Evolution of causality assessment in drug-induced liver injury: From RUCAM to RECAM and emerging Artificial Intelligence applications

Drug-induced liver injury (DILI) remains one of the most challenging adverse drug reactions to diagnose because of its heterogeneous clinical presentation and the absence of a definitive diagnostic test. Accurate causality assessment is therefore essential for patient management, pharmacovigilance, and regulatory decision-making. Over the past four decades, causality assessment has evolved from subjective expert judgment and general adverse drug reaction algorithms to structured liver-specific methods. The introduction of the Roussel Uclaf Causality Assessment Method (RUCAM) in 1993 marked a significant milestone by providing the first standardized and quantitative algorithm for evaluating suspected DILI. Although RUCAM remains the most widely accepted and extensively validated causality assessment tool, its limitations, including manual scoring, inter-observer variability, and restricted compatibility with electronic health records, prompted the development of the Revised Electronic Causality Assessment Method (RECAM). RECAM incorporates clearer definitions, improved standardization, and electronic implementation to enhance reproducibility and facilitate integration into modern clinical practice. More recently, artificial intelligence (AI) has emerged as a promising adjunct for DILI assessment through predictive modelling, automated data extraction, and clinical decision support. However, these technologies require further validation and should complement rather than replace structured causality assessment and clinical expertise. This review summarizes the evolution of DILI causality assessment from RUCAM to RECAM and critically compares their strengths and limitations. Further, it discusses the emerging role of AI in improving diagnostic accuracy and advancing evidence-based DILI management.