Purpose: The Breast Cancer Outcomes Simulator (BCOS), created at Stanford University, was developed for four primary purposes. First, BCOS generates a virtual tumor registry of breast cancer patients diagnosed in the United States since 1975 and, at the individual level, specifies the patient's screening history, mode of cancer detection, adjuvant treatment, and survival. Second, BCOS quantifies the impact of screening mammography and adjuvant therapy on breast cancer mortality trends from 1975 by molecular subtype. Third, BCOS predicts what the incidence and mortality trends would have been had alternative age groups been targeted for screening, had there been changes to the interval between screening examinations, and/or had there been changes to the subgroups targeted for adjuvant therapy. Fourth, BCOS predicts how future trends in breast cancer mortality may be affected by new screening and treatment interventions shown to be beneficial at the clinical trial level.
Overview: The BCOS model aims to reproduce population-level US breast cancer incidence and mortality trends from 1975-2015 (Surveillance, Epidemiology, and End Results [SEER] data) by capturing breast cancer events that involve heterogeneity in disease progression, patient characteristics, compliance with screening, and response to adjuvant treatment.
Compared to prior versions developed in 2009, several changes have been performed to BCOS regarding updates to several base inputs, changes in natural history parameters, and the effect and dissemination of screening and treatment. These changes are described below.
Incorporation of estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2)-specific underlying survival, natural history parameters and treatment efficacy: The current version of BCOS incorporates underlying breast cancer-specific survival by ER and HER2 subtype, as well as their predicted distribution at clinical detection in the absence of screening. Several natural history model parameters and the efficacy of treatment have also been made specific to molecular subtype. Previously, tumor volume doubling times and mammography detection thresholds were age-independent. Based on subsequent work, these parameters were updated and stratified by age at clinical detection. Treatment efficacy was originally modeled assuming proportional hazards, in other words, that the benefits were proportionally distributed across the years following diagnosis. The current BOCS model has been updated to incorporate non-proportional hazard to the effect of treatment depending on the ER status of each breast cancer case.
Updates to base inputs for screening, treatment and other-cause mortality: BCOS has incorporated the changes to the “base inputs” that are shared across all models, including: 1) the update to screening dissemination up to year 2010; 2) the update to treatment dissemination and efficacy up to calendar year 2010; and 3) the update to other-cause mortality for all birth cohorts.
Modeling the effect of menopausal hormonal therapy: The current BCOS model quantifies the effects of menopausal hormonal therapy (MHT) on breast cancer incidence and mortality trends. For this purpose, the BCOS modeling team developed a MHT dissemination model for women that were over age 50 before and after 2002. All parameters related to the MHT modeling were calibrated to reproduce breast cancer incidence trends with increasing MHT use before 2002 and, as validation, predict a rapid decline in incidence after the decline of MHT use in 2002. The team tested the hypothesis that MHT increases tumor growth and decreases mammographic detectability, and found that it is consistent with SEER data.
Update to the background breast cancer age-period-cohort incidence: The current BCOS incorporates an update to the background breast cancer age-period-cohort (APC) incidence. Contrary to other groups, BCOS does not rely on the “base case” input, as these estimates appear to produce an unexpectedly high lifetime breast cancer risk for the younger birth cohorts. Instead, BCOS uses background breast cancer incidence derived from a novel approach developed under the APC framework; the approach explicitly considers the effects of screening and MHT.
References
- Munoz D, Xu C, Plevritis S. A Molecular Subtype-Specific Stochastic Simulation Model of US Breast Cancer Incidence, Survival, and Mortality Trends from 1975 to 2010. Med Decis Making. 2018 Apr;38(1_suppl):89S-98S. [Abstract]
- Plevritis SK, Sigal BM, Salzman P, Rosenberg J, Glynn P. A stochastic simulation model of U.S. breast cancer mortality trends from 1975 to 2000. J Natl Cancer Inst Monogr 2006;(36):86-95.[Abstract]
