Abstract 7860: Fusion gene machine learning models improve clinical outcome prediction of hepatocellular carcinoma.

Abstract Hepatocellular carcinoma is one of the most lethal malignancies for humans. Assessing the clinical outcomes of HCC remains challenging. In this study, we analyzed a panel of 20 fusion genes in 200 hepatocellular carcinoma (HCC) samples to predict the recurrence and survival rates of HCC patients undergoing surgical interventions using machine learning models. The results showed that fusion genes, Milan criteria, serum α-fetal protein (AFP), and pathology grade had moderate predictive accuracy for HCC recurrence. However, the combination of selected fusion genes with these clinical parameters significantly enhanced the prediction accuracy of each parameter. When models of fusion genes were applied to predict the 3-year survival rate of HCC patients, they outperformed the Milan criteria, pathology grade, and serum AFP. The combination of a fusion gene panel with Milan criteria, pathology grade, or serum AFP yielded significantly improved results compared to those produced by these clinical parameters alone. As a result, examining the fusion gene status of HCC samples may hold promise as a new and improved approach to assessing the clinical outcomes of this disease. Citation Format: Jian-Hua Luo, Silvia Liu, Yanping Yu, . Fusion gene machine learning models improve clinical outcome prediction of hepatocellular carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7860.