Novel and accurate measurement of differential protease activity in diagnosed HCC patients compared to non-HCC cirrhotic patients

Method: Patients with chronic liver disease of multiple untreated etiologies, with liver biopsy using Metavir F within 6 months of blood test, were included in a derivation (n = 2159) and a validation (n = 1476 with LSM by TE) sets.Agreement grade was calculated as the absolute difference (AD) between fibrosis (F M ) classifications of FM2G and CM2G.The main outcome was severe fibrosis (Metavir F ≥ 3).Outcome measures were AUROC and accuracy (correct diagnosis).Results: In the derivation set, FM2G accuracy (and prevalences) were according to agreement grade: high (F M AD ≤ 0.5): 83.1% (78.6%), fair (F M AD = 1): 70.2% (16.3%), poor (F M AD >1): 50.0%(5.1%), p < 0.001.CM2G accuracy was, respectively 83.1%, 67.6%, 50.9% ( p < 0.001).Then, agreement was classified as concordant or not for Metavir F ≥ 3. FM2G accuracy (and prevalences) were, concordant: 83.9% (85.9%), discordant: 51.3% (14.1%), p < 0.001.FM2G AUROCs were, respectively: 0.866 and 0.519, p < 0.001.The overall accuracy of FM2G (79.3%) and CM2G (79.0%) were not different ( p = 0.688) but they were complementary (kappa: 0.573).So, a CM/FM algorithm was constructed.Its accuracy was significantly increased to 82.2% ( p < 0.001) vs FM2G or CM2G and different between concordant (85.1%) and discordant (64.1%, p < 0.001) blood tests.In the validation set, previous results were confirmed, e.g.CM/FM accuracy was, concordant blood tests: 79.9%, discordant: 60.5% ( p < 0.001).The concordance prevalence was higher for FM2G vs CM2G (84.9%) than vs LSM (70.9%, p < 0.001).Accuracy was globally, LSM: 78.8% vs CM/FM: 77.0%, ( p = 0.173) or FM2G: 75.5% ( p = 0.011) but in the 70.9% of concordant FM2G/LSM it was, respectively: 88.0% vs 87.4% ( p = 0.210) or 88.0% ( p = 1).The accuracy in concordant tests was however higher in FM2G agreement vs LSM than vs CM2G, e.g.CM/FM: 88.0% vs 79.9% ( p < 0.001). Conclusion:The concordance within a blood test significantly increased the concordance prevalence.The limit was a reduced accuracy; the advantages were a reduced cost and liver biopsy recourse by half, and a simpler use (single time and greater ubiquity).