用于从全片分类器生成组织形态学假设的稀疏概念归因
Sparse concept attribution for histomorphological hypothesis generation from whole-slide classifiers
- Microsoft Research（微软研究院）
- Berlin Institute for the Foundations of Learning and Data (BIFOLD)（柏林学习与数据基础研究所）
- Technische Universität Berlin（柏林工业大学）
- Harvard–MIT Division of Health Sciences and Technology（哈佛-麻省理工学院健康科学与技术系）
- University of Cambridge（剑桥大学）
- Emory University School of Medicine（埃默里大学医学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究提出SCOPE方法，结合病理视觉-语言模型与稀疏概念归因，在MorphoRecoveryBench基准上验证其可从全片分类器生成形态学假设，且计算成本低于密集归因。
AI中文摘要：
组织学图像包含丰富的形态学信息，可为病理过程提供见解。然而，推导形态表型与临床属性相关的假设受到手动图像解读步骤的瓶颈限制。在此，我们证明该过程可通过可解释深度学习实现自动化。我们提出SCOPE方法，该方法结合病理特定视觉-语言模型与针对通用组织形态学概念库的稀疏概念归因，以解释切片级分类器。为衡量此类解释是否恢复已知形态，我们引入MorphoRecoveryBench，这是一个包含7项任务的基准，配有病理学家精心整理的参考描述。在该基准上，密集概念归因与随机基线无差异，而稀疏归因可恢复大量已知形态；对合并切片嵌入进行分解，可在计算成本大幅降低的情况下达到相似的解释正确性。因此，全片分类器的事后解释可大规模生成形态学假设，供专家验证。
英文摘要：
Histology images contain rich morphological information and can provide insights into pathological processes. However, deriving hypotheses relating morphological phenotypes to clinical attributes is bottlenecked by a manual image interpretation step. Here, we demonstrate that this process can be automated through interpretable deep learning. We present SCOPE, a method to interpret slide-level classifiers by combining pathology-specific vision--language models with sparse concept attribution onto a generalist histomorphological concept bank. To measure whether such explanations recover known morphology, we introduce MorphoRecoveryBench, a benchmark of seven tasks with pathologist-curated reference descriptions. On this benchmark, dense concept attribution is indistinguishable from a random baseline, whereas sparse attribution recovers substantial known morphology; decomposing the pooled slide embedding reaches similar explanation correctness at a fraction of the computational cost. Post-hoc interpretation of whole-slide classifiers can thus generate morphological hypotheses at scale, for expert validation.
