ClinicalReTrial: A Self-Evolving AI Agent for Clinical Trial Protocol Optimization

TL;DR

ClinicalReTrial reframes clinical trial optimization as an in silico, iterative protocol redesign problem solved by a self-evolving, multi-agent AI system. A failure-diagnosis agent proposes targeted augmentations, which are validated for safety and evaluated in a simulation environment driven by a predictive model, with rewards guiding hierarchical memory-based learning. The framework demonstrates strong predictive capability (PR-AUC > 0.75) and delivers significant protocol improvements across enrollment, safety, and efficacy failures (mean Δp ≈ 0.057), while maintaining cost-effectiveness (~$0.12/trial) and alignment with real-world redesigns. This approach offers a practical pathway to proactively modify trial designs, potentially accelerating development timelines and reducing costly attrition in drug discovery.

Abstract

Clinical trial failure remains a central bottleneck in drug development, where minor protocol design flaws can irreversibly compromise outcomes despite promising therapeutics. Although cutting-edge AI methods achieve strong performance in predicting trial success, they are inherently reactive for merely diagnosing risk without offering actionable remedies once failure is anticipated. To fill this gap, this paper proposes ClinicalReTrial, a self-evolving AI agent framework that addresses this gap by casting clinical trial reasoning as an iterative protocol redesign problem. Our method integrates failure diagnosis, safety-aware modification, and candidate evaluation in a closed-loop, reward-driven optimization framework. Serving the outcome prediction model as a simulation environment, ClinicalReTrial enables low-cost evaluation of protocol modifications and provides dense reward signals for continuous self-improvement. To support efficient exploration, the framework maintains hierarchical memory that captures iteration-level feedback within trials and distills transferable redesign patterns across trials. Empirically, ClinicalReTrial improves 83.3% of trial protocols with a mean success probability gain of 5.7%, and retrospective case studies demonstrate strong alignment between the discovered redesign strategies and real-world clinical trial modifications.

Figures (6)

• Figure 1: ClinicalReTrial Agent architecture. The system operates through iterative refinement: agents analyze failures, generate augmentations, and receive rewards from the simulation environment. Modifications are distilled into structured knowledge that guides subsequent iterations, enabling progressive improvement.
• Figure 2: Iterative learning dynamics stratified by failure mode. IQR bars span the 25th–75th percentiles (interquartile range) across runs.
• Figure 3: Self-Improving ablation study across 10 trials. Full system (blue) outperforms setups without memory (red) or redesign pool optimization (green). Memory provides early benefits, and pool effects compound over iterations. IQR bars span the 25th–75th percentiles across runs.
• Figure 4: ClinicalReTrial Agent's flowchart on Poor Enrollment failed trial case study (NCT01298752, 2011-02-16), together with the real-world redesign (NCT01591161, 2012-05-02). The Agent's iterative refinement of failure analysis and according modifications are demonstrated.
• Figure 5: Visualization of text segments in the BioBERT encoder's output, illustrating Shapley values derived from Clinical Trials. Shapley values correspond to attention weights, with darker colors indicating higher weights.