This research introduces the Clinical Trial Outcome (CTO) benchmark, a comprehensive repository of ~125,000 labeled clinical trials that enables automated outcome prediction.

• Creates a reproducible, large-scale dataset for training predictive models
• Addresses critical data access limitations in drug development
• Establishes foundation for evidence-based decision-making in clinical research
• Demonstrates potential to reduce costs and accelerate regulatory approvals

This breakthrough matters because it provides the medical industry with standardized data to build AI models that could significantly reduce the substantial costs of drug discovery while improving success rates for new treatments.

Automatically Labeling Clinical Trial Outcomes: A Large-Scale Benchmark for Drug Development