Summary Physical examination and ultrasound B-scan screening are important follow-up procedures in melanoma patients with regional disease. However, they do not allow definite diagnosis of suspicious lesions. Fine-needle aspiration cytology (FNAC) enhances the diagnostic accuracy in such patients but, unfortunately, reaches its technical limits, particularly when very small or necrotic lesions are examined. We therefore tested whether tyrosinase reverse transcription polymerase chain reaction (RT-PCR) of fine-needle aspirates (FNA-PCR) could help to increase diagnostic sensitivity. With clinical follow-up in 69 melanoma patients 81 regional lymph nodes were detected by ultrasound Bscan examination, nine of whom appeared to be palpable. Technically, FNAC was successful in all 81 lymph nodes, while FNA-PCR failed to obtain RNA at detectable levels in two lymph nodes of two patients. Of 79 lesions which have been completely evaluated by B-scan, FNAC and FNA-PCR, 44 proved to be melanoma metastases by histopathology, while the remaining 35 lesions were finally classified as non-specific lymph nodes. Of the 44 melanoma metastases 80% (n = 35) have been detected by B-scan, 90% (n = 39) by FNAC and 100% (n = 44) by FNA-PCR (P < 0.05 vs FNAC, P < 0.005 vs B-scan). In the subclass of lesions with diameters below 10 mm the sensitivities were 72% (n = 13), 78% (n = 14) and 100% (n = 18) respectively. In 35 regional lymph nodes classified as benign lesions, FNAC was always negative while FNA-PCR produced one positive result. Neither of these methods did produce false positive results in 15 control lymph nodes of non-melanoma patients. We conclude, that FNA-PCR might have superior sensitivity as compared to FNAC or ultrasound B-scan, particularly in melanoma lesions with diameters below 10 mm.Keywords: melanoma; ultrasound guided fine needle aspiration cytology (FNAC); tyrosinase reverse transcription polymerase chain reaction (RT-PCR); RT/PCR of fine needle aspiration material (FNA-PCR)