Predictive imaging in abdominal oncology: current trends and future directions
Maxime Barat, Stylianos Tzedakis, Anna Pellat, Ugo Marchese, Anthony Dohan, Romain Coriat, Elliot K Fishman, Linda Chu, Philippe Soyer
Jpn J Radiol. 2025 Oct 30. doi: 10.1007/s11604-025-01900-8. Online ahead of print.
Abstract
While the concept of predictive imaging is not entirely new, advanced analytic tools such as radiomics and machine learning have laid the foundation for a new generation of predictive imaging, which goes far beyond what is visible to the human eye. Predictive imaging has emerged as a transformative tool in abdominal oncology, offering the potential to personalize cancer detection and diagnosis, staging, treatment planning, monitoring and prognostication. A growing trend in predictive imaging is the creation of integrated models that combine multimodal imaging data, clinical parameters, genomic and molecular biomarkers. These integrated models can potentially offer superior prognostic capabilities and better risk stratification than traditional models in patients with abdominal cancers. Predictive imaging powered by radiomics and delta radiomics, artificial intelligence, and multimodal data integration is on the way for reshaping abdominal oncology. When current challenges are overcome, it is presumable that predictive imaging will offer powerful, noninvasive means to guide individualized care for patients with abdominal cancers, translating imaging data into actionable clinical insights. The purpose of this article was to provide an overview of the capabilities of predictive imaging in hepatic, pancreatic, colorectal, and gastric cancers.