Homologous Recombination Deficiency (HRD) and BRCA 1/2 Gene Mutation for Predicting the Effect of Platinum-Based Neoadjuvant Chemotherapy of Early-Stage Triple-Negative Breast Cancer (TNBC): A Systematic Review and Meta-Analysis

Background: Platinum-based agents may benefit patients with triple-negative breast cancer (TNBC) whose tumors are dysfunctional in DNA repair mechanisms associated with the homologous recombination repair (HRR) genes. The purpose of this meta-analysis was to assess the values of BRCA1/2 and homologous recombination deficiency (HRD) in the prediction of the pathological complete response (pCR) rates of patients with TNBC treated with platinum-based neoadjuvant chemotherapy (NAC). Patients and Methods: Patients with TNBC with BRCA or HRD status from platinum-based NAC trials were analyzed. The odds ratios (ORs) with 95% confidence intervals (CI) for the identified studies were calculated. Results: 13 eligible studies between January 2000 and September 2021 were included through systematic literature searches of Embase, PubMed, Cochrane, and Web of Science databases. In 12 trials with BRCA status, 629 of 1266 (49.7%) patients with TNBC achieved pCR with platinum-based NAC, including 134 out of 222 (60.4%) BRCA1/2-mutated patients and 495 out of 1044 (47.4%) BRCA wildtype patients (OR, 1.62; 95% CI, 1.20–2.20). The prevalence of HRD was higher than BRCA1/2 mutations in patients with TNBC (69.2% vs. 17.5%). In six trials with HRD information, pCR rates of HRD-positive patients with TNBC were significantly higher than those of HRD-negative patients with TNBC (241/412, 58.5% vs. 60/183, 32.8%, OR, 3.01; 95% CI, 2.07–4.39, p < 0.001). Conclusions: BRCA1/2-mutated and HRD-positive patients with TNBC could benefit from platinum-based NAC. In the future, a prospective study using unified HRD testing criteria is warranted for further investigation.


Introduction
Triple-negative breast cancer (TNBC) refers to a molecular subtype of breast cancer (BC) with negative estrogen receptor, progesterone hormone receptor, and negative human epidermal growth factor receptor 2, accounting for 15-20% of breast tumors [1]. Patients with TNBC had worse prognoses with relatively inferior disease-free survival (DFS) rates after treatment than other BC subtypes, due to their highly invasive and heterogeneous nature. TNBC treatment is still primarily chemotherapy on account of the scarcity of specific molecular targets. Previous studies have shown that tumor-free survival, as well as overall survival of patients with TNBC who underwent neoadjuvant chemotherapy (NAC) and achieved pathological complete response (pCR), is significantly improved, compared with Eligible studies were identified through systematic literature searches of Embase, PubMed, Cochrane databases, and Web of Science databases using the date limits January 2000 and September 2021, with no language restriction. Meanwhile, relevant references and guidelines were hand searched, and tracked references also included those of studies with supplementary data. The search strategy was conducted using the following keywords: "breast cancer", "triple-negative", "TNBC", "platinum", "neoadjuvant chemotherapy", "BRCA", "HRD", "homologous recombination deficiency", etc. Boolean operators were used for searches, which combined specific keywords and free text terms.
This systematic review and meta-analysis was performed according to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [21].
The full protocol is published on the PROSPERO website in November 2021 with the PROSPERO registration number CRD42021279654.

Data Extraction and Quality Assessment
Study selection, extraction of data onto standard forms according to the PRISMA Statement, and cross-checking were conducted independently by two investigators. In case of disagreements, another researcher would be involved in the discussion until resolved. In the process of screening the literature, the title of the literature was read first, and abstracts and full-text were read further to determine eligibility. If necessary, the authors of the original study could be contacted via e-mail or phone for information that had not been confirmed but was relevant to the study. The risk of bias was assessed using the Cochrane Collaboration tool. The quality assessment included selection bias (generating random sequences and blind attribution), attrition bias (selective reporting of incomplete outcome data), performance and selection bias (blinding of participants, healthcare provider, blinding of outcome assessors), and other possible sources of bias in each study. The risk of bias was classified as low, high, or unclear.
Data extracted included the following: (1) basic information of eligible research: research title, first author, published year, study design, trial phase, published journal, ClinicalTrial.gov (accessed on 29 December 2021) number; (2) basic information of the participants: age, number of patients by BRCA mutational status and HRD status, cancer type and stage; (3) intervention measures: platinum type, frequency of administration, dosage, route; (4) number of patients with pCR; (5) key elements of bias risk evaluation.
HRD positive was defined as having a tumor with a predefined HRD score ≥ cut-off value mentioned in the clinical trial and/or a mutation in BRCA1/2, while HRD negative was an HRD score < cut-off value and/or BRCA 1/2 wildtype.

Statistical Analysis
The primary endpoint for assessing antitumor efficacy was the pCR rate. pCR was confirmed by pathological evaluation. The definitions of pCR in the included studies are shown in Supplemental Table S1. The statistics utilized to compare the treatment effect between two subgroups (BRCA 1/2 mutation vs. BRCA 1/2 wildtype or HRD positive vs. HRD negative) was the odds ratio (OR), with a 95% confidence interval (CI). An OR > 1 indicates a higher pCR rate, whereas an OR < 1 indicates a lower pCR rate in the platinum-based NAC group. A chi-squared test was utilized to compare the differences in pCR rates between the study and control groups.
The results of this study were graphically represented through forest plots. A p-value < 0.05 was considered statistically significant. The Higgins I 2 index was computed to confirm the homogeneity of the study results. A random-effects model (DerSimonian and Laird method) was used when I 2 is above 50%; otherwise, a fixed-effect model (Mantel-Haenszel) method was used. Funnel plots were used to assess publication bias, and based on their results, publication bias in this study was not statistically significant ( Figure 1). Meta-analysis was conducted using Review Manager software (RevMan, version 5.3 for Windows; Cochrane Collaboration, Oxford, UK).
Analyses of the GeparSixto randomized clinical trials were conducted by Hahnen et al., who explored whether BRCA1/2 mutation status affected treatment response in patients with TNBC in 2017, and by Loibl et al., who investigated the HRD status as a predictor of response in 2018 [27,28]. Similarly, analyses of the BrighTNess randomized clinical trials were conducted by Loibl et al., exploring whether BRCA1/2 mutation status affected treatment response in patients with TNBC, and by Telli et al., who investigated the HRD status as a predictor of response in 2018 [26,27].

Association of HRD Status with pCR Rates in Four HRD-Predefined Studies
A total of 6 studies with 241 patients had compared pCR rates of HRD-positive patients, with TNBC (n = 412), compared with 60 of HRD-negative patients (n = 183) ( Figure 3B). The definition of HRD positive was different in these six included studies due to different detection methods. NCT01372579 trial illustrated that 12 patients had HRD-positive results, compared with 14 patients who did not. All patients underwent carboplatin + eribulin. The positive HRD was defined as ≥42, which was derived from the quantitative sum of the telomeric allelic imbalance (TAI), large-scale state transition (LST), and loss of heterozygosity (LOH) scores. Compared with noncarriers, these patients had significantly higher pCR rate (75.0% vs. 14.3%, p = 0.02) [34].
GeparSixto trial included patients with TNBC treated with carboplatin + paclitaxel + doxorubicin + bevacizumab. The HRD-positive outcome was defined as an HRD score of ≥42 based on genome-wide copy number and LOH profiling on tumor DNA. In total, 74 patients with TNBC had HRD-positive scores, and 27 patients had HRD-negative scores. Patients with HRD-positive results had higher pCR rates when compared with the matched controls (64.9% vs. 40.7%, p < 0.001) [27].

Association of HRD Status with pCR Rates in Four HRD-Predefined Studies
A total of 6 studies with 241 patients had compared pCR rates of HRD-positive patients, with TNBC (n = 412), compared with 60 of HRD-negative patients (n = 183) ( Figure  3B). The definition of HRD positive was different in these six included studies due to different detection methods. NCT01372579 trial illustrated that 12 patients had HRD-positive results, compared with 14 patients who did not. All patients underwent carboplatin + eribulin. The positive HRD was defined as ≥ 42, which was derived from the quantitative sum of the telomeric allelic imbalance (TAI), large-scale state transition (LST), and loss of heterozygosity (LOH) scores. Compared with noncarriers, these patients had significantly higher pCR rate (75.0% vs. 14.3%, p = 0.02) [34].
GeparSixto trial included patients with TNBC treated with carboplatin + paclitaxel + doxorubicin + bevacizumab. The HRD-positive outcome was defined as an HRD score of ≥ 42 based on genome-wide copy number and LOH profiling on tumor DNA. In total, 74 patients with TNBC had HRD-positive scores, and 27 patients had HRD-negative scores. Patients with HRD-positive results had higher pCR rates when compared with the matched controls (64.9% vs. 40.7%, p < 0.001) [27]. TBCRC 008 included 18 patients with TNBC whose data of HRD and pCR were accessible. Patients were treated with carboplatin + nab-paclitaxel + vorinostat. HRD positive represented an HRD score higher than or equal to a prespecified threshold of 42 without detailed information of the detection method. Overall, 12 patients had HRD-positive scores, and 6 patients had HRD-negative scores. There was no statistically significant difference in pCR between the two groups (66.7% vs. 16.7%, p = 0.131) [37].
BrighTNess trial also reported the predictive role of HRD testing on platinum response in 329 patients with TNBC [30]. The HRD threshold of ≥42 vs. <42 and ≥33 vs. <33 were both assessed. When a threshold of ≥42 was exploited to define an HRD-positive group, 138 of 225 HRD-positive patients with TNBC achieved pCR, whereas 42 of 104 HRDnegative patients achieved pCR (p < 0.01). However, no relationship between HRD-positive and platinum-sensitive results was observed when an HRD threshold of ≥33 vs. <33 (p > 0.05) was used.
TBCRC 030 trial included 56 patients with TNBC whose HRD results were available. Patients were randomized to receive preoperative cisplatin or paclitaxel. Genomic instability was measured by the HRD assay using next-generation sequencing of formalin-fixed paraffin-embedded tumor tissue, with scores of >33 determined to be HRD positive. However, no significant association was observed between HRD scores and pCR rates to either cisplatin or paclitaxel (p > 0.05) [23].

Discussion
In this meta-analysis, we evaluated the roles of BRCA 1/2 and HRD in the prediction of pCR rates of patients with TNBC treated with platinum-based NAC and obtained encouraging results-namely, higher efficacy of platinum-based NAC was observed in BRCA 1/2-mutated/HRD-positive TNBC, compared with BRCA 1/2-wildtype and HRDnegative TNBC.
Platinum, as a DNA-damaging agent, mainly includes cisplatin and carboplatin. The efficacy of platinum-based NAC in the treatment of patients with TNBC is still controversial [38]. In 2019, the Italian Association of Medical Oncology Guidelines on BC supported the addition of platinum to anthracycline/taxane-based NAC for TNBC, which, based on five included studies, increased the probability of pCR rates but not long-term outcomes [39]. BRCA 1/2 gene was one of the topical, proposed, predictive biomarkers for platinum-based regimens in patients with TNBC. Recently, a meta-analysis including seven clinical trials compared the efficacy of platinum-based NAC for patients with TNBC with or without BRCA mutations. The results showed that there was no statistical difference in pCR rates between the two groups of patients (OR = 1.459, 95%C1, 0.95-2.34, p = 0.082), which are not in line with our results, for the possible reason of a relatively small study sample, with 159 BRCA-mutated cases [40]. In the current study, 134 out of 222 (60.4%) BRCA1/2 mutated patients and 495 out of 1044 (47.4%) BRCA-wildtype patients achieved pCR (OR, 1.62; 95% CI, 1.20-2.20). A statistically significant 13.0% increase in pCR rate was observed in BRCA1/2-mutated patients (p = 0.002), which confirmed that BRCA status could strongly predict the efficacy of platinum-based chemotherapy.
HRD is regarded as the loss of the ability of cells to repair the DNA double-stranded breaks via HR when DNA double-stranded breaks occur. HRD positive has emerged as a potentially useful biomarker for PARP inhibitors (a DNA single-strand repair inhibitor) in patients with ovarian cancer [41]. Therefore, it could be inferred that HRD may also be a biomarker for the potential benefits of other drugs that induce DNA breaks (such as platinum derivatives, alkylating agents, mitomycin C, etc.) [42]. HRD-related biomarkers are mainly divided into three categories: (1) HRR gene mutations: BRCA1/2 and other HRR genes; (2) genomic scar detection: based on array comparative genomic hybridization, single nucleotide polymorphism (SNP) (TAI, LST, LOH), or mutation signal; (3) transcriptional expression markers detection: transcript analysis, protein expression, function analysis, etc. HRD score, the most commonly used HRD assessment method, is most widely applied based on three SNP sites (TAI, LST, LOH) within the genome. Different HRD scores would be obtained through different detection methods.
The prediction of response to platinum-based agents based on HDR in patients with TNBC remains controversial. The TNT phase 3 clinical trial demonstrated that patients with first-line relapse of TNBC and a BRCA 1/2 mutation who underwent carboplatin showed more favorable objective response rates (68% vs. 33%, p = 0.03) and PFS (6.8 vs. 4.4 months, p = 0.002) than those of patients who underwent docetaxel [43]. However, such a benefit was not observed in patients with TNBC with BRCA1 methylation, BRCA1 mRNA-low tumors, or a high score in a Myriad HRD assay [43]. The correlation of HRD with pCR rates of patients with TNBC who underwent platinum-based NAC was further investigated in six HRD-predefined studies [23,27,30,[32][33][34]. Patients with TNBC with HRD-positive results were inclined to respond to platinum-based NAC according to the results of a meta-analysis in the current study (58.5% vs. 32.8%, OR, 3.01; 95% CI, 2.07-4.39, p < 0.001). Patients with TNBC with high HRD scores (≥42 in the GeparSixto, NCT01372579, TBCRCR 008, and BrighTNess study) yielded favorable responses to platinum-based NAC (p < 0.05). However, HRD score was not observed as a predictor of pathological response in TNBC in the TBCRC 030 and BrighTNess trial with an HRD score threshold of 33. These results showed that the HRD threshold of 33 had less potential than the HRD threshold of 42 in identifying specified patient populations who might be derived relatively small benefit from platinum-based NAC. Since the scoring method of HRD score was not uniform in various studies, exploring a suitable and universal HRD definition method warrants future investigation. Moreover, the odds ratio for HRD (3.01) as a predictor of pCR rates of patients with TNBC with platinum-based regimens was much higher than the OR for BRCA status (1.62), which suggested that HRD status as a predictor of response to platinum-based NAC in patients with TNBC resulted in larger variation between groups, compared with BRCA 1/2 mutations. In addition, the rate of HRD positive was higher than BRCA 1/2 mutations in patients with TNBC (69.2% vs. 17.5% with an additional detection rate of 51.7%. This also means that nearly half of patients with TNBC may become the dominant group benefiting from platinum-based agents. Only one trial (TBCRC 008) reported that 3 out of 10 (30.0%) HR-positive, HER2-negative luminal B subtype, HRD-positive patients and 1 out of 20 (5.0%) HR-positive, HER2-negative luminal B subtype, HRD-negative patients achieved pCR [32]. The predictive value of HRD for platinum-containing NAC in patients with HR-positive, HER2-negative luminal B subtype could be further explored in the prospective study in the future.
The long-term survival data of patients with TNBC who underwent platinum-based NAC were provided in two included studies in this meta-analysis. GeparSixto trial demonstrated 3-year DFS and 3-year OS rates were higher in patients with TNBC with HRDpositive results who underwent platinum-based NAC than those who were HRD negative (85.7% vs. 77.8; 92.9% vs. 79.9%, p < 0.05) [27]. NeoSTOP trial demonstrated that pCR status was significantly related to event-free survival (EFS) and OS (p < 0.05) of patients with TNBC who underwent platinum-based NAC on multivariate Cox regression. Estimated 3-year EFS and 3-year OS was 100% and 100% in patients with pCR, compared with 81% and 86% in those without pCR (p < 0.003) [25]. It could be inferred that patients with TNBC with BRCA1/2 mutations or HRD positive may be able to obtain survival benefits from platinum-based NAC from the above studies.
No adverse events stratified by BRCA status or HRD status were provided from included studies in this meta-analysis. A meta-analysis published in 2021 involving 14 studies and 3518 patients with TNBC who received NAC or adjuvant chemotherapy proposed that the addition of platinum was related to increased thrombocytopenia and all-grade neuropathy, while neutropenia and grade 3-4 neuropathy did not increase [4].
Nevertheless, HRD scores with different candidate HR genes were calculated by different methods that differed from institute to institute in these four studies, which may have a slight influence on the results [23,27,32,34]. Second, different trials with different choices of chemotherapy agents might bias results. The chemotherapeutic agents used in our enrolled trials included PARP inhibitors, taxanes, anthracyclines, alkylating agents, and platinum. However, only platinum and PARP inhibitors were most associated with the HRR pathway. Of the six trials that had compared pCR rates of HRD-positive patients with TNBC with those of HRD-negative patients, BrighTNess and PrECOG 0105 trials included PARP inhibitors and platinum simultaneously, and we still included these two studies in this meta-analysis due to relatively limited data of HRD in the prediction of the pCR rates of patients with TNBC currently treated with platinum-based NAC. Thirdly, the definitions of pCR, treatment modalities, evaluation criteria, and the technologies to access tumors were not unified in the included studies, although with modest differences, which might cause a certain bias. Fourth, the BRCA1/2 mutation is currently the most common mutation known to cause HRD. As there are relatively few studies (only six trials enrolled in this meta-analysis) that explored the predictive value of HRD for the efficacy of platinum-based NAC in patients with TNBC, and no available data were provided on the response to platinum-based regimens in patients with HRD but without BRCA1/2 mutation in these six studies, the conclusion was based on patients with HRD (with/without BRCA1/2 mutation). The conclusions of our studies could be verified in patients with HRD but without BRCA1/2 mutation in the future. Fifth, the difference in adverse reactions to platinum-based agents between patients stratified by HRD could not be analyzed due to the loss of documents in the included studies. A prospective study with unified HRD testing criteria is warranted in the future.

Conclusions
BRCA1/2-mutated and HRD-positive patients with TNBC could benefit from platinumbased NAC. A prospective study with HRD testing criteria is warranted in the future.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jpm12020323/s1, Table S1: The definitions of pCR in the included studies.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The datasets developed and analyzed in this study are available from the corresponding author upon reasonable request.