[The application and challenges of multi-modal data fusion based on deep learning in pathology].

Abstract

In recent years, with the rapid development of artificial intelligence technology, the application of deep learning in the field of pathology has been continuously expanding. Particularly, the rise of multimodal data fusion methods has opened up new technical paths for the precise diagnosis, prognosis assessment, and individualized treatment of tumors. By integrating multi-level and multi-source data such as clinical information, pathological omics, molecular omics, and imaging omics, deep learning models can identify potential associated features and key biological mechanisms that are difficult to reveal by a single modality, thereby significantly improving the accuracy of disease classification and the scientific nature of risk stratification. This article systematically reviews the research progress of multimodal data fusion methods based on deep learning in the field of pathology in recent years, focuses on sorting out different types of fusion strategies, evaluates their advantages and challenges in practical clinical applications, and looks forward to future development trends.