevi (AMN), particularly when multiple or uneven in their clinical morphologic features, are considered potential precursors of melanoma or predictors of the development of melanoma.^1- 3 Additional risk factors for melanoma include white race, personal or familial history of melanoma, male sex, age, and fair skin type (ST).⁴ Patients with multiple nevi thus require individualized management that is based on the clinical features of the lesions and the associated risk factors.

The dermoscopic classification of AMN is based on the assessment of 3 main criteria—overall architecture (global pattern), pigment distribution, and color.^5- 7 This classification is of further help in the treatment of patients with multiple nevi because usually a single nevus type is seen in most AMN in the same patient.^7- 8 This is a key point in the differentiation of banal and atypical melanocytic lesions, in line with the concept of the Ugly Duckling sign and the Little Red Riding Hood sign to detect melanoma.^9- 10 According to this comparative approach, the role of dermoscopy in the treatment of patients with multiple nevi is not only in the evaluation of the single lesion but also in the context of a nevi constellation.¹¹ There is evidence that the prevalent nevus pattern seems to be influenced by some patient characteristics, including age and personal history of melanoma.^12- 13

Although fair ST is a recognized risk factor for melanoma, to date there is a lack of information as to whether individuals with different STs are prone to develop different dermoscopic types of nevi. The purpose of the present study was to investigate in a white population whether the prevalent nevus type, as seen at dermoscopy in a given patient, is related to patient ST.

Patients were recruited from 7 specialized pigmented skin lesion clinics in Argentina (Buenos Aires), Austria (Graz), Germany (Konstanz), Italy (Naples), New Zealand (Christchurch), Spain (Barcelona), and the United States (Miami, Fla) between June 1, 2004, and June 30, 2005. The rationale to involve 7 centers worldwide was to reduce a possible selection bias from a prevalent ST in a single geographic area and to obtain a balanced representation of individuals from around the world.

Individuals of any age were eligible for inclusion if they had at least 1 AMN, defined as a melanocytic nevus with a macular component and a diameter greater than 2 mm.¹⁴ Individuals were excluded if they had a history of intentional or prolonged sun exposure in the 4 weeks before the examination or had tanned skin; lesions on the head or neck including the face, acral sites, or mucosae; or features of Spitz nevus, agminated nevus, dermal nevus, recurrent nevus, halo nevus, or congenital nevus, including blue nevus and nevus spilus. Each individual gave oral informed consent for study participation.

Study patients were enrolled until a total of 100 in each of the 7 centers was attained. All patients were white, and their ST was scored according to the Fitzpatrick classification (Table 1).¹⁵ All AMN in each patient were clinically and dermoscopically examined, and 1 representative nevus was selected. The representative nevus was defined as a nevus with a dermoscopic typology (global pattern, pigment distribution, and color) repeatedly seen in the same patient (ie, in >40% of all nevi in a single individual). A digital dermoscopic image of the selected nevus was obtained using a DermLite FOTO lens (3Gen LLC, Dana Point, Calif) or Delta 20 (Heine Optotechnik GmbH & Co, Herrsching, Germany) coupled with a digital camera (Nikon CoolPix 4500; Nikon Corp, Tokyo, Japan) or a standardized digital system (MoleMax; Derma Medical Systems, Vienna, Austria). For each included nevus, patient demographic data including sex, age, total nevus count and location of nevi, and enrolling center were obtained.

All digital images were examined by one of us (I.Z.) blinded to patient ST and were evaluated for the following 3 dermoscopic criteria: global pattern, including globular, globular reticular, globular homogeneous, reticular, reticular homogeneous, and homogeneous; pigment distribution, including uniform, central hyperpigmentation, central hypopigmentation, eccentric hyperpigmentation, eccentric hypopigmentation, and multifocal hypopigmentation or hyperpigmentation; and color, including white, red, light brown, dark brown, blue, gray, and black. Nevi that did not reveal any of the morphologic criteria described above were classified as being nonspecific. When a nevus exhibited more than 1 color, the prevalent color (seen in >50% of the lesion) was considered. Nevi that did not exhibit any of the dermoscopic patterns, images of insufficient quality or that showed only part of the lesion, and lesions that did not meet the inclusion criteria were excluded from the analysis.

For each of the 3 dermoscopic criteria (global pattern, pigment distribution, and color), the frequency distribution among the different STs was calculated by univariate analysis. Differences between proportions were tested using the χ² test.

To test the correlation between ST and each of the 3 dermoscopic criteria, considering age, sex, and enrolling center as potentially confounding factors, the likelihood ratio test comparing logistic regression models including age, sex, and enrolling center with and without ST was used. Because 15 comparisons were made, at this stage of inference, we set the statistical significance level at .003 (.05/15 = .003).

The saturated model with age, sex, and enrolling center was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for the ST categories II through IV, with ST I as the reference category. When the expected frequency for a given dermoscopic criterion was lower than 5, ST categories I and II, and III and IV, were combined for the purpose of multivariate analysis, using categories I and II as the reference group. Because many comparisons were made, with 33 ORs calculated, we used the Bonferroni correction considering significant ORs with P<.002 (Wald test). Setting the statistical significance level at .003, the Wald test for trend was performed where appropriate, using ST as the continuous variable (coded I, II, III, or IV). To assess a possible interaction between ST and enrolling center for each dermoscopic criterion, the likelihood ratio test was used, comparing logistic regression models with and without the interaction term. STATA statistical software was used for all analyses.¹⁶

Of 700 AMN included in the study, 680 were eligible for analysis. Table 1 gives the number of study participants assigned to the 4 STs using the Fitzpatrick classification scale. The study population consisted of 367 female subjects (54.0%) and 313 male subjects (46.0%) with a median age of 34 years (age range, 8-88 years). No significant differences in the age of individuals within the different ST groups were observed. The median age of individuals with ST I was 36.7 years (age range, 11-64 years), with ST II was 32.8 years (age range, 8-72 years), with ST III was 35.0 years (age range, 6-88 years), and with ST IV was 36.7 years (age range, 15-64 years).

The most prevalent ST was ST III, followed by ST II, ST I, and ST IV (Table 1). The total nevus count most frequently observed was 11 to 50 (57.5%), followed by a nevus count of more than 50 in 189 individuals (27.8%) and fewer than 10 in 100 participants (14.7%). Of the 680 representative nevi selected for dermoscopic evaluation, 592 (87.1%) were located on the trunk and 88 (12.9%) were located on the extremities.

Dermoscopic features scored in our 680 AMN are given in Tables 2, 3, and 4. Among the 3 main categories of dermoscopic features (global pattern, pigment distribution, and color), reticular pattern (39.1%, n = 266), uniform pigment distribution (36.2%, n = 245), and light brown color (71.6%, n = 487) were most frequently seen in the overall sample of AMN. The uniform pigment distribution was most prevalent in patients with a total nevi count of fewer than 50, whereas individuals having more than 50 nevi demonstrated a similar prevalence of the uniform pigment distribution, central hyperpigmentation, multifocal hypopigmentation or hyperpigmentation, or central pigmentation (data not shown). Ten nevi were classified as nonspecific.

Tables 2, 3, and 4 give the absolute numbers, percent prevalences, and ORs adjusted by age, sex, and enrollment center.

Of the 6 global patterns, the reticular pattern was the most frequent in all STs (Table 2). Its prevalence increased with increasing darkening of the skin (OR, 2.80; 95% CI, 1.08-7.26; P  = .03, for ST IV compared with ST I). No significant differences were found between the other 5 global patterns and the 4 STs.

Central hypopigmentation was the most frequent type of pigment distribution of nevi in patients with ST I (41.2%; P<.001) (Table 3). Its prevalence gradually decreased with increasing darkening of the skin (Figure 1), and decreased to 7.0% in ST IV (test for trend, P<.001).

Central hyperpigmentation showed an opposite trend, gradually decreasing with increasing lightening of the skin (Figure 1) (test for trend, P<.001), with higher prevalence observed in ST III (OR, 4.45; 95% CI, 1.44-13.74) and ST IV (OR, 13.37; 95% CI, 3.67-48.63), compared with ST I. Although most nevi in ST II and III exhibited uniform pigment distribution (32.4% and 40.3%, respectively), multifocal hyperpigmentation or hypopigmentation was significantly associated with ST II (OR, 5.92; 95% CI, 1.96-17.87) and ST III (OR, 7.24; 95% CI, 2.28-22.92), compared with ST I.

Light brown and dark brown showed remarkable differences in the 4 ST groups (Table 4). Light brown gradually decreased with increasing darkening of the skin (test for trend, P<.001), and dark brown gradually increased with increasing darkening of the skin (test for trend, P<.001) (Figure 1). Most nevi were light brown in ST I and II, whereas dark brown was more frequent in ST III (adjusted OR, 9.32; 95% CI, 2.08-41.71) and ST IV (adjusted OR, 16.72; 95% CI, 3.34-83.63) than in ST I and II.

The most striking result of our study is that individuals with different STs are prone to different nevus types in terms of pigment distribution, color, and, to some extent, global dermoscopic pattern. In ST I, the prevalent nevus type was characterized by central hypopigmentation and light brown color (Figure 2), whereas in ST IV, nevi exhibited central hyperpigmentation and dark brown color (Figure 3). In ST II and III, a significant association was seen with nevi exhibiting multifocal hyperpigmentation or hypopigmentation (Figure 4). Although the pigment network showed higher prevalence in darker STs, this global pattern was the most frequent in all ST groups. This is not surprising given the median age of our patients (34 years). As previously demonstrated, the reticular pattern is the most frequently seen at this age.¹²

In Figure 1, the changing dermoscopic profile for pigment distribution and color is shown for the various STs. The changes are represented by a gradual increasing or decreasing trend among the 4 STs.

The observed differences in pigment distribution and color among the 4 ST groups might be explained by the different dispersion or activation levels of melanosomes in melanocytes and their ratio of pheomelanin to eumelanin production in the various STs.¹⁷ In general, melanosomes in fair STs are clustered and small, less activated, and produce higher levels of pheomelanin (ie, red to yellow pigment). In contrast, in dark STs, melanosomes are homogeneously dispersed, large and activated, and produce higher levels of eumelanin, which is responsible for the dark brown to black coloration.

The particular nevus type we found in ST IV, known as black or hypermelanotic nevus, is a lesion representing the Ugly Duckling sign because it usually draws the clinician's attention because of its dark color.¹⁸ Dermoscopically, this nevus is characterized by a central black blotch (black lamella), corresponding histopathologically to a pigmented parakeratosis, that covers the underlying dermoscopic structures and, thus, may sometimes cause diagnostic difficulty. However, the central black lamella can be easily removed with tape or plaster, enabling visualization of a typical network (reticular pattern) and the diagnosis of a black nevus with increased confidence.

The prevalence of black nevi in dark-skinned individuals might be interpreted as a peculiar biologic mechanism of nevi to protect the melanocytes against UV irradiation, with the black lamella representing a sort of natural sunshade. The sunshade theory of black nevi might be further supported by previous observations regarding UV-dependent dermoscopic changes in melanocytic nevi.^19- 22 The UV-induced changes constitute a significant darkening of color, increased fading of borders, increase or decrease of a prominent pigment network, reduction of hypopigmentation, and development of new black dots or blotches. At electron microscopy, the black dots or blotches correspond to melanosomes in keratinocytes, and at histopathologic analysis to pigmented parakeratosis (ie, transepidermal pigment transfer and elimination), which can be explained by UV-induced increased melanin synthesis of melanocytes, resulting in a subsequent increased pigment transfer between melanocytes and keratinocytes (melanocyte-keratinocyte unit).¹⁹ However, because irregular black dots and blotches may also be associated with melanoma,⁶ it is generally recommended that nevi not be examined after intense UV exposure to avoid unnecessary excision of benign nevi. Because we included only individuals without sun exposure before the examination, it seems remarkable that the predominant nevus type in dark-skinned individuals exhibits the same basic dermoscopic patterns as UV-irradiated nevi, such as dark brown color, prominent pigment network, and the central black blotch. This observation allows, to some extent, the theoretical conclusion that UV exposure may neither cause nor significantly influence the dermoscopic profile of the prevalent nevus type in individuals with a dark ST, but this finding may be explained by an intrinsic genetic or racial/ethnic background.

A possible limitation of our study is the prevalence of nevus patterns in our study population, which might have been biased by the inclusion of a single representative nevus in each individual. We cannot rule out that some study participants might have exhibited additional dermoscopic types of nevi. However, the literature has repeatedly addressed the concept that the predominant nevus type is valid and reliable for the management of individuals with multiple nevi.^7- 8,12 In this context, our study was not based on a previously established and well-defined hypothetical consideration but had a purely descriptive-investigative purpose. The risk that the investigators were prone to select representative nevi to prove or disprove an underlying hypothesis seems to be low. Furthermore, that a single observer analyzed these data might introduce a certain risk of misclassification bias. In the field of morphologic diagnosis, there is always the intrinsic problem related to the human subjective evaluation of criteria, and this problem is difficult to resolve even if more independent observers are involved.

In conclusion, our study demonstrates that individuals with different STs exhibit significant differences in their prevalent dermoscopic nevus type. Our results could be the basis for further research into the cellular mechanisms of UV protection, with particular interest in the black nevus. In addition, further studies are needed to better clarify the role of UV irradiation on nevogenesis and its influence on dermoscopic nevus patterns.

Correspondence: Iris Zalaudek, MD, Department of Dermatology, Medical University of Graz, Auenbruggerplatz 8, 8036 Graz, Austria (iris.zalaudek@meduni-graz.at).

Financial Disclosure: None reported.

Accepted for Publication: July 18, 2006.

Author Contributions: Dr Argenziano had full access to all 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: Zalaudek, Malvehy, Puig, and Soyer. Acquisition of data: Zalaudek, Argenziano, Mordente, Moscarella, Blum, Cabo, Di Stefani, Hofmann-Wellenhof, Langford, Johr, Malvehy, Kolm, Sgambato, and Puig. Analysis and interpretation of data: Zalaudek, Corona, Sera, Blum, Cabo, Malvehy, and Puig. Drafting of the manuscript: Zalaudek and Sera. Critical revision of the manuscript for important intellectual content: Zalaudek, Argenziano, Mordente, Moscarella, Corona, Sera, Blum, Cabo, Di Stefani, Hofmann-Wellenhof, Malvehy, Kolm, Sgambato, Puig, and Soyer. Statistical analysis: Corona and Sera. Obtained funding: Cabo. Administrative, technical, and material support: Zalaudek, Argenziano, Mordente, Moscarella, Cabo, Hofmann-Wellenhof, and Sgambato. Study supervision: Zalaudek, Blum, Cabo, Di Stefani, Malvehy, Puig, Soyer, and Kerl.