Using medical informatics to follow-up a colorectal sessile serrated polyp cohort

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Recent evidence suggests that sessile serrated polyps (SSPs), flat, colonic lesions with architectural abnormalities at the base of crypts and a serrated appearance, may be important precursors to colorectal cancer. Previously, advanced adenomas were the only known polyp precursors for colorectal cancer, and SSPs were clinically grouped with hyperplastic polyps, lesions routinely believed to have no malignant potential. Now, there is growing consensus that SSPs progress to cancer along the "serrated pathway". New guidelines were developed in 2012 recommending complete removal of SSPs and increased colorectal cancer surveillance in patients with a history of SSPs. However, these guidelines acknowledge that these new recommendations are largely based on cross-sectional data, with the only longitudinal study of colorectal cancer risk in patients with SSPs limited by size (n=40). Therefore, the primary objective of this application is to gather preliminary data to estimate the risk of colorectal cancer in a large cohort of patients with clinically diagnosed SSPs. We will address the following specific aims: 1) compare the risk of subsequent colorectal cancer in patients who had SSPs diagnosed at an index colonoscopy to patients with advanced adenomas, and estimate the time to cancer progression in both polyp groups; 2) among those with SSPs at index colonoscopy, evaluate whether risk of incident colorectal cancer varies according to size and location of the index SSP. To accomplish these aims, we propose a cohort study of 2,250 patients evaluated via an index colonoscopy between 2003 and 2011 at the University of Washington Medical Center Gastroenterology Clinic and diagnosed with SSPs (N=750) or advanced adenomas (N=1,500). Using University of Washington's established Microsoft Amalga data aggregation platform, which links electronic medical records data across different sources, we have already identified our study cohort. Additional medical informatics and natural language processing will be used to electronically extract data on patient characteristics, and index polyp size and location. Then, we will electronically follow-up patients for incident colorectal cancer through December 2012 by linking our study cohort to the Puget Sound Surveillance, Epidemiology, and End Results Cancer Registry (SEER), a population based cancer registry covering western Washington State, including the catchment area for the University of Washington Medical Center Gastroenterology Clinic patient population. Cox regression models will be used to estimate hazard ratios and 95% confidence intervals of colorectal cancer risk comparing patients with SSPs to those with advanced adenomas (Aim 1) and to estimate colorectal cancer risk between patients with different polyp characteristics among those with SSPs (Aim 2). This will be the largest longitudinal study of colorectal cancer risk in patients with SSPs to date and the first study in a colonoscopy-based cohort to estimate incidence of colorectal cancer in those with SSPs diagnosed though usual care in the clinical setting.