Validation of a Clinical and Polygenic Risk Prediction Model for Ovarian Cancer in the Nurses' Health Study

Area of Science:

• Gynecologic Oncology
• Genetics
• Epidemiology

Background:

• Ovarian cancer is a leading cause of gynecological cancer mortality due to late diagnosis.
• Current risk stratification relies on pathogenic variant testing, but high-risk variants are rare.
• Clinical models incorporating polygenic risk offer improved risk stratification for ovarian cancer.

Purpose of the Study:

• To cross-validate a refined risk model integrating polygenic risk and clinical factors for ovarian cancer.
• To compare the performance of the refined risk model against traditional clinical risk scores and polygenic risk score models.

Main Methods:

• A nested case-control study design was employed using data from the Nurses' Health Study.
• The refined risk model was evaluated against a clinical risk score and a polygenic risk score model.
• Statistical analysis included odds ratios and area under the curve (AUC) calculations.

Main Results:

• The refined risk model demonstrated an odds ratio of 1.26 (95% CI: 1.01-1.48) and an AUC of 0.56 (95% CI: 0.51-0.61).
• Performance metrics for the clinical risk and polygenic risk models were comparable, with AUCs of 0.54 and 0.55, respectively.
• While population screening has shown stage shifting, a clear mortality benefit has not yet been established.

Conclusions:

• The refined risk model shows potential for improved ovarian cancer risk stratification.
• As ovarian cancer treatments advance, a risk-stratified screening approach is expected to yield greater benefits.
• Further validation and implementation of refined risk models could enhance targeted prevention and early detection strategies.