AI-powered biotherapies: Artificial intelligence and informatics at the intersection of transfusion medicine and...

The rapid evolution of biotherapies-encompassing hematopoietic stem cell transplantation (HSCT), chimeric antigen receptor T-cell (CAR-T) therapies, gene-modified cellular therapies, and CRISPR-based platforms-has fundamentally transformed hematology, oncology, and regenerative medicine. Artificial intelligence (AI) and machine learning (ML) are increasingly recognized as potential central enablers of precision biotherapies, yet their systematic applications across the biotherapy pipeline remain incompletely characterized.

This narrative review synthesizes published literature, registry data, and emerging regulatory frameworks to examine 12 transformative applications of AI/ML and informatics in the biotherapy ecosystem, organized within three thematic domains: (1) precision donor-recipient matching, cell and gene therapy engineering, and outcomes monitoring; (2) AI/ML-enabled simulation, adaptive clinical trials, and quality control; and (3) informatics infrastructure, multi-omics integration, and regulatory science.

AI/ML demonstrates significant potential across the biotherapy pipeline: from advanced HLA donor-recipient matching and CAR construct optimization to manufacturing process analytics, digital twin simulation, automated quality control, and long-term survivorship prediction. Applications span a maturity spectrum from early clinical adoption (HLA matching, manufacturing QC) to largely conceptual stages (digital twins, personalized conditioning). Critical challenges include algorithmic bias, explainability deficits, reproducibility gaps, and evolving data privacy and regulatory frameworks.

AI and informatics are positioned to usher in a new era of precision, data-driven biotherapies. Realizing this potential requires interdisciplinary collaboration, rigorous external validation, equitable dataset representation, and alignment with emerging regulatory standards to ensure safe, transparent, and patient-centered integration into clinical and manufacturing workflows.