A "medically trained" artificial intelligence system at Cleveland Clinic has solved one of medicine's most persistent recruitment problems: finding the right patients for clinical trials without exhausting research staff. The AI tool, called Synapsis AI, identified 29 of 30 eligible patients for a heart disease study that traditional recruitment methods had missed, according to research published in the Journal of Cardiac Failure .

Why Are Clinical Trial Recruitment Methods So Broken?

For decades, finding patients who qualify for clinical trials has relied on a tedious, manual process. Research coordinators manually review patient charts, searching for specific diagnoses or lab results that match trial criteria. This approach works, but it is slow, expensive, and often misses eligible patients who exist outside a hospital's main research centers. The problem is especially acute for rare or underrecognized diseases like transthyretin amyloid cardiomyopathy (ATTR-CM), an age-related form of heart failure that many doctors don't immediately recognize .

Roughly 80% of patient information in electronic medical records lives in unstructured form, buried within clinical notes and imaging reports rather than in organized databases. This means the only way to find eligible patients has been through manual review, a process that consumes significant portions of clinical trial budgets and often fails to identify candidates who don't regularly visit specialized centers .

How Does Synapsis AI Actually Work?

Synapsis AI, developed by Dyania Health, was embedded directly into Cleveland Clinic's electronic medical record system across 18 hospitals and 250 outpatient centers. The system uses large language models (LLMs), which are AI systems trained on vast amounts of medical text, to read and understand clinical notes, lab results, and imaging reports. Unlike traditional database searches that rely on diagnostic codes, the AI can interpret the nuanced language doctors use when describing patient conditions .

In the ATTR-CM study, Synapsis AI reviewed 1,476 patient records and identified 46 potential matches. When tested against external physicians on 100 randomly selected patients, the system achieved 96.2% accuracy in answering 7,700 trial-specific questions across nine different medical domains. Importantly, the AI's reasoning for each decision was 100% accurate and interpretable by physician reviewers, meaning doctors could understand exactly why the system flagged a patient as eligible or ineligible .

The system also correctly excluded 198 of 200 ineligible patients, achieving a 99% negative predictive value, which means it reliably prevented false positives that would waste researcher time .

What Makes This Different From Previous AI Recruitment Attempts?

Speed and scale matter in clinical trials. Using just two graphics processing units (GPUs), specialized computing chips that accelerate AI processing, Synapsis AI screened 30 patients within six days. With more computing resources, that timeframe could shrink to a single day . But the real breakthrough isn't just speed; it's the diversity of patients the AI identified.

Of the 30 patients Synapsis AI found, 36.6% were Black, compared to just 7.1% identified through routine screening. This matters because Black patients in the United States disproportionately carry a genetic variant (V1221 in the TTR gene) that causes hereditary cardiac amyloidosis, yet they remain significantly underrepresented in cardiac amyloid trials. The AI also identified patients who weren't connected to heart failure specialists, expanding recruitment beyond Cleveland Clinic's main campus to people researchers would never have encountered in routine practice .

"Without the involvement of EMR-based tools to help find patients, trial recruitment is somewhat of an idiosyncratic process. If you happen to be seeing the patient and you're involved with the study, then you would think about enrolling," explained Trejeeve Martyn, M.D., staff cardiologist and director of heart failure population health at Cleveland Clinic.

Trejeeve Martyn, M.D., Staff Cardiologist and Director of Heart Failure Population Health, Cleveland Clinic

How Are Health Systems Preparing to Deploy AI in Pathology and Beyond?

The success of AI in clinical trial recruitment is part of a broader transformation across healthcare. Major health systems are now deploying AI across multiple specialties, from pathology to ophthalmology. MedStar Health, a large health system in Maryland and Washington, D.C., recently announced a partnership with PathAI to deploy digital pathology infrastructure across its network of laboratories supporting over 40 pathologists . The deployment includes PathAI's AISight Dx platform, a cloud-based system for managing pathology slides and integrating AI algorithms for tasks like detecting artifacts and identifying tumors .

Similarly, the Bascom Palmer Eye Institute hosted Digital EyeCon 2026, a conference exploring how AI and remote patient monitoring are reshaping ophthalmology and healthcare more broadly. The conference emphasized that successful AI adoption requires more than just technology; it demands alignment between clinical workflows, organizational culture, and ethical oversight .

Steps for Health Systems to Implement AI-Driven Patient Recruitment

• Assess Your Data Infrastructure: Evaluate whether your electronic medical records system can support AI integration and whether your clinical notes are sufficiently detailed for AI systems to extract meaningful information about patient eligibility.
• Partner With Validated Vendors: Select AI platforms that have been vetted by clinical teams and tested in real-world settings, similar to how Cleveland Clinic competitively selected Dyania Health based on both technological innovation and clinical expertise.
• Implement Clinician-in-the-Loop Review: Ensure that AI recommendations are always reviewed by qualified physicians before final eligibility determinations, maintaining both safety and accountability in the recruitment process.
• Monitor for Diversity and Equity: Track whether AI-assisted recruitment actually improves representation of underserved populations, as this requires intentional measurement and adjustment of recruitment strategies.
• Plan for Scalability: Begin with pilot programs in specific departments or disease areas, then expand enterprise-wide once workflows are optimized and staff are trained on new systems.

What Do These Developments Mean for Patients and Drug Development?

The implications extend far beyond recruitment efficiency. When clinical trials can identify and enroll more diverse patient populations, the resulting drugs and treatments are tested on people who actually represent the broader population. This helps prevent the historical problem where treatments approved based on predominantly white, male study populations perform differently in other demographic groups .

For patients with rare diseases like cardiac amyloidosis, AI-assisted recruitment means faster access to potentially life-changing therapies. The ATTR-CM study was evaluating an antibody-based therapy designed to remove amyloid proteins that have already accumulated in the heart, addressing an unmet need that previous treatments couldn't solve . Without AI to identify eligible patients scattered across multiple healthcare systems, such trials would take longer to enroll and might never reach their enrollment goals.

The Cleveland Clinic study also demonstrated that AI systems can identify patients connected to different medical centers within an integrated health system. The AI found eligible patients connected to heart failure cardiologists at an enrolling site in Florida, showing that AI recruitment can leverage the full network of a large health system rather than relying on individual physician referrals .

"By deploying AISight Dx and advanced AI applications, we are modernizing our pathology infrastructure in a way that makes us more connected, more efficient, and better positioned to bring innovative diagnostics into everyday practice," stated Moira Larsen, M.D., Physician Executive Director of MedStar Medical Group Pathology at MedStar Health.

Moira Larsen, M.D., Physician Executive Director of MedStar Medical Group Pathology, MedStar Health

As more health systems adopt AI tools for clinical operations, the technology is moving from experimental pilots to enterprise-wide deployment. Cleveland Clinic began rolling out Synapsis AI across all investigators after a year of testing in cancer and cardiovascular disease areas, and uptake among researchers is accelerating . The regulatory path is also becoming clearer; because final eligibility determinations always require clinician review, AI-assisted recruitment tools don't require FDA approval as standalone devices .

The convergence of AI in clinical trial recruitment, digital pathology, and remote patient monitoring suggests that healthcare's next transformation won't come from a single breakthrough technology, but from systematic integration of AI tools across the entire care delivery pipeline. For patients, this means faster access to new treatments and more representative clinical trials. For researchers and health systems, it means lower recruitment costs and the ability to identify eligible patients who would otherwise remain invisible to the medical system.