Clinical and Dermoscopic Features Associated With Difficult-to-Recognize Variants of Cutaneous Melanoma : A Systematic Review

David Moher and colleagues introduce PRISMA, an update of the QUOROM guidelines for reporting systematic reviews and meta-analyses

To determine the predictive dermoscopic features of amelanotic and hypomelanotic melanoma. A total of 105 melanomas (median Breslow thickness, 0.76 mm), 170 benign melanocytic lesions, and 222 nonmelanocytic lesions lacking significant pigment (amelanotic, partially pigmented, and light colored) were imaged using glass-plate dermoscopy devices and scored for 99 dermoscopic features. Diagnostic models were derived from and tested on independent randomly selected lesions. Predominantly hospital-based clinics from 5 continents. Sensitivity, specificity, and odds ratios for individual features and models for the diagnosis of melanoma and malignancy. The most significant negative predictors of melanoma were having multiple (>3) milialike cysts (odds ratio, 0.09; 95% confidence interval, 0.01-0.64), comma vessels with a regular distribution (0.10; 0.01-0.70), comma vessels as the predominant vessel type (0.16; 0.05-0.52), symmetrical pigmentation pattern (0.18; 0.09-0.39), irregular blue-gray globules (0.20; 0.05-0.87), and multiple blue-gray globules (0.28; 0.10-0.81). The most significant positive predictors were having a blue-white veil (odds ratio,13; 95% confidence interval, 3.9-40.0), scarlike depigmentation (4.4; 2.4-8.0), multiple blue-gray dots (3.5; 1.9-6.4), irregularly shaped depigmentation (3.3; 2.0-5.3), irregular brown dots/globules (3.2; 1.8-5.6), 5 to 6 colors (3.2; 1.6-6.3), and predominant central vessels (3.1; 1.6-6.0). A simple model distinguishing melanomas from all nonmelanomas had a sensitivity of 70% and a specificity of 56% in the test set. A model distinguishing all malignant lesions from benign lesions had a sensitivity of 96% and a specificity of 37%. Conclusion Although the diagnostic accuracy of dermoscopy for melanoma lacking significant pigment is inferior to that of more pigmented lesions, features distinguishing the former from benign lesions can be visualized on dermoscopic evaluation.

This review presents the main challenges encountered when diagnosing unusual variants of malignant melanoma with the aim of raising awareness to allow application of the most appropriate treatment strategies. Although these melanomas are often rare, their misdiagnosis potentially jeopardizes patients’ health and survival, and has medicolegal implications. The clinical and histologic presentations of melanoma vary greatly, and assessment of uncommon melanomas can be difficult for practitioners because of their scarcity and resemblance to other dermatologic entities. The most problematic melanoma types are desmoplastic melanoma, polypoid melanoma, primary dermal melanoma, verrucous malignant melanoma, pigmented epithelioid melanocytoma, mucosal melanoma, follicular melanoma and melanoma with non-melanocytic differentiation. The two most difficult-to-diagnose subtypes of melanoma are the nevoid and the amelanotic melanomas. Some specific attributes of these variants can be more easily recognized with digital dermatoscopy, facilitating early detection and possibly avoiding invasive procedures. Key cases with the most notable clinical, dermatoscopic, and histopathologic features are presented, highlighting the practical issues of making an accurate diagnosis and choosing the best therapy.