Development and Validation of a LASSO-RF-Cox-Derived Nomogram Incorporating Inflammatory Markers to Predict Long-Term Survival in Early-Stage Hepatocellular Carcinoma After Microwave Ablation.

Purpose: This study aimed to identify factors influencing overall survival (OS) in patients with hepatocellular carcinoma (HCC) undergoing microwave ablation (MWA) and to develop and validate a nomogram for predicting 3‑, 5‑, and 8‑year OS. Materials and Methods: Data from 440 patients who received MWA at Beijing Ditan Hospital, Captical Medical University were analyzed using least absolute shrinkage and selection operator (LASSO) regression, random forest (RF), and multivariable Cox regression to identify independent prognostic factors. A prognostic nomogram was constructed and validated. OS was assessed using Kaplan-Meier curves. Results: The variables selected by both LASSO and RF were incorporated into a multivariable Cox regression model, which identified tumor number, tumor size, monocyte-to-lymphocyte ratio (MLR), white blood cell count (WBC), and diabetes as independent risk factors for OS. The predictive accuracy, reliability, and clinical utility of the model were confirmed through Harrell's concordance index (C-index), time-dependent area under the receiver operating characteristic curve (AUC) analysis, calibration curves, and decision curve analysis (DCA). Furthermore, risk stratification based on the model effectively differentiated patients into distinct prognostic subgroups. Conclusion: A nomogram based on LASSO-RF-Cox analysis was developed to predict OS in early-stage HCC patients after MWA. The model demonstrated acceptable predictive performance in internal validation, with moderate discriminative ability. This model may help identify high-risk individuals and facilitate clinical decision-making, particularly in primary care settings due to its simple and readily available predictors.