arXiv:2609.32810 · 评测基准
OpenTumorBoard: A Real-World Benchmark of Multidisciplinary Tumor Board Discussion Trajectories
OpenTumorBoard:多学科肿瘤委员会讨论轨迹的真实世界基准
OpenTumorBoard: A Real-World Benchmark of Multidisciplinary Tumor Board Discussion Trajectories
- 类型:arxiv
- 标识:2609.32810
- 链接:https://arxiv.org/abs/2609.32810
- 主分类:evaluation
- 形态:benchmark
- TLDR:Multidisciplinary tumor boards integrate multimodal clinical observations and longitudinal patient histories through specialist discussions, yet benchmarks rarely capture these real-world trajectories. We introduce OpenTumorBoard, a benchmark with 611 patient cases and 19,157 discussion turns across ten specialist roles, transcribed from 12,534 minutes of publicly available tumor board recordings on YouTube. The benchmark evaluates two settings: SPECIALIST TURN, in which an LLM responds to a clinically significant question posed during a real discussion, and BOARD SIMULATION, in which it gener
- 待LLM分类:否
- 标题中文:OpenTumorBoard:多学科肿瘤委员会讨论轨迹的真实世界基准
- TLDR中文:多学科肿瘤委员会通过专家讨论整合多模态临床观察与纵向患者病史,但现有基准很少捕捉这些真实世界轨迹。我们推出 OpenTumorBoard,该基准包含 611 个病例和 19,157 轮讨论,覆盖十个专家角色,转录自 YouTube 上 12,534 分钟的公开肿瘤委员会录像。该基准评估两种设置:SPECIALIST TURN——LLM 回答真实讨论中提出的临床重要问题;BOARD SIMULATION——LLM 生成……
- 来源文件:
- /inbox/tom/_candidates/2026-10-03-agent-rag-longcontext-candidates.json