Association of Genetic Testing Results With Mortality Among Women With Breast Cancer or Ovarian Cancer

Allison W. Kurian, MD, MSc; Paul Abrahamse, MA; Irina Bondarenko, MS; Ann S. Hamilton, PhD; Dennis Deapen, DrPH; Scarlett L. Gomez, PhD; Monica Morrow, MD; Jonathan S. Berek, MD, MMSc; Timothy P. Hofer, MD, MSc; Steven J. Katz, MD, MPH; Kevin C. Ward, PhD, MPH

Disclosures

J Natl Cancer Inst. 2022;114(2):245-253. 

In This Article

Results

Study Population

Supplementary Figure 1 (available online) shows flow of patients into the analytic cohort, and Table 1 shows characteristics of genetically tested breast (n = 22 495) and ovarian cancer patients (n = 4320). Among breast cancer patients, 58.0% were non-Hispanic White, 14.8% Black, 16.2% Hispanic, 10.7% Asian or Pacific Islander, and less than 1% Native American or Alaskan Native, with a similar distribution in ovarian cancer patients. Approximately one-fifth of patients lived in high-poverty areas and half in low-poverty areas. Breast cancer subtype distribution was 48.7% ER/PR-positive, HER2-negative (n = 10 956), 27.0% HER2-positive (n = 6078), and 24.3% TNBC (n = 5461). Most (71.7%) ovarian cancer patients had serous histology and high (3 and 4, 81.6%) grades. The median follow-up was 41 (range = 1-85) months. For breast cancer patients, genetic testing occurred before diagnosis in 4.6% (n = 1037) and before chemotherapy initiation in 64.0% (n = 14 411); for ovarian cancer patients, these proportions were 2.9% (n = 124) and 18.9% (n = 857), respectively.

Genetic Testing Results, Treatment, and Mortality

Genetic results are summarized in Table 2. PVs were present in 12.6% (n = 1386) of patients with ER/PR-positive, HER2-negative breast cancer; 9.7% (n = 595) with HER2-positive breast cancer; 16.8% (n = 916) with TNBC; and 17.2% (n = 744) with ovarian cancer (Supplementary Table 1, available online). PVs were most common in BRCA1/2. Among breast cancer patients, other common PVs were, with ER/PR-positive, HER2-negative disease (n = 567 other gene PVs): CHEK2 (n = 214), PALB2 (n = 120), ATM (84), BRIP1 (n = 32), and TP53 (n = 22); with HER2-positive disease (n = 363 other gene PVs): CHEK2 (n = 156), ATM (n = 71), TP53 (n = 64), and PALB2 (n = 31); and with TNBC (n = 182 other gene PVs): PALB2 (n = 66), RAD51C (n = 23), BRIP1 (n = 23), CHEK2 (n = 23), ATM (n = 19), and RAD51D (n = 13). Among ovarian cancer patients, other common PVs (n = 174 other gene PVs) were in BRIP1 (n = 35), CHEK2 (n = 27), RAD51C (n = 24), ATM (n = 19), and RAD51D (n = 17).

Treatment receipt is shown according to breast cancer subtype (Table 2) and genetic testing results (Table 3). Death from the diagnosed cancer occurred in 6.0% of breast cancer patients with ER/PR-positive, HER2-negative disease; 4.0% with HER2-positive disease; 14.0% with TNBC; and 28.8% of ovarian cancer patients (Table 2). BRCA1/2 PV carriers were more likely than other patients to receive bilateral mastectomy and debulking surgery (Table 3).

Breast Cancer-specific Mortality

Multivariable model results are shown in Table 4. Among TNBC patients, those with BRCA1 PVs had lower cancer-specific mortality (hazard ratio [HR] = 0.49, 95% confidence interval [CI] = 0.35 to 0.69) vs those testing negative, as did BRCA2 PV carriers (HR = 0.60, 95% CI = 0.41 to 0.89). Equivalent cancer-specific mortality was observed among TNBC patients with other gene PVs (HR = 0.65, 95% CI = 0.37 to 1.13) vs those testing negative. Among patients with HER2-positive or ER/PR-positive, HER2-negative subtypes, there was no association of cancer-specific mortality with genetic test results. Other factors associated with increased cancer-specific mortality included higher stage, surgical procedure other than breast-conserving surgery, higher neighborhood poverty, and Black race, whereas Asian and Pacific Islander race and ethnicity was associated with lower cancer-specific mortality.

Ovarian Cancer-specific Mortality

Multivariable model results are shown in Table 5. Compared with patients testing negative, lower mortality was seen in patients with PVs in BRCA2 (HR = 0.35, 95% CI = 0.25 to 0.49) and other tested genes (HR = 0.47, 95% CI = 0.32 to 0.69) but not with BRCA1 PVs. Other factors associated with higher cancer-specific mortality included older age, Native American and Alaskan Native race and ethnicity, higher stage, and no surgery.

Sensitivity Analysis

All covariates satisfied proportional hazards assumptions except stage (for TNBC and ER/PR-positive, HER2-negative disease), surgery (for TNBC and HER2-positive disease), endocrine therapy (for HER2-positive and ER/PR-positive, HER2-negative disease), grade (for ER/PR-positive, HER2-negative disease), radiation (for TNBC), marital status (for ovarian cancer), and histology (for ovarian cancer). Models were created with interactions between these variables and time: across all models and variables, the coefficients for the primary covariate of test result did not change in statistical significance or size. Models including interactions between test result and all other covariates found no statistically significant effects.

In models weighted for probability of genetic testing (Supplementary Tables 2 and 3, available online), PVs in genes other than BRCA1/2 were associated with lower breast cancer-specific mortality among patients with ER/PR-positive, HER2-negative disease (HR = 0.47, 95% CI = 0.30 to 0.75; Supplementary Table 2, available online). A Fine and Gray analysis accounting for competing mortality found similar results (ER/PR-positive, HER2-negative: HR = 0.63, 95% CI = 0.42 to 0.95; Supplementary Table 4, available online). Results for ovarian cancer patients did not change statistically significantly (Supplementary Tables 3 and 4, available online).

Sensitivity analyses of excluding patients tested after starting treatment (Supplementary Table 5, available online), overall mortality (Supplementary Table 6, available online), and stratifying by stage (Supplementary Table 7, available online) found no substantial difference in results. No PV was associated with higher cancer-specific or overall mortality in any analysis.

processing....