Inter-reader agreement of the Society of Abdominal Radiology-American Gastroenterological Association (SAR-AGA) consensus reporting for key phenotypes at MR enterography in adults with Crohn disease: impact of radiologist experience.
Citation: Abdominal Radiology. 46(11):5095-5104, 2021 11.PMID: 34324038Institution: MedStar Washington Hospital CenterDepartment: RadiologyForm of publication: Journal ArticleMedline article type(s): Journal ArticleSubject headings: *Crohn Disease | *Radiology | Adult | Consensus | Crohn Disease/dg [Diagnostic Imaging] | Humans | Magnetic Resonance Imaging | Observer Variation | Phenotype | Radiologists | Reproducibility of Results | Retrospective Studies | United StatesYear: 2021Abstract: CONCLUSION: There was strong to almost-perfect inter-reader agreement for key CD phenotypes described in the SAR-AGA consensus recommendations including active ileum and proximal small bowel inflammation, stricture, penetrating disease, abscess, and perianal abscess. Areas of lower inter-reader agreement could be targeted for future education efforts to further standardize CD MRE reporting. Dedicated perianal sequences should be included on follow-up MRE. Copyright (c) 2021. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.METHODS: Two experienced and two less-experienced radiologists retrospectively evaluated 99 MRE in CD patients (50 initial MRE, 49 follow-up MRE) performed from 1/1/2019 to 3/20/2020 for the presence of active bowel inflammation (stomach, proximal small bowel, ileum, colon), stricture, probable stricture, penetrating disease, and perianal disease. The MRE protocol did not include dedicated perianal sequences. Inter-rater agreement was determined for each imaging feature using prevalence-adjusted bias-adjusted kappa and compared by experience level.PURPOSE: To assess inter-reader agreement of key features from the SAR-AGA recommendations for the interpretation and reporting of MRE in adult patients with CD, focusing on the impact of radiologist experience on inter-reader agreement of CD phenotypes.RESULTS: All readers had almost-perfect inter-reader agreement (kappa > 0.90) for penetrating disease, abscess, and perianal abscess in all 99 CD patients. All readers had strong inter-reader agreement (kappa: 0.80-0.90) in 99 CD patients for active ileum inflammation, proximal small bowel inflammation, and stricture. Less-experienced readers had significantly lower inter-reader agreement for active ileum inflammation on initial than follow-up MRE (kappa 0.68 versus 0.96, p = 0.018) and for strictures on follow-up than initial MRE (kappa 0.76 versus 1.0, p = 0.027). Experienced readers had significantly lower agreement for perianal fistula on follow-up than initial MRE (kappa: 0.55 versus 0.92, p = 0.008).Originally published: Abdominal Radiology. 2021 Jul 29Abdominal Radiology. 46(11):5095-5104, 2021 11.Fiscal year: FY2022Fiscal year of original publication: FY2022Digital Object Identifier: Date added to catalog: 2021-11-01| Item type | Current library | Collection | Call number | Status | Date due | Barcode |
|---|---|---|---|---|---|---|
| Journal Article | MedStar Authors Catalog | Article | 34324038 | Available | 34324038 |
CONCLUSION: There was strong to almost-perfect inter-reader agreement for key CD phenotypes described in the SAR-AGA consensus recommendations including active ileum and proximal small bowel inflammation, stricture, penetrating disease, abscess, and perianal abscess. Areas of lower inter-reader agreement could be targeted for future education efforts to further standardize CD MRE reporting. Dedicated perianal sequences should be included on follow-up MRE. Copyright (c) 2021. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
METHODS: Two experienced and two less-experienced radiologists retrospectively evaluated 99 MRE in CD patients (50 initial MRE, 49 follow-up MRE) performed from 1/1/2019 to 3/20/2020 for the presence of active bowel inflammation (stomach, proximal small bowel, ileum, colon), stricture, probable stricture, penetrating disease, and perianal disease. The MRE protocol did not include dedicated perianal sequences. Inter-rater agreement was determined for each imaging feature using prevalence-adjusted bias-adjusted kappa and compared by experience level.
PURPOSE: To assess inter-reader agreement of key features from the SAR-AGA recommendations for the interpretation and reporting of MRE in adult patients with CD, focusing on the impact of radiologist experience on inter-reader agreement of CD phenotypes.
RESULTS: All readers had almost-perfect inter-reader agreement (kappa > 0.90) for penetrating disease, abscess, and perianal abscess in all 99 CD patients. All readers had strong inter-reader agreement (kappa: 0.80-0.90) in 99 CD patients for active ileum inflammation, proximal small bowel inflammation, and stricture. Less-experienced readers had significantly lower inter-reader agreement for active ileum inflammation on initial than follow-up MRE (kappa 0.68 versus 0.96, p = 0.018) and for strictures on follow-up than initial MRE (kappa 0.76 versus 1.0, p = 0.027). Experienced readers had significantly lower agreement for perianal fistula on follow-up than initial MRE (kappa: 0.55 versus 0.92, p = 0.008).
English