Publication:

Artificial intelligence in jejunal diverticulosis: emerging applications in imaging recognition, risk stratification and surgical decision support

Jejunal diverticulosis is a rare small-bowel disorder that is frequently under-recognized and may present with severe complications including diverticulitis, perforation, haemorrhage, and intestinal obstruction. Diagnosis relies primarily on cross-sectional imaging, particularly computed tomography (CT) and CT enterography, yet diverticula are often overlooked due to their subtle appearance and the rarity of the disease. In parallel, artificial intelligence (AI) has rapidly advanced across gastrointestinal imaging and clinical decision support, offering new opportunities for automated lesion detection, imaging phenotyping, and risk prediction. Aim of the study was to review current and emerging applications of artificial intelligence relevant to jejunal diverticulosis, with emphasis on imaging recognition, complication risk stratification, and surgical decision support. A structured narrative review was conducted using a literature search informed by the PRISMA framework. Electronic databases including PubMed/MEDLINE, Scopus, and Google Scholar were searched for studies published between January 2000 and February 2026. Search terms included combinations of “jejunal diverticulosis,” “small bowel diverticula,” “artificial intelligence,” “machine learning,” “deep learning,” “capsule endoscopy,” “CT enterography,” and “diverticulitis.” Studies describing AI applications in capsule endoscopy, cross-sectional small-bowel imaging, diverticular disease, and surgical decision support were included. Findings were synthesized narratively due to heterogeneity in study design and methodologies. AI has demonstrated high diagnostic accuracy in gastrointestinal imaging, particularly in capsule endoscopy where convolutional neural networks can detect small-bowel lesions, including diverticula, with very high sensitivity and specificity. In cross-sectional imaging, machine-learning and radiomics models applied to CT enterography can identify and quantify small-bowel pathology and predict clinically relevant outcomes. AI applications in colonic diverticulitis and acute abdominal CT have shown promising performance in differentiating disease entities, predicting complications, and guiding clinical triage. These advances provide a methodological foundation for applying AI to jejunal diverticulosis. Potential applications include automated detection of jejunal diverticula on CT or CT enterography, multimodal risk stratification models integrating clinical and imaging data, and AI-assisted decision support for conservative versus surgical management. Although AI models specifically targeting jejunal diverticulosis have not yet been reported, advances in gastrointestinal imaging analytics and clinical prediction models suggest substantial potential for future development. Multi-institutional data sharing, standardized imaging annotation, and prospective validation studies will be essential to develop reliable AI tools capable of improving early detection, risk assessment, and surgical decision-making in this rare but clinically significant condition.