Abstract
BACKGROUND: Cross-sectional imaging is increasingly used for both initial diagnosis and long-term monitoring of Crohn's disease. The quantitative morphology of the terminal ileum may predict treatment response. OBJECTIVE: We aimed to identify baseline qualitative and quantitative imaging features that are associated with clinical and radiologic treatment response in a large cohort of children with Crohn's disease. MATERIALS AND METHODS: This was a retrospective study of the RISK cohort study in pediatric Crohn's disease. This multicenter study included 1,136 children <18 years from 28 sites in North America. Subjects enrolled with newly diagnosed Crohn's disease who underwent endoscopy with baseline and follow-up CT or MRI were considered for this study. Exclusion criteria were incomplete data or surgical resection prior to follow-up imaging. Imaging analysis included assessing a qualitative terminal ileum (TI) categorical score based on SAR-AGA consensus definitions ((1) normal, (2) inflammation only without luminal narrowing, (3) inflammation with luminal narrowing, or (4) stricture with pre-stenotic dilation ≥3 cm) and quantitative measurements (maximum bowel wall thickness and maximum/minimum lumen diameter). Two endpoints were considered: (1) clinical response (off corticosteroids and quiescent Physician Global Assessment at follow-up imaging) and (2) CT and MRI response (follow-up imaging normalization). Multivariable logistic regression analyses were developed for each endpoint. RESULTS: Ninety-six subjects were included. Clinical response endpoint was achieved in 38% (n=36) of participants, and imaging normalization was achieved in only 20% (n=19) of participants. Follow-up imaging showed disease progression in 24 (25%) patients, 7 (7%) of whom were radiologically normal at baseline (7%). A higher baseline TI categorical score was associated with lower odds of imaging normalization during follow-up (OR 0.4 [0.2, 0.8], P=0.009). Larger TI minimum lumen diameter (OR 1.1 [1.01, 1.3], P=0.04) and smaller maximum bowel wall thickness at baseline (OR 0.8 [0.6, 0.97], P=0.03) were associated with imaging normalization. There were no baseline imaging measurements associated with clinical response. CONCLUSIONS: Baseline increased terminal ileal minimum lumen diameter and decreasing wall thickness were associated with imaging normalization at follow-up, but not clinical response.