PTH-005 Reliability of post-colonoscopy colorectal cancer algorithms. (June 2019)

• Record Type: Journal Article
• Authors: Beaton, David; Rutter, Matthew; Beintaris, Iosif
• Is Part Of: Gut. Volume 68(2019)Supplement 2
• Page Start: A15
• Page End: A16
• Publication Date: 2019-06
• Subjects: Gastroenterology -- Periodicals; 616.33
• DOI: 10.1136/gutjnl-2019-BSGAbstracts.30
• Languages: English
• ISSNs: 0017-5749

Abstract

Abstract : Introduction: Post-colonoscopy colorectal cancer (PCCRC) rate is an important metric of colonoscopy quality. The World Endoscopy Organisation (WEO) recently published a consensus statement advising review of PCCRC cases, to facilitate categorisation of PCCRCs into interval or non-interval cancers and determine their most plausible explanation. However, it is still unclear if the relevant algorithms are reproducible in real life practice, or how the respective plausible explanations are affected by factors such as reported adenoma location. This study aims to test the inter-rater reliability for both categorisation of PCCRC and most plausible explanation of cause; it will also test whether altering the location of previously seen adenomas to neighbouring segments of the colon impacts on most plausible explanation categorisation. Methods: C olorectal cancer (CRC) cases diagnosed from January 2015 to December 2016 in a single trust were identified and each case was cross-referenced with online endoscopy and pathology databases to analyse cases of PCCRC. PCCRC definition was as per WEO consensus statement. Two assessors independently reviewed each PCCRC case, and agreement was measured using Cohen's kappa (k). Results: Among 527 patients with CRC, 35 cases of PCCRC were identified; mean age was 71 years (range 48–87) and 21 (60%) of cases were males. Categorisation of most plausible explanations is seen in table 1 . Consistent most plausible explanation was found in 33/35 cases, k= .903 (95% CI, .839 to .967), p<0.0005. Sub-categorisation into interval and non-interval types was consistent in 21/35 cases, k= .417 (95% CI, .305 to .529), p<0.0005. Altering location of adenomas to adjacent segments did not affect assessment of most plausible explanation in any of the cases. Conclusions: The inter-rater agreement for most plausible explanation of PCCRC is almost perfect, indicating that the WEO algorithm is reproducible in a real-world setting. Inter-rater agreement for categorisation of PCCRC was moderate, most likely reflecting the fact that patient notes were not reviewed in higher detail in this study, making intended interval times not always available to assessors. Altering adenoma location did not affect results, but larger sample sizes are required to assess this further.