NeuroRad Atlas
A living map of what neuroradiology knows — and what it still doesn’t.
NeuroRad Atlas organizes selected Brain/Glioma, Spine, and Head & Neck literature into evidence-graded claims, controversies, and research gaps — so you can see what is clinically supported, what remains unsettled, and where the field should look next.
Currently mapped
Three subspecialties, held to one standard.
Read straight from the live Atlas graph — the same database the map itself runs on, not a brochure.
— nodes · — edges · — papers — live · Supabase
Atlas does not index every paper or every topic. Topics are included when they meet at least two of four criteria: clinical decision-relevance, active disagreement, standardization pressure, and public data availability.
How is this different?
Traditional searches help you find papers. NeuroRad.ai helps you understand the field.
Every paper becomes part of a connected research map — where claims, evidence, disagreements, and unanswered questions are linked together.
The neuroradiology literature, as one connected map — every point a concept, every line a relationship in the evidence. Open the live map →
A new way to think about research
Stop asking which paper to read next. Start asking what the field knows.
“Which paper should I read next?”
“What does the field know?”
Instead of collecting PDFs, you’re exploring a living map of evidence.
From papers to maps.
An example
Follow a claim through the evidence. Get the whole picture, not a tidy answer.
Live from the map · Controversy C04
Can MRI reliably predict MGMT methylation in glioblastoma?
Weak
Evidence grade — the claim that it can is not well supported.
What some models claim
Multiple ML/DL models report AUCs of 0.7–0.9 for MGMT prediction on multiparametric MRI.
What rigorous validation shows
Well-designed studies conclude MRI cannot reliably predict MGMT — external test performance is near chance, and apparent gains trace to data leakage and overfitting.
Why it matters: MGMT status governs temozolomide benefit — a false imaging surrogate could misdirect therapy or trial stratification. So the grade, not just the claim, is the point.
Computed from the cited literature · re-graded as new papers land
See how the evidence splits →
Every claim has a computed evidence grade you can inspect yourself — nothing here is magic.
One topic on the map
Take glioblastoma — everything known about it, connected in one place.
Why NeuroRad.ai
What the textbook says and what the evidence supports aren’t always the same thing.
- where the evidence comes from
- how strong it actually is
- where experts still disagree
- how ideas connect across the literature
- what questions remain unanswered
Who it’s for
For anyone navigating the evidence, at any level.
Residents
learning a topic — see the whole shape of it at once.
Fellows
exploring the literature — follow the disagreements, not just the conclusions.
Practicing neuroradiologists
checking the evidence behind a claim at the workstation.
Researchers
identifying the unanswered questions worth studying.
The map doesn’t go stale.
A monthly engine reads new literature, proposes claim diffs, and re-grades what changed — always under human review before anything commits. As the literature evolves, so does the map. It’s free.
On using this responsibly.
NeuroRad.ai is a research and education instrument. Evidence grades are computed from the literature to show where the field agrees, disagrees, and hasn’t looked — they are not clinical recommendations, and nothing here should drive individual patient management. Read the methods for exactly how each grade is derived.