Development and validation of a new algorithm model for differential diagnosis between Crohn's disease and intestinal tuberculosis: a combination of laboratory, imaging and endoscopic characteristics

Abstract Background Sometimes in clinical practice, it is a great challenge to distinguish Crohn's disease (CD) and intestinal tuberculosis (ITB), we conducted this study identify simple useful algorithm for distinguishing them. Methods We retrospectively reviewed the medical history of patients who were diagnosed as ITB or CD. firstly identified patients, then with CD matched by age, sex, admission time 1:1 ratio. Patients admitted between May 1, 2013 April 30, 2019 regarded training cohort, 2020 validation cohort. used multivariate analysis potential variables, R package rpart build classification regression tree (CART), validated newly developed model. Results In total, cohort included 84 22 patients. Multivariate showed that, positive interferon-gamma release assays (IGRAs), ≥ 4 segments involved, longitudinal ulcer, circular aphthous ulcer confirmed independent discriminating factors. Using these parameters CART model made an overall accuracy rate was 88.64%, sensitivity, specificity, NPV, PPV being 90.91%, 86.36%, 90.48% 86.96%, respectively. Conclusion novel covering laboratory, imaging, endoscopy differentiate good accuracy. Positive IGRAs suggestive ITB, while