HERO:基于假设驱动的组学证据检索用于多任务乳腺癌分析
HERO: Hypothesis-Driven Evidence Retrieval from Omics for Multi-Task Breast Cancer Analysis
AI总结:
提出HERO框架,将组学数据作为形态学假设进行显式检索,通过稀疏通路先验和TF-IDF检索实现可审计的多模态乳腺癌分析,在TCGA-BRCA上取得多项预测任务的最优结果。
摘要:
匹配的多组学数据可以改善基于WSI的生物标志物和预后预测,但大多数现有流水线将组学用作并行特征流或文本上下文,而不是作为显式检索约束。HERO询问观察到的组学是否可以作为可测试的形态学假设:一个稀疏的通路到形态学先验将DNA甲基化和miRNA映射为K维意图向量m(K=16),对结构化10个标题进行TF-IDF检索以选择端点相关区域,余弦门c=cos(m,v)在c<τ时触发确定性缺陷驱动修复。这种闭环设计限制了VLM调用,减少了对基于嵌入的语义匹配的依赖,并使每个检索和验证步骤在词汇上可审计。在TCGA-BRCA(930个WSI,患者级5折交叉验证)上,HERO在ER、PR、HER2、亚型和风险预测中均取得了新的最优结果,优于多模态融合和基于VLM的基线。
英文摘要:
Matched multi-omics can improve WSI-based biomarker and prognosis prediction, but most existing pipelines use omics as a paral lel feature stream or textual context rather than as an explicit retrieval constraint. HERO asks whether observed omics can be a testable mor phology hypothesis: a sparse pathway-to-morphology prior maps DNA methylation and miRNA into a K-dimensional intent vector m (K=16), TF-IDF retrieval over structured 10 captions selects endpoint-relevant regions, and a cosine gate c=cos(m,v) triggers deterministic deficit driven repair when c<τc. This closed-loop design bounds VLM calls, reduces reliance on embedding-based semantic matching, and makes every retrieval and verification step lexically auditable. On TCGA-BRCA (930WSIs, patient-level 5-fold CV), HERO sets new state-of-the-art across ER, PR, HER2, subtype, and risk prediction, outperforming both multimodal fusion and VLM-based baselines.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。