Amy Trentham-Dietz

PhD

Board of Regents of the University of Wisconsin System

Professor

PLCO (Learn more about this study)

PLCO-1922

Jun 25, 2025

POPULATION MODELING TO OPTIMIZE EQUITABLE OVARIAN CANCER OUTCOMES

Most ovarian cancers are discovered at later stages after metastases have occurred and have high rates of recurrence, even if initial treatments are successful. Emerging liquid biopsy screening tests have garnered optimism for addressing deficiencies of conventional screening approaches for detecting ovarian cancer at earlier stages and are rapidly disseminating. Our improved understanding of ovarian cancer etiology has not been leveraged in the setting of early detection, and considerable potential remains for more effective ovarian cancer screening.

In the context of long-standing poor prognosis after an ovarian cancer diagnosis, we hypothesize that recent advances in the understanding of ovarian cancer etiology, applications of artificial intelligence, and development of novel liquid biopsies can be leveraged to improve early detection. Our overarching objective is to use innovative population modeling to drive interventions that will reduce the burden of epithelial ovarian cancer.

We propose 3 Aims: (1) Develop, calibrate, and validate a simulation model of ovarian cancer outcomes for U.S. women. (2) Using data from ovarian cancer screening trials, identify strategies with favorable benefits-to-harms using conventional ovarian cancer screening tests according to risk level of ovarian cancer. (3) Evaluate the comparative effectiveness of emerging ovarian cancer screening technologies including liquid biopsy for reducing ovarian cancer mortality according to a range of risk levels.

We propose to construct the University of Wisconsin Model of Ovarian Cancer (UW-MOCa), a discrete-event micro-simulation model to replicate ovarian cancer epidemiology and outcomes in U.S. women over time. UW-MOCa will simulate the lifetimes of individual women through the interaction of four main model components: population core (birth and competing mortality), ovarian cancer natural history, detection, and treatment/survival. UW-MOCa will use data for model parameters drawing from nationally representative sources. The model will be calibrated against national ovarian cancer incidence, survival, and mortality rates for women aged 15-84 for the years 1992-2025. Separate rates will be estimated for histotypes of ovarian cancer. Validation will use trial data. For Aim 2, we will seek to identify a risk level that results in the number of screen-detected ovarian cancer cases exceeding the number of surgeries following false-positive screening tests. Risk levels will depend on well-established risk factors for ovarian cancer. For Aim 3, the model will generate long-term outcome data based on published and hypothetical test performance for the major available liquid biopsy tests.

By developing, calibrating, and implementing a simulation model of ovarian cancer of U.S. women, we can accelerate the dissemination of ovarian cancer screening in an evidence-based and equitable manner. Simulation modeling as we propose will be a critical approach for integrating multiple sources of evidence, optimizing the deployment of novel screening tests, and avoiding exacerbating patient hardship from false positive tests.