Diagnostic validation of the GAAD algorithm for hepatitis B virus-related hepatocellular carcinoma surveillance in Vietnam

Objectives Chronic hepatitis B infection is a major cause in Vietnam of developing hepatocellular carcinoma (HCC). HCC surveillance for hepatitis B patients is recommended to detect tumors early for effective treatment and survival extension. The novel Elecsys (R) GAAD algorithm (Roche Diagnostic) was generated by using serum PIVKA-II and AFP measurements with age and biological sex factors for more effective HCC screening expectations. This study was conducted to validate the GAAD algorithm for HCC surveillance in Vietnamese hepatitis B patients.Methods A cross-section study with 337 hepatitis B patients divided into 142 HCC and 195 non-HCC including benign liver tumors, liver fibrosis and cirrhosis was performed in Vietnam. The Elecsys (R) GAAD score was computed by Elecsys (R) AFP, Elecsys (R) PIVKA-II and patient's age, gender. Clinical valuation of the GAAD algorithm was accessed by sensitivity, specificity at recommend cut-off, and AUC of receiver operating characteristic curve.Results The GAAD was shown moderate performance with AUC of 0.767 (95 % CI 0,716-0,818) but more effective than single AFP in general HCC detection, and more useful than PIVKA-II in HCC detection at early stage with tumor <= 2 cm and in distinguishing benign liver tumor from HCC. It did not provide more benefit than PIVKA-II to detect HCC at all stages or cirrhosis background.Conclusions The study was shown that GAAD was moderate performance but seemed more effective than single AFP in general HCC surveillance and might be superior to PIVKA-II for detecting early-stage HCC with tumors <= 2 cm, and for distinguishing HCC from benign liver tumors in chronic hepatitis B patients.