A large language model-based AI system deployed at Cleveland Clinic effectively identified diverse, eligible participants for a rare disease clinical trial through EMR screening.
Key Details
- 1Cleveland Clinic and Dyania Health used an AI system (Synapsis AI) to pre-screen for a Phase 3 trial in transthyretin amyloid cardiomyopathy (ATTR-CM).
- 2In one week, the AI reviewed 1,476 records and identified 46 potential matches; 29/30 AI-identified matches were missed by traditional methods.
- 3The AI achieved 96.2% accuracy for 7,700 trial-specific questions and a 99% negative predictive value (NPV) in excluding ineligible patients.
- 4AI-assisted recruitment led to a more diverse patient cohort, with 36.6% Black participants versus 7.1% via routine screening.
- 5Only 60% of AI-identified patients had heart failure specialist care, versus 92.8% found by traditional methods, showing improved outreach.
- 6The system used both structured EMR data and NLP to analyze clinical notes and provided auditable inclusion/exclusion justifications.
Why It Matters
This study showcases the potential of AI-driven EMR review to increase trial enrollment efficiency and address disparities by identifying qualified participants who might otherwise be overlooked, particularly in underrepresented groups. It underlines how imaging and clinical AI advancements are intersecting with research recruitment and population health strategies.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Method Protects Sensitive Data in ECGs While Retaining Clinical Value
Researchers developed an AI model (PP-VAE) that safeguards sensitive personal data in electrocardiograms without sacrificing clinical utility.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.