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

Discussion

We studied short-term cancer-specific mortality associated with germline genetic testing results among 22 495 breast cancer patients and 4320 ovarian cancer patients treated with chemotherapy in the population-based setting of 4 SEER registries comprising the statewide populations of California and Georgia. Consistent with our hypothesis, we found that compared with those testing negative for PVs, TNBC patients with PVs in BRCA1/2 and ovarian cancer patients with PVs in BRCA2 or other genes (notably BRIP1, CHEK2, RAD51C, and ATM) had lower cancer-specific mortality at 41 months' median follow-up time. To our knowledge, this is the first population-based study to report lower short-term cancer-specific mortality associated with germline PVs in genes other than BRCA1/2. These findings can inform discussions about prognosis between patients and their oncologists.

Our findings add to an extensive literature on outcomes of BRCA1/2-associated breast cancer.[13,14] There has been little consensus, with some studies reporting higher[14–16,27–29] and others lower cancer-specific mortality[30] among BRCA1/2 PV carriers vs noncarriers, whereas others found no difference.[17,19,22,23,31] Focusing on higher-risk subtypes and chemotherapy recipients has offered more clarity: some studies found lower cancer-specific mortality among BRCA1/2 PV carriers who had TNBC and/or received chemotherapy.[9,10,18,25,32] These results are consistent with clinical trials such as GeparSixto and INFORM, which showed that BRCA1/2 PV–associated cancers respond well to various chemotherapy regimens.[33,34] Our finding of lower BRCA1/2-associated cancer-specific mortality with TNBC only may reflect its more aggressive biology than other subtypes, which might confer an earlier and greater benefit from chemotherapy. BRCA1/2 PV carriers might also have received more intensive chemotherapy—with more agents and/or of longer duration—than other patients. We found that BRCA1/2 PV carriers more often received bilateral mastectomy (and debulking surgery for ovarian cancer); however, multivariable modeling controlled for surgical procedure, so this variation does not account for the observed results. A limitation is that SEER does not report risk-reducing salpingo-oophorectomy, which may have contributed to the lower breast cancer-specific mortality observed in BRCA1/2 PV carriers. A further consideration is the short median follow-up of this study (41 months), because TNBC is prone to early recurrence and mortality.[35] Longer follow-up is needed to determine whether lower cancer-specific mortality with BRCA1/2 PVs emerges for late-recurring subtypes, such as ER/PR-positive, HER2-negative disease.[36] Future studies should also include cases diagnosed more recently, because poly (ADP-ribose) polymerase [PARP] inhibitors were not approved for BRCA1/2-associated metastatic disease until 2018, and thus their effects are unlikely to be substantial in this 2013–2017 diagnosis cohort.[37–39]

We found no evidence of higher short-term cancer-specific mortality among TNBC patients with PVs in genes other than BRCA1/2. Furthermore, in 2 sensitivity analyses, we observed statistically significantly lower cancer-specific mortality with other gene PVs (most commonly CHEK2, PALB2, and ATM) among patients with ER/PR-positive, HER2-negative breast cancer. A study of CHEK2 PV carriers including all breast cancer subtypes found equivalent cancer-specific mortality to noncarriers in the first 6 years postdiagnosis, but twofold higher mortality afterward.[40] Two hospital-based series found that overall mortality was higher among PALB2 PV carriers than noncarriers.[16,41] The PATTERN trial of adjuvant chemotherapy for TNBC found no difference in disease-free survival between non-BRCA1/2 PV carriers and noncarriers.[23] Although longer follow-up is essential, our early findings suggest that carriers of PVs in genes other than BRCA1/2 are not more likely (and may be less likely) than noncarriers to die of their breast cancer.

In contrast to breast cancer, prior ovarian cancer studies including some in population-based settings more consistently reported lower short-term cancer-specific mortality associated with BRCA1/2 PVs.[11,12,14,42,43] However, longer-term studies reported attenuation or reversal of this advantage.[44,45] Our finding of lower short-term ovarian cancer-specific mortality with BRCA2 PVs is largely consistent with prior studies.[42] Our results may also reflect an emerging contribution from PARP inhibitors, which were approved in 2014.[46] However, longer follow-up is necessary.

We found lower short-term ovarian cancer-specific mortality with PVs in genes other than BRCA1/2 (notably BRIP1, RAD51C, CHEK2, and ATM), consistent with results of the Gynecologic Oncology Group 218 trial.[12] To our knowledge, this has not previously been reported in community practice. High-grade, serous ovarian cancer often has a homologous recombination-deficient phenotype conferring sensitivity to chemotherapy, particularly platinum agents, and PARP inhibitors.[47] However, there is greater treatment responsiveness among ovarian cancer patients who do vs who do not carry BRCA1/2 PVs.[11,48] This enhanced response might also pertain to carriers of PVs in other genes and contribute to their lower cancer-specific mortality.

Our study has limitations. The relatively few deaths and few PVs in each gene limited statistical power to analyze the association of cancer-specific mortality with specific genes; larger subsequent analyses with longer follow-up may achieve this. Although there were many patients from most racial and ethnic groups, there were fewer Native American and Alaskan Natives. Additionally, we lack information on specific chemotherapy agents received. However, because breast cancer trials such as INFORM and GeparSixto found that BRCA1/2 PV carriers responded well regardless of specific drugs used,[33,34] this limitation seems unlikely to affect our conclusions. Although we lack information on PARP inhibitor use, the study period overlaps with their Food and Drug Administration approval, and thus their impact was probably limited, especially for breast cancer.[37,38] Results for patients who underwent clinical genetic testing, potentially because of family cancer history, may not be generalizable to patients who did not; however, we found that a sensitivity analysis accounting for selection into testing offered no evidence of higher cancer-specific mortality, and additional evidence of lower cancer-specific mortality, among PV carriers. We lack data on other prognostic factors including prediagnostic screening, comorbidities, extent of surgical debulking, and metastatic recurrence. As noted previously, the median follow-up time of 41 months is short, yet it encompasses a period that matters to patients as they plan for their immediate future. The study's limitations are balanced by considerable strengths, including a large, diverse, contemporary population-based sample; genetic results obtained directly from testing laboratories; and uniform ascertainment of treatment and mortality data by SEER registries that have near-total capture of all cancers statewide, minimizing selection bias.

This study's results have substantial implications for patients newly diagnosed with breast cancer or ovarian cancer. We found that no PV, whether in BRCA1/2 or another gene, was associated with any increase in short-term cancer-specific or overall mortality among patients treated with chemotherapy. This may help reassure cancer patients that testing positive for a PV does not mean they are more likely to die within the first several years following their cancer diagnosis.

processing....