Artificial intelligence in inflammatory bowel disease: From current evidence, clinical translation, and the road to precision medicine.

Abstract

Artificial intelligence (AI) is rapidly transforming healthcare, supporting disease management and enabling outcome prediction across multiple clinical settings. Inflammatory bowel diseases (IBD) are complex, heterogeneous conditions whose assessment relies on integrating several modalities, including endoscopy, histology, cross-sectional imaging, and omics data, all of which are critical for evaluating disease activity and predicting long-term outcomes. In recent years, numerous AI-based systems have been developed within each of these domains. In IBD endoscopy, deep learning algorithms have demonstrated high accuracy in objectively assessing mucosal inflammation and detecting colitis-associated dysplasia. In histology, AI applications enable automated, standardized evaluation of disease activity, reducing interobserver variability. Similarly, in cross-sectional imaging, AI models have shown promise in characterizing disease severity, identifying complications, and supporting outcome prediction. Beyond individual modalities, machine learning approaches are increasingly being explored to integrate complex clinical, imaging, and multi-omics data to predict disease trajectories and enable precision medicine strategies in IBD. The present review provides an overview of current AI applications across endoscopy, histology, imaging, and omics in IBD, highlighting its potential clinical impact and ability to advance precision medicine strategies through multimodal and multi-omics integration. Moreover, it discusses the main challenges, unmet needs, and limitations that remain barriers to adoption in clinical trials and routine clinical practice.