AI integration for transforming care of patients with liver diseases

Chronic liver disease is a major cause of morbidity and mortality worldwide, accounting for approximately 4% of global deaths. Despite the increasingly complex care of liver disease, conventional scoring systems remain widely used to assess fibrosis and predict clinical outcomes. Although cost-effective and interpretable, these scores have limitations that may contribute to missed diagnoses and delayed treatment. Furthermore, their reliance on discrete laboratory biomarkers may inadequately reflect patients' overall clinical status and may introduce bias across demographic subgroups.

This review aims to provide an overview of current applications of artificial intelligence (AI) in hepatology.

A literature search of PubMed was conducted to identify publications addressing the application of AI in liver disease. The literature was reviewed to examine the use of AI in diagnosis, risk stratification, clinical prioritization, and liver transplantation, as well as its application across liver diseases of different etiologies. Technical challenges and barriers to clinical implementation were also considered.

Artificial intelligence models can integrate multimodal data and capture complex, nonlinear relationships, potentially enabling a more comprehensive assessment of patients' physiological status and disease progression than conventional scoring systems. AI has considerable potential to enhance diagnostic accuracy, prognostic assessment, risk stratification, and decision-making in liver transplantation. However, limitations relating to model interpretability, generalizability, data quality, algorithmic bias, and integration into clinical workflows must be addressed to facilitate safe, equitable, and effective implementation in hepatology.