Metraipolis › archive › dc_271707c9ffec

RadMatch: Auditable Radiology Report Evaluation via Finding-Level Matching — arXiv:2609.01470

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.01470
License (as archived)
CC0-1.0
Archived
2026-10-17T07:13:57.198Z
SHA-256 of the archived text
84e42e0cefdb52abe6bc1d28ac959f24f018c22bff0d91ce1bf62eb9a6cb5eef
Stable URL of this page
https://metraipolis.com/documento/dc_271707c9ffec

How to cite

RadMatch: Auditable Radiology Report Evaluation via Finding-Level Matching — arXiv:2609.01470 — arXiv (en), https://arxiv.org/abs/2609.01470 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_271707c9ffec (Metraipolis archive dc_271707c9ffec, sha256 84e42e0cefdb52abe6bc1d28ac959f24f018c22bff0d91ce1bf62eb9a6cb5eef)
BibTeX
@misc{dc271707c9ffec,
  title = {RadMatch: Auditable Radiology Report Evaluation via Finding-Level Matching — arXiv:2609.01470},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01470},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_271707c9ffec; sha256 84e42e0cefdb52abe6bc1d28ac959f24f018c22bff0d91ce1bf62eb9a6cb5eef},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01470. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As AI systems are increasingly used to draft radiology reports, reliably evaluating their clinical quality remains a critical challenge. Large language model (LLM)-based metrics are now the best-correlated with radiologist judgment, yet they output a single opaque score that neither a clinician nor a model builder can easily interpret or audit. We introduce RadMatch, a multi-stage, LLM-based metric that decomposes report comparison into a structured finding-level matching with significance-aware scoring and error characterization across seven clinical attribute dimensions (status, location, severity, morphology, certainty, longitudinal comparison, and measurement). The main score is the actionable-error count, both interpretable and auditable. Candidate findings are graded correct, partial, or incorrect, and unmatched findings are counted as missed or hallucinated. Triage and actionable safety recall/precision and per-subset views add complementary, deployment-oriented lenses. Across two expert benchmarks, RadMatch is the most clinically aligned metric, matching inter-radiologist agreement on ReXVal and more than doubling the best prior metric on the harder RadEvalExpert. Relying only on few-shot prompting, it is designed to extend to other modalities and anatomies. We will release RadMatch as open-source code with an interactive dashboard for inspecting results.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-02 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.01470 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T07:00:58.991Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).