Annual Scientific Meeting
Sessile serrated polyps (SSPs) are candidate precursors of most interval colorectal cancers. These mostly right-sided lesions are easily missed, often incompletely removed, and may progress rapidly to malignancy. Accurate histologic distinction between precancerous and non-precancerous polyps is key to choosing appropriate intervals for colonoscopy. Among serrated polyps (SPs), precancerous SSPs are histologically similar to non-precancerous hyperplastic polyps, and disparity among expert pathologists is high with independent reviews of the same specimens. We assessed if this discrepancy would be reflected by significant variability in classification rates of HPs and SSPs among pathologists in a real-world academic clinical setting.
Using our Colonoscopy Quality Database (UCICQD) at a large academic medical center, we analyzed prospectively-collected data for all polyps removed and examined between 6/2012 and 12/2016. Polyp classification rates of each pathologist were calculated; association between pathologists and polyp classifications were assessed by unconditional logistic regression for multivariate analysis. The rate of classification of each pathologist was analyzed as polyp size increased.
Results: Of 8,274 in polyps classified by 9 pathologists, 62.5% were described as adenomas. For 3,103 SPs, a mean of 22.0% was classified as SSPs. SSP classification rate varied among pathologists; a statistically significant association was found between pathologists and type of SP (HP vs SSP) (p< 0.0001). The rate increased for all pathologists with increasing polyp size, although variability was high and pathologists maintained their relative positions when comparing SSPs/SPs between polyp sizes (1-3 mm vs > 4 mm), R = 0.8, p = 0.01.
The SP classification rates of pathologists in the right colons are highly variable. The fraction of SPs classified as SSPs increases with polyp size but variability among pathologists remains high. This suggests that each pathologist is influenced by polyp size but has a different histologic threshold for distinguishing HPs and SSPs in the right colon. This data supports the idea that serrated polyps (HPs or SSPs) proximal to the sigmoid colon should be considered "surveillance relevant". This strategy will likely reduce interval cancer rates at the expense of additional colonoscopies in those with true HPs. Novel methodologies are needed to distinguish these morphologically similar but biologically distinct lesions.