Beyond Scalar Scores: Exploring LLM-based Metrics for Clinical Significance Evaluation in Radiology Reports

In three linesStudy on evaluating AI-generated radiology reports. Researchers show existing LLMs over-penalize harmless rephrasings while detecting clinical errors. They train lightweight metrics on Qwen3-8B and MedGemma-4B outperforming 32B medical models, with dataset and metric release planned.

Summary generated by Claude — human-verified