Predicted Chemotherapy Benefit for Breast Cancer Patients With Germline Pathogenic Variants in Cancer Susceptibility Genes

Abstract Breast cancer patients increasingly undergo genetic testing. To examine chemotherapy indications for germline pathogenic variant (PV) carriers, we linked results of germline testing to Georgia and California Surveillance, Epidemiology, and End Results registry records, including 21-gene recurrence score (RS) results, for breast cancer patients diagnosed in 2013-2017. All statistical tests were 2-sided. Patients (N=37 349) had RS results of whom 714 had BRCA1, BRCA2, CHEK2, ATM, PALB2, or Lynch syndrome (MLH1, MSH2, MSH6, PMS2) PVs. For women aged 50 years or older at breast cancer diagnosis, RS often exceeded the chemotherapy benefit threshold (≥26) with BRCA1 (71.7% vs 14.4% with none; P <.001), PALB2 (37.1%; P = .001), and BRCA2 (44.3%; P < .001) PVs. Results were similar for women diagnosed at younger than 50 years of age. PVs in BRCA1, but not BRCA2, PALB2, ATM, CHEK2, or Lynch syndrome genes, were associated with elevated RS on multivariable analysis (P < .001). Results may inform RS testing decisions in breast cancer patients with PVs.

Germline multiple-gene sequencing is increasingly common, and some guidelines endorse testing every breast cancer patient (1,2). Guidelines do not advise using germline results in treatment decisions for early stage breast cancer (3). However, a recent study suggests that some clinicians modify treatment based on germline results (4). This emphasizes the need for better understanding of chemotherapy's benefit for patients with germline pathogenic variants (PVs).
Tumor gene expression profiling characterizes breast cancer prognosis and chemotherapy response, as demonstrated by a randomized trial of the 21-gene recurrence score (RS) in early stage estrogen receptor and/or progesterone receptor (ER/PR)positive, HER2-negative disease (5). We sought to determine chemotherapy indications among PV carriers by linking the results of germline and RS testing to records of female breast cancer patients reported to the statewide, population-based Surveillance, Epidemiology and End Results (SEER) registries of Georgia and California.
We previously published details of the Georgia-California SEER Genetic Testing Linkage Initiative (1,4,6). Briefly, all women diagnosed with breast cancer at age 20 years or older from 2014 to 2017 and reported to the California and Georgia registries by August 30, 2019, were linked with results of germline testing from 4 participating laboratories (Ambry Genetics, Aliso Viejo, CA; GeneDx, Gaithersburg, MD; Invitae, San Francisco, CA; Myriad Genetics, Salt Lake City, UT). Laboratories reported results as provided to the ordering clinician: pathogenic or likely pathogenic (combined for analysis as PV), variant of uncertain significance, and benign or likely benign (combined for analysis as negative) but omitted specific sequence changes because of privacy concerns. SEER provided all other variables including RS, which was also obtained by linking SEER data with the testing laboratory (Exact Sciences, Redwood City, CA) (7).
Patients were included if eligible for RS testing, defined as having stage I-II tumors (T stage 1-2, N stage 0-1) that were ER/ PR-positive and HER2-negative (8). We limited the analysis of germline testing to breast cancer-associated genes in which PVs were detected frequently enough to provide sufficient cases for analysis (n 20): ATM, BRCA1/2, CHEK2, and PALB2. We included the Lynch syndrome (LS)-associated genes as a single category (MLH1, MSH2, MSH6, and PMS2), because they are commonly tested on multiple-gene panels, but their breast cancer association is uncertain (1,3). Women with PVs in other genes were excluded (n ¼ 459).
Two-sided P values were calculated using 1-way Analysis of Variance (ANOVA) tests. Multivariable modeling was used to adjust for potential confounders of the association between germline results and RS, including race and ethnicity, age, tumor size, grade, and lymph node involvement. We examined the proportions of patients aged younger than 50 years with RS of 16 or higher and aged 50 years or older with RS of 26 or higher, the agespecific thresholds for chemotherapy benefit in the TAILORx trial, according to germline results (5). All statistical tests were 2-sided, and a P value of less than .05 was considered statistically significant.
A total of 199 201 women were diagnosed with breast cancer in California and Georgia during 2014-2017. Among these, 108 058 were RS eligible of whom 37 349 (34.5%) had RS results; of these, 11 257 (30.1%) were also linked to germline testing results. A total of 714 women had a PV in ATM, BRCA1/2, CHEK2, PALB2, or LS genes.
Among women diagnosed at age 50 years or older, mean RS was highest among BRCA1 PV carriers (mean RS ¼ 26.7, 95% confidence interval [ Table 1). The association between RS and germline results did not vary by age. On multivariable modeling, PVs in BRCA1 (P < .001) were associated with higher RS. Tumor grade was the most important confounder in the multivariable model, with grade 3 being statistically significantly associated with higher RS and with PVs in BRCA1, BRCA2, and PALB2. The proportion of patients with RS of 26 or higher followed a similar pattern: for carriers of PVs in BRCA1, it was 71.7%, in PALB2 37.1%, and in BRCA2 44.3% compared with 14.4% among patients testing negative (P .001 for each comparison Figure 1). Results followed a similar pattern for women aged younger than 50 years with a threshold RS of 16 or higher (Table 1).
We integrated genetic testing results with data from the population-based SEER registries of Georgia and California to characterize predicted chemotherapy response among germline PV carriers. Our results are generally consistent with prior studies that reported higher mean RS among BRCA1/2 PV carriers; however, we found lower mean RS among BRCA2 PV carriers than in prior studies. This difference may reflect the concentration of patients with worse tumor prognostic features among the tertiary referral samples used in prior studies, in contrast to our population-based sample (9)(10)(11)(12). One previous study reported higher RS and PVs in BRCA1/2, but not in other breast cancer-associated genes combined; however, small numbers limited its statistical power (13). With more ATM, CHEK2, and PALB2 PV carriers, we could analyze these genes separately and detected a trend toward higher RS among PALB2 PV carriers, second in magnitude only to BRCA1. These results build on studies of PALB2 epidemiology that suggest similar cancer spectrum and risk as with BRCA2 PVs (14), perhaps reflecting similar tumor biology and chemotherapy response. The proportion of RS-eligible tumors varies substantially by affected gene, with more ER/PR-negative tumors among women with BRCA1 and PALB2 PVs. Our results suggest that breast cancers among carriers of PVs in ATM and CHEK2, which are relatively common and more typically ER/PR-positive, may behave similarly to those of women testing negative for germline PVs. We found no difference by age in mean RS. However, the TAILORx trial demonstrated that the threshold RS value for chemotherapy benefit varies by age. For patients diagnosed at age younger than 50 years of age, a threshold RS value of 16 or higher should be considered (5,15). Our findings suggest that most RS-eligible BRCA1/2 and PALB2 PV carriers diagnosed with breast cancer at younger than 50 years of age (RS 16: BRCA1 ¼ 89.4%, BRCA2 ¼ 85.3%, PALB2 ¼ 73.4%), and many at age 50 years and older (RS 26: BRCA1 ¼ 71.7%, BRCA2 ¼ 44.3%, PALB2 ¼ 37.1%), have results indicating benefit from chemotherapy. However, it is important to note that the prognostic and predictive value of RS has not been studied among PV carriers, which is a key question for future research.
This study has limitations. There were few carriers of PVs in genes other than ATM, BRCA1/2, CHEK2, PALB2, and the LS genes. We have not yet evaluated cancer recurrence or mortality in association with germline or RS testing results. As in any clinically tested, real-world sample, there is the inherent bias that clinicians selected patients into testing. Although patients were from 2 large states, the cohort may not represent the entire US population. These limitations are balanced by considerable strengths: notably, this is the largest, most racially and ethnically diverse and population-based sample in which this question has been addressed, with detailed germline and RS results obtained from testing laboratories.
BRCA1 PV carriers diagnosed with breast cancer at any age are statistically significantly more likely to have RS that indicates benefit from chemotherapy; those with PALB2 PVs also show a trend toward higher RS. These results may inform RS testing and chemotherapy decisions among breast cancer patients who carry cancer susceptibility gene PVs.

Funding
Research reported in this publication was supported by the National Cancer Institute (NCI) of the National Institutes of Health under award number R01 CA225697 to Stanford University. The collection of cancer incidence data in Georgia was supported by contract HHSN261201800003I, Task Order HHSN26100001 from the NCI and cooperative agreement 5NU58DP006352-03-00 from the Centers for Disease Control and Prevention (CDC). The collection of cancer incidence data used in this study was supported by the

Notes
Role of the funders: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclosures: Allison W. Kurian, MD, MSc, reports research funding to her institution for an unrelated project from Myriad Genetics. The other authors declare no conflicts of interest.

Data Availability
Paul Abrahamse and Allison Kurian are independent of any commercial funder, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.