Key Points
Question
What breast tumor characteristics are associated with rare pathogenic protein truncating or missense variants in breast cancer susceptibility genes?
Findings
In this case-control study involving 46 387 control participants and 42 680 women with a diagnosis of breast cancer, pathology features (eg, tumor subtype, morphology, size, TNM stage, and lymph node involvement) associated with rare germline (likely) pathogenic variants in 9 different breast cancer susceptibility genes were studied. Substantial differences in tumor subtype distribution by gene were found.
Meaning
The results of this study suggest that tumor subtypes differ by gene; these findings can potentially inform guidelines for gene panel testing, risk prediction in unaffected individuals, variant classification, and understanding of breast cancer etiology.
Abstract
Importance
Rare germline genetic variants in several genes are associated with increased breast cancer (BC) risk, but their precise contributions to different disease subtypes are unclear. This information is relevant to guidelines for gene panel testing and risk prediction.
Objective
To characterize tumors associated with BC susceptibility genes in large-scale population- or hospital-based studies.
Design, Setting, and Participants
The multicenter, international case-control analysis of the BRIDGES study included 42 680 patients and 46 387 control participants, comprising women aged 18 to 79 years who were sampled independently of family history from 38 studies. Studies were conducted between 1991 and 2016. Sequencing and analysis took place between 2016 and 2021.
Exposures
Protein-truncating variants and likely pathogenic missense variants in ATM, BARD1, BRCA1, BRCA2, CHEK2, PALB2, RAD51C, RAD51D, and TP53.
Main Outcomes and Measures
The intrinsic-like BC subtypes as defined by estrogen receptor, progesterone receptor, and ERBB2 (formerly known as HER2) status, and tumor grade; morphology; size; stage; lymph node involvement; subtype-specific odds ratios (ORs) for carrying protein-truncating variants and pathogenic missense variants in the 9 BC susceptibility genes.
Results
The mean (SD) ages at interview (control participants) and diagnosis (cases) were 55.1 (11.9) and 55.8 (10.6) years, respectively; all participants were of European or East Asian ethnicity. There was substantial heterogeneity in the distribution of intrinsic subtypes by gene. RAD51C, RAD51D, and BARD1 variants were associated mainly with triple-negative disease (OR, 6.19 [95% CI, 3.17-12.12]; OR, 6.19 [95% CI, 2.99-12.79]; and OR, 10.05 [95% CI, 5.27-19.19], respectively). CHEK2 variants were associated with all subtypes (with ORs ranging from 2.21-3.17) except for triple-negative disease. For ATM variants, the association was strongest for the hormone receptor (HR)+ERBB2− high-grade subtype (OR, 4.99; 95% CI, 3.68-6.76). BRCA1 was associated with increased risk of all subtypes, but the ORs varied widely, being highest for triple-negative disease (OR, 55.32; 95% CI, 40.51-75.55). BRCA2 and PALB2 variants were also associated with triple-negative disease. TP53 variants were most strongly associated with HR+ERBB2+ and HR–ERBB2+ subtypes. Tumors occurring in pathogenic variant carriers were of higher grade. For most genes and subtypes, a decline in ORs was observed with increasing age. Together, the 9 genes were associated with 27.3% of all triple-negative tumors in women 40 years or younger.
Conclusions and Relevance
The results of this case-control study suggest that variants in the 9 BC risk genes differ substantially in their associated pathology but are generally associated with triple-negative and/or high-grade disease. Knowing the age and tumor subtype distributions associated with individual BC genes can potentially aid guidelines for gene panel testing, risk prediction, and variant classification and guide targeted screening strategies.
This case-control study examines tumors associated with breast cancer susceptibility genes in large-scale population- or hospital-based studies.
Introduction
Breast cancer (BC) is a heterogeneous disease; different subtypes are associated with distinct biology, prognosis, and potential for therapy.1,2,3 There is evidence that inherited genetic predisposition contributes to this heterogeneity.4,5 However, data for detailed analysis of tumor pathologies that are associated with most BC susceptibility genes have been limited, particularly in population-based studies. Recent results from 2 large-scale sequencing studies, BRIDGES6 and CARRIERS,7 found evidence of an association with BC risk for germline protein-truncating variants (PTVs) and/or rare missense variants (MSVs) in 9 genes: ATM, BARD1, BRCA1, BRCA2, CHEK2, PALB2, RAD51C, RAD51D, and TP53. Women carrying variants in these genes may be offered enhanced screening, including by magnetic resonance imaging, risk-reducing surgery, chemoprevention, and genetic counselling; knowledge of germline gene variants also affects treatment.8 Intrinsic BC subtypes have been defined on the basis of patterns of gene expression; these include luminal-A, which defines a subset of hormone receptor–positive tumors that are associated with a good 5-year prognosis, and luminal-B, ERBB2-enriched and basal tumors with poorer prognosis.2,9 Gene expression data are not routinely available in diagnostic laboratories, but large-scale epidemiological studies can use subtypes based on immunochemical markers to define intrinsic-like surrogates that are broadly associated with the molecular subtypes.10,11 In this article, we use data from BRIDGES to assess associations between variants in these genes and pathological features of nonmetastasized breast tumors relevant to prognosis and/or distinct therapeutic options. We further quantify the contribution of rare BC susceptibility genes to the development of distinct BC subtypes in women of different ages.
Methods
Studies and Inclusion Criteria
The BRIDGES study included women with BC and unaffected control participants who were participating in the Breast Cancer Association Consortium (https://bcac.ccge.medschl.cam.ac.uk/; eTable 1 in Supplement 1). The analyses presented in this article are based on cases from the subset of population-based or hospital-based studies that were sampled independently of family history, together with population-matched control participants (38 studies). Women aged between 18 and 79 years were included. Pathology information from the first primary invasive BC was considered. Cases in which the index tumor was the second tumor and patients with metastases at initial diagnosis were excluded.12 All studies were approved by the relevant ethical review boards, and participants provided written informed consent.
Laboratory Methods, Variant Calling, and Classification
We focused on 9 genes with evidence of an association with BC.6 We considered PTVs for all 9 genes, and rare (carrier frequency <0.1%) MSVs in BRCA1, BRCA2 and TP53 that were likely pathogenic according to adaptations of the American College of Medical Genetics guidelines.6 Approximately 80% of CHEK2 PTVs were c.1100delC. The TP53 PTV and MSV carriers were considered together. Carriers of BRCA1 and BRCA2 PTVs were excluded from the analyses of other genes. Carriers of PTVs in BRCA1 and BRCA2 and women who harbored a pathogenic variant in more than 1 non-BRCA gene were also excluded. Noncarriers were defined as women without PTVs or MSVs in any of the genes. Further details are provided in the eMethods in Supplement 1.
Tumor Pathology
Pathology information was based on histology and immunohistochemistry results from medical records, rescored whole slides, or tumor microarrays that were curated in the Breast Cancer Association Consortium database, version 12.13,14 Markers included estrogen receptor (ER), progesterone receptor (PR), and erb-b2 receptor tyrosine kinase 2 (ERBB2, formerly known as HER2) status, which was denoted as positive or negative; histological grade (grades 1, 2, and 3); morphology; tumor size (<2, 2-5, or >5 cm); lymph node involvement (yes/no); and TNM stage (I, II, and III). For the purposes of this analysis, we defined 5 clinically relevant intrinsic subtypes based on available immunohistochemistry and grade: HR+ERBB2− low (/intermediate) grade, HR+ERBB2+, and HR+ERBB2− high grade, HR−ERBB2+ and triple negative (TN). Grades 1 and 2 were considered low-grade and grade 3 high-grade disease (eTable 2 in Supplement 1).11,12,15,16
Statistical Analysis
Analyses were based on estimating the odds ratios (ORs) associated with carrying any PTV (or pathogenic MSV) in each gene. First, complete-case analyses based on all available data were conducted. Case-control analyses were used to estimate the OR for developing a tumor of a particular subtype according to single markers and case-only analyses to evaluate the evidence for differences by subtype. Logistic regression was used for binary characteristics and multinomial logistic regression for multicategory tumor characteristics. For multicategory outcomes, a model in which the log(OR) varied linearly with the outcome level was also fitted. Analyses were adjusted for age (defined as age at diagnosis for patients and age at interview for control participants) and country of origin of the study.
To evaluate heterogeneity of risk by intrinsic tumor subtypes, we first imputed missing pathology variables using Multiple Imputation by Chained Equations. Intrinsic subtypes were constructed for each of 100 imputed data sets, and the results of multinomial regression for each imputed data set were pooled. We also compared these data with results obtained after imputing tumor pathology using an expectation-maximization (EM) algorithm (eMethods in Supplement 1).17 We investigated interactions with age for each gene according to tumor subtype by including an age x variant product term in the model and also estimated the proportion of BC cases, by age-group and intrinsic subtype, for pathogenic variants in each gene.
Associations between (likely) pathogenic variant carrier status and tumor size and lymph node status were evaluated. Analyses were also conducted that included size, lymph node status, and intrinsic subtype in the same model and PR status and the HR-positive subtypes in the same model.
Gene-specific cumulative risks for each subtype were calculated by combining age-specific OR estimates with 2016 UK population incidence rates as a baseline and accounting for competing risk of not developing BC of a different subtype. Age-specific and gene-specific subtype proportions for tumor subtypes included in the risk prediction algorithm BOADICEA18 were also calculated (eMethods in Supplement 1).
Analyses were conducted using RStudio, version 1.2.5033 (RStudio); Stata, version 14.2 (StataCorp); and GFortran. Statistical significance was set at P < .05.
Results
Study Characteristics
The study comprised 46 387 control participants and 42 680 women with a diagnosis of BC from 22 countries, with mean (SD) ages at interview and diagnosis of 55.1 (11.9) and 55.8 (10.6) years, respectively (eTable 3 in Supplement 1). Numbers of variant carriers by gene are shown in eTable 4 in Supplement 1 and patterns of missingness in pathology data in eTable 5 in Supplement 1 and eTable 1 in Supplement 2; for ER status, 18%; grade, 18%; PR status, 32%; and ERBB2 status, 43% of data were missing. There was no association between missingness and genotype.
Single marker analyses were based on complete data (eFigure 1 in Supplement 1 and eTables 2 and 3 in Supplement 2). The remaining analyses were carried out following imputation of missing data.
Distribution of Intrinsic Tumor Subtypes and Age Trends
The PTVs in all 9 BC genes showed evidence of variation in the ORs among the 5 intrinsic subtypes (Figure 1, Figure 2; eFigures 2-5 and eTables 6 and 7 in Supplement 1; eTable 4 in Supplement 2). For BRCA1 PTV carriers, the OR was highest (OR, 55.32; 95% CI, 40.51-75.55) for TN disease, much lower for HR+ERBB2− low-grade disease (OR, 3.26; 95% CI, 2.21-4.80) and HR+ERBB2+ disease (OR, 2.27; 95% CI, 1.16-4.45), and intermediate for HR+ERBB2− high-grade and HR−ERBB2+ disease (OR, 13.5 [95% CI, 9.16-19.90] and OR, 9.85 [95% CI, 5.71-17.02], respectively). Associations between BRCA2 PTVs and intrinsic subtypes were more homogeneous across subtypes, with higher ORs associated with HR+ERBB2− high-grade disease (OR, 11.53; 95% CI, 8.92-14.90) and TN tumors (OR, 10.07; 95% CI, 7.61-13.32). For ATM, the association was strongest for HR+ERBB2− high-grade tumors (OR, 4.99; 95% CI, 3.68-6.76). CHEK2 PTVs were associated with similar ORs with all subtypes except TN, for which there was no evidence of association. PALB2 PTVs were associated with all subtypes, but with higher ORs for HR+ERBB2− high-grade (OR, 9.43; 95% CI, 6.24-14.25) and TN disease (OR, 8.05; 95% CI, 5.17-12.53).
Figure 1. Association Odds Ratios (ORs) for Protein-Truncating Variant Carrier Status in Breast Cancer Susceptibility Genes BRCA1, BRCA2, ATM, and CHEK2 and Intrinsic Subtypes of Breast Cancer.
Multiple Imputation by Chained Equations imputation was conducted as described in the Methods and intrinsic subtypes constructed for each imputed data set. Multinomial logistic regression was conducted with intrinsic subtypes as the outcome variable, adjusting by age at diagnosis/interview and country, and the results of these analyses were pooled. These results are also shown in eTable 4 in Supplement 2. HR indicates hormone receptor.
Figure 2. Association Odds Ratios (ORs) for Protein-Truncating Variant Carrier Status in Breast Cancer Susceptibility Genes PALB2, BARD1, RAD51C, and RAD51D and Intrinsic Subtypes of Breast Cancer.
Multiple Imputation by Chained Equations imputation was conducted as described in the Methods and intrinsic subtypes constructed for each imputed data set. Multinomial logistic regression was conducted with intrinsic subtypes as the outcome variable, adjusting by age at diagnosis/interview and country, and the results of these analyses were pooled. These results are also shown in eTable 4 in Supplement 2. HR indicates hormone receptor; NA, not applicable.
The PTVs in RAD51C, RAD51D, and BARD1 were most strongly associated with TN disease (OR, 6.19 [95% CI, 3.17-12.12]; OR, 6.19 [95% CI, 2.99-12.79]; and OR, 10.05 [95% CI, 5.27-19.19], respectively). RAD51D PTVs were also associated with HR+ERBB2− high-grade tumors.
Similar to PTVs, carriers of BRCA1 MSV were strongly enriched for TN disease (eTable 4 in Supplement 2 and eFigure 6 in Supplement 1). The ORs were higher than those for PTVs for all of the intrinsic subtypes, although these differences were not statistically significant. For BRCA2, the ORs for MSVs and PTVs were similar. TP53 variants were associated with HR−ERBB2+ and HR+ERBB2+ but not TN disease.
A decline in the ORs with increasing age was observed for BRCA1 and BRCA2; this trend was similar for all subtypes (OR, 0.96 per year for both genes; P = 7.05 × 10−7 and 3.14 × 10−11 for BRCA1 [95% CI, 0.94-0.98] and BRCA2 [95% CI, 0.95-0.97], respectively; eTable 4 in Supplement 2). The ORs also declined with age for CHEK2, but the trend was much weaker. There was no evidence of a decline in the ORs for ATM, BARD1, RAD51C, or RAD51D, but the confidence limits for the last 3 genes were wide.
We further stratified HR-positive subtypes by PR expression to determine whether carrier status was associated with PR. For BRCA1, the ORs were lowest for ER+, PR+ tumors compared with other categories (eTable 8 in Supplement 1). Consistent with this observation, BRCA1 PTV carriers were more likely to be PR negative, even after adjusting for intrinsic subtype. There was also some weak evidence for BRCA2 PTVs and PR negativity, but no evidence for the other genes.
Association Between Breast Cancer Susceptibility Genes and Other Prognostic Factors
The PTVs in BRCA2, CHEK2, and PALB2 were associated with larger tumor size, lymph node involvement, and higher stage at diagnosis (eFigure 1 in Supplement 1). The individual associations with larger tumor size and lymph node involvement remained significant after adjusting for intrinsic subtypes. The association between PTVs in all 9 genes with intrinsic subtypes remained similar after including size and lymph node status in the model (eTable 5 in Supplement 2).
For each gene, most BCs were carcinoma no special type (ductal carcinoma); in aggregate, 71% of tumors in carriers and 68% in noncarriers were ductal carcinoma. BRCA1 tumors were less likely to be lobular than ductal (OR, 0.40; 95% CI, 0.25-0.63) but more likely to be medullary than nonmedullary (OR, 5.24; 95% CI, 3.34-8.22) (eTables 2 and 3 in Supplement 2). TP53 tumors were more likely to be mixed lobular and ductal than ductal carcinoma (OR, 7.01; 95% CI, 3.04-16.17; P = 5 × 10−6). Otherwise, tumors associated with variations in the other BC genes were not enriched for any particular morphology.
Prevalence of Pathogenic Variants According to Subtypes and Age
We assessed the association between rare variants in BC susceptibility genes with the burden of disease in women of different ages (eFigures 7-12 and eTable 7 in Supplement 1). Together, the 9 genes were associated with 14.4% of all tumors in women 40 years or younger but less than 4% in women older than 60 years. Among younger women, the prevalence of variants combined was higher among women with TN and HR+ERBB2− high-grade tumors than those with other subtypes. The highest prevalence (27.3%) was among women 40 years or younger with TN tumors, mainly driven by BRCA1 (eFigure 7 in Supplement 1). The combined prevalence of pathogenic variants was close to or exceeded 10% for all subtypes in women younger than 40 years and for TN and HR+ERBB2− high-grade disease in women aged 40 to 59 years. Although TP53-related tumors comprised only a small proportion of ERBB2-positive disease, approximately 70% of TP53 tumors among women 40 years or younger were ERBB2-positive.
Age-Specific Cumulative Risk of Developing Intrinsic BC Tumor Subtypes
Estimated cumulative risks according to intrinsic subtypes are shown in Figure 3 and Figure 4. The estimated risk for TN tumors was highest for BRCA1 (40% by age 80 years), and 7% to 12% for BRCA2, BARD1, PALB2, RAD51C, and RAD51D. In contrast, the highest risks for HR+ERBB2− low-grade disease were associated with BRCA2 (22%) followed by PALB2 and CHEK2.
Figure 3. Estimates of Cumulative Risks of Breast Cancer by Age and Hormone Receptor (HR)–Positive Subtype for Protein-Truncating Variants in 8 Breast Cancer Susceptibility Genes.
Age-, gene-, and subtype-specific cumulative risks were calculated as described in the Methods and eMethods in Supplement 1. Baseline incidence rates were derived from UK breast cancer incidence rates for 2016 (https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/breast-cancer/incidence-invasive).
Figure 4. Estimates of Cumulative Risks of Breast Cancer by Age and Hormone Receptor (HR)–Negative Subtype for Protein-Truncating Variants in 8 Breast Cancer Susceptibility Genes.
Age-, gene-, and subtype-specific cumulative risks were calculated as described in the Methods and eMethods in Supplement 1. Baseline incidence rates were derived from UK breast cancer incidence rates for 2016 (https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/breast-cancer/incidence-invasive).
Discussion
This case-control study evaluated the pathology of BCs developing in carriers of PTVs and/or rare MSVs in 9 BC susceptibility genes: ATM, BARD1, BRCA1, BRCA2, CHEK2, PALB2, RAD51C, RAD51D, and TP53 in a large multicenter collaborative study comprising population-based and hospital-based studies. The pattern of intrinsic subtypes and markers of tumor aggressiveness differed between carriers of variants in individual BC susceptibility genes and noncarriers. As expected,19,20 BC in BRCA1 carriers were strongly enriched for TN tumors, with TN disease representing approximately 60% of all tumors and approximately 70% of tumors in women 40 years and younger. However, the risks for all other subtypes were also increased (ORs, 2.27-13.5). For BRCA2, the highest ORs were for HR+ERBB2− high-grade and TN disease, which was consistent with the strong association with ERBB2 negativity.21,22 The most common subtype (43% of cases) was HR+ERBB2− low (/intermediate)–grade disease, but a clear excess of TN disease (approximately 18% of tumors) was apparent, even at younger ages. Subtype-specific associations for BRCA1 and BRCA2 MSVs were similar to those for PTVs in the corresponding genes.
Although the ORs were lower, the pattern of intrinsic subtypes for PALB2 carriers was very similar to that for BRCA2 carriers (Figures 1 and 2), with variation in both genes being associated with ERBB2 negativity and TN disease.23 This similarity may reflect the closely associated functions of PALB2 and BRCA2 in the DNA damage response.24
Conversely, the profile of intrinsic subtypes associated with BARD1 carriers was similar to BRCA1 carriers, with an excess of TN tumors (40% of cases), albeit the overall risk was much lower. Consistent with this observation, Bard1 and Brca1 knockout mice have similar phenotypes.25 BARD1 and BRCA1 proteins form a stable complex, the heterodimer coordinating a range of cellular pathways to maintain genomic stability. Although BRCA1 requires BARD1 for stability and tumor suppressor functions, BARD1 also plays distinct roles in cell cycle progression.25,26
Carriers of PTVs in ATM and CHEK2 were more strongly associated with ER-positive disease, but this study highlights some differences. For ATM, the association was particularly strong for HR+ERBB2− high-grade tumors (OR, 4.99; 95% CI, 3.68-6.76), with weaker associations for the other HR-positive subtypes (although HR+ERBB2− low-grade tumors were still the most common). An association with the luminal B subtype has been reported previously in a small data set (n = 28).27 CHEK2 was associated with a similar OR for all the HR-positive subtypes and increased risk of HR−ERBB2+, but not TN disease. ATM plays a central role in the activation of DNA damage response and cell cycle checkpoint control, while CHEK2 is involved downstream of ATM in cell cycle arrest, apoptosis, and DNA repair.28,29
RAD51C and RAD51D are known ovarian cancer susceptibility genes and are more recently associated with BC,6,30 in particular with TN disease.4,31,32 In the present study, ORs for TN disease for PTVs in both genes were approximately 6.0. The subtype distribution of RAD51C is similar to RAD51D, reflecting their closely associated functions. We did observe an excess of HR+ERBB2− high-grade tumors in RAD51D but not RAD51C carriers; however, the numbers of PTV carriers were small and we cannot exclude the subtype distributions being similar.
TP53 tumors were strongly enriched for ERBB2-positive subtypes (46% of cases), which was consistent with earlier studies in patients with or without Li-Fraumeni syndrome33,34 and examination of patients identified by multigene panel testing.35 We also observed an association with mixed ductal and lobular morphology, tumors that comprise distinct but clonally related morphological components.36
Pathogenic PTVs and MSVs in these 9 BC susceptibility genes were disproportionately associated with more aggressive BC, particularly among younger women. Carriers of rare genetic variants in the 9 genes constituted almost a third of women who received a diagnosis at or younger than 40 years of TN disease and approximately 16% of women with HR+ERBB2− high-grade disease. All genes except CHEK2 were more strongly associated with high-grade disease. Across genes, 27% to 72% of tumors were grade 3 (eTable 2 in Supplement 2). Previously studies have suggested that tumors in carriers of rare PTVs are larger23,37,38,39,40 and more likely to be identified as interval rather than screen-detected cancers.37 In the present study, BRCA2-, CHEK2-, and PALB2-associated tumors were larger and more likely to be lymph node positive.
Despite the strong enrichment of TN disease for many of the genes, most carriers will still develop HR+ disease. With the exception of BRCA1, the most common subtype for all genes was HR+ERBB2− low (/intermediate)–grade disease (Figures 3 and 4). However, these absolute risk projections indicate average subtype-specific risks, while individual risk prediction should also consider polygenic modifiers, family history, and lifestyle and reproductive factors, as well as the risk of developing cancers at other sites.41 The age- and subtype-specific risk estimates (eTable 6 in Supplement 2 and eFigure 13 in Supplement 1) may be used to refine BC risk prediction algorithms, such as BOADICEA.41
These results may also inform guidelines for eligibility for gene panel sequencing and BC surveillance in the general population. The combined prevalence of pathogenic variants in any of the 9 genes reached 10% for TN cases in those who received a diagnosis when younger than 60 years and HR+ERBB2− high-grade and HR+ERBB2+ cases in those who received a diagnosis at 40 years or younger (in HR−ERBB2+ cases, the prevalence was 9.4%). These are slight underestimates of the true frequency because some variants deleterious to gene function, notably large gene rearrangements, will have been missed in the targeted sequencing.6
Tumor characteristics can also be used in determining whether variants of uncertain significance are likely to be pathogenic based on the assumption that the tumor characteristics of pathogenic variants of uncertain significance will be similar to known pathogenic variants.42 Therefore, these data should improve the precision of variant classification algorithms and extend them to a larger set of genes.
Strengths and Limitations
The strengths of this study are its large sample size (42 680 cases and 46 387 control participants) and sampling of cases independent of family history, while most earlier investigations have involved women who were ascertained in genetics clinics and selected based on family history, genotype, or pathology. The large sample size allowed us to obtain unbiased estimates of ORs and age interaction effects, while the sampling framework provided results that are particularly relevant as gene panel testing becomes applied at a general population level. Cases and control participants underwent sequencing on the same platform and using a single variant calling algorithm. We analyzed a comprehensive set of variables and their associations. Finally, the results found using 2 different imputation methods, Multiple Imputation by Chained Equations and an EM algorithm were consistent (eTable 9 in Supplement 1).
Despite the large size of this study, the sample size with complete pathology data was still limited for some genes. For example, ERBB2 status was missing for approximately 43% of samples, although missingness is likely to be random with respect to genotype, and imputation methods performed well. There was also minor heterogeneity in definition of stage, grade, and cutoffs for ER, PR, and ERBB2 across studies. The subtypes defined by immunohistochemical markers do not align perfectly with intrinsic subtypes defined by expression profiles, such as PAM-50,43,44 but such data are not available in large-scale epidemiological studies or routine practice. Finally, most participants were of European descent, and larger studies of women from other racial and ethnic groups will be important.
Conclusions
This case-control study suggests that rare variants in BC susceptibility genes display marked heterogeneity with respect to tumor phenotype, but also similarities between genes that are consistent with known biological functions. This present study provides detailed quantification of subtype-specific BC risks; these can potentially improve risk prediction models and breast cancer prevention strategies.
eMethods.
eTable 1. Description of studies included in this analysis
eTable 2. Immunohistochemistry and tumor grade-based surrogates for five intrinsic breast cancer subtypes
eTable 3. Numbers of cases and controls, and age distributions, by country of origin
eTable 4. Numbers of variant carriers by breast cancer susceptibility gene
eTable 5. Cross tabulation of ER, PR, HER2 and Grade data
eTable 6. Distribution of intrinsic tumor subtypes in women of all ages and in different age groups, by breast cancer susceptibility gene.
eTable 7. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women of different age groups at diagnosis
eTable 8. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes refined by PR expression
eTable 9. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes of breast cancer following imputation using an EM algorithm.
eFigure 1. Case-only analysis of phenotypic markers and prognostic features by gene (complete case analysis)
eFigure 2. Frequency histogram of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in 9 genes.
eFigure 3. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged ≤40 years.
eFigure 4. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged 41-60 years.
eFigure 5. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged >60 years.
eFigure 6. Association odds ratios for MSV carrier status in BRCA1, BRCA2, TP53 and intrinsic subtypes of breast cancer.
eFigure 7. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged ≤40 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 8. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged 41-60 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 9. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged >60 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 10. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged ≤40 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 11. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged 41-60 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 12. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged >60 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 13. Smoothed proportions of subtypes used in BOADICEA for PTVs in ATM, BARD1, BRCA1, BRCA2, CHEK2, PALB2, RAD51C and RAD51D.
eTable 1. Missing data by gene/mutation status and age at diagnosis
eTable 2. Number of cases by subtype of breast cancer, as defined by individual tumor markers, and PTV/MSV carrier status for each gene
eTable 3. Odds ratios for association between PTV and MSV carrier status and subtypes of breast cancer as defined by individual tumor markers (complete case analyses)
eTable 4. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes of breast cancer following MICE imputation– analyses with and without an age interaction term
eTable 5. Association between breast cancer genes and intrinsic subtypes - Joint models of intrinsic subtype, tumor size and lymph node involvement
eTable 6. Smoothed proportions of subtypes used in BOADICEA for PTVs in ATM, BARD1, BRCA1, BRCA2 CHEK2, PALB2, RAD51C and RAD51D
References
- 1.Sørlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A. 2001;98(19):10869-10874. doi: 10.1073/pnas.191367098 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sørlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A. 2003;100(14):8418-8423. doi: 10.1073/pnas.0932692100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sørlie T. Molecular portraits of breast cancer: tumour subtypes as distinct disease entities. Eur J Cancer. 2004;40(18):2667-2675. doi: 10.1016/j.ejca.2004.08.021 [DOI] [PubMed] [Google Scholar]
- 4.Hu C, Polley EC, Yadav S, et al. The contribution of germline predisposition gene mutations to clinical subtypes of invasive breast cancer from a clinical genetic testing cohort. J Natl Cancer Inst. 2020;112(12):1231-1241. doi: 10.1093/jnci/djaa023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhang H, Ahearn TU, Lecarpentier J, et al. ; kConFab Investigators; ABCTB Investigators; EMBRACE Study; GEMO Study Collaborators . Genome-wide association study identifies 32 novel breast cancer susceptibility loci from overall and subtype-specific analyses. Nat Genet. 2020;52(6):572-581. doi: 10.1038/s41588-020-0609-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dorling L, Carvalho S, Allen J, et al. ; Breast Cancer Association Consortium . Breast cancer risk genes—association analysis in more than 113,000 women. N Engl J Med. 2021;384(5):428-439. doi: 10.1056/NEJMoa1913948 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hu C, Hart SN, Gnanaolivu R, et al. A population-based study of genes previously implicated in breast cancer. N Engl J Med. 2021;384(5):440-451. doi: 10.1056/NEJMoa2005936 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Tung NM, Boughey JC, Pierce LJ, et al. Management of hereditary breast cancer: American Society of Clinical Oncology, American Society for Radiation Oncology, and Society of Surgical Oncology guideline. J Clin Oncol. 2020;38(18):2080-2106. doi: 10.1200/JCO.20.00299 [DOI] [PubMed] [Google Scholar]
- 9.Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature. 2000;406(6797):747-752. doi: 10.1038/35021093 [DOI] [PubMed] [Google Scholar]
- 10.Goldhirsch A, Wood WC, Coates AS, Gelber RD, Thürlimann B, Senn HJ. Strategies for subtypes—dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Ann Oncol. 2011;22(8):1736-1747. doi: 10.1093/annonc/mdr304 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Curigliano G, Burstein HJ, Winer EP, et al. ; St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2017 . De-escalating and escalating treatments for early-stage breast cancer: the St. Gallen International Expert Consensus Conference on the Primary Therapy of Early Breast Cancer 2017. Ann Oncol. 2017;28(8):1700-1712. doi: 10.1093/annonc/mdx308 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Burstein HJ, Curigliano G, Loibl S, et al. ; Members of the St. Gallen International Consensus Panel on the Primary Therapy of Early Breast Cancer 2019 . Estimating the benefits of therapy for early-stage breast cancer: the St. Gallen International Consensus Guidelines for the Primary Therapy of Early Breast Cancer 2019. Ann Oncol. 2019;30(10):1541-1557. doi: 10.1093/annonc/mdz235 [DOI] [PubMed] [Google Scholar]
- 13.Schmidt MK, Hogervorst F, van Hien R, et al. Age- and tumor subtype–specific breast cancer risk estimates for CHEK2*1100delC carriers. J Clin Oncol. 2016;34(23):2750-2760. doi: 10.1200/JCO.2016.66.5844 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Broeks A, Schmidt MK, Sherman ME, et al. ; Genica Network; kConFab; Australian Ovarian Cancer Study . Low penetrance breast cancer susceptibility loci are associated with specific breast tumor subtypes: findings from the Breast Cancer Association Consortium. Hum Mol Genet. 2011;20(16):3289-3303. doi: 10.1093/hmg/ddr228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dai X, Li T, Bai Z, et al. Breast cancer intrinsic subtype classification, clinical use and future trends. Am J Cancer Res. 2015;5(10):2929-2943. [PMC free article] [PubMed] [Google Scholar]
- 16.Guiu S, Michiels S, André F, et al. Molecular subclasses of breast cancer: how do we define them? the IMPAKT 2012 Working Group Statement. Ann Oncol. 2012;23(12):2997-3006. doi: 10.1093/annonc/mds586 [DOI] [PubMed] [Google Scholar]
- 17.Zhang H, Zhao N, Ahearn TU, Wheeler W, García-Closas M, Chatterjee N. A mixed-model approach for powerful testing of genetic associations with cancer risk incorporating tumor characteristics. Biostatistics. 2021;22(4):772-788. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lee AJ, Cunningham AP, Kuchenbaecker KB, Mavaddat N, Easton DF, Antoniou AC; Consortium of Investigators of Modifiers of BRCA1/2; Breast Cancer Association Consortium . BOADICEA breast cancer risk prediction model: updates to cancer incidences, tumour pathology and web interface. Br J Cancer. 2014;110(2):535-545. doi: 10.1038/bjc.2013.730 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lakhani SR, Van De Vijver MJ, Jacquemier J, et al. The pathology of familial breast cancer: predictive value of immunohistochemical markers estrogen receptor, progesterone receptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2. J Clin Oncol. 2002;20(9):2310-2318. doi: 10.1200/JCO.2002.09.023 [DOI] [PubMed] [Google Scholar]
- 20.Mavaddat N, Barrowdale D, Andrulis IL, et al. ; HEBON; EMBRACE; GEMO Study Collaborators; kConFab Investigators; SWE-BRCA Collaborators; Consortium of Investigators of Modifiers of BRCA1/2 . Pathology of breast and ovarian cancers among BRCA1 and BRCA2 mutation carriers: results from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA). Cancer Epidemiol Biomarkers Prev. 2012;21(1):134-147. doi: 10.1158/1055-9965.EPI-11-0775 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Evans DG, Lalloo F, Howell S, Verhoef S, Woodward ER, Howell A. Low prevalence of HER2 positivity amongst BRCA1 and BRCA2 mutation carriers and in primary BRCA screens. Breast Cancer Res Treat. 2016;155(3):597-601. doi: 10.1007/s10549-016-3697-z [DOI] [PubMed] [Google Scholar]
- 22.Honrado E, Osorio A, Palacios J, Benitez J. Pathology and gene expression of hereditary breast tumors associated with BRCA1, BRCA2 and CHEK2 gene mutations. Oncogene. 2006;25(43):5837-5845. doi: 10.1038/sj.onc.1209875 [DOI] [PubMed] [Google Scholar]
- 23.Heikkinen T, Kärkkäinen H, Aaltonen K, et al. The breast cancer susceptibility mutation PALB2 1592delT is associated with an aggressive tumor phenotype. Clin Cancer Res. 2009;15(9):3214-3222. doi: 10.1158/1078-0432.CCR-08-3128 [DOI] [PubMed] [Google Scholar]
- 24.Park JY, Zhang F, Andreassen PR. PALB2: the hub of a network of tumor suppressors involved in DNA damage responses. Biochim Biophys Acta. 2014;1846(1):263-275. doi: 10.1016/j.bbcan.2014.06.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Irminger-Finger I, Ratajska M, Pilyugin M. New concepts on BARD1: regulator of BRCA pathways and beyond. Int J Biochem Cell Biol. 2016;72:1-17. doi: 10.1016/j.biocel.2015.12.008 [DOI] [PubMed] [Google Scholar]
- 26.Tarsounas M, Sung P. The antitumorigenic roles of BRCA1-BARD1 in DNA repair and replication. Nat Rev Mol Cell Biol. 2020;21(5):284-299. doi: 10.1038/s41580-020-0218-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Renault AL, Mebirouk N, Fuhrmann L, et al. ; CoF-AT; GENESIS; kConFab . Morphology and genomic hallmarks of breast tumours developed by ATM deleterious variant carriers. Breast Cancer Res. 2018;20(1):28. doi: 10.1186/s13058-018-0951-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ahn JY, Schwarz JK, Piwnica-Worms H, Canman CE. Threonine 68 phosphorylation by ataxia telangiectasia mutated is required for efficient activation of Chk2 in response to ionizing radiation. Cancer Res. 2000;60(21):5934-5936. [PubMed] [Google Scholar]
- 29.Cai Z, Chehab NH, Pavletich NP. Structure and activation mechanism of the CHK2 DNA damage checkpoint kinase. Mol Cell. 2009;35(6):818-829. doi: 10.1016/j.molcel.2009.09.007 [DOI] [PubMed] [Google Scholar]
- 30.Yang X, Song H, Leslie G, et al. Ovarian and breast cancer risks associated with pathogenic variants in RAD51C and RAD51D. J Natl Cancer Inst. 2020;112(12):1242-1250. doi: 10.1093/jnci/djaa030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Shimelis H, LaDuca H, Hu C, et al. Triple-negative breast cancer risk genes identified by multigene hereditary cancer panel testing. J Natl Cancer Inst. 2018;110(8):855-862. doi: 10.1093/jnci/djy106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Couch FJ, Hart SN, Sharma P, et al. Inherited mutations in 17 breast cancer susceptibility genes among a large triple-negative breast cancer cohort unselected for family history of breast cancer. J Clin Oncol. 2015;33(4):304-311. doi: 10.1200/JCO.2014.57.1414 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Melhem-Bertrandt A, Bojadzieva J, Ready KJ, et al. Early onset HER2-positive breast cancer is associated with germline TP53 mutations. Cancer. 2012;118(4):908-913. doi: 10.1002/cncr.26377 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Masciari S, Dillon DA, Rath M, et al. Breast cancer phenotype in women with TP53 germline mutations: a Li-Fraumeni syndrome consortium effort. Breast Cancer Res Treat. 2012;133(3):1125-1130. doi: 10.1007/s10549-012-1993-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Fortuno C, Mester J, Pesaran T, et al. ; Li-Fraumeni Exploration (LIFE) Consortium . Suggested application of HER2+ breast tumor phenotype for germline TP53 variant classification within ACMG/AMP guidelines. Hum Mutat. 2020;41(9):1555-1562. doi: 10.1002/humu.24060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.McCart Reed AE, Kutasovic JR, Nones K, et al. Mixed ductal-lobular carcinomas: evidence for progression from ductal to lobular morphology. J Pathol. 2018;244(4):460-468. doi: 10.1002/path.5040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Li J, Ugalde-Morales E, Wen WX, et al. Differential burden of rare and common variants on tumor characteristics, survival, and mode of detection in breast cancer. Cancer Res. 2018;78(21):6329-6338. doi: 10.1158/0008-5472.CAN-18-1018 [DOI] [PubMed] [Google Scholar]
- 38.Cybulski C, Kluźniak W, Huzarski T, et al. ; Polish Hereditary Breast Cancer Consortium . Clinical outcomes in women with breast cancer and a PALB2 mutation: a prospective cohort analysis. Lancet Oncol. 2015;16(6):638-644. doi: 10.1016/S1470-2045(15)70142-7 [DOI] [PubMed] [Google Scholar]
- 39.Cybulski C, Górski B, Huzarski T, et al. CHEK2-positive breast cancers in young Polish women. Clin Cancer Res. 2006;12(16):4832-4835. doi: 10.1158/1078-0432.CCR-06-0158 [DOI] [PubMed] [Google Scholar]
- 40.Deng M, Chen HH, Zhu X, et al. Prevalence and clinical outcomes of germline mutations in BRCA1/2 and PALB2 genes in 2769 unselected breast cancer patients in China. Int J Cancer. 2019;145(6):1517-1528. doi: 10.1002/ijc.32184 [DOI] [PubMed] [Google Scholar]
- 41.Lee A, Mavaddat N, Wilcox AN, et al. BOADICEA: a comprehensive breast cancer risk prediction model incorporating genetic and nongenetic risk factors. Genet Med. 2019;21(8):1708-1718. doi: 10.1038/s41436-018-0406-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Spurdle AB, Couch FJ, Parsons MT, et al. ; ABCTB Investigators; EMBRACE Group; GENICA Network; HEBON Group; kConFab Investigators . Refined histopathological predictors of BRCA1 and BRCA2 mutation status: a large-scale analysis of breast cancer characteristics from the BCAC, CIMBA, and ENIGMA consortia. Breast Cancer Res. 2014;16(6):3419. doi: 10.1186/s13058-014-0474-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Parker JS, Mullins M, Cheang MC, et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol. 2009;27(8):1160-1167. doi: 10.1200/JCO.2008.18.1370 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Szymiczek A, Lone A, Akbari MR. Molecular intrinsic versus clinical subtyping in breast cancer: a comprehensive review. Clin Genet. 2021;99(5):613-637. doi: 10.1111/cge.13900 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
eTable 1. Description of studies included in this analysis
eTable 2. Immunohistochemistry and tumor grade-based surrogates for five intrinsic breast cancer subtypes
eTable 3. Numbers of cases and controls, and age distributions, by country of origin
eTable 4. Numbers of variant carriers by breast cancer susceptibility gene
eTable 5. Cross tabulation of ER, PR, HER2 and Grade data
eTable 6. Distribution of intrinsic tumor subtypes in women of all ages and in different age groups, by breast cancer susceptibility gene.
eTable 7. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women of different age groups at diagnosis
eTable 8. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes refined by PR expression
eTable 9. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes of breast cancer following imputation using an EM algorithm.
eFigure 1. Case-only analysis of phenotypic markers and prognostic features by gene (complete case analysis)
eFigure 2. Frequency histogram of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in 9 genes.
eFigure 3. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged ≤40 years.
eFigure 4. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged 41-60 years.
eFigure 5. Frequency distribution of intrinsic subtypes among noncarriers and carriers of PTVs and MSVs in the 9 genes, in women aged >60 years.
eFigure 6. Association odds ratios for MSV carrier status in BRCA1, BRCA2, TP53 and intrinsic subtypes of breast cancer.
eFigure 7. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged ≤40 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 8. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged 41-60 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 9. Prevalence of PTV and MSV in breast cancer susceptibility genes by intrinsic subtypes of breast cancer among women aged >60 at diagnosis (A) HR+ HER2- low-grade (B) HR+ HER2- (C) HR+ HER2- high-grade (D) HR- HER2+ (E) TN.
eFigure 10. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged ≤40 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 11. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged 41-60 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 12. Prevalence of PTV and MSV in breast cancer susceptibility genes by tumor grade among women aged >60 years at diagnosis with (A) grade 1 (B) grade 2 and (C) grade 3 breast cancer.
eFigure 13. Smoothed proportions of subtypes used in BOADICEA for PTVs in ATM, BARD1, BRCA1, BRCA2, CHEK2, PALB2, RAD51C and RAD51D.
eTable 1. Missing data by gene/mutation status and age at diagnosis
eTable 2. Number of cases by subtype of breast cancer, as defined by individual tumor markers, and PTV/MSV carrier status for each gene
eTable 3. Odds ratios for association between PTV and MSV carrier status and subtypes of breast cancer as defined by individual tumor markers (complete case analyses)
eTable 4. Odds ratios for association between PTV and MSV carrier status and intrinsic subtypes of breast cancer following MICE imputation– analyses with and without an age interaction term
eTable 5. Association between breast cancer genes and intrinsic subtypes - Joint models of intrinsic subtype, tumor size and lymph node involvement
eTable 6. Smoothed proportions of subtypes used in BOADICEA for PTVs in ATM, BARD1, BRCA1, BRCA2 CHEK2, PALB2, RAD51C and RAD51D