Crash Lab

research paper

Why We Built Radiology's Last Exam: The Gap Nobody Was Measuring

Dr. Suvrankar Datta2 Oct 20251 min read

The benchmark problem

If a model claims clinical usefulness, it should be measured against the clinicians it intends to assist or replace. That standard was missing.

About the author

Dr. Suvrankar Datta

CRASH Lab is an interdisciplinary research group building responsible healthcare AI from Ashoka University.