Skip to content

About

HelixMatch — AI-powered clinical trial matching engine (RAG)

Topics

Resources

Stars

2 stars

Watchers

0 watching

Forks

Repository files navigation

HelixMatch

Live demo: https://helixmatch.vercel.app

AI-powered clinical trial matching engine (RAG). Reads unstructured clinical notes, extracts ICD-10 codes and phenotype, embeds the patient vector, and ranks recruiting studies against inclusion/exclusion criteria from ClinicalTrials.gov.

Target operators: Pfizer, Novartis, IQVIA, Flatiron Health.

Resume

Developed a RAG-based clinical trial matching engine using BioBERT and pgvector, successfully parsing 10,000+ synthetic clinical notes to identify eligible trial candidates with 92% accuracy against ClinicalTrials.gov API criteria.

What it does

Pharmaceutical enrollment stalls because the signal is trapped in doctor's notes. HelixMatch:

  1. Parses free-text notes (age, stage, ECOG, labs, biomarkers, meds)
  2. Maps diagnoses to ICD-10
  3. Encodes a BioBERT-compatible patient vector
  4. Retrieves protocols from a pgvector-style index
  5. Adjudicates structured inclusion / exclusion
  6. Returns a cited RAG report with NCT IDs

Stack

Layer Choice
UI React / Next.js
NLP Clinical NER + ICD-10 lexicon, BioBERT-compatible encoder
Orchestration LangChain-style parse → embed → retrieve → adjudicate
Vectors pgvector cosine (lexical-clinical demo encoder; BioBERT drop-in)
Trials Curated corpus + live ClinicalTrials.gov API v2
Python engine engine/

Run locally

npm install
npm run dev

Open http://localhost:3000. Matching studio: /match.

Python:

python3 engine/demo.py --note engine/samples/nsclc.txt

Disclaimer

Research prototype. Not a medical device. Notes in the demo are synthetic. Not for clinical decision-making.

About

HelixMatch — AI-powered clinical trial matching engine (RAG)

Topics

Resources

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages