从组织病理学获得人工智能支持的虚拟空间蛋白质组学,用于在肺癌中发现可解释的生物标志物
Zhe Li1, Yuchen Li1, Jinxi Xiang1
1Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, USA.
Nature medicine
|January 6, 2026
概括
我们开发了HEX,这是一种AI模型,可以从标准病理学幻灯片中预测空间蛋白质表达. 这种人工智能驱动的方法提高了癌症预后和免疫治疗反应预测,为精准医学提供了一个可扩展的工具.
科学领域:
- 生物医学成像学 生物医学成像学
- 计算病理学计算病理学
- 蛋白质组学是指蛋白质组学.
背景情况:
- 空间蛋白质组学提供了高分辨率的蛋白质表达映射,但面临着成本和可扩展性等临床翻译挑战.
- 标准的组织病理学幻灯片缺乏详细的蛋白质表达数据,这些数据对于了解疾病至关重要.
- 需要将AI整合到病理学中,以弥合成像和分子洞察力之间的差距.
研究的目的:
- 开发和验证一个人工智能模型 (HEX) 用于从血素和乙素 (H&E) 染色基因病理学幻灯片计算生成空间蛋白质组学概况.
- 评估HEX在预测蛋白质表达和提高癌症预后和预测准确性的性能.
- 探索整合人工智能衍生的虚拟空间蛋白质组与H&E图像的实用性,以改善临床决策.
主要方法:
- 开发了HEX,这是一款在819,000个组织病理学图像上训练的AI模型,并匹配了来自382个瘤样本的蛋白质表达数据.
- 验证了HEX在预测40种不同生物标志物的准确性,并将其性能与替代方法进行比较.
- 实施了多式联运数据集成方法,将H&E图像与AI衍生的虚拟空间蛋白质组合起来.
主要成果:
- HEX准确地预测了40种生物标志物的表达,比H&E图像中蛋白质预测的现有方法表现出更高的性能.
- 在6个独立的非小细胞肺癌队列 (2298名患者) 中,使用HEX的多模式整合提高了22%的预后准确性和24-39%的免疫治疗反应预测.
- 确定了特定的瘤免疫,例如响应者中的T辅助和细胞毒性T细胞共定位,以及非响应者中的免疫抑制细胞聚合物,以预测治疗结果.
结论:
- HEX提供了一种低成本,可扩展的方法,用于从标准基因病理学幻灯片分析空间生物学.
- 与H&E成像集成的AI衍生虚拟空间蛋白质学显著提高了对患者结果和治疗反应的预测.
- 赫克斯促进了可解释生物标志物的发现和临床翻译,为精准医学铺平了道路.
更多相关视频
09:19Spatial Profiling of Protein and RNA Expression in Tissue: An Approach to Fine-Tune Virtual Microdissection
Published on: July 6, 2022
5.3K
13:01Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
Published on: June 3, 2022
4.4K
相关概念视频
02:43Mouse Models of Cancer Study
Mice have long served as models for studying human biology and pathology because of their phylogenetic and physiological similarity with humans. They are also easy to maintain and breed in the laboratory, and hence, many inbred strains are now available for research. Studies on mice have contributed immeasurably to our understanding of cancer biology.
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
02:43Mouse Models of Cancer Study
Mice have long served as models for studying human biology and pathology because of their phylogenetic and physiological similarity with humans. They are also easy to maintain and breed in the laboratory, and hence, many inbred strains are now available for research. Studies on mice have contributed immeasurably to our understanding of cancer biology.
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...