[论文] Health-System-Wide Mapping of Demographic Information in Routine Clini...

论文概要

研究领域: CV
作者: Shengjia Chen, Ruchika Verma, Kevin Clare, Paula Restrepo, Jannes Jegminat
发布时间: 2026-11-06
arXiv: 2611.09851

中文摘要

人口统计学相关信息在常规临床组织病理学中、跨器官和疾病状态的可检测程度仍不清楚。本文分析了来自257,322名患者、25个器官的130万张苏木精-伊红染色全切片图像，使用病理学基础模型表示和多实例学习，系统绘制与性别、年龄组和自报种族相关的信息图谱。在69个器官-人口组合中，人口属性被广泛恢复：跨器官平均AUROC分别为性别0.828、年龄组0.733、自报种族0.662，且跨器官存在显著差异。人口类别平衡对性能影响相对较小，在诊断文本聚类内评估显示轻微衰减。在基于诊断分组训练下，人口相关信号在多个器官仍可检测：胃组织中的性别（AUROC 0.865）、乳腺中的年龄组（0.816）、皮肤中的自报种族（0.728）。基于文献的分析为这些场景生成了候选的组织形态学和上下文解释。这些发现提供了常规组织病理学中人口相关信息的系统图谱，有助于调查其来源、评估潜在捷径学习，并开发更具可解释性的计算病理学模型。

原文摘要

The extent to which demographic-associated information is detectable across organs and disease states in routine clinical histopathology remains unclear. Here, we analyzed 1.3 million hematoxylin and eosin whole-slide images from 257,322 patients across 25 organs, using pathology foundation-model representations and multiple-instance learning to systematically map information associated with sex, age group, and self-reported race. Across 69 organ-demographic combinations, demographic attributes were broadly recoverable, with pan-organ mean AUROCs of 0.828 for sex, 0.733 for age group, and 0.662 for self-reported race, and substantial variation across organs. Demographic-class balancing produced relatively small changes in performance, while evaluation within diagnosis-text clusters showed ...

自动采集于 2026-11-10

#论文 #arXiv #CV #小凯