Integrating multimodal data to advance predictive biomarkers for cancer immunotherapy outcomes

The field of cancer immunotherapy has reshaped modern oncology, with immune checkpoint inhibitors and cellular therapies producing unprecedented and durable responses in some patients. Yet, response heterogeneity and immune-related toxicities remain major challenges, underscoring the need for reliable biomarkers to guide patient selection, therapeutic monitoring, and risk management. Conventional single-modality biomarkers—whether molecular, histological, or imaging-based—often capture only partial aspects of the tumor–immune–host interaction, limiting their generalizability across populations and clinical settings. In parallel, the growing availability of multimodal data—including radiological imaging, digital pathology, tissue and liquid biopsy omics, and longitudinal clinical records—offers new opportunities for integrative modeling. Advances in machine learning, representation learning, and cross-domain harmonization now make such data fusion increasingly feasible, but systematic validation, interpretability, and clinical translation remain lagging behind.This Research Topic aims to accelerate the development of multimodal data fusion strategies that yield robust, generalizable, and clinically actionable biomarkers for cancer immunotherapy. We invite contributions addressing the fundamental question of how heterogeneous data sources can be integrated to better predict therapeutic benefit, toxicity risk, and dynamic immune responses. Particular interest lies in methodological innovations that handle missing modalities, temporal heterogeneity, and cross-institutional variability, as well as studies emphasizing rigorous evaluation and clinical relevance. Work that explicitly enhances transparency, reproducibility, and real-world applicability—through open data pipelines, standardized reporting, or prospective validation—is especially encouraged.This Research Topic focuses on the integration of multimodal data toward clinically useful biomarker discovery in immunotherapy. Contributions should prioritize fusion methodology and its validation across modalities and clinical contexts. To gather further insights into multimodal biomarker research in cancer immunotherapy, we welcome articles addressing, but not limited to, the following themes:• Fusion architectures, representation learning, and multimodal foundation models• Radiopathomics and imaging–omics integration; modeling of longitudinal and multi-timepoint biomarkers• Robust modeling under missing modalities, weak supervision, domain shift, or privacy constraints• Interpretability, uncertainty estimation, calibration, and clinical decision support• External and prospective validation, reproducibility, and translation into clinical workflowsPlease note: Manuscripts consisting solely of bioinformatics or computational analysis of public genomic or transcriptomic databases which are not accompanied by validation (independent cohort or biological validation in vitro or in vivo) are out of scope for this section and will not be accepted as part of this Research Topic. Topic editors: Zhen Zhang, Lujun Zhao Source: https://www.frontiersin.org/research-topics/81855/integrating-multimodal-data-to-advance-predictive-biomarkers-for-cancer-immunotherapy-outcomes