Interobserver agreement in the reporting of colorectal polyp pathology among bowel cancer screening pathologists in Wales

Abstract

To assess the interobserver agreement in the reporting of colorectal polyps among histopathologists participating in the Welsh Bowel Cancer Screening (BCS) programme. Twelve benign polyps representative of BCS cases were identified from pathology files and reported by 28 BCS histopathologists using proforma sheets. The level of agreement between the participants and a gold standard was determined using kappa (κ) statistics. A moderate level of agreement was achieved in the reporting of polyp type [κ = 0.45; 95% confidence interval (CI) 0.34-0.59] and adenomatous lesions were distinguished from non-adenomatous lesions in 96% of cases. Substantial agreement was obtained in distinguishing low- and high-grade dysplasias (κ = 0.67; 95% CI 0.50-0.86), but there was only fair agreement in reporting excision margin status (κ = 0.24; 95% CI 0.07-0.43) with frequent use of the 'uncertain' category. Significant issues included categorizing serrated lesions, recognizing focal high-grade dysplasia and epithelial misplacement, and apparent overdiagnosis of villous change in adenomas. Interobserver variability in some aspects of reporting colorectal polyps by BCS pathologists is suboptimal, with a potential impact upon patient management and the efficient running of the screening service. Approaches to addressing this are discussed.