{"paper":{"title":"Can Language Models Identify Side Effects of Breast Cancer Radiation Treatments?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Grounding large language models in clinician-curated lists improves reliability when listing side effects of breast cancer radiation treatments.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Danielle S. Bitterman, Daphna Spiegel, Natalie Seah, Thomas Hartvigsen","submitted_at":"2026-05-08T20:02:49Z","abstract_excerpt":"Accurately communicating the side effects of cancer treatments to cancer survivors is critical, particularly in settings such as informed consent, where clinicians must clearly and comprehensively convey potential treatment toxicities. However, this task remains challenging due to clinical knowledge deficits about adverse treatment effects and fragmentation across electronic health record (EHR) systems. Large language models (LLMs) have the potential to assist in this task, though their reliability in oncology survivorship contexts remains poorly understood. We present a deployment-oriented st"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Grounding outputs in clinician-curated side effect lists substantially improves reliability and robustness while constraints on the number of side effects generated reduce precision.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The clinician-curated reference derived from informed consent documents at two academic centers accurately and comprehensively captures all relevant toxicities broken down by frequency and temporal onset.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LLMs under-recall rare and long-term radiation side effects in breast cancer, show prompt sensitivity, and improve when outputs are grounded in clinician-curated references.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Grounding large language models in clinician-curated lists improves reliability when listing side effects of breast cancer radiation treatments.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"446d3367ba24f3ed510331417a6217df3796784f7ccccfad6e0ed97f12fa37d4"},"source":{"id":"2605.08439","kind":"arxiv","version":2},"verdict":{"id":"4771daa1-0735-4b58-8e16-525c3bf7ff65","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T02:34:30.458924Z","strongest_claim":"Grounding outputs in clinician-curated side effect lists substantially improves reliability and robustness while constraints on the number of side effects generated reduce precision.","one_line_summary":"LLMs under-recall rare and long-term radiation side effects in breast cancer, show prompt sensitivity, and improve when outputs are grounded in clinician-curated references.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The clinician-curated reference derived from informed consent documents at two academic centers accurately and comprehensively captures all relevant toxicities broken down by frequency and temporal onset.","pith_extraction_headline":"Grounding large language models in clinician-curated lists improves reliability when listing side effects of breast cancer radiation treatments."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.08439/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T15:01:17.898159Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T11:08:09.549339Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5d87e99d4f49590c54d2ebea262b7a73e2c2f1d56704eb8707b73814b33ce6f5"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}