我国科学家构建人工智能框架 助力乳腺癌等疾病治疗

Core Insights - A Chinese research team has achieved groundbreaking results in the field of spatial multi-omics data integration, developing the SpatialEx AI framework to address the "data island" issue, which provides new pathways for precise diagnosis and treatment of complex diseases like breast cancer and Parkinson's disease [1][2] Group 1: Technological Breakthroughs - The technology features two major breakthroughs: first, it utilizes a histological image-based foundational model to directly interpret deep molecular information at the single-cell level from standard histological stained images [2] - Second, it employs an innovative "omics cycling module" to seamlessly integrate different single-omics data originally distributed across adjacent slices using histological images as a universal "bridge" [2] Group 2: Applications and Implications - In breast cancer research, the technology demonstrated superior analytical capabilities, successfully constructing a "whole slide view" of the tissue and accurately identifying subtle differences in the immune microenvironment that even pathologists find challenging to distinguish [2] - In the Parkinson's disease model, the technology revealed the co-variation patterns of dopamine-related genes and metabolites in brain injury areas by integrating metabolomics and transcriptomics data, providing a new dimension for disease mechanism research [2] - This achievement offers a Chinese solution to the spatial omics "data island" problem and is expected to significantly advance the application of spatial multi-omics technology in precision medicine, drug development, and clinical diagnostics [2]