Background: Early diagnosis of recurrence may provide the opportunity for cure among colon cancer survivors. Randomized studies comparing post-treatment surveillance strategies suggest that patients receiving more frequent testing are more likely to undergo curative treatment upon recurrence. However, uncertainty about the cancer-specific mortality benefit of intensive surveillance has led to variation in surveillance recommendations across professional societies. We developed a simulation model to predict outcomes for different surveillance strategies within a hypothetical population of survivors. Methods: The Colorectal Cancer Surveillance and Recurrence (CRCSuRe) model generates natural history events in a simulated population of colon cancer survivors based on patient-level data from the Clinical Outcomes of Surgical Therapy (COST) Trial [NEJM.2004;350(20):2050-59]. Individuals are at time-varying risks for progression to recurrence, recurrence symptoms, metastatic recurrence, and death. Any schedule of follow-up testing using carcinoembryonic antigen (CEA), computed tomography (CT), and colonoscopy can be simulated in a population with defined age and stage distribution. We used CRCSuRe to compare three distinct surveillance regimens—all of which complied with one or more published guidelines—and no surveillance in a population with mean age of 65 years and equal proportions of stage II and III disease. Results: Predicted clinical outcomes are shown in the table. Conclusions: Increased intensity of surveillance is associated with more frequent detection of resectable recurrences and improved survival. However, modest predicted survival differences suggest that risk-adjusted surveillance strategies are needed to improve patient outcomes.

* In months if not stated; yrs = years.