SPATIAL CONCORDANCE OF LEFT AND RIGHT COLON FINDINGS IN PEDIATRIC PATIENTS WITH ULCERATIVE COLITIS AFTER TREAMTENT

Abstract

Abstract OBJECTIVES Achieving endoscopic remission has become the preferred clinical endpoint in the care of inflammatory bowel disease (IBD) patients. This requires repeat colonoscopy in patients with ulcerative colitis (UC) after initiation of therapy. Recent adult data suggests findings in the left colon on flexible sigmoidoscopy can be highly predictive of the findings if a full colonoscopy was performed; however, pediatric data is lacking for this association. The primary aim our study to assess the accuracy of findings on the left colon with the right colon for repeat endoscopy in pediatric patients with UC. METHODS This was a retrospective chart review of pediatric patients with UC seen at a single, academic center between 2010-2021. Patient demographics, clinical disease activity, laboratory values and endoscopic findings were collected. The Mayo Endoscopy Score (MES) was used to assess the findings in the right and left colon by a single reader (JM). Histologic findings were assessed using the Geobes index by a pathology resident (AKA) and over-read by an attending pathologist (SS). Weighted kappa correlation was used to assess concordance between left and right colon endoscopic and histologic findings. Receiver operator curve (ROC) analysis was used to assess the accuracy of left-sided findings predicting right-sided findings. RESULTS We analyzed repeat endoscopies from 57 patients, of which 31 (54.4%) were male, mean age at UC diagnosis was 13.5 ± 4.1 years, 53 (93%) were Caucasian. Most common medication was oral mesalamine (67.2%) and most common distribution was pancolitis (53.4%). Mean time between index and repeat colonoscopy was 11.3 months. On repeat colonoscopy, endoscopic findings on the left colon moderately correlated with findings on the right colon [(κ = 0.53 (0.30-0.75)], similarly histologic findings for the left and right colon had a moderate correlation [(κ = 0.41 (0.17-0.64)]. Area under the ROC (AUC) was 0.89 for endoscopic remission anywhere in the colon and 0.98 for complete endoscopic healing anywhere in the colon. Longer interval between the index and follow-up colonoscopy did not affect odds of concordance. CONCLUSIONS We found that endoscopic and histologic severity correlated between the right and left colon for repeat assessment after initiation of biologic therapy. Left sided findings were highly accurate for endoscopic remission and complete endoscopic healing anywhere in the colon. This data suggest a flexible sigmoidoscopy could be feasible in lieu of full colonoscopy when assessing for endoscopic remission in pediatric patients with UC, to reduce the burden to the family and healthcare system. Further, prospective studies with a central reading component for histology and endoscopy are needed to validate these findings.