The Risk of Recurrence due to Incomplete Resection of 10-20mm Polyps

Abstract

Introduction: Post-colonoscopy colorectal cancer in a colon segment of a previously removed polyp has been used to adjudicate incomplete resection (IR) as the underlying etiology. Although incomplete removal of large (10-20mm) non-pedunculated polyps is common (17% in the CARE study), the clinical importance of residual polyp tissue remains unclear. Aim: To estimate the risk of metachronous neoplasia due to IR (MIR) of neoplastic 10-20mm polyps. Methods: We built our analysis based on the assumption that MIR can be estimated as the difference between metachronous polyps found in a segment with a previously removed polyp and segments without any previous polyps (Risk of metachronous neoplasia=Risk of de novo lesion+Risk of missed lesion+Risk of IR). We abstracted data on all patients who had a large (10-20mm) polyp resection at two academic medical centers from 2000 to 2012 and a first surveillance colonoscopy within 0.5 to 5 years. Our primary outcome was the MIR following resection among all colon segments with a prior single 10-20mm neoplastic polyp. Metachronous neoplasia was defined as the presence of at least one neoplastic polyp in a colon segment at follow-up. We further examined if polyp morphology, mode of resection, and histology were associated with incomplete resection. Results: We identified 1039 patients (mean age 62, 70.3% men) with a median of 36 months to their surveillance exam (IQR 28). The absolute risk of metachronous neoplasia for segments without any neoplastic polyps at baseline was 10.3%, compared to 23.3% for segments with a single large polyp at baseline (P<0.001) (Figure), resulting in a 13.0% risk of MIR. This risk was higher for large sessile polyps (18.2%) than large pedunculated polyps (3.8%, P<0.001; Table). MIR was less likely following en-bloc resection (9.2%) than piecemeal resection (28.0%, P<0.001). Polyp histology only affected MIR following piecemeal resection: 15.6% for sessile serrated adenomas, 25.0% for tubular adenomas, and 42.1% for advanced adenomas.FigureTable: Table. Risk of metachronous neoplasia in colon segments with a prior single 10-20mm neoplastic polyp in reference to colon segments without any prior neoplastic polypConclusion: Applying a novel approach, this longitudinal study showed that incomplete resection of 10-20mm neoplastic polyps contributed 13% to metachronous neoplasia. This estimate increases to 18% for non-pedunculated large polyps, similar to the reported 17% in the cross-sectional CARE study. Our findings provide a strong argument that incomplete resection matters, although a field effect cannot be completely excluded.