超越病灶：从反应性中枢神经系统组织进行疾病识别
Seeing Beyond the Lesion: Disease Recognition from Reactive CNS Tissue
- Institute of Neuropathology, University Hospital Münster（明斯特大学医院神经病理研究所）
- Institute for Geoinformatics, University of Münster（明斯特大学地理信息研究所）
- Faculty of Mathematics and Computer Science, University of Münster（明斯特大学数学与计算机科学学院）
- Institute for Machine Learning in Medicine (Focus Area Psychiatry), University of Münster（明斯特大学医学机器学习研究所（精神病学重点领域））
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究针对颅内活检常获非病灶组织致无法诊断的问题，以四种病理基础模型为编码器，在多实例学习框架下验证，发现可从非病灶反应性组织恢复疾病信号，需控制抽样混杂因素。
AI中文摘要：
在相当比例的颅内活检样本中，抽样误差会导致仅获取反应性、非病灶性的脑实质，从而无法对潜在疾病做出诊断。我们以245张来自186名经确诊下游诊断患者的全切片图像为数据，将四种病理基础模型（UNI2-h、Virchow2、Prov-GigaPath、H-optimus-0）作为冻结的块编码器，在基于共享注意力的多实例学习框架内进行基准测试。我们首先证明，粗略的疾病类别预测在很大程度上可仅通过切片大小重现；在将分类限制为常见组织类别内的三种更精细诊断区分后，该混杂因素不再能解释性能，但在置换检验下（全程p≤10⁻⁴），疾病标签仍可被预测至高于随机水平。令人惊讶的是，所有基础模型编码器的性能在统计学上无显著差异，表明恢复这些微弱的形态学特征并不受限于当前的块表示。带符号的实例贡献图及专家评审进一步测试预测证据是否定位于反应性实质，而非组织抽样期间引入血液等抽样诱导偏差。这些结果表明，通过仅基于来源的基准进行获取捷径审计，是计算病理学基准中必要的控制措施；且在移除该混杂因素后，弱监督模型仍可从传统上被视为非诊断性的组织中恢复疾病信号。
英文摘要：
Sampling error yields exclusively reactive, non-lesional brain parenchyma in a significant proportion of intracranial biopsies, leaving the underlying disease undiagnosed. We benchmark four pathology foundation models (UNI2-h, Virchow2, Prov-GigaPath, H-optimus-0) as frozen patch encoders within a shared attention-based multiple-instance learning framework using 245 whole-slide images from 186 patients with confirmed downstream diagnoses. We first show that coarse disease-category prediction can be reproduced largely from slide size alone. After restricting classification to three finer diagnostic distinctions within common tissue categories, this confound no longer explains performance, yet disease labels remain predictable above chance under permutation testing (p $\le 10^{-4}$ throughout). Surprisingly, performance is statistically indistinguishable across all foundation-model encoders, suggesting that recovering these weak morphological signatures is not limited by current patch representations. Signed instance-contribution maps and expert review further test whether predictive evidence localizes to reactive parenchyma rather than sampling-induced bias like blood introduced during tissue sampling. These results position acquisition-shortcut auditing via a provenance-only baseline as a necessary control in computational-pathology benchmarks, and show, once that confound is removed, that weakly supervised models still recover disease signal from tissue conventionally regarded as non-diagnostic.
