Clinically equivocal melanocytic skin lesions with features of regression: a dermoscopic-pathological study

Background  Benign melanocytic skin lesions may be difficult to differentiate from melanoma both clinically and dermoscopically. One of the most confounding dermoscopic features, commonly seen in melanoma but in our experience also in melanocytic naevi, is represented by the so-called blue–white structures (BWS). Objectives  To evaluate diagnostic significance and histopathological correlates of BWS seen by dermoscopy in a series of clinically equivocal melanocytic skin lesions that were excised. Methods  Patients were recruited from six specialized pigmented lesion clinics in Austria, Italy and Spain over a period of 9 months. All consecutive patients showing one or more melanocytic lesions with BWS, but not classified as melanoma dermoscopically, were included. Each lesion was photographed clinically and dermoscopically. All images were reviewed by one of us and the degree, type and location of BWS evaluated for each lesion. A panel of four experienced dermatopathologists independently reviewed all specimens for diagnosis and one of them evaluated presence and degree of melanosis and/or fibrosis. The main outcome measures were the percentage and histopathological correlates of lesions with different degree, type and location of BWS. Results  All included lesions with BWS (n = 158) showed partial or focal regression histopathologically. One hundred and thirty-five (85·4%) lesions were diagnosed as melanocytic naevi (complete histopathological interobserver agreement), whereas 23 (14·6%) were defined as equivocal because at least one of four pathologists diagnosed the given lesion as melanoma. Only one lesion was diagnosed as melanoma by all four pathologists. The majority of naevi exhibited blue areas (84·4%) with a central distribution (57%) and involving < 50% of the lesion surface (89·6%). By contrast, 78·3% of equivocal lesions revealed a combination of white and blue areas with an irregular distribution (60·9%) and involving > 50% of the lesion surface (47·8%). Conclusions  Using degree and type of BWS, an algorithm was constructed that can be applied for the management of lesions exhibiting dermoscopic features of regression.