Mount Sinai Health System, one of the largest academic medical centers in the United States, has become one of the first major institutions to deploy a fully integrated, system-wide AI platform specifically engineered for real-time cancer clinical trial matching.

The platform internally referred to as TrialMatch AI was developed in partnership with Mount Sinai’s in-house AI team, the Institute for Healthcare Delivery Science, and select commercial AI vendors. It went live across the entire health system in early January 2026 after a successful multi-month pilot in the Tisch Cancer Institute.

How TrialMatch AI Works

Continuously monitors newly diagnosed cancer patients and those with disease progression across all Mount Sinai hospitals and outpatient sites.

Extracts structured data (stage, biomarkers, prior therapies) and uses advanced NLP to interpret unstructured clinical notes, pathology reports, radiology, and genomic sequencing results.

Matches patients against an up-to-date database of open trials (internal, national cooperative group, and industry-sponsored) using complex eligibility logic.

Generates ranked match lists with confidence scores and rationale explanations for oncologists and research coordinators.

Integrates directly into the Epic electronic health record via a custom-built sidebar module, enabling one-click referral to the clinical research team.

Early Impact (Pilot Data – Q4 2025)

4–6× increase in the number of patients identified as potentially eligible compared to manual screening.

Time to first match reduced from an average of 9–14 days to under 48 hours in most cases.

Enrollment rate among flagged patients increased by ~35% (preliminary).

Significant reduction in coordinator time spent on manual chart review.

Dr. Cardinale B. Smith, Chief Medical Officer of the Tisch Cancer Hospital, commented: “Clinical trial enrollment has historically been one of the most labor-intensive and inefficient parts of oncology care. TrialMatch AI is changing that equation allowing us to match more patients faster while reducing burnout among our research staff.”

The platform was designed with explainability and bias mitigation at its core. Every match recommendation includes a transparent rationale trace, and the models undergo continuous monitoring for performance drift across diverse patient populations.

Mount Sinai plans to expand the system to all solid tumor and hematologic malignancy disease groups by mid-2026 and is already exploring extensions into non-oncology trials (rare diseases, cardiology).

This deployment positions Mount Sinai as one of the most advanced academic health systems in the United States in operationalizing AI for clinical research acceleration.