Precise diagnosis of well-differentiated hepatocellular carcinoma (HCC) is sometimes difficult to establish. Telomerase activity was examined by telomeric-repeat-amplification protocol (TRAP) in 37 HCC nodules smaller than 3 cm in diameter, including 24 fine-needle-aspiration biopsy specimens, 22 non-tumor chronic-liver-disease tissues (9 chronic hepatitis and 13 liver cirrhosis) and 3 normal liver tissues. Telomerase activity was assayed by serially diluted samples and quantitated by using an internal telomerase assay standard (ITAS). Telomerase activity was detected in all HCC and in 11 of 22 non-tumor chronic-liver-disease tissues. Normal liver samples had undetectable telomerase activity. Cut-off level of telomerase activity for its practical usage in HCC diagnosis was tentatively set for 0.6 microg liver protein/assay at 10-cell equivalent activity of a gastric-cancer cell line, MKN-1. This level was twice the highest activity in non-tumor chronic liver disease therefore, telomerase activity in all non-tumor liver samples was below this level. The telomerase-positive incidence exceeding this cut-off level was 73% (11/15) in well-differentiated HCC, 94% (16/17) in moderately differentiated HCC and 100% (5/5) in poorly differentiated HCC. Well-differentiated HCC showed low positivity by other diagnostic markers. 21% by AFP, 0% by PIVKA-II and 13% by angiography. The detection of telomerase activity may thus be a useful additional tool for precise and early diagnosis of small differentiated HCC, even when diagnosis is inconclusive by conventional techniques.