Emerging advances in multimodal AI in cancer immunotherapy: from multi-scale data integration to clinical decision support.

Authors

Wenjie Zhang, Zeyu Luo, Kaijie Liu, Yuzhu Wang, Yongying Hou, Xiaotan Su, Xisong Wu, Cheng Miao, Qin Liu, Zongsheng He, Qiaoqiao Zhang, Guohua Wang, Xiuwu Bian, Bin Wang

Published in

Science bulletin. Sep 08, 2026. Epub Sep 08, 2026.

Abstract

Immunotherapy is increasingly reshaping the therapeutic landscape of clinical oncology by inducing effective tumor regression and durable responses in various types of human cancers. However, a substantial fraction of patients resist immunotherapy, or even experience relapse after an initial response. Clinically, the conventional biomarkers often exhibit limited predictive accuracy and poor transferability across tumor types. These limitations highlight the urgent need for more robust, data-driven strategies to decipher tumor-immune dynamics and guide personalized treatment. Recently, artificial intelligence (AI) has emerged as a transformative paradigm in immuno-oncology, offering scalable solutions to decode high-dimensional and heterogeneous biomedical data. The evolution of AI can be broadly divided into six key stages: early machine learning (ML) classifiers, unsupervised representation learning, graph-based modeling of cell-cell interactions, multimodal foundation and biomedical language models, reinforcement and active learning paradigms, and AI virtual cells. At the molecular level, AI facilitates the integration of genomics, transcriptomics, proteomics, and spatially resolved single-cell omics, supporting fine-grained reconstruction of the tumor immune microenvironment (TIME). At the clinical interface, the application of deep learning (DL) in liquid biopsy, digital pathology, and radiologic imaging has advanced response prediction, biomarker discovery, and early detection of immune-related adverse events (irAEs), driving personized immunotherapy. This review synthesizes recent progress in AI-guided precision immunotherapy across molecular, cellular, and phenotypic area. By outlining key methodological advances and translational opportunities, we highlight that multimodal AI frameworks can pave the way toward the more interpretable, robust, and individualized immunotherapy strategies for cancer treatment.

PMID:
42791141
Bibliographic data and abstract were imported from PubMed on 26 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.