Population balance models capture the behavior of a heterogeneous population by describing variation among individuals. These models represent the continuum between purely empirical models and mechanistic model where details are often lumped into measurable parameters without explicitly accounting for detailed mechanisms. In the models of colorectal cancer (CRC), a patient’s CRC risk varies based on his age, demographics, and clinical history. The tools available on the cceHUB – which were built in the cceHUB shared development environment - allow a user to run the population balance models and explore the dynamics of CRC development for a target population or individual patient.
Three population-based tools have been developed for cceHUB:
The colorectal cancer incidence prediction model is a stochastic simulation of CRC incidence with age. This model describes the accumulation of somatic mutations within a single cell for genes commonly altered in colorectal adenomas and carcinomas. The model predictions are compared to the Indiana population and scenarios based on race and gender parameters/data are available.
Browse the collection of tool runs for parameters such as male/female and black/white for the population of the state of Indiana.
Monte Carlo Simulations of Colorectal Cancer Development
The sequential mutation model of Nordling (1953) demonstrates that a series of roughly six somatic mutations captures the observed linear log (CRC incidence) versus log(age) relationship.
The model of Luebeck and Moogavkar (2002) includes the abnormal proliferation of an adenoma (which develops after a series of somatic mutations) as an intermediate step to carcinoma.
The MISCAN-COLON model of Loeve et al. (1999) describes the natural history of the adenoma-carcinoma sequence. Small adenomas develop at an age-dependent rate and then make discrete transitions to larger sized adenomas or to and among CRC stages.
Since each Monte Carlo trial is a random re-creation of an individual patient’s lifetime, the behavior of a population can be approximated by observing the results of multiple Monte Carlo trials.
Browse the collection of tool runs for predictions of patients with colorectal cancer or advanced adenoma at each CRC screen.
Patient-Specific Colonic Neoplasia Incidence
The tool patient-specific colonic neoplasia incidence predicts the likelihood of CRC and other colonic neoplasia for an individual patient based on his demographic risk factors and endoscopic history. Risk factors for CRC differ between patients and can include characteristics such as gender, age, and family history of CRC. In addition, clinical risk factors such as finding multiple or advanced adenomas at a baseline colonoscopy have shown even stronger ties to CRC risk than the demographic characteristics.
Browse the collection of tool runs for predicting the likelihood of CRC and other colonic neoplasia based on risk factors and health history.
credits
Eric Sherer, e-Enterprise Center, Purdue University; Veteran's Administration CoE on Implementing Evidence-based Practice; Indiana University Center for Health Services and Outcomes Research; Regenstrief Institute, Inc.
references
CO Nordling, “A new theory on cancer-inducing mechanism.” British Journal of Cancer, 7: 68-72, 1953.
EG Luebeck and SH Moolgavkar, “Multistage carcinogenesis and the incidence of colorectal cancer.” PNAS, 99: 15095-15100, 2002.
F Loeve, R Boer, GJ vn Oortmarssen, M van Ballegooijen, and JDF Habbema, “The MISCAN-COLON simulation model for the evaluation of colorectal cancer screening.” Computers and Biomedical Research, 32: 13-33, 1999.
Cite this work
If you reference this work in a publication, please cite as follows:
Eric Sherer (2010), "Population-Based Models," http://ccehub.org/resources/289.
