AI-based patient identification support system  to streamline the identification of eligible CAR-T candidates.

Three spanish centers participated, demonstrating the system’s adaptability across different hospital settings.

The system achieved a 96.43% sensitivity rate, proving its effectiveness in detecting eligible patients.

The AI algorithm screened over 4 million patient records, identifying 34 potential candidates, with a 79.4% accuracy confirmed by hematologists.

Chimeric Antigen Receptor T-cell (CAR-T) therapy has revolutionized the treatment of certain hematologic cancers, such as lymphomas and leukemias, providing new hope for patients with limited options. However, early identification of suitable candidates and efficient management until CAR-T infusion remain significant challenges.

This pilot study, conducted by the Spanish Hematopoietic Transplant and Cellular Therapy Group (GETH-TC) in collaboration with IOMED, introduces an artificial intelligence (AI)-driven system using real-world data (RWD) to improve patient identification for CAR-T therapy, reducing missed candidates and accelerating treatment timelines.

CAR-T therapy is a life-changing option for patients with refractory hematologic malignancies. However, identifying candidates in a timely manner and ensuring a smooth process until infusion is critical for treatment success. Traditional selection methods can be slow and error-prone, highlighting the need for innovative solutions to optimize this process. The primary goal was to reduce missed CAR-T candidates and accelerate infusion timing through an artificial intelligence (AI)-based patient identification support system.

IOMED implemented its AI-powered Data Space Platform that extracts clinical concepts from free-text medical records and EHRs, harmonizing data according to the OMOP-CDM model. This plataform allowed participating hospitals to rapidly identify patients meeting predefined criteria for CAR-T therapy, improving efficiency and accuracy in patient selection.