Glioma Copilot
Glioma Copilot helps a neuro-oncology clinician and, through them, the patient - decide whether a clinical trial fits this specific glioma patient.
Core trick
A deterministic WHO CNS5 rule engine sets diagnosis, and each trial criterion is marked met, not met, or unknown with a source citation.
Tags
Description
Glioma Copilot helps a neuro-oncology clinician and, through them, the patient - decide whether a clinical trial fits this specific glioma patient. It is not a trial finder: clinicians already know the trials. The hard, still-manual part is fit. Eligibility criteria and patient charts are both free text, so after filtering, a clinician verifies every criterion by hand , since the 2021 WHO reclassification, a stale diagnosis label can silently misroute eligibility.
I built a copilot that reads both sides and produces a per-criterion fit table , met, not-met, or unknown, each source-cited , turning unknowns into a workup checklist. Its principle: Claude is never the source of truth. Diagnosis runs a deterministic WHO CNS5 rule engine; drugs are grounded in RxNorm/ChEMBL; the literature layer is generated with Claude Science and audited against PubMed (23/23, zero hallucinated). A three-agent loop drafts, an Opus verifier rewrites over-claims, and an independent Opus auditor re-derives eligibility blind and challenges the table.
What we found: with real de-identified TCGA patients against real strict trials, almost nothing is a clean "eligible" , so the tool refuses to fake a green match and shows honest unknowns instead. And it does what no trial tool does ,it brings the patient into the decision with a plain-language explanation, a curated FAQ, and a shared-decision handout. It shows the clinician why a trial fits, or doesn't
README excerpt
# Glioma Copilot
**An AI copilot that helps neuro-oncology clinicians review, verify, assess-fit, explain, and document clinical-trial options for one specific glioma patient — and hands that patient a plain-language, shared-decision handout.**
For **neuro-oncologists and tumor boards** (primary user) and, through them, **glioma patients and families** (via a doctor-issued handout). Built for **Built with Claude: Life Sciences** (Cerebral Valley × Anthropic × Gladstone Institutes), Builder Track.
> **This is not a "trial finder."** Clinicians already know the trials; coordinators handle recruitment. The hard, unautomated part is *fit*: for a candidate trial and a specific patient, which eligibility criteria are met, what's missing, which biomarkers matter, what logistical barriers and uncertainties remain — with **every claim source-grounded and independently verified** — then explained in plain language and documented for a shared decision. The goal is never "AI recommends Trial A." It is "AI shows *why* Trial A may or may not fit this patient, for the clinician to review."
---
## The problem
**1. After filtering, a clinician still verifies every criterion by hand.** ClinicalTrials.gov stores eligibility as a single free-text blob — there is no `min_kps`, `max_age`, or `required_biomarker` field. The same threshold is written a hundred ways ("KPS ≥ 60" vs "ECOG 0–1" — a *different scale* — vs "ambulatory, able to care for self"), mixed with fuzzy criteria ("adequate or