The Simulation Model of Colorectal Cancer (SimCRC; University of Minnesota and Massachusetts General Hospital) describes discrete event state transitions in continuous time. State transitions can depend on the age, sex, and race of the individual, lesion location within the large intestine, and calendar time. For each simulated individual, SimCRC first generates a time of birth and a time of death from causes other than colorectal cancer (CRC). Next, SimCRC generates adenomas within the individual, with the age of onset for each adenoma drawn from a cumulative probability function. The probability function depends on sex, race, age, and an individual risk index that captures whether a person tends to produce more (or fewer) adenomas than average. Individual adenoma risk is simulated from a Truncated Normal distribution (restricted to positive values). Once initiated, each adenoma is assigned a location in the large intestine according to the location distribution of adenomas in autopsy studies.
SimCRC simulates three adenoma categories (small [1-5 mm], medium [6-9 mm], and large [10+ mm]) and in six locations. All adenomas start small and can transition through larger size categories. Each adenoma is randomly assigned an adenoma-specific growth index, based on a Truncated Normal distribution, which allows between-adenoma variability in the probability of transitioning to larger sizes. The timing of transitions between adenoma size categories depends on age, sex, location (proximal colon, distal colon, rectum), and (optionally) eight modifiable risk factors described below.
Medium and large adenomas may progress to preclinical CRC, although most will not in an individual’s lifetime. Progression depends on sex, race, and adenoma location. SimCRC can be (optionally) implemented to allow progression to preclinical CRC to depend on eight risk factors and birth year. Progression through preclinical cancer stages I through stage IV depends on initial stage and location. At each preclinical stage, SimCRC simulates whether there is transition to symptomatic and clinically detected cancer as a function of stage and preclinical cancer location. After clinical detection, SimCRC simulates the survival time to death from CRC using relative survival estimates from the Surveillance, Epidemiology, and End Results (SEER) program, following an approach that is common across the three models.1 CRC survival time depends on age, calendar year, stage at detection, cancer location (colon or rectum), sex, and (optionally) race (black or white). For individuals with synchronous CRCs at the time of diagnosis, SimCRC uses the stage-specific survival of the cancer with the highest stage. For all individuals with CRC, their death date is set to the earliest simulated death date (either due to CRC death or other causes).
Risk Factor Modeling: SimCRC includes an optional risk factor module that allows individual-level risk factors to influence CRC incidence by affecting both the age-specific incidence of adenomas and adenoma progression to CRC. The SimCRC risk module includes body mass index (BMI) (kg/m2), smoking (nonsmoker, or smoker with number of cigarettes per day), and red meat consumption (servings per day) as factors related to an increased risk for adenomas and/or progression to CRC, and physical activity (met-hours per week), fruit and vegetable consumption (servings per day), multivitamin use (yes/no), aspirin use (yes/no), and hormone replacement therapy (yes/no) as factors related to a decreased risk for adenomas and/or adenoma progression to CRC. This module tracks risk factor values for each simulated person over time based on analyses of multiple waves of the National Health and Nutrition Examination Survey (NHANES). The SimCRC team utilized a combination of simulation modeling and epidemiologic analysis2 to decompose the effect of risk factors on adenoma risk and progression to cancer (unobserved states). They empirically estimated the effect of CRC risk factors on these unobserved states using data from the Nurses’ Health Study (NHS) and the Health Professionals’ Follow-up Study (HPFS) to derive risk functions that describe the relationship between the CRC risk factors and: 1) the development of an adenoma; and 2) the progression of an adenoma to preclinical CRC. Evidence from NHS and HPFS suggests that three of the risk factors (aspirin use, multivitamin use, and smoking) act primarily on the initial adenoma development because CRC incidence is associated with exposure to these risk factors 10 to 15 years earlier. 3,4,5,6,7,8,9,10
Post-CRC-diagnosis Model: SimCRC has a separate post-CRC-diagnosis model that can be run as a stand-alone model or can be linked with the natural history model. In this model, each simulated cancer patient faces a risk of developing a metastatic recurrence, which may or may not be amenable to treatment. Only patients with non-resectable metastatic disease can die from CRC. The model parameters (e.g., risk of developing metastatic recurrence) have been estimated through calibrations to match SEER relative survival curves by stage, location, age (<75, ≥75), and period (to reflect changes over time). This framework allows for a more accurate reflection of the use of chemotherapy for metastatic recurrences.
References
- Rutter CM, Johnson EA, Feuer EJ, Knudsen AB, Kuntz KM, Schrag D. Secular trends in colon and rectal cancer relative survival. J Natl Cancer Inst. 2013;105(23):1806-13.
- Morris JM. Using Monte Carlo disease simulation to empirically estimate the effect of aspirin use on progression of colorectal cancer. [Doctoral dissertation], Harvard University; 2000.
- Giovannucci E, Stampfer MJ, Colditz GA, et al. Folate, methionine, and alcohol intake and risk of colorectal adenoma. J Natl Cancer Inst. Jun 2 1993;85(11):875-884.
- Giovannucci E, Colditz GA, Stampfer MJ, et al. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. women. J Natl Cancer Inst. Feb 2 1994a;86(3):192-199.
- Giovannucci E, Rimm EB, Stampfer MJ, et al. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men. J Natl Cancer Inst. Feb 2 1994b;86(3):183-191.
- Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio A, Willett WC. Aspirin use and the risk for colorectal cancer and adenoma in male health professionals. Ann Intern Med. Aug 15 1994c;121(4):241-246.
- Giovannucci E, Egan KM, Hunter DJ, et al. Aspirin and the risk of colorectal cancer in women. N Engl JMed. Sep 7 1995;333(10):609-614.
- Giovannucci E, Stampfer MJ, Colditz GA, et al. Multivitamin use, folate, and colon cancer in women in the Nurses' Health Study. Ann Intern Med. Oct 1 1998;129(7):517-524.
- Chan AT, Giovannucci EL, Schernhammer ES, et al. A prospective study of aspirin use and the risk for colorectal adenoma. Ann Intern Med. Feb 3 2004;140(3):157-166.
- Wei EK, Giovannucci E, Wu K, et al. Comparison of risk factors for colon and rectal cancer. Int J Cancer. Jan 20 2004;108(3):433-442.
