The Role of Artificial Intelligence in Cancer Diagnostic Imaging Across Six Domains: An Extensive Umbrella Review (Preprint)

BACKGROUND Cancer remains one of the leading causes of death worldwide, responsible for nearly 10 million deaths annually. Despite major advances in therapy, mortality continues to rise due to delayed diagnosis and unequal access to high-quality screening. Artificial Intelligence (AI) has rapidly emerged as a transformative technology in medical imaging, with growing evidence of its capacity to enhance diagnostic accuracy, optimize workflow efficiency, and support clinical decision-making. However, the evidence base is dispersed across multiple cancer types and imaging modalities, necessitating an integrated synthesis to assess AI’s true clinical readiness and the implementation challenges it faces OBJECTIVE This umbrella review aimed to comprehensively synthesize evidence from published systematic reviews and meta-analyses on the role of AI in cancer diagnostic imaging, identifying demonstrated benefits, limitations, and future research priorities METHODS A comprehensive search of PubMed, Embase, Scopus, Web of Science, and Cochrane Library identified systematic reviews and meta-analyses published between January 2020 and September 2025. Eligible reviews examined AI-based imaging applications for cancer screening, detection, diagnosis, or prognostication and reported relevant clinical, technical, or efficiency outcomes. Data were extracted using a standardized template, methodological quality was appraised with the CASP checklist, and findings were synthesized narratively and thematically to identify overarching domains of evidence. Two reviewers independently assigned each review to one or more domains; disagreements were resolved through discussion, with a third reviewer consulted where needed RESULTS Fifty-three systematic reviews covering thirteen cancer types and multiple imaging modalities identified six key domains of AI application in cancer imaging: 1) Screening and Early Detection, radiologist-level accuracy in breast, lung, and gastrointestinal cancers; 2) Diagnostic Support, improved lesion detection, classification, and staging; 3) Risk Prediction and Stratification, enhanced risk estimation using radiomics and deep learning; 4) Workflow and Efficiency, automated triage and segmentation reduced reporting time by up to 50%; 5) Prognostic and Predictive Functions, accurate prediction of treatment response and survival; and 6) Implementation and Methods, limited validation, transparency, and bias oversight. CONCLUSIONS AI demonstrates strong potential to enhance the accuracy, efficiency, and prognostic insight of cancer imaging. However, widespread clinical translation depends on robust validation, transparent reporting, and ethical governance. Standardized evaluation frameworks and prospective multi-center studies are essential to move AI from experimental promise to reliable, equitable clinical practice.