Computerized tomography-based predictive model for differentiation of Crohn's disease from intestinal tuberculosis

Area of Science:

• Gastroenterology
• Radiology
• Infectious Diseases

Background:

• Intestinal tuberculosis (ITB) and Crohn's disease (CD) share similar clinical and radiological features.
• Distinguishing between ITB and CD using computed tomography (CT) requires further investigation.

Purpose of the Study:

• To compare CT features of ITB and CD.
• To develop a predictive model for differentiating ITB and CD.

Main Methods:

• Retrospective analysis of CT scans from ITB and CD patients.
• Radiologist blinded to diagnosis identified specific CT findings.
• A predictive model was developed based on significant differentiating features.

Main Results:

• Univariate analysis revealed significant differences in colonic involvement, skip lesions, and lymph node size.
• Multivariate analysis identified ileocecal involvement, long-segment involvement, and lymph node size (≥1 cm) as key differentiators.
• A risk score model demonstrated good specificity and positive predictive values for both ITB and CD.

Conclusions:

• CT features, including long-segment involvement, ileocecal involvement, and lymph nodes ≥1 cm, can effectively differentiate ITB and CD.
• The developed predictive model offers a valuable tool for distinguishing these conditions.