智能胃肠病理学的子专业特定基础模型
Subspecialty-specific foundation model for intelligent gastrointestinal pathology.
摘要(中文译文)
胃肠道（GI）疾病构成重大临床负担，而传统组织病理学存在主观性强和可重复性有限的问题。尽管现有的计算病理学基础模型通常在“广而浅”的基准测试中跨多个亚专科进行验证，但很少在真实世界场景中展现深度临床实用性。为解决这一问题，我们开发了Digepath——一个专聚焦于高影响力胃肠道病理学的疾病专用基础模型。我们的方法采用两阶段迭代优化：首先，在来自210,043张H&E染色切片的超过3.53亿个多尺度图像块上进行预训练；其次，在471,443个专家标注区域上进行微调，平衡肿瘤与非肿瘤样本，以增强在全切片图像中病理稀疏情况下的病变感知能力。Digepath在胃肠道病理学的33项系统性下游任务中的32项上达到最先进性能——包括诊断、分子分型和生存预后——展现出稳健的泛化能力。此外，我们将其能力整合到一个基于智能体的临床推理框架中，支持端到端智能诊断工作流程，为真实世界部署铺平道路。
Abstract
Gastrointestinal (GI) diseases pose a major clinical burden, yet conventional histopathology suffers from subjectivity and limited reproducibility. While existing computational pathology foundation models are often validated across many subspecialties in "broad but shallow" benchmarks, they rarely demonstrate deep clinical utility in real-world scenarios. To address this, we develop Digepath-a disease-specialized foundation model focused exclusively on high-impact GI pathology. Our approach employs a two-stage iterative optimization: first, pretraining on over 353 million multi-scale patches from 210,043 H&E-stained slides; second, fine-tuning on 471,443 expert-annotated regions, balancing tumor and non-tumor samples to enhance lesion perception amid sparse pathology in whole-slide images. Digepath achieves state-of-the-art performance on 32 of 33 systematic downstream tasks in GI pathology-including diagnosis, molecular profiling, and survival prognosis-demonstrating robust generalization. Moreover, we integrate its capabilities into an agent-based clinical reasoning framework that supports end-to-end intelligent diagnostic workflows, paving the way for real-world deployment.
如何引用
AMA
Lianghui Zhu, Xitong Ling, Minxi Ouyang, Xiaoping Liu, Tian Guan, Mingxi Fu, et al. Subspecialty-specific foundation model for intelligent gastrointestinal pathology.. NPJ digital medicine. 2026; doi:10.1038/s41746-026-02684-5.
APA
Lianghui Zhu, Xitong Ling, Minxi Ouyang, Xiaoping Liu, Tian Guan, Mingxi Fu, et al (2026). Subspecialty-specific foundation model for intelligent gastrointestinal pathology.. NPJ digital medicine. https://doi.org/10.1038/s41746-026-02684-5
GB/T 7714
Lianghui Zhu, Xitong Ling, Minxi Ouyang, Xiaoping Liu, Tian Guan, Mingxi Fu, et al. Subspecialty-specific foundation model for intelligent gastrointestinal pathology.[J]. NPJ digital medicine, 2026 doi:10.1038/s41746-026-02684-5.
在 PubMed增强版 阅读全文
翻译全文 · AI 文献问答 · PDF 原文下载 · 引用导出