Background: Measurement of serum alpha-fetoprotein (AFP) and abdominal ultrasound (US) examination are used for the early detection of hepatocellular carcinoma (HCC) in chronic liver disease patients. However, the accuracy and usefulness of these tests in a clinical setting in the United States of America have not been clarified.
Methods: We conducted a 7-year prospective surveillance study by using both AFP and US to detect HCC in 602 patients with chronic viral hepatitis. Our main goal was to determine the optimal test for detection of early HCC. We also assessed the clinical outcome of HCC patients identified during this time period.
Results: Thirty-one cases of HCC were detected. Serum AFP levels were elevated in 74% of HCC patients, but was also high in 10% of patients who did not develop HCC. The positive predictive value for AFP to detect HCC was only 12% or less for all AFP cut-off values, and the maximum joint sensitivity and specificity as determined by receiver operator characteristic analysis was approximately 65 and 90%, respectively. Abdominal US identified all 31 cases of HCC. The positive predictive value for US examinations to detect HCC was 78%, while the sensitivity and specificity was 100 and 98%, respectively. After detection of HCC, 24 (77%) patients died within a mean of 16.7 +/- 19.4 months.
Conclusions: Our study indicates that US examination was more accurate in detecting HCC. Because of its poor predictive value and low sensitivity, serum AFP should not be used as the only test for screening and surveillance for HCC.