Scholarlink

<p>Large language models (LLMs) show strong performance on clinical benchmarks, yet their reasoning reliability in real-world oncology care remains unclear. We evaluated LLM reasoning on authentic oncology notes using a novel hierarchical error taxonomy across two retrospective cohorts spanning breast, pancreatic, and prostate cancer. GPT-4 produced reasoning errors in 23.1% of note interpretations, the majority reflecting cognitive bias patterns. Errors were more frequent in recommendation tasks and were strongly associated with guideline-discordant recommendations and lower clinician-rated clinical impact scores. Confirmation bias, anchoring bias, and omission errors were most strongly linked to potentially harmful outputs. Compared to GPT-4, GPT-5.1 demonstrated reduced error rates and improved clinical performance but retained structured reasoning failure patterns. Automated LLM-based evaluators detected error presence but failed to reliably classify subtypes, and a preliminary self-mitigation strategy yielded only modest improvement. Endpoint accuracy alone may mask clinically meaningful reasoning failures. Therefore, evaluating and monitoring reasoning fidelity should be a prerequisite for safe deployment of LLMs in oncology decision support.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Structured reasoning failures compromise LLM interpretation of clinical oncology notes

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月30日

查看全文
  • Matthew W. Kenaston,
  • Umair Ayub,
  • Mihir Parmar,
  • Priya Kumar,
  • Muhammad Umair Anjum,
  • Syed Arsalan Ahmed Naqvi,
  • Aadel A. Chaudhuri,
  • Yousef Zakharia,
  • Elisabeth I. Heath,
  • Tanios S. Bekaii-Saab,
  • Cui Tao,
  • Eliezer M. Van Allen,
  • Ben Zhou,
  • YooJung Choi,
  • Chitta Baral,
  • Irbaz Bin Riaz

摘要

Large language models (LLMs) show strong performance on clinical benchmarks, yet their reasoning reliability in real-world oncology care remains unclear. We evaluated LLM reasoning on authentic oncology notes using a novel hierarchical error taxonomy across two retrospective cohorts spanning breast, pancreatic, and prostate cancer. GPT-4 produced reasoning errors in 23.1% of note interpretations, the majority reflecting cognitive bias patterns. Errors were more frequent in recommendation tasks and were strongly associated with guideline-discordant recommendations and lower clinician-rated clinical impact scores. Confirmation bias, anchoring bias, and omission errors were most strongly linked to potentially harmful outputs. Compared to GPT-4, GPT-5.1 demonstrated reduced error rates and improved clinical performance but retained structured reasoning failure patterns. Automated LLM-based evaluators detected error presence but failed to reliably classify subtypes, and a preliminary self-mitigation strategy yielded only modest improvement. Endpoint accuracy alone may mask clinically meaningful reasoning failures. Therefore, evaluating and monitoring reasoning fidelity should be a prerequisite for safe deployment of LLMs in oncology decision support.