P115 Low grade dysplasia prognosis in the 21st century – a large multi-centre retrospective cohort study

Abstract

Introduction Recent advances in ulcerative colitis (UC) endoscopic surveillance such as high-definition imaging and greater chromoendoscopy (CE) use have led to an increase in detection and resection of visible dysplasia. An updated study of prognosis of low grade dysplasia (LGD) is needed to address uncertainty as to the accuracy of progression rates based on historical studies. Methods This retrospective cohort study involved four UK IBD centres. Hospital pathology databases were searched between 1 January 2001 and 30 December 2018 to identify adult patients with UC who had their first LGD diagnosis diagnosed within the extent of colitis. Only patients followed up with at least one colonoscopy or colectomy by 30 August 2019 were included. The study end point was time to high grade dysplasia or cancer, i.e. advanced neoplasia (AN), or end of follow-up. Kaplan-Meier and Cox proportional hazards model survival analyses were performed. Results 460 patients met the inclusion criteria and were followed up for a median of 4.1 years (2,2201 patient-years). 77% of patients had CE surveillance. Complete endoscopic resection was achieved in 94% and 64% of the polypoid and non-polypoid LGD respectively. Incidence rate of AN per 100 patient-years was 1.0 (95% CI 0.6–1.7) after endoscopic resection of polypoid LGD, 2.5 (95% CI 1.3–4.4) after resection of non-polypoid LGD resection and 8.9 (95% CI 6.0–12.6) if the LGD was unresected. Figure 1 demonstrates the cumulative incidence of AN according to LGD visibility and resectability. On multivariate analysis, predictors of AN progression were visible LGD of 1 cm diameter or more [Hazard ratio (HR) 2.5; 95% CI 1.4–4.45; p=0.002], multifocality [HR 1.8; 95% CI 1.1–3.2; p=0.046] and incomplete endoscopic resection, including invisible dysplasia [HR 5.8; 95% CI 3.3–10.1; p Conclusions This is the largest study this century to examine prognosis of LGD based on endoscopic features. Incidence of AN is low if visible LGD is endoscopically resected but large size, multifocality and co-existent invisible LGD increase this risk and should be considered when decision-making.