Coris Match reads the chart and the image. Imaging AI, a clinical-notes LLM, and discrete EHR data turn your full patient population into a verified, coordinator-ready candidate list.
The criteria that decide a patient — imaging findings, disease severity, prior treatment — live in scans and free-text notes, not in structured fields alone. Code-based screening quietly misses qualifying patients across every indication.
Eligibility is scattered across scans, notes, and structured fields. We built a model for each — and run them together.
Purpose-trained models read medical images directly. The library is designed to grow into new modalities and areas over time.
A language model reads free-text clinical notes the way a coordinator would — surfacing eligibility signals that never make it into structured fields.
We pull structured variables — diagnoses, medications, labs, demographics, and visit history — to apply inclusion and exclusion rules precisely.
One platform, room to grow. The same engines are built to extend to new biomarkers and areas over time — so the platform can grow alongside your pipeline.
On a live trial, Coris Match surfaced far more trial-ready candidates than the site's manual, referral-based process — found and verified in minutes.
See the results in a demotrial-ready candidates than manual referral surfaced
minutes from raw data to a verified call list
candidates from the list are in screening today
Purpose-built and iterated for the trial's exact eligibility criteria — not a generic recruiter pointed at a new problem.
Structured data, free-text notes, and image-derived criteria run in one workflow, so nothing falls between disconnected tools.
Every match is explainable and confirmed by a clinician before it reaches a coordinator.
Designed to surface every eligible patient, then confirmed in a fast review step — hours of chart work compressed to minutes.
Chief Medical Officer · Founder
Chief Technology Officer · Founder
Chief Science Advisor · Founder