Community Approach Targeting Cirrhosis and Hepatocellular Carcinoma (CATCH): 4AGP, a new indirect biomarker-based algorithm that can predict risk of liver-related outcomes.

Background and Aim: Transient elastography with liver stiffness measurement (LSM) is known to correlate with liver fibrosis and liver-related outcomes. 4AGP is a new algorithm that uses indirect biomarkers to determine those who are at risk of elevated LSM (> 12.5 kPa). It is calculated using alpha-fetoprotein, albumin, aspartate aminotransferase, age, sex (gender), and platelet count. Although 4AGP correlates with elevated LSM in patients with chronic hepatitis C (CHC) virus infection, it is unknown whether it also predicts liver-related outcomes. We aimed to determine whether LSM or 4AGP, or other indirect biomarkers such as aspartate aminotransferase to platelet ratio index (APRI), Fibrosis-4 (FIB-4), or Forns index, can best predict liver-related outcomes in a cohort of patients with CHC. Method(s): The Community Approach Targeting Cirrhosis and Hepatocellular Carcinoma (CATCH) study is a prospective study that recruited patient with CHC. Baseline LSM and indirect biomarkers were assessed (all were before treatment), along with outcomes (hepatocellular carcinoma [HCC], decompensation, or liver-related death). Survival curves were analyzed using a log-rank test, along with receiver operator characteristic (ROC) curves and Cox proportional hazard ratios. Optimal cut-off points were calculated, accounting for sensitivity, specificity, false positives and negatives, as well as the area under the ROC curve (AUROC). Result(s): A total of 1049 patients with CHC were recruited between October 2014 and June 2018. Mean follow-up was 3.4 +/- 0.7 years, indicating 3631 patient-years. Patients' mean age was 45.1 years (19-82), and 31.3% were female. Seven patients developed HCC; the best predictors of HCC development were, in order, 4AGP, Forns index, FIB-4, LSM, and APRI (Fig. 1a). An LSM of 21.3 kPa was the optimal cut-off point for predicting HCC, with a sensitivity of 85.7%, specificity of 92.2%, and a hazard ratio (HR) of 30 (95% CI, 5.7-150). A 4AGP value of -2.81 had