Mo1666 INTERVAL COLORECTAL CANCERS ARE INFREQUENT AND OCCUR DUE TO BOTH MISSED AND INCOMPLETELY RESECTED POLYPS

Abstract

Although colonoscopy reduces colorectal cancer (CRC) risk, interval CRCs (iCRCs) still occur. Although uncommon, iCRCs can inform areas wherein colonoscopy quality improvements must occur. To determine the incidence and potential causes of iCRC. We performed a retrospective cohort study of all patients who underwent colonoscopy from April 1996 to March 2017 at a single academic medical center. Patients who developed an iCRC were identified through a review of our institutional data warehouse, an integrated database of clinical and research information from all patients receiving treatment through our healthcare system. An iCRC was defined as a cancer diagnosed 6-60 months after index colonoscopy; an early iCRC was defined as cancer first diagnosed 6-36 months after index colonoscopy. We manually reviewed all charts to confirm appropriateness for study inclusion and abstract patient, procedure, endoscopist, and cancer characteristics. A total of 252,256 colonoscopies were performed over the study period. We identified 203 patients diagnosed with iCRC with a median age at index colonoscopy of 65 (IQR 58-74). The overall iCRC rate was 0.12% and the early iCRC rate was 0.06%. This rate did not significantly improve over the study period (r=-0.3, p=0.4). However, the iCRC rates did vary significantly among providers performing ≥500 colonoscopies and correlated inversely with provider adenoma detection rates (ADR; Figure 1). Patients undergoing procedures by endoscopists in the highest ADR quartile (36-52%) had an iCRC risk five times lower (HR 0.21, 95% CI 0.08-0.52) than those undergoing colonoscopy by an endoscopist in the lowest quartile (11-19%). The iCRC-related mortality rate was 23% for both overall and early iCRC. Proximal (cecum to transverse colon) cancers accounted for a majority (54%) of iCRCs and rectal/rectosigmoid cancers represented 22% of iCRCs. One-third of cancers were diagnosed at an advanced stage (III and IV). In 42% of patients, gastrointestinal symptoms prompted an early repeat colonoscopy which identified the iCRC. One-third of iCRC cases (34%) were detected during surveillance exams. Incomplete polypectomy was the confirmed etiology for iCRC in 7.4% of cases and a possible etiology (prior polypectomy at same or adjacent colon segment) in 25% of cases. The most common modifiable etiology of iCRC during the study period was a missed polyp (65%). However, incomplete polypectomy was associated with 7-32% of iCRCs. Furthermore, higher provider ADRs were significantly associated with lower iCRC and early iCRC rates. Colonoscopy quality improvement measures aimed at reducing iCRC incidence should therefore target both improving polyp detection and resection.Table 1. Summary of iCRC patient, tumor, and endoscopic characteristics.View Large Image Figure ViewerDownload Hi-res image Download (PPT)