S1294 Comparing the Accuracy of Computed Tomography Enterography to Balloon-Assisted Enteroscopy in the Evaluation of Small Bowel Crohn’s Disease

Introduction: Evaluation of small bowel Crohn’s Disease (CD) beyond the reach of standard endoscopy often relies on cross-sectional imaging such as computed tomography enterography (CTE) and small bowel endoscopy. Balloon-assisted enteroscopy (BAE) is frequently utilized in this clinical setting as it allows for mucosal visualization, tissue acquisition, and therapeutic capabilities. The accuracy of CTE in evaluating features of small bowel CD in both naïve and post-surgical bowel remains unclear compared to BAE. Methods: Patients with an established diagnosis of small bowel CD who underwent a CTE and BAE within a 6-month period between 2011 and 2023 were reviewed. Findings of active inflammation, long-segment disease, skip-segments, strictures, and presence of high-grade strictures (HGS) were extracted from both CTE and BAE studies and analyzed using BAE as the gold standard. Standard of care and expert interpretations by a radiologist with expertise in small bowel imaging were compared, with expert interpretations used for final analyses. Results: A total of 76 patients with 99 corresponding unique CTE and BAE pairings were identified. CTE was most sensitive for active inflammation (84.2% (74.4 – 91.3) and strictures (92.1% (83.6 - 97.1)) and most specific for long segment inflammation (88.9% (78.4 – 95.4)) and HGS (91.2% (80.7 - 97.1)). CTE had low sensitivity for HGS (52.5%) and long segment inflammation (58.3%), and low specificity for strictures (65.2%). NPV was low overall for active inflammation (50.0%). In subgroup analyses, CTE demonstrated improved detection of active inflammation (sensitivity: 86.1%, specificity: 100%) in naïve bowel and, in post-surgical bowel, sensitivity for HGS improved modestly to 59.3% with a high specificity of 92.3%. Conclusion: CTE demonstrated reasonable accuracy for the use of positive identification of active inflammation and fibrostenotic lesions to guide clinical decision-making, but may be insufficient to rule out active inflammation and reliably detect HGS. The accuracy of CTE for detecting small bowel CD features differed between naïve and post-surgical bowel. Overall, CTE remains complementary to BAE in the assessment and management of small bowel CD (see Figure 1, Table 1).Figure 1.: A) Computed tomography enterography imaging of a patient with a small bowel stricture and signs of chronicity including upstream dilated bowel loops. B) Intra-procedure image of the same stricture identified in 1A taken during balloon-assisted enteroscopy. Table 1. - Computed Tomography Enterography (CTE) compared to Balloon Assisted Enteroscopy (BAE) for the assessment of small bowel Crohn’s Disease for A) All included studies and B) Following removal of confounders including BAE with dilation performed before CTE, treatment changes between CTE and BAE, and non-traversable stricture on BAE post-dilation. C) Patients with naïve small bowel and D) Patients with post-surgical bowel. 95% confidence intervals are reported in parentheses under each value Active Inflammation Long Segment Inflammation Skip-Segments Presence of Strictures High-Grade Strictures A) Sensitivity 84.2 (74.4 - 91.3) 58.3 (40.8 - 74.5) 82.6 (68.6 - 92.2) 92.1 (83.6 - 97.1) 52.4 (36.4 - 68.0) Specificity 76.5 (50.1 - 93.2) 88.9 (78.4 - 95.4) 75.5 (61.7 - 86.2) 65.2 (42.7 - 83.6) 91.2 (80.7 - 97.1) B) Sensitivity 80.0 (66.3 - 90.0) 50.0 (28.2 - 71.8) 86.7 (69.3 - 96.2) 93.0 (80.9 - 98.5) 60.9 (38.5 - 80.3) Specificity 83.3 (51.6 - 97.9) 95.0 (83.1 - 99.4) 81.3 (63.6 - 92.8) 68.4 (43.5 - 87.4) 87.2 (72.6 - 95.7) C) Sensitivity 86.1 (72.1 - 94.7) 57.1 (37.2 - 75.5) 85.7 (63.7 - 97.0) 97.0 (84.2 - 99.9) 40.0 (16.3 - 67.7) Specificity 100 (29.2 - 100) 83.3 (58.6 - 96.4) 72.0 (50.6 - 87.9) 61.5 (31.6 - 86.1) 90.3 (74.3 - 98.0) D) Sensitivity 82.1 (66.5 - 92.5) 62.5 (24.5 - 91.5) 80.0 (59.3 - 93.2) 88.4 (74.9 - 96.1) 59.3 (38.8 - 77.6) Specificity 71.4 (41.9 - 91.6) 91.1 (78.8 - 97.5) 78.6 (59.1 - 91.7) 70.0 (34.8 - 93.3) 92.3 (74.9 - 99.1)

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.