Chemotherapy-related agranulocytosis as a predictive factor for germline BRCA1 pathogenic variants in breast cancer patients: a retrospective cohort study

DOI: https://doi.org/https://doi.org/10.57187/smw.2023.40055

Introduction

Individuals who are heterozygous for pathogenic variants in BRCA1 and BRCA2 genes (gBRCA) face an increased lifetime risk of developing breast cancer and ovarian cancer. Early identification of gBRCA pathogenic variants is essential as it can lead to tailored interventions and potentially influence patients’ prognosis [1–4]. Recent advances in next-generation sequencing significantly increased the accessibility and workflow of BRCA genetic testing; however, the identification of gBRCA pathogenic variants remains challenging, with financial, psychosocial, and legal implications. Current guidelines do not recommend routine gBRCA molecular analysis in all breast cancer patients; instead, they encourage offering genetic counselling to individuals at risk of harbouring deleterious gBRCA variants. In addition, this intervention facilitates a useful informed risk/benefit discussion with patients [5–7].

Several factors influence the probability of carrying gBRCA1/gBRCA2 pathogenic variants, with young age at diagnosis being the most important [8]. Criteria based on tumour characteristics and individual and family history have been published by knowledgeable societies to select patients for genetic counselling referral [5–7]. A variety of risk-prediction models have been developed to assist clinicians in their informed decision-making to better identify a “high-risk” subset of breast cancer patients selected to undergo gBRCA genetic testing, such as the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA), the pathology-adjusted version of the Manchester score (MSS3), and BRCAPRO^TM [9–17]. By accurately calculating the area under the curve on receiver operating characteristic (AUROC), complex statistical algorithms in BOADICEA and BRCAPRO assess an individual’s risk of harbouring a pathogenic variant in breast cancer susceptibility genes. Today, these risk-prediction models are largely routinely implemented in genetic counselling practice. Each model has advantages and limitations and need to be interpreted in the context of the studied populations in which they have been validated.

Breast cancer occurring in gBRCA heterozygotes accounts for approximatively 3–5% of all breast cancers [18–20]. breast cancers related to gBRCA1 tend to be of higher grade and mostly present as the triple-negative molecular subtype (TNBC) [21–23]. Haplo-insufficient BRCA neoplastic cells have an impaired ability to repair DNA double-stranded breaks (DSBs) through homologous recombination mechanisms [24]. Neoadjuvant and adjuvant chemotherapy regimens used in breast cancer usually include alkylating agents and anthracyclines, which both induce DNA-DSBs. Because haematological cells in gBRCA heterozygotes harbour one defective allele, it has been hypothesised that gBRCA haematological cells could exhibit increased sensitivity to agents inducing DNA-DSBs (e.g., chemotherapy and radiotherapy). Several groups attempted to elucidate this question using mixed breast cancer populations and heterogeneous methodologies, leading to inconsistent results [25–36]. Analysing neutrophils counts at their nadir (mid-cycle), our group recently reported that gBRCA1 heterozygotes but not gBRCA2 heterozygotes with breast cancer tend to be associated with a significantly higher rate of developing grade 4 and febrile agranulocytosis during their first cycle of chemotherapy [35]. Our observation is consistent with a recent report highlighting increased haematological toxicity in gBRCA1 but not gBRCA2 heterozygotes [36]. Furthermore, gBRCA1 heterozygotes require more granulocyte colony-stimulating factor (G-CSF) support than non-heterozygotes [35]. The aim of this study was to determine whether agranulocytosis occurring during the first cycle (C1) of (neo-)adjuvant chemotherapy (days 7–14) could predict the likelihood of harbouring a pathogenic gBRCA1 variant.

Patients and methods

Patient cohort

We analysed the files of all breast cancer patients referred for genetic counselling and gBRCA1/gBRCA2 testing between January 1998 and December 2017 at the Unit of Oncogenetics, Hôpitaux Universitaires de Genève (HUG), Switzerland. As per edited Swiss criteria [37], patients were referred for counselling if they presented with i) newly diagnosed ovarian cancer, ii) family history of breast cancer and/or ovarian cancer, or iii) age ≤60 years with TNBC or age <40 years independent of breast cancer subtype. The Geneva Ethics Committee approved the research protocol in 2015 (CCER 15-158), and an amendment to the study protocol was obtained in 2017. Written informed consent was obtained from living patients and a waiver was applied for deceased patients. The research was conducted in accordance with the Declaration of Helsinki and Swiss regulations for human research.

Patients included in the present analysis met the following inclusion criteria: (1) non-metastatic breast cancer, (2) having received (neo-)adjuvant chemotherapy, and (3) available blood count between day 7 and 14 after their first cycle of chemotherapy.

The exclusion criteria were (1) metastatic disease at diagnosis, (2) absence of (neo-)adjuvant chemotherapy, (3) absence of genetic testing, and (4) no follow-up.

Data collection

Clinical data (age at diagnosis, chemotherapy regimen and timing [neoadjuvant or adjuvant]), tumour characteristics (TNM stage, grade, oestrogen/progesterone receptors, HER2 status), and laboratory results (blood count, including neutrophils) at baseline and day 7–14 after C1 were collected from patients’ medical records in a dedicated deidentified case report form. The key list was accessible only by the principal investigator. No software libraries, frameworks, or packages were used in the current work. Haematological toxicity was graded according to the Common Terminology Criteria for Adverse Events version 5.0 [38], with agranulocytosis defined as absolute neutrophil count <0.5 × 10⁹ cells/L. Febrile neutropenia was defined as absolute neutrophil count <1 × 10⁹ cells/L and fever >38.3 °C.

Pedigree data were collected from the genetic counselling charts, provided in standard linkage format with year of birth, age at last follow-up or death, and details of any cancer diagnosis (cancer site, age at diagnosis) for each family member. When age was missing but the year of birth was provided, age was calculated as years from the date of birth to the date the pedigree was drawn for living probands and the date of death for deceased probands. Cancers were recorded only when there was no ambiguity in patients’ charts and age information was available. The following family history characteristics were collected for first-, second-, and third-degree relatives: (1) ≥1 breast cancer at age 40 years, (2) ≥3 female relatives with breast cancer aged ≤50 years, (3) ≥1 ovarian cancer at any age, (4) ≥1 breast cancer associated with ovarian cancer at any age (in the same relative or not), and (5) ≥1 bilateral breast cancer at any age. First-degree relatives were considered as direct filiation (biological parents and children), and second-degree relatives included aunts/uncles, grand-daughters/grand-sons, and grand-parents. Third-degree relatives include cousins and other family members.

Prediction of comprehensive genetic risk models was generated from pedigree data of genetic counselling charts: i) BOADICEA v. 3.0, a computer-based risk assessment incorporating detailed personal risk factors, pathological molecular markers, and individual and familial information from both lineages up to the third relative generation degree [39], and ii) MSS3, a paper-based risk model including several individual and familial characteristics from only one lineage, not applicable to individuals ofAshkenazi Jewish descent [14]. BOADICEA risk was prospectively calculated for probands enrolled between 2016 and 2017 and retrospectively for all other probands. MSS3 score was retrospectively generated for all probands; each variable had a dedicated numerical weight, with the sum providing a score in points that was converted into a percentage probability of finding a gBRCA mutation in a breast cancer patient. There is no formal consensus on the thresholds to apply for gBRCA1/gBRCA2 mutation testing [12, 15]. In the current study, we choose a 10% pre-test cut-off (15 points equivalent for MSS3) in patients diagnosed with breast cancer. Germline molecular blood results of BRCA1 and BRCA2 variants were classified as pathogenic according to the ENIGMA BRCA1/2 Gene Variant Classification Criteria. Women with variants of uncertain significance were considered non-heterozygotes [40, 41].

Endpoints

The primary endpoint was to assess the proportion of gBRCA1 pathogenic variants in a non-metastatic breast cancer population presenting with agranulocytosis during C1 of chemotherapy. Secondary endpoints included comparison of the predictive value of individualised parameters (TNBC, agranulocytosis after C1) and both risk models (BOADICEA v. 3.0 and MSS3).

Statistics

This was an exploratory analysis; no sample size calculation was performed. Proportions were calculated for binary data (e.g., family history, agranulocytosis, febrile neutropenia) and categorical data (e.g., TNM stage, grade, chemotherapy regimen), whereas median and interquartile range were calculated for continuous data (age). Comparison of categorical data was performed using the χ² test or Fisher’s exact test, when appropriate. Continuous variables were compared using the Kruskal–Wallis test. Patient characteristic frequencies were summarised according to gBRCA status and were compared pair by pair (gBRCA1 heterozygotes vs. non-gBRCA1 heterozygotes; gBRCA2 heterozygotes vs. non-gBRCA2 heterozygotes). Missing data or inapplicable responses were excluded when calculating p values. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using logistic regression. Variables were added to the model if the two-sided significances level was less than 0.05 in the univariate analysis. A stepwise backward selection procedure removing terms with P > 0.05 was performed to assess the independent contribution of each studied factor to the outcome variables. All probability values were two-tailed, and the level of significance was set at 0.05. Calculations of sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were performed for mid-cycle agranulocytosis (cut-off 0.5 G/L neutrophils), BOADICEA (cut-off 10%), and MSS3 (cut-off 15 points) individually in univariate analysis. Variables yielding p values less than 0.1 by univariate analysis were retained for the multivariate analysis. ORs were generated for variables that retained significance in the multivariate analysis. All analyses were conducted using STATA software (version 14.0 SE, Stata Corporation, College Station, Texas, USA).

Results

Patient characteristics

Among 892 women diagnosed with breast cancer who were referred for genetic counselling between 1998 and 2017 to the Unit of Oncogenetics at HUG, 322 met the inclusion criteria (figure S1). A total of 307 patients were included in the present analysis: 248 (81.1%) were BRCA1/2 non-heterozygotes, 32 (10.4%) were gBRCA1 heterozygotes, and 27 (8.8%) were gBRCA2 heterozygotes. Our cohort include roughly one-third of breast cancer patients referred for genetic counselling from private practice and two-third from the university hospital. Patient demographics, tumour characteristics, chemotherapy regimens, and family history are summarised in table 1. Approximatively one-third of breast cancer patients were treated with neoadjuvant chemotherapy, and two-third received adjuvant chemotherapy. Nearly all patients (>95%) received alkylating agents, and 91% of non-heterozygotes, 94% of gBRCA1 heterozygotes, and 75% of gBRCA2 heterozygotes received anthracyclines. The median age at diagnosis was comparable between gBRCA heterozygotes and non-heterozygotes (40 years). TNBC phenotype was significantly more frequently observed in gBRCA1 heterozygotes (68.8%) compared with gBRCA2 heterozygotes (30.8%) and non-heterozygotes (25.8%) (p <0.001). Carriers of BRCA1 pathogenic variants were more likely to have a positive personal and/or familial history of breast cancer and/or ovarian cancer. No familial history was reported by 34% of gBRCA1 and 38% of gBRCA2 heterozygotes.

Occurrence of mid-cycle agranulocytosis

Compared with non-heterozygotes, the incidence of agranulocytosis occurring during C1 was higher among gBRCA1 heterozygotes (45.8% versus 16.8%; p = 0.002) than among gBRCA2 heterozygotes (20.8% versus 16.8%; p = 0.576) (Table 2). Febrile neutropenia occurred more frequently in gBRCA1 heterozygotes (28%; p = 0.001) than gBRCA2 heterozygotes (8%; p = 0.641) compared with non-heterozygotes (5.5%). Only 10 patients (3.3%) received primary prophylaxis with G-CSF during C1 (one gBRCA1heterozygote, one gBRCA2 heterozygote, and eight non-heterozygotes). Clinical factors and therapeutic agents that may predict agranulocytosis were evaluated via multivariate analysis; however, none of these were statistically significant in our cohort (supplemental table S1 and supplemental table S2).

Prediction of gBRCA1 pathogenic variants

The median BOADICEA BRCA1 score was 9.7 (2.5–56.6) among gBRCA1 heterozygotes compared with 0.7 (0.3–2.8) for non-heterozygotes (p <0.001). This difference remained statistically significant when applying a pre-test probability threshold of 10. Median BOADICEA BRCA2 score was not statistically different between gBRCA2 heterozygotes (2.6 [1.0–7.0]) and non-heterozygotes (1.8 [0.8–4.0]; p = 0.113). MSS3 scores were significantly higher in both gBRCA1 (27.5 [20–38]; p <0.001) and gBRCA2 heterozygotes (20 [14–28]; p <0.001) compared with non-heterozygotes (13 [9–17]). This difference remained significant when a 15-point cut-off was applied (equivalent to a 10% pre-test probability threshold). Scores per mutational status are summarised in table 2. AUROCs were generated per mutational status for BOADICEA and MSS3 (supplemental table S3).

Univariate and multivariate analyses

Following univariate analysis, nearly all individual and familial characteristics (TNBC subtype, agranulocytosis during C1, febrile neutropenia during C1, family history of one relative with breast cancer diagnosed before 41 years of age, ≥3 relatives with breast cancer diagnosed before 51 years of age, or one relative with a positive history of ovarian cancer with or without breast cancer) were significantly associated with gBRCA1 mutation status (table 3). Only two variables (≥3 relatives with breast cancer diagnosed <51 years of age and bilateral breast cancer) were significant for gBRCA2 status. Age at diagnosis was not significant in our population (based on the Swiss criteria for referral to genetic counselling). All factors listed above remained statistically independently associated with gBRCA1 mutation statusfollowing multivariate analysis (table 4). TNBC subtype was the strongest individual predictive factor for gBRCA1 mutation (odds ratio [OR]: 15.6 [4.3–56.6]; p <0.001). Other factors were the presence of three or more relatives with breast cancer diagnosed before 51 years of age (OR: 14.1 [2.7–74.7]; p = 0.002), one relative with breast cancer diagnosed before 41 years of age (OR: 12.8 [3.2–51.5]; p <0.001), agranulocytosis during C1 (OR: 6.1 [1.9–19.3]; p = 0.002), and a positive history of ovarian cancer (OR: 4.1 [1.0–17.1]; p = 0.05). Due to the small population size, it was not possible to estimate the OR for febrile neutropenia and the presence of breast cancer or ovarian cancer in family history.

Performance of individual and combined parameters

We evaluated the individual performance of agranulocytosis during C1, TNBC status, and BOADICEA BRCA1 and MSS3 scores to predict gBRCA1 pathogenic variants in our cohort by calculating sensitivity, specificity, PPV, and NPV (table 5). Furthermore, to determine whether the risk evaluation could be improved, we attempted to combine agranulocytosis during C1 with either TNBC status, BOADICEA BRCA1 score, or MSS3 score (scores being considered positive with a ≥10% threshold). Adding agranulocytosis information to TNBC status, BOADICEA BRCA1 score, or MSS3 score improved PPV from 23.4% (15.3–33.3%) to 50% (26.0–74.0%), 50% (31.3–68.7%)to 77.8% (40.0–97.2%), and 18.2% (12.4–25.4%) to 37% (19.4–57.6%), respectively (table 5).

Discussion

In the general population, the frequency of gBRCA1 and gBRCA2 pathogenic variants varies between 0.18% and 0.34%, respectively [42]. The presence of an increased number of relatives affected by breast cancer, ovarian cancer, prostate, or pancreatic cancers occurring at a young age usually alerts physicians to a potential hereditary breast-ovarian cancer syndrome [43]; however, most breast cancers arise in individuals with no obvious family history. The lifetime cumulative risk of developing breast cancer in gBRCA heterozygotes is substantially increased compared with the general population, estimated at 65% in gBRCA1 heterozygotes and 45% in gBRCA2 heterozygotes by the age of 70 years [2]. Identification of gBRCA pathogenic variants is of prime importance to implement specific medical management, such as cascade genetic screening of close relatives, specific individual surveillance [5–7], chemopreventive approaches [44–53], and personalised therapeutic strategies [4, 54, 55–60]. Even though genetic testing has been demonstrated to be cost-effective [61–65], for financial reasons, gBRCA1/gBRCA2 genetic testing is restricted to newly diagnosed breast cancer patients with a high likelihood of harbouring a deleterious genetic variant [5–7].

Taking advantage of the previously demonstrated susceptibility of haematopoietic cells in gBRCA1 breast cancer heterozygotes after DNA-damaging chemotherapies [66, 67], we hypothesised that agranulocytosis occurring during C1 could predict gBRCA1 mutation.We chose to investigate agranulocytosis after C1 to avoid the bias introduced with the use of secondary G-CSF prophylaxis. In line with our previously reported results, the incidence of agranulocytosis during C1 was higher among gBRCA1 heterozygotes, confirming the sensitivity of gBRCA1 haploinsufficient haematopoietic cells to DNA-damaging agents [35]. This could be due to the requirement of BRCA1 for haematopoietic stem cell function and normal haematopoiesis [67]. Following multivariate analysis, we found that agranulocytosis during C1 was an independent predictive factor for gBRCA1 but not gBRCA2 pathogenic variants. This finding reinforces the weight of individual pathological and biological features in predicting gBRCA1 pathogenic variants [68–71].

Our cohort was representative of Geneva’s breast cancer population and was relatively well-balanced in terms of age, stage, and therapeutic regimens. In line with the reported literature, we noted an enrichment of high-grade tumours and the TNBC subtype among gBRCA1 heterozygotes [68–71]. A median age of 40 years in our study underlined the fact that young age at diagnosis is the strongest predictive factor of genetic counselling and gBRCA testing [72]. In a study conducted by Beck et al., age ≤45 years at the time of breast cancer diagnosis and age ≤60 years at the time of TNBC diagnosis facilitated the detection of 9.3% and 9.7% of gBRCA pathogenic variants, respectively [73]. In a similar population to ours, Grindedal et al. found that age of onset <40 years or TNBC status facilitated the identification of 32–34% of gBRCA heterozygotes, highlighting the importance of these individualised parameters in gBRCA risk prediction [74].

In addition to well-known risk criteria [5–7, 37], several models can predict a subject’s individual probability of harbouring a gBRCA pathogenic variant and were demonstrated to be superior to counsellors’ estimated probability [7, 10–18]. The most recent versions of these risk algorithms integrate the tumour pathological characteristics of the index case, leading to increased discrimination performance [12, 15]. However, these models still rely on the fulfilment of a detailed familial history; therefore, their value is limited in the absence of any positive familial history [21]. As in previous cohorts, the BOADICEA model performed well in its prediction of gBRCA1 pathogenic variant detection (AUC:0.71) and relatively poorly for gBRCA2 pathogenic variants (AUC: 0.53) [10, 66–67, 75]. This could be explained by the fact that breast cancer in gBRCA1 heterozygotes is mainly TNBC, whereas gBRCA2 breast cancer tends to resemble sporadic breast cancer and is predominantly hormone receptor positive. Consequently, genetic predictive models considering pathological characteristics are usually better at predicting the risk of gBRCA1 than gBRCA2 [75]. With AUCs of 0.79 for gBRCA1 and 0.66 for gBRCA2, theMSS3 model demonstrated slightly better performance than the BOADICEA model in our cohort of breast cancer patients, achieving comparable results to those reported in the literature [13, 14]. This discrepancy between the predictive risk models could be due to the presence of a high proportion of gBRCA heterozygotes lacking a positive family history in our cohort (34% of gBRCA2 and 38% of gBRCA1 heterozygotes) [74]. This finding reflects the low birth rate and tendency for smaller family sizes in Switzerland and in developed countries in general. Furthermore, young women are less likely to have close relatives who are affected due to their age [72, 76]. MSS3 tests both lineages, taking in account the highest score, and incorporates proportionally more individual cancer-based information, relying less on family history. Therefore, MSS3 is expected to perform well in breast cancer patients lacking a strong family history.

The strengths of our study are its novelty and methodology. Systematic blood count check performed at mid-cycle after C1 (days 7–14) does not always reflect a precise nadir of neutrophils; however, this time window allows for the capture of most agranulocytosis events and eliminates potential reporting bias. With its inclusion of patients referred from a public university hospital and the private sector of care, our cohort is representative of the Swiss breast cancer population. As chemotherapy regimens vary in their impact on the incidence of haematological toxicity, we deliberately restricted our patient population to non-metastatic breast cancer treated with neoadjuvant or adjuvant chemotherapy. This choice made our studied population quite homogenous, with nearly all breast cancer patients receiving both anthracyclines and alkylating agents in the first cycle of chemotherapy. Moreover, risk-prediction scores were uniformly retrospectively calculated with the latest version of the Manchester scoring system (MSS3) [14] and the BOADICEA model version 3 [39] based on available clinical and pathological information, with only few missing data. Finally, all patients benefited from a complete BRCA1 and BRCA2 sequencing, thereby avoiding potential selection bias that could be induced by analysis of founder pathogenic variants alone.

In addition to its small sample size and inherent retrospective nature, our study had several limitations. First, only breast cancer patients meeting Swiss criteria underwent genetic counselling and gBRCA1/gBRCA2 testing. This led to recruitment bias, with substantial enrichment of gBRCA heterozygotes in our cohort, reflected by a mean age of 40 years. By applying Swiss guidelines, we likely missed patients whose carrier pre-test probability was too low to undergo genetic testing; therefore, our reported NPV and PPV values could not be generalised to an unselected breast cancer population. Second, the study was conducted before the implementation of 12-gene panel testing; therefore, it is important to investigate whether carriers of other genes (e.g., FANC), are at increased risk of acute haematological toxicity. This is a timely question given the new standard therapeutic regimen for TNBC (pembrolizumab, carboplatin, anthracyclines, and cyclophosphamide), which is associated with substantial haematological toxicity according to the recently reported KEYNOTE-552 trial results [77].

Conclusion

BRCA1 haploinsufficiency confers an increased sensitivity to DNA damaging agents in breast cancer patients, and physicians should be aware that these patients are at increased risk of developing severe acute haematological toxicity and febrile neutropenia. Although germline genetic testing is increasingly used in routine practice for young breast cancer and/or TNBC patients, our results suggest that agranulocytosis that develops during C1 in non-metastatic breast cancer patients could be an independent predictive factor of interest for gBRCA1 pathogenic variant detection.

References

1. Kuchenbaecker KB , Hopper JL , Barnes DR , Phillips KA , Mooij TM , Roos-Blom MJ , et al.; BRCA1 and BRCA2 Cohort Consortium . Risks of Breast, Ovarian, and Contralateral Breast Cancer for BRCA1 and BRCA2 Mutation Carriers. JAMA. 2017 Jun;317(23):2402–16. https://doi.org/10.1001/jama.2017.7112  

2. Antoniou A , Pharoah PD , Narod S , Risch HA , Eyfjord JE , Hopper JL , et al.  Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case Series unselected for family history: a combined analysis of 22 studies [published correction appears in Am J Hum Genet. 2003 Sep;73(3):709]. Am J Hum Genet. 2003 May;72(5):1117–30. https://doi.org/10.1086/375033  

3. De Talhouet S , Peron J , Vuilleumier A , Friedlaender A , Viassolo V , Ayme A , et al.  Clinical outcome of breast cancer in carriers of BRCA1 and BRCA2 mutations according to molecular subtypes [published correction appears in Sci Rep. 2020 Nov 2;10(1):19248]. Sci Rep. 2020 Apr;10(1):7073. https://doi.org/10.1038/s41598-020-63759-1  

4. Tutt AN , Garber JE , Kaufman B , Viale G , Fumagalli D , Rastogi P , et al.; OlympiA Clinical Trial Steering Committee and Investigators . Adjuvant Olaparib for Patients with BRCA1- or BRCA2-Mutated Breast Cancer [published online ahead of print, 2021 Jun 3]. N Engl J Med. 2021 Jun;384(25):2394–405. https://doi.org/10.1056/NEJMoa2105215  

5. Daly MB , Pilarski R , Yurgelun MB , Berry MP , Buys SS , Dickson P , et al.  NCCN Guidelines Insights: Genetic/Familial High-Risk Assessment: Breast, Ovarian, and Pancreatic, Version 1.2020. J Natl Compr Canc Netw. 2020 Apr;18(4):380–91. https://doi.org/10.6004/jnccn.2020.0017  

6. NICE Guidelines Committee . Familial Breast Cancer: Classification, Care and Managing Breast Cancer and Related Risks in People with a Family History of Breast Cancer. London, UK: National Institute for Health and Care Excellence. Published date: 25 June 2013 Last updated: 20 November 2019. www.nice.org.uk/guidance/cg164  

7. Paluch-Shimon S , Cardoso F , Sessa C , Balmana J , Cardoso MJ , Gilbert F , et al.; ESMO Guidelines Committee . Prevention and screening in BRCA mutation carriers and other breast/ovarian hereditary cancer syndromes: ESMO Clinical Practice Guidelines for cancer prevention and screening [published correction appears in Ann Oncol. 2017 Jul 1;28(suppl_4):iv167-iv168]. Ann Oncol. 2016 Sep;27 suppl 5:v103–10. https://doi.org/10.1093/annonc/mdw327  

8. Chen S , Parmigiani G . Meta-analysis of BRCA1 and BRCA2 penetrance. J Clin Oncol. 2007 Apr;25(11):1329–33. https://doi.org/10.1200/JCO.2006.09.1066  

9. Antoniou AC , Pharoah PP , Smith P , Easton DF . The BOADICEA model of genetic susceptibility to breast and ovarian cancer. Br J Cancer. 2004 Oct;91(8):1580–90. https://doi.org/10.1038/sj.bjc.6602175  

10. Antoniou AC , Cunningham AP , Peto J , et al.  The BOADICEA model of genetic susceptibility to breast and ovarian cancers: updates and extensions [published correction appears in Br J Cancer. 2008 Jun 17;98(12):2015. Passini, B [corrected to Pasini, B]]. Br J Cancer. 2008;98(8):1457–66. https://doi.org/10.1038/sj.bjc.6604305  

11. Lee A , Mavaddat N , Wilcox AN , Cunningham AP , Carver T , Hartley S , et al.  BOADICEA: a comprehensive breast cancer risk prediction model incorporating genetic and nongenetic risk factors [published correction appears in Genet Med. 2019 Feb 21]. Genet Med. 2019 Aug;21(8):1708–18. https://doi.org/10.1038/s41436-018-0406-9  

12. Evans DG , Eccles DM , Rahman N , Young K , Bulman M , Amir E , et al.  A new scoring system for the chances of identifying a BRCA1/2 mutation outperforms existing models including BRCAPRO. J Med Genet. 2004 Jun;41(6):474–80. https://doi.org/10.1136/jmg.2003.017996  

13. Evans DG , Lalloo F , Cramer A , Jones EA , Knox F , Amir E , et al.  Addition of pathology and biomarker information significantly improves the performance of the Manchester scoring system for BRCA1 and BRCA2 testing. J Med Genet. 2009 Dec;46(12):811–7. https://doi.org/10.1136/jmg.2009.067850  

14. Evans DG , Harkness EF , Plaskocinska I , Wallace AJ , Clancy T , Woodward ER , et al.  Pathology update to the Manchester Scoring System based on testing in over 4000 families. J Med Genet. 2017 Oct;54(10):674–81. https://doi.org/10.1136/jmedgenet-2017-104584  

15. Gail MH , Brinton LA , Byar DP , Corle DK , Green SB , Schairer C , et al.  Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst. 1989 Dec;81(24):1879–86. https://doi.org/10.1093/jnci/81.24.1879  

16. Parmigiani G , Berry D , Aguilar O . Determining carrier probabilities for breast cancer-susceptibility genes BRCA1 and BRCA2. Am J Hum Genet. 1998 Jan;62(1):145–58. https://doi.org/10.1086/301670  

17. Tyrer J , Duffy SW , Cuzick J . A breast cancer prediction model incorporating familial and personal risk factors. Stat Med. 2004 Apr;23(7):1111–30. https://doi.org/10.1002/sim.1668  

18. Rebbeck TR , Mitra N , Wan F , Sinilnikova OM , Healey S , McGuffog L , et al.; CIMBA Consortium . Association of type and location of BRCA1 and BRCA2 mutations with risk of breast and ovarian cancer. JAMA. 2015 Apr;313(13):1347–61. https://doi.org/10.1001/jama.2014.5985  

19. Li J , Wen WX , Eklund M , Kvist A , Eriksson M , Christensen HN , et al.  Prevalence of BRCA1 and BRCA2 pathogenic variants in a large, unselected breast cancer cohort. Int J Cancer. 2019 Mar;144(5):1195–204. https://doi.org/10.1002/ijc.31841  

20. Lang GT , Shi JX , Hu X , Zhang CH , Shan L , Song CG , et al.  The spectrum of BRCA mutations and characteristics of BRCA-associated breast cancers in China: screening of 2,991 patients and 1,043 controls by next-generation sequencing. Int J Cancer. 2017 Jul;141(1):129–42. https://doi.org/10.1002/ijc.30692  

21. Peshkin BN , Alabek ML , Isaacs C . BRCA1/2 mutations and triple negative breast cancers. Breast Dis. 2010;32(1-2):25–33. https://doi.org/10.3233/BD-2010-0306  

22. Lakhani SR , Reis-Filho JS , Fulford L , Penault-Llorca F , van der Vijver M , Parry S , et al.; Breast Cancer Linkage Consortium . Prediction of BRCA1 status in patients with breast cancer using estrogen receptor and basal phenotype. Clin Cancer Res. 2005 Jul;11(14):5175–80. https://doi.org/10.1158/1078-0432.CCR-04-2424  

23. Lakhani SR , Jacquemier J , Sloane JP , Gusterson BA , Anderson TJ , van de Vijver MJ , et al.  Multifactorial analysis of differences between sporadic breast cancers and cancers involving BRCA1 and BRCA2 mutations. J Natl Cancer Inst. 1998 Aug;90(15):1138–45. https://doi.org/10.1093/jnci/90.15.1138  

24. Roy R , Chun J , Powell SN . BRCA1 and BRCA2: different roles in a common pathway of genome protection. Nat Rev Cancer. 2011 Dec;12(1):68–78. https://doi.org/10.1038/nrc3181  

25. Drooger JC , Heemskerk-Gerritsen BA , Smallenbroek N , Epskamp C , Seynaeve CM , Jager A . Toxicity of (neo)adjuvant chemotherapy for BRCA1- and BRCA2-associated breast cancer. Breast Cancer Res Treat. 2016 Apr;156(3):557–66. https://doi.org/10.1007/s10549-016-3777-0  

26. Huszno J , Budryk M , Kołosza Z , Nowara E . The influence of BRCA1/BRCA2 mutations on toxicity related to chemotherapy and radiotherapy in early breast cancer patients. Oncology. 2013;85(5):278–82. https://doi.org/10.1159/000354834  

27. Huszno J , Budryk M , Kołosza Z , Nowara E . The risk factors of toxicity during chemotherapy and radiotherapy in breast cancer patients according to the presence of BRCA gene mutation. Contemp Oncol (Pozn). 2015;19(1):72–6. https://doi.org/10.5114/wo.2015.50014  

28. Tomao F , Musacchio L , Di Mauro F , Boccia SM , Di Donato V , Giancotti A , et al.  Is BRCA mutational status a predictor of platinum-based chemotherapy related hematologic toxicity in high-grade serous ovarian cancer patients? Gynecol Oncol. 2019 Jul;154(1):138–43. https://doi.org/10.1016/j.ygyno.2019.04.009  

29. Weitzner O , Yagur Y , Kadan Y , Beiner ME , Fishman A , Ben Ezry E , et al.  Chemotherapy Toxicity in BRCA Mutation Carriers Undergoing First-Line Platinum-Based Chemotherapy. Oncologist. 2019 Dec;24(12):e1471–5. https://doi.org/10.1634/theoncologist.2019-0272  

30. Kotsopoulos J , Willows K , Trat S , Kim RH , Volenik A , Sun P , et al.   BRCA mutation status is not associated with increased hematologic toxicity among patients undergoing platinum‐based chemotherapy for ovarian cancer. Int J Gynecol Cancer. 2018 Jan;28(1):69–76. https://doi.org/10.1097/IGC.0000000000001144  

31. Shanley S , McReynolds K , Ardern-Jones A , Ahern R , Fernando I , Yarnold J , et al.; Royal Marsden NHS Foundation Trust . Acute chemotherapy-related toxicity is not increased in BRCA1 and BRCA2 mutation carriers treated for breast cancer in the United Kingdom. Clin Cancer Res. 2006 Dec;12(23):7033–8. https://doi.org/10.1158/1078-0432.CCR-06-1246  

32. Sucheston LE , Zhao H , Yao S , Zirpoli G , Liu S , Barlow WE , et al.  Genetic predictors of taxane-induced neurotoxicity in a SWOG phase III intergroup adjuvant breast cancer treatment trial (S0221). Breast Cancer Res Treat. 2011 Dec;130(3):993–1002. https://doi.org/10.1007/s10549-011-1671-3  

33. Badora-Rybicka A , Budryk M , Nowara E , Starzyczny-Słota D . Treatment related toxicity in BRCA1-associated epithelial ovarian cancer - is DNA repairing impairment associated with more adverse events? Contemp Oncol (Pozn). 2016;20(5):381–4. https://doi.org/10.5114/wo.2016.64597  

34. Egloff H , Jatoi A . Do ovarian cancer patients with a family history of cancer (suspected BRCA1 or BRCA2 mutation) suffer greater chemotherapy toxicity? Cancer Invest. 2016 Nov;34(10):531–5. https://doi.org/10.1080/07357907.2016.1242011  

35. Friedlaender A , Vuilleumier A , Viassolo V , Ayme A , De Talhouet S , Combes JD , et al.  BRCA1/BRCA2 germline mutations and chemotherapy-related hematological toxicity in breast cancer patients. Breast Cancer Res Treat. 2019 Apr;174(3):775–83. https://doi.org/10.1007/s10549-018-05127-2  

36. Furlanetto J , Möbus V , Schneeweiss A , Rhiem K , Tesch H , Blohmer JU , et al.  Germline BRCA1/2 mutations and severe haematological toxicities in patients with breast cancer treated with neoadjuvant chemotherapy. Eur J Cancer. 2021 Mar;145:44–52. https://doi.org/10.1016/j.ejca.2020.12.007  

37. Chappuis PO , Bollinger B , Bürki N , et al.  Swiss guidelines for counselling and testing: genetic predisposition to breast and ovarian cancer. Schweiz Arzteztg. 2017;98(2122):682–4. https://doi.org/10.4414/saez.2017.05502  

38. NCI Common Terminology Criteria for Adverse Events V5 . 0. https://ctep.cancer.gov/protocoldevelopment/electronic_applications/docs/CTCAE_v5_Quick_Reference_8.5x11.pdf (Accessed on August 2021). 

39. BOADICEA Research Tool Software version 3.0, available via the World Wide Web https://pluto.srl.cam.ac.uk/cgi-bin/bd3/v3/bd.cgi (Accessed on February 2021). 

40. Spurdle AB , Healey S , Devereau A , Hogervorst FB , Monteiro AN , Nathanson KL , et al.; ENIGMA . ENIGMA—evidence-based network for the interpretation of germline mutant alleles: an international initiative to evaluate risk and clinical significance associated with sequence variation in BRCA1 and BRCA2 genes. Hum Mutat. 2012 Jan;33(1):2–7. https://doi.org/10.1002/humu.21628  

41. ENIGMA (Evidence‐based Network for the Interpretation of Germline Mutant Alleles) . http://www.enigmaconsortium.org (Accessed on February 2021) 

42. Manickam K , Buchanan AH , Schwartz MLB , et al. Exome Sequencing-Based Screening for BRCA1/2 Expected Pathogenic Variants Among Adult Biobank Participants. JAMA Netw Open. 2018;1(5):e182140. Published 2018 Sep 7. 

43. Collaborative Group on Hormonal Factors in Breast Cancer . Familial breast cancer: collaborative reanalysis of individual data from 52 epidemiological studies including 58,209 women with breast cancer and 101,986 women without the disease. Lancet. 2001 Oct;358(9291):1389–99. https://doi.org/10.1016/S0140-6736(01)06524-2  

44. Narod SA , Brunet JS , Ghadirian P , Robson M , Heimdal K , Neuhausen SL , et al.; Hereditary Breast Cancer Clinical Study Group . Tamoxifen and risk of contralateral breast cancer in BRCA1 and BRCA2 mutation carriers: a case-control study. Lancet. 2000 Dec;356(9245):1876–81. https://doi.org/10.1016/S0140-6736(00)03258-X  

45. Phillips KA , Milne RL , Rookus MA , Daly MB , Antoniou AC , Peock S , et al.  Tamoxifen and risk of contralateral breast cancer for BRCA1 and BRCA2 mutation carriers. J Clin Oncol. 2013 Sep;31(25):3091–9. https://doi.org/10.1200/JCO.2012.47.8313  

46. Gronwald J , Robidoux A , Kim-Sing C , Tung N , Lynch HT , Foulkes WD , et al.; Hereditary Breast Cancer Clinical Study Group . Duration of tamoxifen use and the risk of contralateral breast cancer in BRCA1 and BRCA2 mutation carriers. Breast Cancer Res Treat. 2014 Jul;146(2):421–7. https://doi.org/10.1007/s10549-014-3026-3  

47. Friebel TM , Domchek SM , Rebbeck TR . Modifiers of cancer risk in BRCA1 and BRCA2 mutation carriers: systematic review and meta-analysis. J Natl Cancer Inst. 2014 Jun;106(6):dju091. https://doi.org/10.1093/jnci/dju091  

48. Domchek SM , Friebel TM , Singer CF , Evans DG , Lynch HT , Isaacs C , et al.  Association of risk-reducing surgery in BRCA1 or BRCA2 mutation carriers with cancer risk and mortality. JAMA. 2010 Sep;304(9):967–75. https://doi.org/10.1001/jama.2010.1237  

49. Heemskerk-Gerritsen BA , Menke-Pluijmers MB , Jager A , Tilanus-Linthorst MM , Koppert LB , Obdeijn IM , et al.  Substantial breast cancer risk reduction and potential survival benefit after bilateral mastectomy when compared with surveillance in healthy BRCA1 and BRCA2 mutation carriers: a prospective analysis. Ann Oncol. 2013 Aug;24(8):2029–35. https://doi.org/10.1093/annonc/mdt134  

50. Metcalfe K , Gershman S , Ghadirian P , et al.  Contralateral mastectomy and survival after breast cancer in carriers of BRCA1 and BRCA2 mutations: retrospective analysis. BMJ. 2014;348:g226. Published 2014 Feb 11. https://doi.org/10.1136/bmj.g226  

51. Greene MH , Piedmonte M , Alberts D , Gail M , Hensley M , Miner Z , et al.  A prospective study of risk-reducing salpingo-oophorectomy and longitudinal CA-125 screening among women at increased genetic risk of ovarian cancer: design and baseline characteristics: a Gynecologic Oncology Group study. Cancer Epidemiol Biomarkers Prev. 2008 Mar;17(3):594–604. https://doi.org/10.1158/1055-9965.EPI-07-2703  

52. Finch AP , Lubinski J , Møller P , Singer CF , Karlan B , Senter L , et al.  Impact of oophorectomy on cancer incidence and mortality in women with a BRCA1 or BRCA2 mutation. J Clin Oncol. 2014 May;32(15):1547–53. https://doi.org/10.1200/JCO.2013.53.2820  

53. Marchetti C , De Felice F , Palaia I , Perniola G , Musella A , Musio D , et al.  Risk-reducing salpingo-oophorectomy: a meta-analysis on impact on ovarian cancer risk and all cause mortality in BRCA 1 and BRCA 2 mutation carriers. BMC Womens Health. 2014 Dec;14(1):150. https://doi.org/10.1186/s12905-014-0150-5  

54. Byrski T , Huzarski T , Dent R , Gronwald J , Zuziak D , Cybulski C , et al.  Response to neoadjuvant therapy with cisplatin in BRCA1-positive breast cancer patients. Breast Cancer Res Treat. 2009 May;115(2):359–63. https://doi.org/10.1007/s10549-008-0128-9  

55. Fong PC , Boss DS , Yap TA , Tutt A , Wu P , Mergui-Roelvink M , et al.  Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers. N Engl J Med. 2009 Jul;361(2):123–34. https://doi.org/10.1056/NEJMoa0900212  

56. Tutt A , Robson M , Garber JE , Domchek SM , Audeh MW , Weitzel JN , et al.  Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and advanced breast cancer: a proof-of-concept trial. Lancet. 2010 Jul;376(9737):235–44. https://doi.org/10.1016/S0140-6736(10)60892-6  

57. Sandhu SK , Schelman WR , Wilding G , Moreno V , Baird RD , Miranda S , et al.  The poly(ADP-ribose) polymerase inhibitor niraparib (MK4827) in BRCA mutation carriers and patients with sporadic cancer: a phase 1 dose-escalation trial. Lancet Oncol. 2013 Aug;14(9):882–92. https://doi.org/10.1016/S1470-2045(13)70240-7  

58. Rodler ET , Kurland BF , Griffin M , Gralow JR , Porter P , Yeh RF , et al.  Phase I Study of Veliparib (ABT-888) Combined with Cisplatin and Vinorelbine in Advanced Triple-Negative Breast Cancer and/or BRCA Mutation-Associated Breast Cancer. Clin Cancer Res. 2016 Jun;22(12):2855–64. https://doi.org/10.1158/1078-0432.CCR-15-2137  

59. Audeh MW , Carmichael J , Penson RT , Friedlander M , Powell B , Bell-McGuinn KM , et al.  Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and recurrent ovarian cancer: a proof-of-concept trial. Lancet. 2010 Jul;376(9737):245–51. https://doi.org/10.1016/S0140-6736(10)60893-8  

60. Litton JK , Scoggins M , Ramirez DL , et al.  A feasibility study of neoadjuvant talazoparib for operable breast cancer patients with a germline BRCA mutation demonstrates marked activity. NPJ Breast Cancer. 2017;3:49. Published 2017 Dec 6. https://doi.org/10.1038/s41523-017-0052-4  

61. Tuffaha HW , Mitchell A , Ward RL , Connelly L , Butler JR , Norris S , et al.  Cost-effectiveness analysis of germ-line BRCA testing in women with breast cancer and cascade testing in family members of mutation carriers. Genet Med. 2018 Sep;20(9):985–94. https://doi.org/10.1038/gim.2017.231  

62. Slade I , Hanson H , George A , Kohut K , Strydom A , Wordsworth S , et al.; MCG programme . A cost analysis of a cancer genetic service model in the UK. J Community Genet. 2016 Jul;7(3):185–94. https://doi.org/10.1007/s12687-016-0266-4  

63. Sun L , Brentnall A , Patel S , Buist DS , Bowles EJ , Evans DG , et al.  A Cost-effectiveness Analysis of Multigene Testing for All Patients With Breast Cancer [published online ahead of print, 2019 Oct 3]. JAMA Oncol. 2019 Oct;5(12):1718–30. https://doi.org/10.1001/jamaoncol.2019.3323  

64. Eccleston A , Bentley A , Dyer M , Strydom A , Vereecken W , George A , et al.  A Cost-Effectiveness Evaluation of Germline BRCA1 and BRCA2 Testing in UK Women with Ovarian Cancer. Value Health. 2017 Apr;20(4):567–76. https://doi.org/10.1016/j.jval.2017.01.004  

65. Asphaug L , Melberg HO . The Cost-Effectiveness of Multigene Panel Testing for Hereditary Breast and Ovarian Cancer in Norway. MDM Policy Pract. 2019 Feb;4(1):2381468318821103. https://doi.org/10.1177/2381468318821103  

66. Dhawan MS , Bartelink IH , Aggarwal RR , Leng J , Zhang JZ , Pawlowska N , et al.  Differential Toxicity in Patients with and without DNA Repair Mutations: Phase I Study of Carboplatin and Talazoparib in Advanced Solid Tumors [published correction appears in Clin Cancer Res. 2018 Feb 15;24(4):985]. Clin Cancer Res. 2017 Nov;23(21):6400–10. https://doi.org/10.1158/1078-0432.CCR-17-0703  

67. Mgbemena VE , Signer RA , Wijayatunge R , Laxson T , Morrison SJ , Ross TS . Distinct Brca1 Mutations Differentially Reduce Hematopoietic Stem Cell Function. Cell Rep. 2017 Jan;18(4):947–60. https://doi.org/10.1016/j.celrep.2016.12.075  

68. Lakhani SR , Van De Vijver MJ , Jacquemier J , Anderson TJ , Osin PP , McGuffog L , 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 May;20(9):2310–8. https://doi.org/10.1200/JCO.2002.09.023  

69. Atchley DP , Albarracin CT , Lopez A , Valero V , Amos CI , Gonzalez-Angulo AM , et al.  Clinical and pathologic characteristics of patients with BRCA-positive and BRCA-negative breast cancer. J Clin Oncol. 2008 Sep;26(26):4282–8. https://doi.org/10.1200/JCO.2008.16.6231  

70. Mavaddat N , Barrowdale D , Andrulis IL , Domchek SM , Eccles D , Nevanlinna H , 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 Jan;21(1):134–47. https://doi.org/10.1158/1055-9965.EPI-11-0775  

71. Goodwin PJ , Phillips KA , West DW , Ennis M , Hopper JL , John EM , et al.  Breast cancer prognosis in BRCA1 and BRCA2 mutation carriers: an International Prospective Breast Cancer Family Registry population-based cohort study. J Clin Oncol. 2012 Jan;30(1):19–26. https://doi.org/10.1200/JCO.2010.33.0068  

72. Vig HS , McCarthy AM , Liao K , Demeter MB , Fredericks T , Armstrong K . Age at diagnosis may trump family history in driving BRCA testing in a population of breast cancer patients. Cancer Epidemiol Biomarkers Prev. 2013 Oct;22(10):1778–85. https://doi.org/10.1158/1055-9965.EPI-13-0426  

Supplementary tables and figure

Figure S1 Patient selection flow diagram.