Pathologic assessment of melanoma sentinel nodes: a role for molecular analysis using quantitative real-time reverse transcription-PCR for MART-1 and tyrosinase messenger RNA

Abstract

Purpose: Molecular analysis of melanoma sentinel nodes (SN) is sensitive, but poorly specific because metastases cannot be distinguished from benign nevus inclusions (BNI). We investigated whether quantitative reverse transcription-PCR (RT-PCR) detection of MART-1 and tyrosinase mRNAs could improve this specificity and contribute to SN assessment.

Experimental design: Two hundred twenty SNs from 95 melanoma patients analyzed by extensive immunohistopathology and real-time quantitative RT-PCR.

Results: Using histopathology, SNs and patients were allotted to three diagnostic groups: (a) metastasis positive, (b) BNI positive, and (c) melanocyte-free. Median MART-1 and tyrosinase mRNA levels in SNs were significantly different in patients with metastasis compared with patients with BNIs (P < 0.05) and patients without melanocytic lesions (P < 0.001). However, a "gray-zone" was observed where distinction, based on mRNA levels, could not be made between the three groups. For both genes, the highest mRNA level recorded in each RT-PCR-positive patient was positively correlated with Breslow's tumor thickness. For SNs with metastases, tumor burden was significantly correlated to the mRNA level. Using the presence of a MART-1 RT-PCR signal to detect patients with metastases, a sensitivity of 100% and a negative predictive value of 100% were achieved when extensive immunohistology was used as reference.

Conclusions: Quantitative RT-PCR MART-1 and tyrosinase mRNA analysis cannot be used alone for SN diagnosis because of its poor specificity for melanoma metastasis. However, in approximately one third of cases without RT-PCR evidence of MART-1 expression, extensive histopathologic SN investigation is not necessary, thus substantially reducing the cost of SN analysis. The level of melanocyte-associated mRNA is associated with both tumor thickness and tumor burden as measured histopathologically, suggesting that this may be of prognostic value.