Pitfalls of PRAME Immunohistochemistry in a Large Series of Melanocytic and Nonmelanocytic Lesions With Literature Review

Abstract

Abstract Preferentially expressed antigen in melanoma (PRAME) immunohistochemistry is currently used to facilitate distinction of benign and malignant melanocytic proliferations. We hypothesized that evaluation of 1 institution's experience with PRAME labeling in a large number of consecutive cases might elucidate additional strengths and potential pitfalls and reveal base rates of positivity versus negativity in 1 academic practice. Pathology reports for all specimens on which PRAME labeling was performed at our institution between January 2021 and May 2022 were retrieved from our database. Eighty percent of conventional malignant melanomas were labeled diffusely positive with PRAME; there were no significant differences in mean age, sex, site, Breslow depth, ulceration status, or American Joint Committee on Cancer pathological tumor stage when comparing diffusely PRAME-positive malignant melanomas with those that lack diffuse labeling. Although no banal melanocytic nevi were labeled with PRAME, 13% of dysplastic nevi were diffusely PRAME positive, with junctional proliferations, severe atypia, male gender, and older age being associated with PRAME positivity. As some but not all ambiguous melanocytic lesions in which malignancy could not be excluded based on morphology alone were diffusely PRAME positive, PRAME's accuracy in predicting malignancy remains unclear to the authors; further study is needed to assess the precision to which PRAME immunohistochemistry can separate benign borderline lesions from their malignant counterparts. Among nonmelanocytic lesions, some poorly differentiated tumors, including atypical fibroxanthomas, can be PRAME positive. This series underscores the importance of clinicopathologic correlation and shows that diffuse PRAME positivity is highest in conventional malignant melanomas (∼80%, or 8 of 10 lesions), is seen in about half of challenging borderline lesions at our institution, and can be observed in lesions diagnosed as dysplastic nevi by our group (∼10% or 1 in 10 lesions), as well as in rare poorly differentiated malignancies.