Reliability of MR Enterography Features for Describing Fibrostenosing Crohn Disease
Florian Rieder,Mark E Baker,David H Bruining,Jeff L Fidler,Eric C Ehman,Shannon P Sheedy,Jay P Heiken,Justin M Ream,David R Holmes 3rd,Akitoshi Inoue,Payam Mohammadinejad,Yong S Lee,Stuart A Taylor,Jaap Stoker,Guangyong Zou,Zhongya Wang,Julie Rémillard,Rickey E Carter,Ronald Ottichilo,Norma Atkinson,Mohamed Tausif Siddiqui,Venkata C Sunkesula,Christopher Ma,Claire E Parker,Julian Panés,Jordi Rimola,Vipul Jairath,Brian G Feagan,Joel G Fletcher,Stenosis Therapy and Anti-Fibrosis Research (STAR) Consortium
DOI: https://doi.org/10.1148/radiol.233039
IF: 19.7
Radiology
Abstract:Background Clinical decision making and drug development for fibrostenosing Crohn disease is constrained by a lack of imaging definitions, scoring conventions, and validated end points. Purpose To assess the reliability of MR enterography features to describe Crohn disease strictures and determine correlation with stricture severity. Materials and Methods A retrospective study of patients with symptomatic terminal ileal Crohn disease strictures who underwent MR enterography at tertiary care centers (Cleveland Clinic: September 2013 to November 2020; Mayo Clinic: February 2008 to March 2019) was conducted by using convenience sampling. In the development phase, blinded and trained radiologists independently evaluated 26 MR enterography features from baseline and follow-up examinations performed more than 6 months apart, with no bowel resection performed between examinations. Follow-up examinations closest to 12 months after baseline were selected. Reliability was assessed using the intraclass correlation coefficient (ICC). In the validation phase, after five features were redefined, reliability was re-estimated in an independent convenience sample using baseline examinations. Multivariable linear regression analysis identified features with at least moderate interrater reliability (ICC ≥0.41) that were independently associated with stricture severity. Results Ninety-nine (mean age, 40 years ± 14 [SD]; 50 male) patients were included in the development group and 51 (mean age, 45 years ± 16 [SD]; 35 female) patients were included in the validation group. In the development group, nine features had at least moderate interrater reliability. One additional feature demonstrated moderate reliability in the validation group. Stricture length (ICC = 0.85 [95% CI: 0.75, 0.91] and 0.91 [95% CI: 0.75, 0.96] in development and validation phase, respectively) and maximal associated small bowel dilation (ICC = 0.74 [95% CI: 0.63, 0.80] and 0.73 [95% CI: 0.58, 0.87] in development and validation group, respectively) had the highest interrater reliability. Stricture length, maximal stricture wall thickness, and maximal associated small bowel dilation were independently (regression coefficients, 0.09-3.97; P < .001) associated with stricture severity. Conclusion MR enterography definitions and scoring conventions for reliably assessing features of Crohn disease strictures were developed and validated, and feature correlation with stricture severity was determined. © RSNA, 2024 Supplemental material is available for this article. See also the article by Rieder and Ma et al in this issue. See also the editorial by Galgano and Summerlin in this issue.