PTU-038 Predictors of adenoma detection at colonoscopy after bowel scope surveillance

    loading  Checking for direct PDF access through Ovid

Abstract

Introduction

Flexible sigmoidoscopy (bowel scope) reduced colorectal cancer incidence and mortality in a population aged 55–64.[1]Patients progressed to colonoscopy based on ‘high risk’ features (Table 1).[1] Based on these pivotal findings, the UK bowel scope (BS) surveillance programme was introduced in 2013 to individuals aged 55. The Wolverhampton bowel cancer screening centre was the first UK site to fully roll out the programme. The correlation between BS findings and subsequent colonoscopy has not previously been evaluated in this specific cohort.

Method

We prospectively collated data from all BS patients at our centre and identified those undergoing colonoscopy between August 2013–2016. We assessed conversion rates, compliance with BS protocol and correlated endoscopic and histological findings to identify predictors for detection of pathology at colonoscopy. Univariate analysis was performed using Pearson’s chi2.

Results

11 711 bowel scopes were performed, with an adenoma detection rate (ADR) of 8.5%, and conversion to colonoscopy in 421 patients (3.6%). 386 were included for analysis after excluding incomplete colonoscopy/histology. All patients were aged 55 (64.8% male). Additional ADR at colonoscopy was 35.2%, with malignant diagnoses in 1.5% (all detected at BS). The adenoma miss rate at BS was 5.2%. On univariate analysis (Table 1), polyp ≥10 mm was the only indication associated with increased ADR at colonoscopy (OR 2.13, p<0.001). Additional predictors identified included villous (not tubulovillous) histology (OR 4.41, p=0.02), and male gender (OR 2.35, p<0.001). These factors also significantly predicted new ≥10 mm adenoma. 57 (14.8%) underwent colonoscopy outside protocol, which reduced ADR (OR 0.29, p=0.03). After excluding high risk indications, changing the conversion criteria from any villous to villous only histology altered sensitivity from 27.2% to 83.3%, and specificity from 84.5% to 80.5%.

Conclusion

At BS, male gender, ≥10 mm polyps, and villous histology are predictors of proximal colonic pathology. Further analyses are required to clarify the benefits of converting low-risk tubulovillous adenomas at BS to colonoscopy.

Disclosure of Interest

None Declared

Related Topics

    loading  Loading Related Articles