The potential role of CT enterography and gastrointestinal ultrasound in the evaluation of anti-tubercular therapy response of intestinal tuberculosis: a retrospective study

BACKGROUND: Accurate evaluation of anti-tubercular therapy (ATT) responses is crucial for both diagnosis and treatment of intestinal tuberculosis (ITB). Little is known about the role of cross-sectional imaging techniques in ITB follow-up assessment. We aimed to investigate the accuracy of cross-sectional imaging modalities, CT enterography (CTE) and gastrointestinal ultrasound (GIUS), in the evaluation of ATT responses in ITB patients. METHODS: Patients diagnosed with ITB and followed up by CTE and/or GIUS were retrospectively searched in the databases. Clinical, imaging, laboratory and endoscopic data were collected at baseline and the first follow-up visit. Responses were graded as good, partial and no response based on protocols described in the literature and by our institution. CTE evaluation was based on changes in the lesion area, mural thickness, enhancement patterns and lymph nodes, while GIUS evaluation was based on changes in bowel wall morphology and the Limberg score. Clinical evaluation was used as the gold-standard evaluation method, which was determined by a comprehensive impression of endoscopic changes along with symptomatic improvement and laboratory tests, with imaging results masked. RESULTS: Twenty patients with ITB were enrolled in our study. The first follow-up time was from 2 to 12 months (average 6 months). According to the gold standard evaluation, 11 patients were evaluated as having a good ATT response, while 9 had a partial response. A total of 18 patients were followed up by CTE, while 7 were followed up by GIUS, depending on medical and/or financial considerations. The accuracy of CTE and GIUS was 83% (15/18) and 85.7% (6/7), respectively. The sensitivity, specificity, PPV and NPV of CTE were 88.9, 77.8, 80 and 87.5%, respectively. Moreover, the sensitivity, specificity, PPV and NPV of GIUS were 100, 50, 83.3 and 100%, respectively. By combining the results of CTE and GIUS results, the overall accuracy was 90%, with sensitivity and specificity of 91.7 and 87.5%, respectively. CONCLUSION: To our knowledge, this is the first study exploring the accuracy of the cross-sectional imaging modalities CTE/GIUS in the evaluation of ATT responses. Our results indicated their promising application prospect in clinical practice as a non-invasive and cost-effective approach. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12876-019-1030-0) contains supplementary material, which is available to authorized users.