Using Medical Informatics To Evaluate The Risk Of Colorectal Cancer In Patients With Clinically Diagnosed Sessile Serrated Polyps

Background : Recent research suggests that in addition to advanced adenomas, sessile serrated polyps (SSPs) may be important precursors for proximal colon cancer. In this study, we conducted the first large cohort study to evaluate the risk of colorectal cancer (CRC) in patients diagnosed with SSPs through usual care. Methods : The University of Washington Medical Center (UWMC) uses a comprehensive electronic medical records (EMR) system to track patient demographics and health-related information, including diagnoses and procedures for all patients. We used procedure codes to identify a cohort of patients receiving colonoscopies at UWMC from 2003-2013. Natural language processing of text in the final diagnosis section of the pathology report was used to characterize the type of polyps present at each colonoscopy procedure, including non-advanced conventional adenomas, advanced conventional adenomas (defined as having villous histology, high-grade dysplasia, or a diameter ≥ 10 mm), and sessile serrated polyps. These colonoscopy records were then linked to the Puget Sound Surveillance, Epidemiology, and End Results Cancer Registry (SEER) and subsequent EMR data to identify incident CRCs occurring through December 31, 2014 within this cohort. Those who lived outside of the SEER catchment area or who had prior colectomy, inflammatory bowel disease, or CRC were excluded from analyses. Cox proportional hazards regression models were used to calculate adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) comparing the risk of CRC in each polyp group to those who were polyp-free at an index colonoscopy. HR estimates were adjusted for age, sex, race/ethnicity, smoking status, and body mass index. Results : From 2003 through 2013, 32,136 colonoscopies were performed at UWMC, and 17,424 colonoscopies from 14,846 patients met the study inclusion criteria. Of these patients, 8,908 were polyp-free, 4,145 had only non-advanced conventional adenomas, 927 had advanced conventional adenomas and no SSPs, 314 had ≥1 SSP and ≥1 conventional adenoma, and 552 had SSPs and no conventional adenomas at an index colonoscopy. Median follow-up time in the study cohort was 5.8 years, and 66 incident colorectal cancers occurred during the follow-up period. The risk of incident CRC in those with advanced conventional adenomas at their index colonoscopy was significantly higher than those who were polyp-free (HR=3.9; CI: 1.9-7.7). However, there was not a statistically significant difference in the risk of incident CRC between those who were polyp-free at their index colonoscopy and those who had only non-advanced conventional adenomas (HR=1.5; CI: 0.8-2.6), SSPs with conventional adenomas (HR=2.0; CI: 0.5-8.7), or SSPs without conventional adenomas (HR=1.3; CI: 0.3-5.5). Discussion : Despite recent evidence from cross-sectional studies suggesting that SSPs are high-risk precursors for a subset of colorectal cancers, our results indicate that the risk of CRC in patients with clinically-diagnosed SSPs is similar to the risk of CRC in those with non-advanced adenomas. Additional longitudinal studies of SSPs diagnosed through usual care are needed to inform guidelines for the surveillance of patients with SSPs. This abstract is also being presented as Poster B01. Citation Format: Andrea Burnett-Hartman, Polly A. Newcomb, Chan X. Zeng, Yingye Zheng, John M. Inadomi, Christine Fong, Melissa P. Upton, William M. Grady. Using medical informatics to evaluate the risk of colorectal cancer in patients with clinically diagnosed sessile serrated polyps. [abstract]. In: Proceedings of the AACR Special Conference on Colorectal Cancer: From Initiation to Outcomes; 2016 Sep 17-20; Tampa, FL. Philadelphia (PA): AACR; Cancer Res 2017;77(3 Suppl):Abstract nr PR05.