COBRAS (Ottawa) University of Ottawa

Purpose

Simulate the natural history of bladder cancer, including tumor growth, symptoms, diagnosis and survival after diagnosis, to evaluate screening and surveillance strategies.

Overview

Bladder cancer is a significant public health concern, especially for older adults.^1,2 Despite advancements in therapy, incidence rates have remained relatively steady, partly because of population aging and persistent exposure to key risk factors, such as tobacco use and certain occupational hazards.³ Early detection and effective treatment can improve outcomes, but questions remain about allocating resources and tailoring interventions to specific patient groups.⁴ The Cancer of the Bladder in R Analytic Simulator (COBRAS) was created to address these challenges by offering a detailed, data-driven view of how bladder cancer arises, progresses, and can be identified, prevented, and treated in the United States.⁵

COBRAS is a discrete-event simulation (DES) model that begins by simulating individuals starting at birth. These simulated individuals are free of bladder tumors until a cancer-initiating event occurs. The model accounts for an array of influences on disease risk and progression, including demographic factors (age, sex, race), smoking behaviors, and other exposures. Once tumors develop, COBRAS distinguishes between different non–muscle-invasive stages (low-grade Ta, high-grade Ta, Tis, T1), advanced muscle-invasive stages (T2–T4), and potential nodal or metastatic involvement. As individuals age, tumors may progress from one stage to another, sometimes undetected until symptoms appear. The model also accommodates screening strategies, surveillance schedules, and potential interventions that alter the course of disease detection and treatment.

To ensure the reliability of its estimates, COBRAS is rigorously calibrated to national data from the Surveillance, Epidemiology, and End Results (SEER) program, focusing on age- and stage-specific incidence rates. This calibration employs a Bayesian method that generates parameters, making the model outputs consistent with observed data accounting for their uncertainty and expert inputs. As a result, COBRAS closely reproduces real-world patterns in bladder cancer incidence while quantifying the uncertainty surrounding inputs such as progression rates and symptom onset rates. Once the COBRAS model is calibrated, it was validated to the lifetime risk measure from SEER data. The model also estimates that the average “dwell time” from tumor onset to symptomatic detection is just over one year, emphasizing the aggressiveness of this cancer and the potential for earlier detection to improve outcomes.

COBRAS is designed to be computationally efficient. It can simulate realistic populations – on the order of millions of individuals – in seconds on a typical laptop.6 This capability is critical when exploring how small changes in risk factors or screening practices might translate into substantial shifts in public health outcomes at the national level. Decision-makers can use COBRAS to compare alternative screening regimens, assess the impact of more intensive surveillance for high-risk patients, and identify the most promising strategies for reducing the burden of bladder cancer. For instance, policymakers could test how tighter tobacco control measures or targeted screening in older populations might influence bladder cancer incidence and survival in the U.S. population over the next several decades.

Public health impact

By modeling the natural history of bladder cancer, calibrating the model to observed data, and simulating the consequences of varied early detection and treatment interventions, COBRAS supports informed policy decisions to optimize resource use for the cancer control of bladder cancer. As the population continues to age and risk factors for bladder cancer, such as smoking, remain commonplace, simulation models like COBRAS can inform prevention and treatment strategies that can reduce bladder cancer morbidity and mortality. Comparative modeling involving COBRAS enables both short- and long-term projections of bladder cancer burden and helps evaluate the feasibility and impact of screening and surveillance interventions in the general population and high-risk subgroups.

References

1. NCI. National Cancer Institute: SEER Cancer Stat Facts: Bladder Cancer 2019
2. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer Statistics, 2021. CA Cancer J Clin . 2021;71(1):7-33. Epub 2021/01/13. doi: 10.3322/caac.21654.
3. Lenis AT, Lec PM, Chamie K, Mshs MD. Bladder Cancer: A Review. JAMA. 2020;324(19):1980-91. Epub 2020/11/18. doi: 10.1001/jama.2020.17598.
4. Jalal H, Kang S, Trikalinos TA. Comparative Modeling of the Burden of Bladder Cancer in the United States. 2025.
5. Garibay-Treviño DU, Jalal H, & Alarid-Escudero F. A Fast Nonparametric Sampling Method for Time to Event in Individual-Level Simulation Models. Med Decis Making. 2025 Feb;45(2):205-213. doi: 10.1177/0272989X241308768. Epub 2025 Jan 5