AI-Enabled Real-World Evidence in Oncology: A Statistical Perspective for Regulatory Decisions.

Artificial intelligence (AI) has the potential to strengthen real-world evidence (RWE) for regulatory decision-making, but its contribution varies by application and methodological maturity. RWE remains limited by challenges in data quality, population selection, treatment characterization, outcome assessment, and statistical methodology. Machine learning and generative AI (genAI), combined with causal inference frameworks, may address these challenges. We review applications, limitations, including reproducibility, bias, transportability, uncertainty quantification, and regulatory acceptability, and methodological priorities for fit-for-purpose AI-enabled RWE.