A new study published in Therapeutic Advances in Gastroenterology has highlighted the growing potential of artificial intelligence (AI) combined with intestinal ultrasound (IUS) in improving the early diagnosis of Crohn’s disease, particularly among patients presenting with chronic diarrhea. Researchers from The First People’s Hospital of Foshan reported that machine learning models integrating ultrasound findings with routine clinical data demonstrated remarkably high diagnostic accuracy, underscoring the expanding role of AI-assisted decision-making in gastroenterology.

Crohn’s disease is a chronic inflammatory bowel disorder characterized by recurrent inflammation of the gastrointestinal tract. Since symptoms such as chronic diarrhea are often nonspecific, diagnosis is frequently delayed, increasing the risk of complications including strictures, fistulas, abscesses, and progressive bowel damage. Investigators noted that timely diagnosis remains one of the greatest challenges in the management of inflammatory bowel disease.

The retrospective study evaluated 1,119 patients with chronic diarrhea between January 2024 and July 2025. Among them, 178 patients were diagnosed with Crohn’s disease, while 941 had other gastrointestinal disorders. Researchers analyzed clinical records, laboratory findings, and intestinal ultrasound features using five machine learning models, including XGBoost, random forest, logistic regression, multilayer perceptron, and support vector machine models.

Ultrasound parameters assessed in the study included bowel wall thickness, intestinal wall layering, vascular blood flow, mesenteric fat proliferation, and lymph node enlargement. These findings were combined with laboratory indicators such as C-reactive protein, white blood cell count, fecal occult blood, platelet count, and other inflammatory markers to develop multimodal prediction models.

Among the tested systems, the XGBoost model integrating ultrasound and clinical data demonstrated the best overall performance, achieving a diagnostic accuracy of 94%, sensitivity of 89%, and specificity of 96%. Researchers stated that the model showed strong discriminatory ability in distinguishing Crohn’s disease from other gastrointestinal disorders. Ultrasound-only models also demonstrated strong performance even without laboratory data. Logistic regression models based solely on intestinal ultrasound features achieved an area under the curve (AUC) close to 0.90, with sensitivity exceeding 85%. The findings suggest that intestinal ultrasound may be particularly valuable in settings where laboratory results are delayed or rapid bedside assessment is required.

The study further emphasized the practical advantages of intestinal ultrasound, which is noninvasive, inexpensive, repeatable, and widely accessible. Researchers suggested that combining AI with ultrasound could facilitate faster screening, improve diagnostic standardization, and support clinical decision-making in both tertiary care centers and community healthcare settings.

Investigators believe that the proposed AI-assisted diagnostic framework may help clinicians identify Crohn’s disease earlier while reducing unnecessary referrals and missed diagnoses. They also noted that the system could serve as the foundation for future automated clinical decision-support platforms capable of assisting gastroenterologists in real-time patient assessment.

Supporting these findings, a previous study by Ceccato et al. explored the role of AI in precision medicine for inflammatory bowel disease and reported promising applications in disease prognosis and prediction of treatment response. However, the authors emphasized that wider clinical adoption would require further prospective validation studies.

Overall, the findings represent an important step toward precision medicine in inflammatory bowel disease. As AI technologies continue to evolve, the integration of machine learning with noninvasive imaging techniques may significantly transform the diagnosis and monitoring of chronic gastrointestinal disorders. The study reinforces the growing potential of AI-driven healthcare tools to improve diagnostic accuracy, streamline clinical workflows, and enable earlier intervention for patients living with Crohn’s disease.

 

References
1. Huang R, Zhong M, Hong C, Liu Z, Chen K, Li X, Lu Q, Liu Q, Li Z, Jiang C, Jian G. Revolutionizing Crohn’s disease detection: integrating AI with intestinal ultrasound for superior diagnosis. Therapeutic Advances in Gastroenterology. 2026 May;19:17562848261446568.

2. Ceccato HD, e Silva TA, Genaro LM, Silva JF, de Souza WM, Oliveira PD, de Azevedo AT, Ayrizono MD, Leal RF. Artificial intelligence use for precision medicine in inflammatory bowel disease: a systematic review. American Journal of Translational Research. 2025 Jan 15;17(1):28.