Background and AimPredictors of neutrophil-to-lymphocyte ratio (NLR) and traditional clinical variables for hepatocellular carcinoma (HCC) prognosis after locoregional therapies were useful while exhibited modest prognostic performances. We dig out the potential of circulating immune features for HCC prognosis prediction.Methods244 patients with early-stage HCC who were treated with thermal ablation and performed the peripheral blood mononuclear cells (PBMCs) tests were included. Patients were randomly assigned in 3:1 ratio to discovery (n = 183) and validation (n = 62) sets. Three models, including clinical (Clin-model), NLR-Clin-model and Immune-NLR-Clin-model were constructed using Cox regression model. Concordance index (c-index), integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were used for performance evaluation.ResultsThe Immune-NLR-Clin-model exhibited the best performance of 0.706 (95% CI:0.644-0.768) and 0.702 (95% CI:0.566-0.837) in discovery and validation sets, respectively. At 36-month prediction, the IDI and continuous-NRI show trend of improvement, with the IDI was 0.050 (95%CI: -0.5%-12.5%) (