Review identifies challenges and opportunities for predictive models in CAR T-cell therapy to improve outcomes.
Key Points
This review aims to evaluate existing predictive models for toxicities associated with CAR T-cell therapy and explore their limitations and potential improvements.
Reviewed predictive models for CAR T-cell toxicities
Assessed strengths and challenges in existing modeling approaches
Discussed future directions for improving predictive accuracy
Identified small sample sizes and poor data quality as major limitations for model reproducibility
Highlighted the need for better biomarker integration and context-specific modeling
Emphasized the role of federated learning in enhancing collaborative data sharing