Dermoscopy of Lentigo Maligna Melanoma

Report of 125 Cases

P. Pralong; E. Bathelier; S. Dalle; N. Poulalhon; S. Debarbieux; L. Thomas


The British Journal of Dermatology. 2012;167(2):280-287. 

In This Article


LMM, particularly in facial skin, exhibits different dermoscopic patterns from those observed in melanoma occurring elsewhere on nonacral skin. To date, only Stolz and coworkers have previously specifically studied LMM dermoscopic patterns on the face.[6,7,14] This study is the first to investigate them further with the addition of original criteria, in a large population including extrafacial LMM. Our results confirm that the main classical dermoscopic features of extrafacial melanoma cannot be used for the diagnosis of LMM and suggest that LMM-specific dermoscopic features should be well known in order to diagnose this skin cancer accurately at an early stage.

The semiological value of the criteria of Stolz et al. is confirmed in our study as they were present in more than eight in 10 lesions of our series. Although several case reports in the literature have shown the relevance of the classical Stolz criteria,[21–24] the series presented herein is the first to confirm them in a large population. The classical Stolz criteria of follicular invasion were also frequently observed in in situ LMM (excepted obliterated hair follicles), therefore the observation of these features could be useful for early diagnosis, as suggested in the literature.[24] The presence of pigmented rhomboidal structures was the most frequent criterion in our study (69%). It was correlated with facial localization, which suggests its specificity for facial skin anatomy. The presence of pigmented follicular openings was observed in > 50% of cases and was the most frequent feature of in situ LMM. Asymmetrical pigmented follicular openings were not found significantly more frequently than symmetrical ones. An annular-granular pattern was observed in 40% of cases. The distinction between annular-granular pattern and other types of granulation (peppering, for example) was sometimes difficult. Distribution of the dots and associated signs must be considered. The importance of the dots distribution and associated signs in the analysis of dermoscopic granularity has been demonstrated in extrafacial melanomas by Braun et al.[25] They studied granularity in different types of extrafacial melanomas and observed granularity in 26·5% of benign lesions and 93·5% of melanomas. Granularity located at the periphery of the lesion, irregularly distributed or observed in association with red or white colours, was highly statistically significantly in favour of the diagnosis of extrafacial melanoma. Obliterated hair follicles were rarely observed (13%). However, when present, obliterated hair follicles as well as pigmented rhomboidal structures were strongly associated with invasive forms of LMM. This correlation is in accordance with the Stolz progression model, for which pigmented rhomboidal structures and obliterated hair follicles correspond to the two later stages of epidermal invasion by LMM.[6,7]

Four additional original dermoscopic criteria have been investigated in our study: increased vascular network, red rhomboidal structures, target-like images and darkening at dermoscopic examination. Increased vascular network density inside the lesion and presence of red rhomboidal structures were observed in > 40% of cases. These two new vascular-linked dermoscopic features have never been described previously. Increased vascular network density within the lesion was present since the in situ stage of development of LMM, whereas the presence of red rhomboidal structures was correlated to invasive forms. These two criteria could be linked to the development of tumour-induced neovascularization. These two new vascular criteria for the diagnosis of melanoma did not belong to the classical standardized list of vascular structures seen in dermoscopy. In a study reviewing vascular structures in skin tumours, Argenziano et al.[15] separated two groups of vascular features: abnormalities of the vessels (comma-like, hairpin-like, linear-irregular, arborizing, dotted and globular) and reddish coloration (milky-red areas and erythema, respectively localized in a elevated part and regression areas). The recognition of distinctive vascular structures may be helpful in dermoscopy, especially when the classical pigmented dermoscopic structures are absent. A target-like pattern was present in 41% of lesions. It may correspond to a specific form of pilar infundibulum invasion; however, its exact pathogeny and its dermoscopic-pathological correlation must be studied. Interestingly, this feature was also observed with a relatively high frequency in noninvasive LMM, in which few other criteria are present. With a more uncommon frequency (of 25%), darkening at dermoscopic examination was rarer, but can sometimes be helpful.

The presence of a large number of colours could be predictive for melanoma thickness.[26] In our study, in situ LMMs were more often one- or two-coloured and always presented < 4 standardized dermoscopic colours. In contrast, invasive LMM more frequently presented > 5 standardized colours. The presence of a large number of colours is therefore also in favour of a deeply invasive LMM. Vertical growth-associated signs (ulceration, black structureless areas, blue palpable areas) have also been showed to be deep invasion markers in our series of LMM.

The main limitation of this study is that it is not comparative. It indicates the frequency of dermoscopic criteria in a given condition but neither their sensitivity nor their specificity. To date, only two comparative studies about LMM dermoscopic criteria have been published. In 2000, Shiffner et al.[6] showed the existence of LMM-specific dermoscopic criteria. However, more recently in 2010, Akay et al.[5] discussed the specificity of these criteria. In their study, the presence of at least one of the classical Stolz criteria for LMM was used to select a series of 20 LMMs and 69 actinic lentigos. However, acknowledging the fact that the relative prevalence of these two conditions is by far much lower than 20/69 in the general population, we personally believe that if the observation of at least one criterion defined by Stolz et al. selects 3·45 false-positive diagnoses of melanoma for one true positive, this proportion is very acceptable.

In conclusion, LMM presents specific dermoscopic findings, and classical extrafacial dermoscopic criteria for melanoma are not very helpful for its diagnosis. This study confirms the value of the classical follicular invasion criteria defined by Stolz et al. and demonstrates the utility of original new features, pending validation in a study inclusive of benign lesions. Interestingly, these new features are also frequently observed and may facilitate early detection as well as differential diagnosis of LMM: increased vascular density, red rhomboidal structures, target-like pattern and, yet more rarely, darkening at dermoscopic examination. Annular-granular structures, dark outlining of hair follicles, increased vascular density within the lesion and target-like images are present in histopathologically superficial lesions and could be used for early diagnosis and biopsy targeting. In contrast, rhomboidal hyperpigmentation, obliterated hair follicles, a large number of dermoscopic standardized colours (black, white, red, dark brown, light brown, blue, grey) and red rhomboidal structures are associated with invasive LMM and can be used as predictors of thicker lesions. Furthermore, classical melanoma vertical growth phase-associated dermoscopic signs such as blue palpable areas, ulceration and black structureless areas are associated with thicker tumours with poorer prognosis.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: