Multimodal and Trustworthy Explainable AI for Next-Generation Diagnosis and Decision-Making in Gastrointestinal Cancers

Gastrointestinal (GI) cancers — including gastric, colorectal, esophageal, pancreatic, and hepatobiliary malignancies — remain among the leading causes of cancer-related morbidity and mortality worldwide. Late-stage diagnosis, tumor heterogeneity, and variable treatment response continue to drive this burden. Artificial intelligence (AI) has shown substantial promise in improving early detection, risk stratification, and treatment planning by integrating diverse data sources, including endoscopic and radiologic imaging, histopathology, genomic and transcriptomic profiles, and electronic health records (EHRs).Two barriers still limit the clinical translation of AI in GI oncology. The first is the reliance on fragmented, single-modality models that fail to capture the full complexity of tumor biology and patient-level heterogeneity. The second is the limited transparency and trust in AI decision-making, which constrains clinician confidence, regulatory approval, and patient acceptance — particularly for “black-box” deep learning models.This Research Topic brings together research on multimodal data fusion and on explainable, trustworthy AI frameworks, with the aim of advancing next-generation diagnostic and decision-support systems for GI cancers and closing the gap between algorithmic performance and real-world clinical adoption. It seeks to advance clinically interpretable, multimodal AI systems for GI cancer diagnosis and decision-making, to foster interdisciplinary collaboration between AI researchers, oncologists, radiologists, and pathologists, to highlight strategies that improve clinician and patient trust in AI-driven GI cancer care, and to identify current gaps and future directions for translating explainable multimodal AI into clinical practice.We welcome Original Research, Review, and Methods articles on themes including, but not limited to:1. Multimodal AI integrating imaging (endoscopy, CT/MRI, PET), histopathology, genomics and multi-omics, and clinical/EHR data for GI cancer diagnosis and prognosis2. Explainable AI (XAI) methods — for example attention maps, SHAP, and counterfactual explanations — for interpretable GI cancer risk prediction and decision support3. Trustworthy AI frameworks addressing robustness, fairness, uncertainty quantification, and bias mitigation in GI oncology4. AI-assisted early detection and risk stratification of gastric, colorectal, esophageal, pancreatic, and hepatobiliary cancers5. Clinical validation, prospective studies, and real-world deployment of multimodal or explainable AI tools in GI cancer care6. Federated learning and privacy-preserving approaches for multi-institutional GI cancer data integration7. Human–AI collaboration and clinician trust in AI-assisted GI cancer diagnosis and treatment decisions8. Benchmarking, standardization, and regulatory considerations for explainable AI models in GI oncologyPlease note: Manuscripts consisting solely of bioinformatics, computational analysis, or predictions of public databases which are not accompanied by validation (independent clinical or patient cohort, or biological validation in vitro or in vivo, which are not based on public databases) are not suitable for publication in this journal. Topic editors: Dr. Hari Mohan Rai, Anwar Ghani (SMIEEE), Syed Atif Moqurrab, Abdul Razaque Source: https://www.frontiersin.org/research-topics/83199/multimodal-and-trustworthy-explainable-ai-for-next-generation-diagnosis-and-decision-making-in-gastrointestinal-cancers