To evaluate the role of plasma des-γ-carboxyprothrombin in the early diagnosis of hepatocellular carcinoma, we simultaneously studied both des-γ-carboxyprothrombin activities by staphylocoagulase method and des-γ-carboxyprothrombin antigen levels by enzyme immunoassay in 39 patients with early stage hepatocellular carcinoma (tumor size <3 cm in 21 patients, 3 to 5 cm in 18 patients); 68 patients had large hepatocellular carcinoma and 54 patients had chronic hepatitis or cirrhosis. Des-γ-carboxyprothrombin levels by staphylocoagulase method (X) and enzyme immunoassay method (Y) on the same plasma specimens of hepatocellular carcinoma patients showed a linear correlation (Y = 0.15X - 10.5, r = 0.533, n = 67, p < 0.001). Elevated des-γ-carboxyprothrombin activities were present in 10 of 21 patients (47.6%) with hepatocellular carcinoma <3 cm, 66.7% of 18 with hepatocellular carcinoma 3 to 5 cm, 67.6% of 68 with hepatocellular carcinomas >5 cm and 27.8% of 54 with chronic hepatitis or cirrhosis. The plasma des-γ-carboxyprothrombin levels did not correlate with the tumor size or serum α-fetoprotein levels. Plasma des-γ-carboxyprothrombin and serum α-fetoprotein measurements were comparable in the diagnosis of hepatocellular carcinoma because 22 (56.4%) and 21 (53.8%) of 39 patients with hepatocellular carcinoma <5 cm had increased des-γ-carboxyprothrombin and α-fetoprotein levels, respectively. Up to 77% had an abnormal elevation in either marker. The des-γ-carboxyprothrombin antigen levels by enzyme immunoassay method are more specific to hepatocellular carcinoma than des-γ-carboxyprothrombin activities by the staphylocoagulase method: the positive rates of elevated des-γ-carboxyprothrombin antigen levels in the patients with hepatocellular carcinoma <3 cm, 3 to 5 cm, >5 cm and chronic hepatitis or cirrhosis were 19.0%, 55.6%, 66.2% and 14.8%, respectively. We conclude that des-γ-carboxyprothrombin assay may serve as a complementary tool in the diagnosis and follow-up of hepatocellular carcinoma and is particularly useful in patients with normal or low α-fetoprotein levels. Because of higher specificity to hepatocellular carcinoma and less tedious assay procedures, the enzyme immunoassay method is more practical than the staphylocoagulase method in the measurement of des-γ-carboxyprothrombin. For early detection of hepatocellular carcinoma, measurement of plasma des-γ-carboxyprothrombin levels only is not sensitive enough. (HEPATOLOGY 1990;11:481–488.)