Ellis Spencer, Song Steven, Reiman Derek, Hui Xuan, Zhang Renyu, Shahriar Mohammad Hasan, Argos Maria, Kamal Mohammed, Shea Christopher R, Grossman Robert L, Khan Aly A, Ahsan Habibul
Departments of Pathology and Family Medicine, University of Chicago, Chicago, Illinois.
Department of Computer Science, University of Chicago, Chicago, Illinois.
Cancer Epidemiol Biomarkers Prev. 2025 Jul 1;34(7):1080-1088. doi: 10.1158/1055-9965.EPI-25-0132.
Early and precise diagnosis is vital to improving patient outcomes and reducing morbidity. In resource-limited settings, cancer diagnosis is often challenging due to shortages of expert pathologists. We assess the effectiveness of general-purpose pathology foundation models (FM) for the diagnosis and annotation of nonmelanoma skin cancer (NMSC) in resource-limited settings.
We evaluated three pathology FMs (UNI, PRISM, and Prov-GigaPath) using deidentified NMSC histology images from the Bangladesh Vitamin E and Selenium Trial to predict cancer subtype based on zero-shot whole-slide embeddings. In addition, we evaluated tile aggregation methods and machine learning models for prediction. Lastly, we employed few-shot learning of PRISM tile embeddings to perform whole-slide annotation.
We found that the best model used PRISM's aggregated tile embeddings to train a multilayer perceptron model to predict NMSC subtype [mean area under the receiver operating characteristic curve (AUROC) = 0.925, P < 0.001]. Within the other FMs, we found that using attention-based multi-instance learning to aggregate tile embeddings to train a multilayer perceptron model was optimal (UNI: mean AUROC = 0.913, P < 0.001; Prov-GigaPath: mean AUROC = 0.908, P < 0.001). We finally exemplify the utility of few-shot annotation in computation- and expertise-limited settings.
Our study highlights the important role FMs may play in confronting public health challenges and exhibits a real-world potential for machine learning-aided cancer diagnosis.
Pathology FMs offer a promising pathway to improve early and precise NMSC diagnosis, especially in resource-limited environments. These tools could also facilitate patient stratification and recruitment for prospective clinical trials aimed at improving NMSC management.
早期准确诊断对于改善患者预后和降低发病率至关重要。在资源有限的环境中，由于缺乏专业病理学家，癌症诊断往往具有挑战性。我们评估通用病理学基础模型（FM）在资源有限环境中对非黑色素瘤皮肤癌（NMSC）进行诊断和注释的有效性。
我们使用来自孟加拉国维生素E和硒试验的去识别化NMSC组织学图像评估了三种病理学FM（UNI、PRISM和Prov-GigaPath），以基于零样本全切片嵌入预测癌症亚型。此外，我们评估了切片聚合方法和用于预测的机器学习模型。最后，我们采用PRISM切片嵌入的少样本学习来进行全切片注释。
我们发现最佳模型使用PRISM的聚合切片嵌入来训练多层感知器模型以预测NMSC亚型[受试者操作特征曲线下的平均面积（AUROC）= 0.925，P < 0.001]。在其他FM中，我们发现使用基于注意力的多实例学习来聚合切片嵌入以训练多层感知器模型是最优的（UNI：平均AUROC = 0.913，P < 0.001；Prov-GigaPath：平均AUROC = 0.908，P < 0.001）。我们最终举例说明了少样本注释在计算和专业知识有限的环境中的效用。
我们的研究强调了FM在应对公共卫生挑战中可能发挥的重要作用，并展示了机器学习辅助癌症诊断在现实世界中的潜力。
病理学FM为改善早期准确的NMSC诊断提供了一条有前景的途径，尤其是在资源有限的环境中。这些工具还可以促进患者分层和招募，以进行旨在改善NMSC管理的前瞻性临床试验。
