Poly(ADP-Ribose) Polymerase Inhibitors (PARPi) for Patients With Advanced Breast Cancer: a Meta-analysis of Randomized Controlled Trials

BACKGROUND: Poly(ADP–ribose) polymerase (PARP) inhibitors are new class of drugs that are currently being studied in several malignancies. However, datas about the ecacy and safety of the PARP inhibitors are limited. Therefore, we conducted a meta-analysis of randomized controlled trials (RCT) in patients with breast cancer. METHODS: Pubmed/Medline, Embase, Cochrane Library, and abstracts presented at the annual meeting of the American Society of Clinical Oncology (ASCO) were searched for articles published from 2000 to June 2018.Summary incidences and the RR, HR with 95% condence intervals, were calculated by using a random-effects or xed-effects model. RESULTS: The summary HR indicated PARPi was not associated with OS (HR=0.83, 95%CI 0.66–1.06, Z=1.49, P=0.14), while it could signicantly improve PFS ande time to deterioration (TTD) of global health status/quality of life(GHS/QoL) as compared with traditional standard therapy, the HR was 0.60(95%CI 0.50-0.72; Z=5.52, P (cid:0) 0.00001) and 0.4 (95%CI 0.29–0.54,z=5.80 ,p=0.000),respectively.The RR of grade 3 or more anemia ,fatigue and headache was 3.02 (95% CI, 0.69–13.17;p = 0.14,,I 2 =90%),0.77 (95%CI, 0.34–1.73;p=0.52,I 2 =7%) and 1.13 (95% CI,0.30–4.18;p=0.86,I 2 =0%),respectively. CONCLUSION: The ndings of this meta-analysis showed that PARPi has no signicant effect on OS, while it could signicantly improve in PFS and TTD of GHS/QoL for patients with advanced or metastatic breast cancer.Furthermore,our ndings also demonstrated that the PARPi treatment is connected with an increased risk of grade 3 or more anemia adverse events.


Introduction
Breast tumor is the most popular female malignancy, with an estimated 1 million new cases diagnosed in 2012 (a quarter of all cancers [1].In 2016, there were approximately 246 660 women diagnosed with breast tumor within America. Recent decades have shown signi cant progress with the therapy of breast tumor. Human epidermal growth factor receptor type 2 (HER2)-targeted therapies and Endocrine treatment have received a substantial improvement in overall survival(OS) in patients with HER2-positive and hormone receptor(HR) positive breast tumor [1]. Despite extensive research, triple-negative (estrogen receptor (ER),progesterone receptor (PR) negative,and HER2-negative) breast cancers (TNBC), still lack speci c methods for therapy [3].At present, the mainstay of treatment for triple-negative breast cancer and BRCA associated triple-negative is still chemotherapy.
Poly (adenosine diphosphate-ribose) polymerase (PARP) is a novel therapeutic target in cancer,which can bind to DNA breaks and attract DNA repair proteins to the site of damage to facilitates DNA repair [4].Poly (adenosine diphosphate-ribose) polymerase (PARP) is a novel therapeutic target in cancer,which can bind to DNA breaks and attract DNA repair proteins to the site of damage to facilitates DNA repair [5].Those people whose breast cancer susceptibility genes (BRCA1 and BRCA2) mutations,have a high risk of developing breast tumor [6].
PARP inhibitors have played a crucial role in the repair of DNA single-strand breaks by the base excision repair (BER) pathway [7]. PARP inhibitors are oral small-molecule inhibitors of PARP enzymes 1, 2 ,and 3,which can improve the clinical activity of DNA-damaging Chemotherapy by blocking DNA damage repair [8]. PARP inhibition could cause "synthetic lethality" in BRCA mutant tumor cells,which leads to profound cytotoxicity to tumor cells while sparing healthy cells [4,9].Because only the tumors in BRCA 1/2 mutated patients are defective in homologous recombination,therapy with PARP inhibitors appears to be highly tumour-speci c [4].
Olaparib is a newfangled of PARP inhibitor which can induce synthetic lethality in homozygous BRCAde cient cells [10].Olaparib was permitted to treat advanced ovarian cancer patients with malignant germline BRCA-mutated, who have been treated with three or more prior lines of chemotherapy [11].Furthermore the SOLO1 trial showed that BRCA1/2 mutation and newly diagnosed advanced ovarian cancer patients who have accepted platinum-based chemotherapy, the use of maintenance therapy with olaparib would provide a substantial bene t concerning progression free survival(PFS) [12].The SOLO1 trial ,which approved olaparib as maintenance therapy of ovarian cancer patients who were sensitive to platinum therapy.
The OlympiAD trial demonstrated that compared to standard therapy,olaparib provided a remarkable bene t among patients with a germline BRCA mutation and HER2-negative metastatic breast tumor [13].OlympiAD trial that gave the approval of olaparib by the FDA for the therapy of patients with BRCA-mutated HER2-negative metastatic breast tumor patients [14].
Talazoparib, is another oral inhibitor of PRAP.The phase III trial EMBRACA showed that compared to standard chemotherapy, single-agent talazoparib provided a signi cant bene t concerning PFS among patients with germline BRCA1/2 mutation advanced breast cancer.Talazoparib was recently approved as the therapy of people with malignant or suspected malignant germline BRCA-mutated, HER2-negative, locally advanced or metastatic breast cancer by FDA and the European Medicines Agency (EMA) [15,16].
As a new drug, it is vital to understand its adverse effects pro le better to optimize its use. However, data about the e cacy and safety of the PARPi are limited. Few studies have investigated the PARPi in any systematic way. Therefore, the purpose of this meta-analysis is to review recent research to evaluate the e cacy and safety of PARP inhibitors.

Data source
A literature search was independently performed by two investigators using electronic databases including Pubmed/Medline, Embase and Cochrane Library, using the following search keywords: "Olaparib" or "PARPi" or "Talazoparib" and"Randomized" and "Advanced Breast Cancer" restricted to randomized controlled trials (RCT) published in English. We also searched abstracts and virtual meeting presentations from the American Society of Clinical Oncology (http://www.asco.org/ASCO) conferences held between January 2000 and September 2019. What , s more, we searched the clinical trial registration website (http://www.ClinicalTrials.gov) to obtain information on prospective registration trials. When more than one publication was identi ed from the same clinical trial, we used the most recent or complete report of that trial. When data could not be determined, we would contact the authors.

Selection criteria
Studies should meet the following criteria of eligibility: 1) articles published in English language, 2) prospective phase II or III randomized controlled trials (RCT) design in advanced breast cancer patients; 3) participants assigned to treatment with PARPi or control (placebo, chemotherapy or other therapies)and,4)studies with available safety data reporting adverse events.

Data extraction and quality assessment
Two experienced investigators independently performed data extraction [17]. Different variables from each eligible study were extracted: rst author's name, publication year, trial phase, number of enrolled patients, dosing schedules used in PARPi and control arms, CTCAE (Common Terminology Criteria for Adverse Events) version, median progression-free survival, median overall survival,number of patients for analysis and number of the selected adverse events. We assessed the methodological quality of the included trials using the Cochrane Collaboration tool [18]. Studies were graded as having a "low risk ", "high risk",or "unclear risk" of bias across the seven speci ed do-mains. We also used the seven-point Jadad ranking system that included randomization, double-blinding, and withdrawals, a practice in agreement with other meta-analyses done in this context [19].

Statistical analysis
In this meta-analysis, the software Stata 15.0 and review manager 5.3 as used to pool the results of eligible studies. The H.R. with it,s 95% CI was used to evaluate the in uence strength of the PARPi on the prognosis of advanced breast cancer patients for each study. Q statistics and I 2 statistics evaluated the heterogeneity among the independent studies [20], and it is considered to be heterogeneous if P-value 0.1 or I 2 50 %. The xed-effects model would be adopted if there was no heterogeneity among studies; otherwise, we would use the random-effects model [21]. Publication bias was evaluated through funnel plots and quanti ed by the Begg [22]and Egger tests. A symmetrical plot and P-value of Egger,s test ( 0.05) suggested that there was no publication bias in all studies [23].

Search results
The search strategy yielded a total of 130 publications. Fig. 1 provides an overview of the literature search and study selection. After removal of duplicates, 82 studies remained for the screening of titles and abstracts. Of these, 31 potentially relevant publications were selected for full-text reading. After the full-text screening, 3 RCTs reported in 8 publications were included in the systematic review [24][25][26][27][28][29][30][31], and all of them provided su cient data for quantitative synthesis of the overall survival of advanced breast cancer patients by meta-analysis.

Study quality
All included trials were randomized, with 1 phase 2 trials [30] 2 phase 3 trials [24,29] were published in full manuscript, and one trial was presented at the ASCO meeting [32]. Using the Cochrane Collaboration tool for risk of bias classi cation, we found the quality of the included studies to be generally good and fair( g.2 ). Trials were also ranked for the Jadad score, with all trials categorized as being of good quality (achieving a score of 4 or more) [19]. All three trials reported adverse events according to the National Cancer Institute's CTCAE version 3 or 4 criteria.

Patients
The baseline characteristics of the enrolled studies are presented in table 1. A total of 926 patients were available for the meta-analysis (PARPi: 587; controls/placebo: 339)with median age ranging from 44 to 50 years. The most commonly mutated gene was BRCA1, in 757 of patients, followed by BRCA2 in 329. All three included studies reported safety data on adverse events and were included in the analysis. In one trial, randomization was between chemotherapy/control and PARPi olaparib [33]. Another study, PARPi Talazoparib, was compared to Standard-Therapy for advanced breast cancer [24].Since there are few randomized controlled trials related to PARPi monotherapy for advanced breast cancer, to improve the reliability of our conclusion,we added another trial that PARPi Veliparib was combined with Carboplatin/Paclitaxel [30]. All of the three included studies provided OS and PFS data of patients with breast cancer. (table1).

Data synthesis
The primary endpoint of OS in patients with advanced breast cancer was evaluated in 926 patients.No signi cant heterogeneity was found (chi-squared=0.54,df=2,P=0.76,I 2 =0%), and xed effects model was utilized to calculate the pooled HR and 95% CI. The combined results indicated that PARPi was associated with lower mortality (HR 0.83, 95%CI 0.66-1.06, Z=1.49, P=0.14 de cient quality of evidence). Patients with advanced breast cancer who received PARPi had 17% lower mortality than those who received other Standard-Therapy during the study period, but this difference was not statistically signi cant ( g 3.) The primary endpoint of progression-free survival in patients with advanced breast cancer was evaluated in 926 patients. No signi cant heterogeneity was found (chi-squared=2.55,df=2,P=0.28,I 2 =22%), so we used xed-effects model to calculate the pooled HR and 95% CI.The pooled results demonstrated that PARPi group could signi cantly decrease the risk of death from patients with advanced breast cancer compared with the control group, which was only 60% of the control group(HR 0.60,95%CI 0.50 to 0.72; Z=5.52, P 0.00001) and the results were statistically signi cant. ( g4.) Two studies assessed the time to deterioration (TTD) of global health status/quality of life(GHS/QoL),the HR for TTD was 0.4 (95%CI 0.29-0.54, Supplementary Fig.1).No signi cant heterogeneity was found( P= 0.671,I 2 =0%),the pooled HR suggested that PARPi treatment can lead to a statistically signi cant delay in TTD in patients with advanced breast cancer .

Sensitivity Analysis
The root of heterogeneity was assessed by sensitivity analysis, and there was not apparent heterogeneity was detected as described above. Besides, the pooled HR was not affected by any individual study, as shown in sensitivity analyses, which demonstrated that the results of our meta-analysis were robust and stable( g.5.)

Publication bias
No evidence of publication bias was detected for OS or PFS of PARPi group by either the Begg's or the Egger's test (for PFS of PARPi group Begg's P =0.296 and Egger's P = 0.116; for OS of PARPi group Begg's P = 1.000 and Egger's P = 0.653) [22,23].

Discussion
The meta-analysis identi ed 3 RCTs involved 926 advanced or metastatic breast cancer patients. The summary HRs indicated PARP inhibitors were not associated with OS (HR= 0.83, 95%CI 0.66-1.06, Z=1.49, P=0.14 shallow quality of evidence), while it could signi cantly improve PFS compared to patients treated with traditional standard therapy(HR= 0.60,95%CI 0.50 -0.72; Z=5.52, P 0.00001).Moreover, according to results of global health status/quality of life(GHS/QoL),we also observed a signi cant delay in time to deterioration (TTD) compared to patients who were treated with traditional standard chemotherapy.
Concerning all-grade and grade 3 or more adverse events,there are not statistically remarkable differences between PARP inhibitors and control group.The relative risk (RR) of all-grade and grade 3 or more adverse events compared to control group was 1.02 (95% CI,0.99-1.04; p= 0.16) and 0.90 (95% CI, 0.78-1.03; p = 0.13) (Fig. 8). signi cantly increased risk of all-grade and grade 3 or more anemia events associated with PARPi treatment in advanced or metastatic breast cancer patients. The RR of all-grade and grade 3 or more anemia was RR=2.09 95% CI, 1.10-3.97;p=0.02,I 2 =82%) and 5.88 (95% CI, 3-10.36;p 0.00001,I 2 =7%),respectively.Whereas there was signi cant heterogeneity in all-grade anemia.
Our ndings also suggest that the PARP inhibitors treatment are connected with a growing risk of allgrade headache ,the pooled RR was 1.29 (95% CI,1.01-1.65;p =0.04,I 2 =0%),the results were statistically signi cant.
MEDIOLA Trial ,a Phase II study of olaparib and durvalumab in patients with germline BRCA-mutated metastatic breast tumor.The updated results of MEDIOLA Trial showed that the most common grade 3 or more adverse events were anemia ,neutropenia and pancreatitis.And the data also suggest that patients with two prior lines of chemotherapy had shorter median PFS ,median duration of response (mDoR), and median OS ,lower objective response rate (ORR) than those with fewer prior lines of chemotherapy [34].Furthermore,the updated results in the phase 3 EMBRACA and phase 2 ABRAZO trial showed that a higher TALA exposure was related to a higher risk of anemia and thrombocytopenia [35].All these ndings support the conclusion of this meta-analysis.
About EORTC QLQ-C30 symptoms and functioning subscales,the EMBRACA phase III trial showed that compared to the PCT arm,patients who treated with TALA signi cantly delayed TTD in all symptoms [25].Meanwhile,there is similar result in the OlympiAD trial,except vomiting/nausea symptom score compared with TPC arm [36].
Our work has some limitations. First, there is signi cant heterogeneity between PARPi and all-grade fatigue,we failed to nd the source of heterogeneity.Several reasons may account for the heterogeneity;1) Differences in PARP inhibitors type and the dose may lead to heterogeneity; 2)fatigue is very subjective[37],the evaluated patient-reported outcomes in this regard seem to be better. Furthermore, we used adverse events data reported according to the NCI CTCAE grading system that also includes the subjectivity of the investigators when grading the event. All of these may partly explain the signi cant heterogeneity between studies. Second, randomized controlled trials about PARP inhibitors for patients with advanced breast cancer are too limited to perform a subgroup analysis study.
In conclusion,the ndings of this meta-analysis showed that PARP inhibitors have no signi cant effect on OS, while it could signi cantly improve in PFS and TTD of GHS/QoL for patients with advanced or metastatic breast cancer. Furthermore, this meta-analysis also suggested that PARP inhibitors are associated with a signi cant increase in the risk of all-grade headache and anemia adverse events .Our ndings demonstrated that the PARP inhibitors treatment is connected with an increased risk of grade 3 or more anemia adverse events.

Declarations
Funding This research did not receive any speci c grant from funding agencies in the public, commercial, or notfor-pro t sectors.

Disclosure
The authors declared no con icts of interest.     Relative risk of all-grade (a) and grade 3 or more (b) adverse events associated with PARPi versus non-PARPi.

Figure 7
Relative risk of all-grade anemia (a) ,fatigue (b) and headache (c)events associated with PARPi versus placebo/control in strati ed analysis after excluding trial thatPARPi was in combination with chemotherapy(a1 )and( b1).

Figure 8
Relative risk of grade 3 or more anemia (a) fatigue (b) and headache (c)events associated with PARPi versus placebo/control in strati ed analysis, (a1)when evaluating only trials that PARPi was given as monotherapy.