Supplementary Material for: Improved Prediction of Survival by a Risk Factor-Integrating Inflammatory Score in Sorafenib-Treated Hepatocellular Carcinoma

Background and Aims: Inflammation affects progression of hepatocellular carcinoma (HCC). We therefore postulate that systemic inflammatory markers could help to predict prognosis in HCC patients receiving sorafenib therapy. Methods: Overall survival (OS) of HCC patients receiving palliative sorafenib treatment was correlated with the neutrophil to lymphocyte ratio (NLR), platelet to lymphocyte ratio (PLR), C-reactive protein to albumin ratio (CAR), Glasgow prognostic score (GPS) and the modified GPS (mGPS) along with clinicopathological parameters. Predictors of OS were assessed by multivariable Cox regression and receiver operating characteristics and area under the curve (ROC-AUC) analyses. Results: Patients receiving sorafenib (n = 120) for advanced HCC (Barcelona Clinic Liver Cancer stage C) were explored by retrospective analysis. Findings were subsequently validated by a second HCC cohort (n = 113) receiving sorafenib at two independent treatment centers. Multivariable assessment across these HCC cohorts confirmed a stable association of CAR (p ≤ 0.001), GPS (p ≤ 0.01) and mGPS (p ≤ 0.004) with OS. This study also identified Eastern Cooperative Oncology Group (ECOG) performance score (p p = 0.002) as prognostic factors and uncovered an inconsistent OS association of NLR and PLR in HCC patients. Additional combined analysis of ECOG, portal thrombosis and GPS within an extended score (GPS-EP) was associated with OS (p = 0.021), which was confirmed within the validation cohort (p = 0.001). In sorafenib-treated HCC, the ROC-AUC value for the prediction of 12-month survival was 0.761 (CAR >/≤0.37 cut-off, p p p 2 point cut-off. Conclusions: Inflammatory scores obtained before sorafenib treatment initiation are associated with OS in advanced HCC. Their combination with other risk factors improves prediction of 3- and 12-month survival, which could guide treatment decisions in selected patient subgroups