Pigmented lesions are a common reason for consultation in routine clinical practice, and the number of patients consulting for these lesions has multiplied in recent years thanks to melanoma prevention campaigns and other factors. Melanoma in situ has a better prognosis than invasive melanoma, and therefore early diagnosis results in longer survival and lower morbidity and mortality. The ABCD (Asymmetry, Border irregularity, Color, Diameter) algorithm or rule has been used in clinical practice for years,1 but it is not adequate for pigmented lesions and does not always distinguish between melanocytic and nonmelanocytic lesions. Furthermore, it is not always reliable in the evaluation of small lesions.2 Dermoscopy, which is a widely used, cheap, and simple technique, has led to enormous advances in the diagnosis of pigmented skin lesions. Dermoscopes are equipped with a conventional or polarized light source that allows lesions to be studied at a higher magnification and are associated with a 10% to 30% improvement in diagnostic accuracy in clinical settings.3,4 The aim of the current study was to describe the dermoscopic features of a series of melanomas and determine which features are associated with melanomas in situ and which are associated with invasive melanomas.

We performed a retrospective study of 200 primary cutaneous melanomas diagnosed at the dermatology department of Hospital General Universitario Gregorio Marañón in Madrid, Spain; all the tumors had been examined by dermoscopy prior to surgical excision between January 1, 2007 and December 31, 2009. The diagnosis was confirmed by histology in all cases. The preoperative diagnosis was established using both clinical criteria (ABCD rule) and dermoscopic criteria, using the standardized two-stage diagnostic procedure proposed by the virtual Consensus Net Meeting on Dermoscopy.5 All the dermoscopic images were taken using the DermLite dermoscope (magnification ×10) and evaluated by 2 operators (CCB and JAAI). Note was also taken of epidemiological characteristics (patient sex, age), clinical features (skin phototype and size and location of the melanoma), and histologic features (subtype, Breslow thickness, and Clark level).

We performed a descriptive study of the dermoscopic features of the melanomas analyzed and investigated their association with epidemiological, clinical, and histologic variables. The χ2 test was used to test for associations between qualitative variables. When this was not possible (i.e., for variables with few cases), we simply examined correlations to test for possible differences. Statistical significance was set at a P value of less than .05. The data were analyzed using descriptive and analytical techniques in SPSS, version 13.0 for Windows.

The diagnosis of thin melanomas (<1mm) has increased in recent years, probably partly due to the combined effects of melanoma prevention campaigns and the use of dermoscopy.6–8 Melanomas tend to have multiple colors, with most lesions showing 3 or more colors when examined by dermoscopy.9 The multicomponent pattern is the most common dermoscopic feature described in most studies of melanoma.10–12 This pattern has the strongest association with melanoma (odds ratio of 4.3) and the detection of different colors and structures should therefore alert to a possible diagnosis of malignancy.5,10–14 The reticular pattern in melanomas is an atypical pattern, with a broad, irregular mesh, typically seen at the periphery of the lesion. Subtle changes in the characteristics of the pigment network can help to establish a diagnosis of early-stage melanoma.15 Melanoma should also be suspected when faced with a nonspecific pattern (absence of dermoscopic features) combined with red colors or multiple vascular structures. The nonspecific pattern has been observed in between 7% and 15% of melanomas, depending on the series.10,16 Lesions exhibiting this pattern should be excised for histologic examination.

Homogeneous areas were the most common dermoscopic feature in our series of melanomas, and have been described in between 65% and 80% of melanomas in other series.10,12 They are areas of diffuse pigmentation, with no other identifiable structures. The distribution of these structureless areas is important, as a regular diffuse distribution is associated with benign melanocytic lesions, while an irregular, focal distribution is more characteristic of malignant lesions.5,10,11

In our series, white-blue structures were more common in invasive melanomas than in melanomas in situ. These structures correspond to regression areas and are frequently associated with melanoma, although they may also be seen in certain benign lesions with regression. Most experts, however, recommend the excision of lesions with white-blue structures occupying over 10% of the tumor area.17–20

The pigment network is the most characteristic dermoscopic feature of melanocytic lesions.5 The thickness of the network lines, together with size and spacing of the mesh holes, helps to differentiate between the typical pigment network (thin, uniformly spaced mesh) seen in benign melanocytic lesions, such as junctional nevi and solar lentigines, and the atypical pigment network (thick, prominent lines and irregular holes) seen in malignant melanocytic lesions. The atypical pigment network was observed in over 50% of melanomas in our series, which is very similar to previously reported rates.10,12 The atypical network is the most common dermoscopic finding in melanomas in situ, and is key for the diagnosis of very early stage melanoma. Slight changes in the pigment network of thin melanomas can be the first dermoscopic sign of a possible diagnosis of melanoma.15

Irregularly sized and distributed dots and globules are observed by dermoscopy in up to 44% of melanomas, but all the dots and globules in our series had a uniform size and distribution. These irregular structures are observed in both benign and malignant melanocytic lesions, and their distribution (regular or irregular), location (in the center or at the edge of the lesion), and color all play an important role in determining the nature of the lesion.11,12

Other structures included in many of the diagnostic algorithms for melanoma, such as the blue-white veil and radial streaming, were less common in our series. The blue-white veil was observed in 22% of lesions, and was more common in invasive melanomas than in melanomas in situ. This veil has been described as being highly specific for melanoma,5 but it may also be seen in other nonmelanocytic lesions such as basal cell carcinoma.22,23 Furthermore, the proportion of lesions exhibiting this sign varies considerably from one study to the next, suggesting possibly that it is a somewhat subjective marker, characterized by high interobserver variability and a lack of reproducibility.5,10,12 Nevertheless, it is important to correctly interpret this sign, as it has been particularly linked to invasive melanomas. Radial streaming was uncommon in our series, and the projections were irregularly distributed in all cases. Most cases were observed in superficial spreading melanomas. Radial streaming and pseudopods are an important dermoscopic finding for the diagnosis of melanocytic lesions.11,21 Regularly distributed projections at the periphery of the lesion, giving rise to what is known as the starburst pattern, are a characteristic feature of Reed/Spitz nevus, while irregular or asymmetric projections are highly suggestive of melanoma in the radial growth phase.

Sixty percent of the melanomas in our series had vascular patterns (Fig. 3). Milky-red areas and linear irregular and dotted vessels were more common in invasive melanomas than in melanomas in situ. All the vascular structures identified were more common in nodular and su-perficial spreading melanomas. Vascular structures can be difficult to identify and may be missed in more heavily pigmented lesions due to the presence of other structures. Dotted and linear irregular vessels are the most common vessels seen in melanomas, and invasive melanomas in particular.24,25 The presence of these vessels, together with milky red areas, should raise suspicion of melanoma.

Figure 3.

Vascular structures. A, Milky-red areas corresponding to increased vascularization in an invasive melanoma. B, Large, linear irregular vessels in a nodular melanoma with a Breslow thickness of 2.9mm.

(0.11MB).

Melanoma of the face has several distinctive dermoscopic features not seen in other types of melanoma (Fig. 4).26–29 In our series, at least 1 of the 4 Stolz criteria was present in 81% of LMMs. We did not detect any dermoscopic features associated with benign lesions, such as milia-like cysts, yellow opaque areas, or light brown fingerprint-like structures. Asymmetrically pigmented follicular openings were more common in melanomas in situ than in invasive melanomas of the face. Based on the Stolz progression model for LMM, these openings would be the earliest dermoscopic signs of facial melanoma. In the next stage, blue-slate-gray dots would join to form short lines that eventually form rhomboidal structures. Finally, atypical melanocytes would invade the hair follicles and the rhomboidal structures would become thicker and more prominent until the follicular openings were completely obliterated; this stage is viewed as homogenous pigment areas on dermoscopy.26 The presence of a single Stoltz criterion is not sufficient for the diagnosis of malignancy, as these criteria can also be seen in benign lesions. Of the 4 criteria, asymmetrically pigmented follicular openings and rhomboidal structures have been described as being more characteristic of LMM and are rarely found in other pigmented lesions.11,29 Blue-gray dots may also be observed in other pigmented lesions of the face.28 They correspond to areas of regression and can be found in all lesions that exhibit histologic regression. One distinguishing feature is that the annular-granular pattern and blue-gray dots tend to be more homogeneous and regular in these more “benign” lesions than in LMM. Additionally, these lesions do not have asymmetrically pigmented follicular openings or isobar structures, features that correspond to the invasion of hair follicles by atypical melanocytes. These dermoscopic findings may be useful for distinguishing between LMM and other pigmented lesions of the face.

Figure 4.

Typical dermoscopic features of lentigo maligna melanoma of the face. A, Asymmetrically pigmented follicular openings and annular-granular pattern. B, Rhomboidal structures. C, Homogeneous facial areas. D, Target structures corresponding to the invasion of the hair follicles by atypical melanocytes (ovoid).

(0.26MB).

Acral lentiginous melanoma was uncommon in our study. The most common global patterns were the parallel crest pattern in palmoplantar melanoma and the linear irregular pattern in melanoma of the nail. Our findings are similar to previous reports of acral lentiginous melanoma in the literature.30

Several dermoscopic structures were more common in invasive melanomas than in melanomas in situ (Fig. 5). Blue-gray, red, and white colors, a combination of these colors, and the multicomponent pattern were more common in invasive melanomas than in melanomas in situ. The most common local dermoscopic finding in melanomas in situ was the atypical pigment network. All the other dermoscopic features were found in just a small proportion of melanomas in situ. Irregularly distributed dots and globules, white shiny structures, homogeneous areas, the reverse pigment network, white-blue structures, the blue-white veil, and vascular patterns and structures were more common in invasive melanomas that in melanomas in situ.

Figure 5.

Several dermoscopic characteristics are more common in invasive melanoma than in melanoma in situ. A, Blue-gray, red, and white colors are more common in invasive melanoma. B, A blue-white veil in an invasive melanoma. C, White shiny structures. D, Reverse pigment network.

(0.23MB).

One of the limitations of this study is a possible lack of objectivity in interpreting the dermoscopic findings as this was a retrospective study in which there was a known histologic confirmation of melanoma in all cases. Furthermore, some lesions that were not examined by dermoscopy prior to surgical excision and histologic study (because of a low or nonexistent suspicion of melanoma) may have been excluded.

In conclusion, dermoscopy permits the visualization of structures that can facilitate the clinical diagnosis of pigmented lesions. While melanoma must always be confirmed histologically, we identified certain dermoscopic features that were more common in invasive melanomas than in melanomas in situ that could be used as a complementary diagnostic tool.

The authors declare that they have no conflicts of interest.