Abstract:Hepatocellular carcinoma (HCC) is one of the most prevalent malignancies worldwide, and accurate prognostic assessment and treatment planning are vital for improving patient outcomes. Conventional prognostic models, which rely on limited clinicopathological parameters, often fail to capture the profound heterogeneity of HCC. In recent years, artificial intelligence (AI)-particularly machine learning and deep learning-has driven a paradigm shift in precision oncology by leveraging its powerful capabilities in data mining and pattern recognition. This review provides a comprehensive overview of recent advances in AI for prognostic assessment and treatment optimization in HCC, with an emphasis on key methodologies such as radiomics and multi-modal data integration. It further discusses the clinical potential and challenges of AI in predicting postoperative recurrence, evaluating therapeutic response, and supporting individualized treatment decisions, while also outlining future directions in this rapidly evolving field. The review aims to inform and facilitate the clinical translation of AI technologies into the management of HCC.