Reverse transcriptase polymerase chain reaction for the detection of tyrosinase-mRNA-positive cells in peripheral blood of melanoma patients, as a possible marker of hematogenous dissemination, has demonstrated varying detection rates. This study examined the sensitivity and reproducibility of the technique using a protocol of multiple polymerase chain reaction to determine circulating melanocytic cells. For each of the 123 melanoma patients included in this study, four nested polymerase chain reactions were performed from two blood specimens requiring both polymerase chain reactions from at least one blood sample to be positive to consider a patient as positive. Thus, a definitive result was obtained in 98% of the cases, whereas only 1.6% lacked conclusive findings. Thus, we found a correlation between the tyrosinase detection rate and the clinical stage. Circulating tyrosinase-mRNA-positive cells were detected in 13% of patients with primary tumor, 17% with regional skin/lymph node metastasis, and 44% with distant metastasis. Positivity also correlated with known melanoma progression markers such as gender, tumor thickness, and histologic type. Positive results were obtained more frequently in (i) men compared with women, (ii) patients with thick primary melanomas (> 4 mm: 38%) compared with those with thinner tumors (1.1-4 mm, 22%; < or = 1 mm, 5%), and (iii) patients with nonclassifiable (38%), nodular (34%), and occult primary melanomas (30%) compared with those with acrolentiginous (17%), superficial spreading (9%), or lentigo maligna melanoma (0%). These findings suggest that detection of tyrosinase-mRNA-positive cells in peripheral blood is not an adequate marker for identifying melanoma patients with distant metastasis. Reverse transcriptase polymerase chain positivity in early melanoma stages, however, as corresponding to other prognostic parameters, may indicate increased risk for the development of hematogenous metastasis and may be of value as a progression marker.