Association of Polygenic-Based Breast Cancer Risk Prediction With Patient Management.

Abstract

Purpose: Breast cancer (BC) risk prediction is more accurate when clinical and polygenic factors are combined (combined risk score [CRS]), but little is known about how CRS results affect real-world patient management.

Methods: Deidentified medical and pharmacy claims data were linked with Tyrer-Cuzick (TC) and CRS results and evaluated for BC risk management. Patients were divided into subcohorts on the basis of lifetime risk predicted by CRS and by TC ("+": ≥20% risk, "-": <20%): CRS+ TC+, CRS+ TC-, CRS- TC+, and CRS- TC-. Claims data related to screening mammography (SM) in patients younger than 40 years, breast magnetic resonance imaging (MRI), and genetic counseling (GC) were compared 360 days before and after CRS testing. Differences in pre- and post-CRS management were evaluated using McNemar tests, and post-CRS management of subcohorts was compared using multivariable logistic regression.

Results: After CRS testing, the CRS+ TC+, CRS+ TC-, and CRS- TC+ subcohorts had 1.6-2.2-fold increases in SM in patients younger than 40 years (all P < .02) and 4.7-5.6-fold increases in breast MRI (all P < .001). The CRS+ TC+ and CRS+ TC- subcohorts had 1.9-2.3-fold increases in GC (both P < .001). SM in those younger than 40 years, breast MRI, and GC did not increase in the CRS- TC- subcohort. After CRS testing, compared with the CRS- TC- subcohort, the CRS+ TC+, CRS+ TC-, and CRS- TC+ subcohorts had significantly higher odds of receiving SM before age 40 years (odds ratio [OR], 3.80-5.19), breast MRI (OR, 11.55-23.09), and GC (OR, 2.03-2.91; all P < .001).

Conclusion: Patients with ≥20% lifetime risk predicted by either CRS or TC were more likely to receive enhanced management compared with those who had <20% lifetime risk, suggesting that clinicians considered the CRS in BC risk management.