P277 Interobserver agreement of current and new proposed endoscopic scores for postoperative recurrence in Crohn’s disease

Abstract

Abstract Background The modified Rutgeerts’ score (mRS) is widely used for the assessment of endoscopic postoperative recurrence (ePOR) in Crohn’s disease (CD) after ileocolic resection to guide therapeutic decisions. To improve the validity and prognostic value of this endoscopic assessment, two new scores have been proposed. This study assessed the interobserver agreement of the current (mRS) and new endoscopic scores for ePOR in CD. Methods Sixteen academic and non-academic IBD specialists assessed endoscopic videos (n=71) of postoperative CD patients (n=66) retrieved from nine Dutch centers. Each video was assessed for the degree of inflammation by four gastroenterologists using the mRS, REMIND score (separate score of anastomosis and neoterminal ileum) and anatomic score (separate score of lesions at the anastomotic line, ileal body, ileal inlet, neoterminal ileum, colonic and/or ileal blind loop). Interobserver agreement was assessed using Fleiss’ weighted kappa. Results Fleiss’ weighted kappa for the mRS was 0.67 (95% confidence interval [CI] 0.59–0.74). The weighted kappa for the REMIND score was 0.73 (95% CI 0.65–0.80) for lesions in the neoterminal ileum and 0.46 (95% CI 0.35–0.58) for anastomotic lesions. In the anatomic score, the weighted kappa for lesions in the ileal body, ileal inlet, neoterminal ileum, colonic and ileal blind loop was 0.61 (95% CI 0.49–0.73), 0.63 (95% CI 0.54–0.72), 0.61 (95% CI 0.49–0.74), 0.83 (95% CI 0.62–1.00) and 0.68 (95% CI 0.46–0.89), and 0.44 (95% CI 0.32–0.55) for lesions at the anastomotic line. Conclusion The interobserver agreement of the mRS is substantial. Similarly, the interobserver agreement is substantial for lesions in the neoterminal ileum according to both the REMIND and anatomic score, whereas only moderate for anastomotic lesions. Since therapeutic decisions in clinical practice are based on these assessments and these scores are used as outcome measure in clinical studies, further improvement of agreement is essential.