P11 Utilisation and reproducibility Of WEO PCCRC algorithms in a real-world setting

Introduction Colorectal cancer (CRC) diagnosed following a colonoscopy in which no CRC is found is termed Post-Colonoscopy CRC (PCCRC). The World Endoscopy Organisation (WEO) consensus statements recommend review of individual PCCRC cases, including categorisation of cases into interval/non-interval CRCs, and root cause analysis (RCA) to determine most plausible explanation. Our study aim was to test the usability, reproducibility and outcomes of the WEO categorisation. Methods All CRC cases diagnosed from January 2015 to December 2016 in a single NHS trust were identified. Each was cross-referenced with local endoscopy and pathology databases. Cases where non-diagnostic colonoscopy was performed prior to CRC diagnosis were included. All colonoscopies going back to 2007 (when endoscopy reporting system introduced) were reviewed. Each CRC was entered into a spreadsheet, with headings based on WEO RCA checklist for PCCRCs. We performed 2 separate assessments: (1) RCA to identify WEO most plausible explanation for PCCRC; and (2) WEO PCCRC subtype categorisation, which looks at screening/surveillance intervals (table 1). Inter-observer agreement was measured using Cohen’s kappa (k). Cases with inter-rater variation were analysed further using patient notes and then discussed by a panel to determine causes of variation and attempt to reach consensus. Results Among 527 patients with CRC, 48 PCCRCs were identified. In 32 cases, the prior colonoscopy occurred within 4 years of CRC diagnosis (table 1). Median age was 73 (range 48–93), 56% of cases were male. Consistent most plausible explanation was found in 31/32 (97%) cases, showing almost perfect agreement (k=0.94). Categorisation into interval and non-interval types was consistent in 37/48 (77%) cases, showing substantial agreement (k=0.67). Following panel discussion, consensus was reached for most plausible explanation and categorisation of PCCRC in all cases. 66% of PCCRCs within 4 years of diagnosis were attributed to ‘possible missed lesion, prior examination adequate’. 73% of cases were categorised as Non-interval Type B or C. Full results are in table 1. Conclusion Using readily available data, PCCRC cases can be categorised by most plausible aetiology with almost perfect inter-rater agreement. Categorisation of PCCRC subtype was more challenging, reflecting uncertainties in surveillance intervals and endoscopic plans. Further discussion, with additional clinical information, led to agreement in all cases. The majority of PCCRC cases were categorised as ‘probable missed lesions following an adequate colonoscopy’. The high number of Non-interval type B PCCRCs suggests a significant proportion of PCCRC cases could be avoided with adherence to recommended surveillance intervals.