Metraipolis › archive › dc_723d3f97e61e
OpenMTB-Audit: Exposing Over-Refusal and Clinical Expert Perspectives in LLM-Based Molecular Tumor Board Safety Evaluation — arXiv:2610.01497
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.01497
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T12:05:22.903Z
- SHA-256 of the archived text
9beb0c0e449914c9d91ab8acba0fd6e3060a3cb790e1bbe170b06a3f0fafc8f9- Stable URL of this page
https://metraipolis.com/documento/dc_723d3f97e61e
How to cite
OpenMTB-Audit: Exposing Over-Refusal and Clinical Expert Perspectives in LLM-Based Molecular Tumor Board Safety Evaluation — arXiv:2610.01497 — arXiv (en), https://arxiv.org/abs/2610.01497 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_723d3f97e61e (Metraipolis archive dc_723d3f97e61e, sha256 9beb0c0e449914c9d91ab8acba0fd6e3060a3cb790e1bbe170b06a3f0fafc8f9)
BibTeX
@misc{dc723d3f97e61e,
title = {OpenMTB-Audit: Exposing Over-Refusal and Clinical Expert Perspectives in LLM-Based Molecular Tumor Board Safety Evaluation — arXiv:2610.01497},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.01497},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_723d3f97e61e; sha256 9beb0c0e449914c9d91ab8acba0fd6e3060a3cb790e1bbe170b06a3f0fafc8f9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.01497. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Molecular tumor boards integrate genomic findings, clinical context, and therapeutic evidence to support precision oncology. As AI enters this workflow, a key safety challenge is distinguishing truly unsupported recommendations from evidence-supported options that still require oncologist review because of incomplete information, poor ECOG performance status, or other clinical caveats. We introduce OpenMTB-Audit, an open-source benchmark of 500 synthetic non-small cell lung cancer cases spanning five adversarial error categories and four safety labels: Supported, Partially Supported, Unsupported, and Insufficient Information. Across eight large language model configurations, we identify pervasive over-refusal: all LLM configurations failed to retain the Partially Supported label in 83.3-100% of true Partially Supported cases, achieving high aggregate safety scores through label collapse rather than clinically calibrated reasoning. To address this limitation, we developed MTB-AuditAgent, a deterministic seven-module framework separating evidence verification, missing-information detection, safety classification, and abstention. It reduces over-refusal to 6.7% and achieves 91.2% accuracy (95% CI: 88.6-93.6%). A two-oncologist annotation study found disagreement concentrated at the boundary between information sufficiency and treatment optimization, underscoring the need to preserve clinically meaningful distinctions. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.01497 — 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 2026-10-06T10:21:27.956Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).