Purpose: The University of Wisconsin Breast Cancer Simulation (UWBCS) model simulates breast cancer in a population over time, generating cancer registry-like data sets. By manipulating parametric input assumptions about natural history, screening, and treatment, the model can be used to address a number of important policy questions related to breast cancer screening and treatment.
Overview: UWBCS is a discrete-event microsimulation that uses a systems engineering approach to replicate breast cancer epidemiology in the US population over time. This population-based model simulates the lifetimes of individual women through the interaction of four main model components: breast cancer natural history, detection, treatment and mortality. Briefly, each woman enters the model at age 20 and ages in 6-month time cycles. Risk of breast cancer onset is a function of age and calendar year. If onset occurs, cancer progresses through disease stages following a stochastic growth curve. Cancer can be detected by routine clinical finding or by screening. Upon detection, stage- and age-specific treatment is applied. Women with cancer are at risk of death from breast cancer. All women face competing mortality risks from other causes. A brief description of each module follows.
Natural History: Breast cancer inception is a function of a woman's race, birth year, and age, and accounts for secular trends in risk. Cancers are assumed to grow according to a stochastic Gompertz-type model. Tumor spread is described by a Poisson process determined by tumor size and growth rate. Tumors are assigned a Surveillance, Epidemiology, and End Results (SEER) historical stage (in situ, localized, regional, or distant) based on tumor size and lymph node involvement at the time of diagnosis in the model. A fraction of in situ and early localized invasive cancers are assumed to be of low malignant potential and not to pose a lethal threat to women.
Detection: Breast cancer can be detected either mammographically or by routine clinical finding. Mammography sensitivity and the likelihood of clinical detection are functions of age and tumor size, as well as calendar year to account for improvements in technology and increased breast cancer awareness. Sensitivity has been calibrated to match observed estimates. Mammography utilization can follow actual age-specific US screening patterns or fixed criteria by starting/ending ages, frequency, and population participation.
Treatment: Upon diagnosis, all women are assumed to receive standard treatment. Adjuvant therapy follows observed US dissemination patterns. Treatment effectiveness is a function of age, stage, estrogen receptor (ER)/human epidermal growth factor receptor 2 (HER2) status, and receipt of adjuvant treatment, and is modeled independently of the cancer detection method. An effective treatment is assumed to halt breast cancer progression.
Risk of Death: Each woman is assigned a date of death due to non-breast cancer causes based on US birth cohort-based life tables, with breast cancer removed as a cause of death. For women who have progressed to distant-stage breast cancer, a date of death from breast cancer is assigned based on observed SEER cancer survival. The timing and cause of death are determined by the earlier of the two dates of death (breast cancer or other causes).
Model Output and Execution: The interplay of the above modules defines the individual life histories of simulated women. The UWBCS model is calibrated to SEER breast cancer incidence and National Center for Health Statistics (NCHS) breast cancer mortality from 1970-2010 and cross-validated against data from the Wisconsin Cancer Reporting System. Model output is highly customizable in its level of detail and can include underlying disease states as well as observed clinical outcomes by age, race, and calendar year. The model also has current capabilities to track costs, resource use, and quality of life associated with screening and treatment. The model uses the variance reduction technique of common random numbers to reduce random variation in model outcomes. This feature also allows the model to conduct counterfactual experiments for the direct calculation of quantities such as overdiagnosis and lead-time. The model, programmed in C++, runs on both Microsoft Windows and UNIX platforms.
References
- Alagoz O, Ergun MA, Cevik M, et al. The University of Wisconsin breast cancer epidemiology simulation model: an update. Med Decis Making. 2018 Apr;38(1_suppl):99S-111S. [Abstract]
- Fryback DG, Stout NK, Rosenberg MA, Trentham-Dietz A, Kuruchittham V, Remington PL. The Wisconsin Breast Cancer Epidemiology Simulation Model. J Natl Cancer Inst Monogr 2006;(36):37-47. [Abstract]