A prognostic model for advanced colorectal neoplasia recurrence
Following colonoscopic polypectomy, US Multisociety Task Force (USMSTF) guidelines stratify patients based on risk of subsequent advanced neoplasia (AN) using number, size, and histology of resected polyps, but have only moderate sensitivity and specificity. We hypothesized that a state-of-the-art statistical prediction model might improve identification of patients at high risk of future AN and address these challenges.
Data were pooled from seven prospective studies which had follow-up ascertainment of metachronous AN within 3–5 years of baseline polypectomy (combined n = 8,228). Pooled data were randomly split into training (n = 5,483) and validation (n = 2,745) sets. A prognostic model was developed using best practices. Two risk cut-points were identified in the training data which achieved a 10 percentage point improvement in sensitivity and specificity, respectively, over current USMSTF guidelines. Clinical benefit of USMSTF versus model-based risk stratification was then estimated using validation data.
The final model included polyp location, prior polyp history, patient age, and number, size and histology of resected polyps. The first risk cut-point improved sensitivity but with loss of specificity. The second risk cut-point improved specificity without loss of sensitivity (specificity 46.2 % model vs. 42.1 % guidelines, p < 0.001; sensitivity 75.8 % model vs. 74.0 % guidelines, p = 0.64). Estimated AUC was 65 % (95 % CI: 62–69 %).
This model-based approach allows flexibility in trading sensitivity and specificity, which can optimize colonoscopy over- versus underuse rates. Only modest improvements in prognostic power are possible using currently available clinical data. Research considering additional factors such as adenoma detection rate for risk prediction appears warranted.
KeywordsPolyp surveillance Risk stratification Epidemiology Colorectal cancer Colorectal polyps
Area under the receiver operating characteristic curve
Body mass index
L1-regularized logistic regression model
Net reclassification improvement
Receiver operating characteristic
US Multisociety Task Force
This work was supported in part by Public Health Service Grants, CA-41108, CA-23074, CA95060, CA37287, CA104869, CA23108, CA59005, CA26852, and 5R01CA155293 from the National Cancer Institute. Funding for the Veteran’s Affairs Study was supported by the Cooperative Studies Program, Department of Veterans Affairs. The project described was also supported by a pilot grant from the UCSD Department of Family Medicine and Public Health (Liu, PI), as well as in part by Merit Review Award number 1 I01 HX001574-01A1 (Gupta, PI) from the United States Department of Veterans Affairs Health Services Research and Development Service of the VA Office of Research and Development. The views expressed in this article are those of the author(s) and do not necessarily represent the views of the Department of Veterans Affairs.
L.L., K.M., M.E.M., and S.G. contributed to study concept and design. L.L., K.M., and S.G. drafted the manuscript. L.L. and K.M performed the statistical analysis. L.L., K.M., J.A.B., D.A.L., A.J.C., M.G., M.E.M., and S.G. obtained the funding All authors contributed to acquisition, analysis, and interpretation of data, critically revised the manuscript for important intellectual content, and approved the final version of the article, including the authorship list.
Compliance with ethical standards
Conflict of interest