Article Text
PDF
Abstract
Introduction The Bowelscope screening programme offers “one-off” sigmoidoscopy to individuals at the age of 55. Those deemed at high risk following sigmoidoscopy are referred for colonoscopy. An adenoma with a villous or tubulovillous component is deemed high risk and should instigate colonoscopy.
Determining whether a polyp is tubular or tubulovillous is subjective and prone to inter-observer variation. The BCSP pathology guidelines state “it is accepted that a neoplastic villus almost defies definition”. This review was designed to explore the variability of reporting villous morphology in the national Bowel Cancer Screening Program (BCSP) screening centres in Yorkshire, Humber and North Eastern (Y, H and NE) regions and to investigate the level of agreement between individual BCSP pathologists.
Method 32 consultant pathologists active in the BCSP in Y, H and NE region viewed H and E glass slides from 12 bowel cancer screening polyps and categorised each as tubular, tubulovillous or villous. Following viewing, each case was presented to the group with a brief discussion of the features. Pathologists then used individual electronic voting buttons to classify each case into one of the three categories. Following the workshop, lead pathologists at the BCSP screening units in the Y, H and NE region provided the relative proportions of tubular, tubulovillous and villous adenomas reported in their screening centre in the previous 12 months.
Results The graph below shows the percentage of pathologists classifying each of the twelve polyps in one of the three categories. No single case achieved consensus, with many showing a spread of classification little better than chance. Agreement was best on whether a polyp was a pure villous adenoma or not.
Across the Y, H and NE region the proportion of adenomas reported as “tubulovillous”, ranged from 5.5 to 25%, with a median of 17%. These figures may disguise more marked individual variation as the figures from each screening centre represent the combined output of multiple individual pathologists and multiple departments. Pure villous adenomas were rare, median 0.6% of all adenomas, range 0.1 to 1.8%.
Conclusion Use of “villosity” to predict need for colonoscopy is prone to marked inter-observer variation. Using villosity may lead to unacceptable variation in colonoscopy rates with the potential for unpredictable overburden on resources. Better definitions of what constitutes a villus are required, along with training to improve consistency in recognition, before it can be used as a reliable risk predictor in a national screening program.
Disclosure of interest None Declared.
