ABSTRACT

Aim

Approximately 5-10% of breast cancer (BC) cases are hereditary, most frequently associated with germline variants in homologous recombination repair (HRR) genes such as BRCA1/2. However, non-BRCAHRR gene alterations, including ATM gene and CHEK2, may also influence tumor biology and clinical outcomes. This study aimed to evaluate the clinical relevance of germline homologous recombination deficiency (HRD)-related variants in BC and to compare the clinicopathological features and survival outcomes between BRCA and non-BRCA carriers.

Methods

A retrospective cohort of 148 BC patients with germline HRD-related variants identified by next-generation sequencing between 2018 and 2022 was analyzed. Variants were classified according to the American College of Medical Genetics and Genomics/Association for Molecular Pathology 2015 criteria. Clinical, pathological, and survival data were assessed using descriptive statistics and Kaplan-Meier survival analysis with the Log-Rank test.

Results

Of 148 patients (mean age 45.2±10.1 years), 80 (54%) carried BRCA variants and 68 (46%) non-BRCA variants, most frequently ATM and CHEK2. Pathogenic or likely pathogenic variants were more frequent in BRCA carriers (77.5% vs. 58.8%, p=0.014). Disease-free survival (DFS) did not differ significantly between BRCA and non-BRCA groups (p=0.42). Prophylactic mastectomy and oophorectomy were performed significantly more often in BRCA carriers (p<0.05).

Conclusions

Although DFS was comparable between BRCA and non-BRCA carriers, relapse was more frequent in BRCA1 and pathogenic variant carriers. These results emphasize the clinical importance of integrating germline HRR gene analysis into personalized surveillance and management strategies in BC.

Introduction

Breast cancer (BC) is the most frequent type of cancer among women, and approximately 5-10% of BC cases are attributable to hereditary factors. BRCA1/2 are tumor suppressor genes that belong to the homologous recombination repair (HRR) family, which plays a crucial role in repairing deoxyribonucleic acid (DNA) double-strand breaks, which are harmful to cells. In addition to BRCA1 and BRCA2, the HRR gene group includes genes such as ATM, CHEK2, PALB2 (partner and localizer of BRCA2), and RAD51 (RAD51 recombinase). If a pathogenic variant is found in any of the HRR genes, the proteins encoded by these genes cannot perform their normal functions, resulting in homologous recombination deficiency (HRD) [1].

Women carrying pathogenic variants in BRCA1 have a lifetime risk of BC of 50-80% and of ovarian cancer of 20-50%.
For the BRCA2 gene, the lifetime risks are 50% for BC and 20% for ovarian cancer [2]. In addition to BRCA germline variants, alterations can be detected in other genes associated with HRD, including those with high penetrance (CDH1, TP53, PTEN) and those with moderate penetrance (PALB2, ATM, CHEK2, BRCA1, BRIP1) [3, 4]. These variants have been reported to increase BC risk by up to fourfold compared with the general population [5]. ATM and CHEK2 variants, which, along with the PALB2 variant, play a role in DNA repair, are important in the context of hereditary BC [6, 7]. In addition to known pathogenic variants, variants of uncertain clinical significance (VUS) may be detected [8]. Individuals with these variants may develop other malignancies alongside BC, presenting with substantially different pathological and clinical profiles [9]. Current guidelines emphasize that recommendations for treatment or preventive surgery based on these types of variants should not be made [10]. However, evaluating germline HRD-related genes is crucial for improving clinical outcomes through personalized therapies, such as targeted poly (ADP-ribose) polymerase (PARP) inhibitors for various cancers, particularly breast and ovarian cancers [11, 12].

Although several studies have collected data on various malignancies associated with BRCA variants in our country [11, 12], no study has clinically and pathologically evaluated BRCA and non-BRCA variants together, particularly among BC patients. This study aims to evaluate the clinical utility of germline variant analysis in BC for diagnosis and follow-up, and to determine its contribution to understanding the clinical significance of non-BRCA variants.

Methods

Study Design

This study had a retrospective design. All procedures and measurements in this study were performed according to the 1964 Helsinki Declaration or comparable ethical standards. Ethical board approval was granted by the Ethical Board for Clinical Studies at University of Health Sciences Türkiye, Tepecik Education and Research Hospital, with the following protocol number: approval number: 2021/05-37, date: 17.05.2021.

Patients

Data from patients diagnosed with BC, found to have germline HRD gene variants, and referred to the medical oncology clinic at University of Health Sciences, İzmir Tepecik Training and Research Faculty of Medicine Hospital were screened retrospectively between 2018 and 2022. The inclusion criteria were: (1) a confirmed diagnosis of BC, (2) age over 18 years, and (3) the presence of pathogenic germline variants. Patients whose clinical data were unavailable or whose germline variant analysis results were incomplete were excluded from the study. A total of 148 patients met the eligibility criteria and were included in the final analysis.

Genetic Testing

Peripheral blood samples were collected from patients for germline DNA extraction. DNA isolation was performed using standard commercial kits, following the manufacturer’s instructions. The extracted DNA was analyzed with next-generation sequencing to identify germline variants in genes related to HRR and hereditary cancer predisposition. Only variants in HRR-associated genes were considered for analysis. Targeted sequencing included BRCA1, BRCA2, APC, ATM, BARD1, BLM, BMPR1A, BRIP1, CDH1, CDK4, CDKN2A, CHEK2, EPCAM, MLH1, MRE11, MSH2, MSH6, MUTYH, NBN, PALB2, PMS2, PRSS1, PTEN, RAD50, RAD51C, RAD51D, SLX4, SMAD4, STK11, TP53, and VHL. Sequence data were processed through a web-based bioinformatics pipeline (Genomize Seq Analysis v16.7.2) using the GRCh37 (hg19) human reference genome. All coding regions and exon-intron boundaries (±20 bp) were analyzed, ensuring at least 20× coverage depth across 100% of targeted areas. Detected variants were filtered based on a minor allele frequency<5% in population databases such as gnomAD, 1000 Genomes, ESP6500, and ExAC. Variant classification followed the ACMG/AMP 2015 guidelines [13]. Pathogenic and likely pathogenic variants were confirmed by Sanger sequencing before reporting. VUS were noted but not considered actionable for clinical decisions in accordance with current recommendations. The analysis excluded 3’ and 5’ untranslated regions, pseudogenes, and other homologous regions where accurate alignment was not feasible. Copy number variations, large deletions or duplications, repeat expansions, and mosaic variants were beyond the detection limits of this method.

Statistical Analysis

The data were expressed as mean and standard deviation for metric variables and as counts and percentages for nominal variables. Normality was assessed using Kolmogorov-Smirnov test and by examining Skewness and Kurtosis values. Univariate survival analysis was performed using the Kaplan-Meier method and the Log-Rank test for dichotomous data. Survival outcome was based on “death” for overall survival (OS) or “relapse” for disease-free survival (DFS) during the follow-up period, respectively. Data were analyzed at the 95% confidence level, and a p-value of 0.05 was considered significant. IBM Statistical Package for the Social Sciences version 27 (Chicago, IL, USA) was used for the analysis.

Results

This study included 148 patients with BC. Only one patient was male. The mean age of the patients was 45.18±10.11 years. The great majority of patients (n=104, 70.3%) were under the age of 40 at the time of diagnosis. 107 of 147 patients were premenopausal. The vast majority of patients (87.1%) had invasive ductal carcinoma. Nearly all patients had histological grades IIIII (97%). Thirteen of 148 patients had malignancies other than BC. By molecular subtype, 35 of 148 patients had triple-negative BC. Over half of the patients reported no HRD-related cancer (n=83, 56.5%). 23.4% and 19.7% of patients underwent prophylactic mastectomy and oophorectomy, respectively. The clinical and demographic characteristics of patients are shown in Table 1.

The median follow-up time of patients was 50.46 months (interquartile range, 25-75: 33.80-81.10). During the follow-up period, 8 of 148 patients (5.4%) died of BC, while 26 (17.6%) experienced a relapse. 80 out of 148 patients had BRCA-related variants. The percentage distribution of BRCA1 and BRCA2 variants within the BRCA-related group was 46.8% and 53.2%, respectively. In the non-BRCA group, the most frequent types of variants were ATM (n=25, 38.5%) and CHEK2 (n=24, 36.9%). Among variants from 148 patients, 102 (68.9%) were classified as likely pathogenic or pathogenic (LP/P), while the remainder were classified as VUS. Comparison of the distribution of variant pathogenicity between BRCA and non-BRCA groups revealed a significant difference (χ²=5.985, p=0.014): 77.5% vs 58.8% LP/P in BRCA and non-BRCA groups, respectively. The frameshift variant was most frequent in the BRCA group (n=25, 31.6%), whereas the missense variant was most frequent in the non-BRCA group (n=40, 58.8%). 80.8% of patients who relapsed had an LP/P variant; however, there was no significant association between relapse and pathogenicity (χ²=2.068, p=0.150), with relapse proportions of 80.8% versus 19.2% in the LP/P and VUS groups, respectively. There were also significant differences between BRCA and non-BRCA groups with respect to prophylactic mastectomy (χ²=7.201, p=0.027; 30.8% vs 14.9% in the BRCA and non-BRCA groups, respectively) and prophylactic oophorectomy [χ²(2)=9.904, p=0.007; 27.5% vs 10.4% in the BRCA and non-BRCA groups, respectively]. Among patients who relapsed, 54.2% had a BRCA1 variant; nearly 70% of those patients had missense or frameshift variants. The Comparison of the clinical characteristics of patients with and without BRCA mutations is shown in Table 2.

[χ²(1)=0.633, p=0.426], prophylactic mastectomy [χ²(1)=1.678, p=0.190], prophylactic oophorectomy [χ²(1)=1.506, p=0.223], histological grade [χ²(2)=0.931, p=0.621], having HRD related cancer [χ²(1)=0.103, p=0.747], menopausal state [χ²(1)=0.048, p=0.822], age groups (<40<) [χ²(1)=0.038, p=0.843] except for variant pathogenity [LP/P vs VUS, χ²(1) =3.907, p=0.048]. Though the non-BRCA group showed higher median DFS time than the BRCA group (47.83 vs 37.13 months), the Log-Rank test was not significant [χ²(1)=0.915, p=0.333]. The DFS also did not differ significantly between pathogenicity groups (LP/P vs P) [χ²(1)=1.402, p=0.235]. Menopausal status, age groups, HRD-related cancer history, prophylactic mastectomy, and oophorectomy were also not significantly associated with DFS (p>0.050).

Discussion

In this retrospective cohort of 148 BC patients, with a median follow-up of more than four years, we observed no significant difference in DFS between BRCA carriers and non-BRCA carriers. Although the non-BRCA group showed a longer median DFS than the BRCA group, this difference did not reach statistical significance. Similarly, relapse was more frequently associated with BRCA1 LP/P variants, but this association was not statistically significant. Prophylactic mastectomy and oophorectomy were significantly more common among BRCA carriers, reflecting current clinical practice guidelines [14, 15]. The distribution of variants in our cohort was consistent with previous studies, showing a higher proportion of BRCA-related variants and a predominance of invasive ductal carcinoma and luminal molecular subtypes [16]. In addition, the most frequently observed variants in the non-BRCA group were ATM and CHEK2 [2, 7]. Age, menopausal status, and histological grade were not significantly associated with DFS in our cohort, a finding that may be explained by the predominance of advanced-stage disease and the relatively small sample size.

Our findings are consistent with several prior studies reporting comparable survival outcomes between BRCA and non-BRCA carriers [17]. Liu et al. [18] reported that patients with BRCA variants have lower OS rates based on analysis of data from nearly 36,000 patients. While some reports suggest that BRCA1 variants may be linked to a worse prognosis due to their association with aggressive tumor biology [17, 19], others have found no significant differences in long-term survival [20, 21]. Nevertheless, it should be noted that BRCA variant carriers have a lifetime risk of BC exceeding 50% [22]. The lack of statistical significance in our study might have originated from limited sample size, heterogeneity in treatment modalities, or the impact of novel therapeutic approaches, including the increasing use of PARP inhibitors and platinum-based chemotherapy, which may mitigate the adverse prognostic impact of BRCA variants. Moreover, data from the literature [23] suggest that many confounding factors beyond BRCA status may influence OS. In our study, nearly half of the patients who experienced relapse were BRCA1 carriers, a finding consistent with previous reports [17, 19].

The distribution of variant pathogenicity in our cohort is also noteworthy. A higher proportion of LP/P variants was detected in BRCA carriers than in non-BRCA patients, suggesting potential clinical implications for genetic counseling and risk-adapted management [14, 15]. On the other hand, the most common variant types differed between groups: frameshift variants predominated in BRCA carriers, whereas missense variants predominated in non-BRCA carriers. Similarly, Yazıcı et al. [24] reported that frameshift variants were the most frequent type, accounting for 63.5% of BRCA1/2 carriers in their study. From a clinical perspective, our results support the notion that genetic testing should not only focus on BRCA variants but also encompass non-BRCA variants such as ATM and CHEK2, which were frequent in our cohort. Toss et al. [25] also reported that these two variants are associated with a higher risk of developing BC. In our study, nearly 40% of patients within groups with ATM and CHEK2 variants had positive family history, which is comparable yet lower than the rates of over 60% reported by Toss et al. [25]. Given the increased risk of BC in ATM and CHEK2 variant carriers, the National Comprehensive Cancer Network recommends annual mammography beginning at age 40 [10]. Moreover, since those variants are also reported to be associated with other malignancies, such as ovarian, colorectal, and kidney cancers, individuals with a positive family history should be thoroughly investigated [26]. Consistent with this observation, six of our patients with other malignancies—most commonly ovarian cancer—carried BRCA variants, while four carried CHEK2 variants. Although the prognostic implications of these variants remain uncertain, their detection has potential value for individualized surveillance strategies and familial risk assessment [25]. Metcalfe et al. [15] reported that nearly 50% of BRCA carriers in their study underwent risk-reducing mastectomy (RRM), and that the risk of death from BC was approximately 1% within 15 years following RRM. The higher rate of prophylactic surgery among BRCA carriers in our study underscores adherence to guideline-based preventive strategies, which may contribute to long-term outcomes regardless of differences in DFS.

In conclusion, our study demonstrated no significant DFS differences between BRCA carriers and non-BRCA carriers in BC, though relapse events were more frequent among carriers of BRCA1 and LP/P variants. These findings highlight the complexity of interpreting genetic variants in BC and the importance of integrating molecular, clinical, and therapeutic factors into prognostic assessment. Larger, multicenter studies with longer follow-up are warranted to validate these observations and to clarify the prognostic impact of BRCA and non-BRCA variants. These findings emphasize the need for broader genetic testing panels in routine clinical practice, particularly in young BC patients without BRCA1/2 variants. Future studies integrating germline and somatic HRD profiles with treatment outcomes could help define molecular prognostic models for BC.

Study Limitations

The strengths of this study include a relatively long follow-up period and a detailed molecular characterization of variants. Several limitations should be acknowledged in this study. The single-center design, relatively small sample size, and inclusion of only one male patient limit the generalizability of our results. Moreover, subgroup analyses (e.g., BRCA1 vs BRCA2, LP/P vs VUS) may have been underpowered to detect small but clinically relevant differences. In addition, since nearly all patients in our study were white women, the generalizability of our results to other ethnic groups may be limited.

Conclusion

Our findings suggest that while DFS did not significantly differ between BRCA and non-BRCA carriers, relapse was more common among BRCA1 pathogenic variant carriers. This highlights the biological and clinical heterogeneity within HRR gene alterations. Incorporating germline HRR gene testing into the diagnostic and follow-up process can support more individualized risk assessment, surveillance, and preventive strategies for BC patients.