OncoTriad-QA：泛癌推理的放射-病理-基因组学基准
原标题：OncoTriad-QA: A Patient-Level Radiology-Pathology-Genomics Benchmark for Pan-Cancer Reasoning
AI 摘要
arXiv 发布了一项名为 OncoTriad-QA 的基准，用于评估医学大语言模型和视觉语言模型在泛癌推理中的能力，该基准整合了放射学、病理学、基因组学和临床元数据。研究团队还提出了参考模型 OncoVLM，在微调后其性能显著优于 MedGemma-4B，平均提升 10.7 分。该基准包含 86.1k 个问题，覆盖 32 个癌症队列的 9,281 个 TCGA 病例，为多模态癌症问答提供了训练和评估资源。
正文节选
Computer Science > Computation and Language Title:OncoTriad-QA: A Patient-Level Radiology-Pathology-Genomics Benchmark for Pan-Cancer Reasoning View PDF Abstract:Cancer diagnosis and characterization require integrating complementary evidence from radiology, pathology, genomics, and clinical metadata. However, most medical large language model (LLM) and vision-language model (VLM) benchmarks focus on isolated modalities or narrow image-text tasks, leaving patient-level oncology asses