Metachronous Neoplasias Arise in a Higher Proportion of Colon Segments From Which Large Polyps Were Previously Removed, and Can be Used to Estimate Incomplete Resection of 10–20 mm Colorectal Polyps

Abstract

Incomplete resection of polyps could be an important cause of post-colonoscopy colorectal cancer. However, it is difficult to study progression of incompletely removed polyps or their clinical importance. We aimed to estimate incomplete polyp resection using risk of metachronous neoplasia per colon segment. We performed a retrospective study of 1031 patients (6186 colon segments) who initially underwent resection of a large (10-20 mm) neoplastic polyp at 2 academic medical centers (from 2000 through 2012) and then underwent a subsequent colonoscopy within 0.5 to 5 years. We determined the proportions of metachronous neoplasia in colon segments from which a single large neoplastic polyp was removed and in segments without prior neoplasia. We then used the absolute difference in proportions between these groups to estimate the rate of incomplete resection. Our analysis assumed that development of metachronous neoplasia in each colon segment was the consequence of a newly grown polyp, a previously missed polyp, or an incompletely removed polyp. Metachronous neoplasia was detected in 177 of 757 segments (23.4%) with a single large polyp, and in 438 of 4232 segments (10.3%) without any neoplasia at baseline colonoscopy (P < .001). Resections were therefore estimated to be incomplete in 13.0% of segments (95% CI, 9.8-16.2). This proportion was greater for sections with non-pedunculated polyps (18.3%; 95% CI, 14.2-22.5) than pedunculated polyps (3.5%; 95% CI, -0.7 to -11.3; P < .001). A higher proportion of piecemeal resections appeared to be incomplete (28.0%; 95% CI, 20.2-35.7) than of en bloc resections (9.2%; 95% CI, 5.9-12.5) (P < .001). No differences in incomplete resection were associated with polyp histology. Metachronous neoplasia arises in a significantly higher proportion of colon segments from which a polyp was previously removed. Based on these data, we estimate that 13% of all large polyps are incompletely resected and 18% of large non-pedunculated polyps are incompletely resected. These findings indicate that incomplete resection could be a risk factor for later development of neoplasia. Segment metachronous neoplasia might be used as a marker of resection quality.