Optimization and robustness of blood tests for liver fibrosis and cirrhosis

ObjectivesTo optimize the performance and feasibility of fibrosis blood tests and evaluate their robustness.Design and methodsThe derivation population included 1056 HCV patients with liver biopsy and blood markers. Validation populations included 984 patients with various viral hepatitis causes, and Fibroscan and/or liver biopsy and/or blood markers.ResultsThe bootstrap method validated the markers of the original FibroMeter2G, but not those of Fibrotest and Hepascore, and provided a hyaluronate-free FibroMeter3G. AUROCs for significant fibrosis were: FibroMeter2G: 0.853 vs. FibroMeter3G: 0.851, p = 0.489. Compared to FibroMeter2G, FibroMeter3G had a significantly higher patient rate with predictive values ≥ 90% for significant fibrosis. Accuracy for fibrosis stage classification was: Fibrotest: 37.9%, FibroMeter2G: 74.9%, and FibroMeter3G: 86.9% (p < 10− 3).ConclusionThe bootstrap method validated FibroMeter2G and provided a cheaper and more feasible hyaluronate-free FibroMeter3G with comparable performance. Compared to binary diagnosis, fibrosis stage classification increased discrimination, with an increased accuracy to 87% for FibroMeter3G.