S764 Yield of Dysplasia Surveillance Biopsies in Inflammatory Bowel Disease

Abstract

Introduction: Inflammatory bowel disease (IBD) dysplasia surveillance protocols call for a representative number of untargeted or 'random' biopsies to increase the odds of detection of 'invisible' or flat dysplasia. In current practice, the benefit of untargeted surveillance biopsies over a targeted surveillance approach with or without adjunctive chromoendoscopy is unclear. Our study retrospectively examines outcomes of endoscopic dysplasia surveillance at a single institution and compares yields of targeted and untargeted biopsies for the detection of dysplasia from both chromoendoscopy (CE) and traditional high definition white light (HD-WLE) examinations. Methods: A retrospective chart review of all IBD patients (ulcerative colitis, Crohns, and indeterminate colitis) undergoing dysplasia surveillance at a single center between 1/2015 and 10/2021 was performed. Patients younger than 18 and those with a pre-existing diagnosis of dysplasia or cancer were excluded. All surveillance exams done for each patient during the study period were reviewed. Presence or absence of endoscopic and microscopic inflammation, use of targeted biopsies or polypectomy, number of untargeted biopsies, and presence or absence of dysplasia from both targeted and untargeted biopsies were noted for both HD-WLE and CE exams. All instances of ‘invisible’ dysplasia were confirmed by a second reviewer following manual review of endoscopy and pathology reports. Results: 200 IBD patients (96 UC, 94 Crohns, 10 indeterminate) underwent 492 endoscopic surveillance procedures (199 CE, 293 WLE). 11,094 untargeted biopsies and 242 targeted biopsies or polypectomies were performed. Invisible dysplasia was detected 9 times for a yield of 0.06% per untargeted biopsy. 37.2% of endoscopically targeted biopsies or polypectomies were positive for dysplasia (39.6% CE, 33.9% WLE) and accounted for 90.9% of dysplasia detected during the study period. Conclusion: Random biopsies are a low yield means for detection of dysplasia, accounting for < 10% of dysplastic lesions observed during the study period. No patients with ‘invisible’ dysplasia had PSC, opted for colectomy, or had evidence of progression on follow up examination. Given the high yield of targeted biopsy and polypectomy, and low likelihood of missed high risk lesions if untargeted biopsies are not performed, it may be reasonable to emphasize high quality endoscopic examination and targeted biopsies/polypectomies and to de-emphasize the role of untargeted random biopsies.