Conference Presentation · AMIA 2026
Explainable, precision clinical trial matching built on an open oncology patient model — connecting cancer patients with potentially life-saving trials in minutes, with full transparency into every eligibility decision.
Clinical trial enrollment has stagnated at roughly 2% of eligible cancer patients — not because trials are scarce, but because matching patients to them is opaque, slow, and requires clinical expertise most patients don't have access to. EXACT was built to change that.
This presentation introduces EXACT (EXtracting Attributes from Clinical Trials), a stateless, open-source matching engine that evaluates every eligibility criterion for every trial in a catalog and returns a per-criterion trace — each criterion marked passed, failed, or indeterminate, with the patient value and trial threshold behind each verdict. No black-box scores. No opaque rankings. Every decision is legible to the patient and their care team.
EXACT runs on top of CTOMOP, an open oncology extension to the OMOP Common Data Model, giving it a structured, standardized view of the patient record. Together they form the open foundation beneath HealthKey's personal health record — and the engine behind CancerBot, which has matched patients across multiple myeloma, follicular lymphoma, and breast cancer trials.
EXACT on GitHub — open-source, MIT licensed.
AMIA 2026 — Session #95115 — American Medical Informatics Association annual conference.
HealthKey — the personal health record built on CTOMOP and EXACT.