CT-DeltaBench for Longitudinal Medical Imaging Difference Reporting
August 13, 2026
CT-DeltaBench is a new benchmark for evaluating vision-language models on their ability to compare serial CT scans and report interval changes. It utilizes patient-level splitting to ensure rigorous evaluation of temporal disease evolution detection.
HOW THIS AFFECTS YOU
●
researcherUse this benchmark to test how well your vision-language models handle temporal reasoning in medical contexts.
●
healthThis advances the development of tools capable of tracking disease progression over time via automated reporting.