PRAME expression by immunohistochemistry and reverse transcription quantitative PCR in conjunctival melanocytic lesions-a comprehensive clinicopathologic study of 202 cases and correlation of cytogenetics with PRAME expression in challenging conjunctival melanocytic lesions.

This study examined PRAME (preferentially expressed antigen in melanoma) expression by immunohistochemistry and reverse transcription quantitative PCR (RT-qPCR) in 202 histologically unequivocal conjunctival melanocytic lesions: 76 nevi, 29 benign melanoses, 25 low-grade conjunctival intraepithelial melanocytic lesions (LGCMIL), 26 high-grade conjunctival melanocytic intraepithelial lesions/in-situ melanoma (HGCMIL), and 46 invasive melanomas. PRAME score 0 was seen in 96% of nevi (73/76), 96% of benign melanoses (28/29), and 88% of LGCMIL (22/25). PRAME score 4 was seen in 50% HGCMIL (13/26) and 76% invasive melanomas (35/46). PRAME score 4 had a sensitivity of 50% and specificity of 100% in differentiating between HGCMIL and benign melanosis/LGCMIL. PRAME score 4 had a sensitivity of 76% and specificity of 100% in differentiating between melanoma and nevi. Relative quantification of PRAME mRNA expression by RT-qPCR was performed on 49 cases (24%): 17 nevi, 3 benign melanoses, 5 LGCMIL, 9 HGCMIL, and 15 invasive melanomas. The analysis generated two distinct groupings with 'high' relative PRAME expression for the HGCMIL and invasive melanoma and 'low/zero' expression for nevi, benign melanosis, and LGCMIL. Thirty-three challenging conjunctival melanocytic lesions that had previous fluorescence in situ hybridization (FISH) analysis were studied: 18 nevi, 12 melanomas in a nevus, 2 nevoid melanomas, and 1 in-situ melanoma. All nevi (100%) showed concordance between negative FISH and PRAME (scores 0-3). Four of 13 melanomas (31%; in-situ, invasive, isolated, and in association with nevus) showed concordance between positive FISH and PRAME score 4. In conclusion, PRAME score 4 has 100% specificity for the diagnosis of HGCMIL and melanoma. PRAME is limited in its sensitivity in the evaluation of challenging melanocytic lesions.