A Meta-Analysis of Efficacy and Safety of PD-1/PD-L1 Inhibitors in Triple-Negative Breast Cancer

Objective To evaluate the efficacy and safety of programmed cell death receptor-1 (PD-1)/programmed cell death-ligand protein-1 (PD-L1) inhibitors in triple-negative breast cancer (TNBC) to provide a treatment basis for TNBC. Methods Published case-control studies on PD-1/PD-L1 inhibitors in the treatment of TNBC were retrieved from PubMed, Embase, and Cochrane Library databases, and collected data were processed by RevMan 5.4. Results A total of 7 studies with 4340 study subjects were obtained, including 2092 PD-L1-negative cases, 1375 PD-L1-positive cases, and 847 PD-L1 unidentified cases. The use of PD-1/PD-L1 inhibitors showed no significant impact on patients' progression-free survival (PFS) and overall survival (OS). The use of PD-1/PD-L1 inhibitors in the PD-L1-positive subgroup significantly improved patients' PFS and OS. Treatment with PD-1/PD-L1 inhibitors presented no significant effect on the incidence of adverse events (AEs) but increased the risk of AE grade ≥3 and severe AEs (SAEs). Conclusion PD-1/PD-L1 inhibitors are effective in the treatment of TNBC, which is strongly correlated with the expression of PD-L1; patient selection and clinical application require further investigation and verification.


Introduction
Breast cancer is the uncontrolled proliferation of breast epithelial cells induced by multiple oncogenic factors [1]. Its early manifestations include breast lumps, nipple overflow, and enlarged axillary lymph nodes, and distant metastasis of the disease in the advanced stage may develop multiorgan lesions, which is life-threatening [2]. e molecular types of breast cancer include Ki-67, estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2), which provide diagnostic and prognostic evidence and guidance for treatment [3]. Triplenegative breast cancer (TNBC) is a breast cancer with the inexpression of ER, PR, and HER2, which accounts for 12% of all breast cancer cases and is characterized by a young age of onset, high invasiveness, and easy recurrence and metastasis [4]. e 5-year survival of early-stage breast cancer exceeds 90%, that of patients with early-stage TNBC decreases to 77%, and that of patients with advanced TNBC remains only 14% [5]. e current treatments for TNBC are severely circumscribed owing to the insufficiency of effective therapeutic targets. e main treatment for advanced TNBC relies on chemotherapy with traditional anthracyclines, paclitaxel, and platinum [6] to prolong patients' survival, but its efficiency still leaves much to be desired. Besides, intensive chemotherapy may also seriously compromise patients' quality of life [7].
Programmed death protein-1 (PD-1) and its ligand (PD-L1) inhibitors are immune sentinel monoclonal antibodies with outstanding action range, depth, and durability of response. Immunotherapy with PD-1/PD-L1 inhibitors is a new and highly appreciated anticancer immunotherapy that reactivates the body's immune system to defend against cancer cells by blocking the PD-1/PD-L1 signaling pathway [8,9]. Currently, the marketed PD-1 inhibitors including nivolumab, pembrolizumab, and tislelizumab are mainly used for the treatment of melanoma and non-small-cell lung cancer, and their efficacy against renal cell carcinoma, bladder cancer, and Hodgkin's lymphoma is still under large-scale clinical trials [10,11]. Atezolizumab, durvalumab, and avelumab are PD-L1 inhibitors that have been clinically approved for the treatment of urothelial carcinoma, and many other drugs are currently undergoing early clinical trials [12,13]. Currently, targeted PD-1/PD-L1 inhibitors have been ratified for the treatment of hepatocellular carcinoma, renal cancer, and advanced non-smallcell lung cancer, and atezolizumab is the first PD-L1 antibody ratified for advanced TNBC. PD-1/PD-L1 inhibitors combined with chemotherapy have been reported to show great potential as a new TNBC therapy [14]. However, the outcome assessment of TNBC patients treated with PD-1/ PD-L1 inhibitors still warrants further confirmation. Accordingly, this study was to evaluate the clinical efficacy of patients with TNBC receiving PD-1/PD-L1 inhibitors by meta-analysis to provide guidance for the application of PD-1/PD-L1 inhibitors in patients with advanced TNBC.

Literature Retrieval.
A literature search in PubMed, Embase, and Cochrane Library was conducted from database establishment to August 30, 2021, in the language of both English and Chinese, with the search terms of ("PD-1" or "PD-L1" or "PD-1/PD-L1" or "programmed cell death 1" or "programmed cell death-ligand 1") and ("breast cancer" or "mammary cancer" or "Breast Carcinoma" or "Breast Tumor" or "TNBC") and their corresponding Chinese terms. e search was conducted twice to avoid omission. (2) study subjects: pathologically diagnosed breast cancer with clear ER, PR, and HER2 negative confirmed by immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH); (3) interventions: after randomization, at least two groups were divided, one of which received immune-targeted therapy combined with radiotherapy; (4) study indicators: one or more of overall survival (OS), progression-free survival (PFS), and adverse events (AEs); (5) scientific and standardized study design, with clear grouping and interventions, and well-documented followup data.

Exclusion Criteria.
e exclusion criteria were as follows: (1) nonrandomized controlled clinical trials or nonprimary studies; (2) unavailability of relevant outcome indicators such as OS, PFS, and AEs; (3) inclusion of fewer than 20 patients.

Screening of the Included Literature.
e data were searched by two researchers, and literature management was performed using Endnote. Duplicate literature was excluded, and initial screening of the search results was carried out with regard to the titles, abstracts, and full text, followed by the inclusion of the literature according to the inclusion criteria. e Newcastle-Ottawa Scale was used to evaluate the quality of the included literature, and the decision for inclusion was independently assessed by a third investigator in case of disagreement between the two investigators.

Data Extraction.
Data extraction and collation were performed independently by two investigators, including authors, time of publication, subject type, intervention method, use or absence of blinding, and randomization. OS, PFS, and AEs were considered the main effect measures for meta-analysis.

Risk of Bias.
e risk of bias of the included literature was assessed in terms of the following six dimensions: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias), and was categorized into three types: low risk, high risk, and uncertain risk.

Statistical
Analyses. RevMan 5.4 was used to organize and analyze the data, and the generic inverse variance was selected for the hazard ratio (HR) data to compute and record the log [HR] and the corresponding SE. Dichotomous data types were selected for data such as adverse events. e heterogeneity of the included literature was evaluated by the Cochrane Q test and I 2 test. I 2 � 0 and P > 0.1 in both subgroups indicate no heterogeneity in the included studies whose analyses were subjected to a fixed-effect model. I 2 > 0 and P < 0.1 in both subgroups indicate heterogeneity in the included studies whose analyses were subject to a randomeffect model. Egger's and Begg's tests were used to assess publication bias.

Results
Of 1538 original papers retrieved, 1491 papers were excluded, and 47 were preliminarily included after reviewing the abstracts and excluding duplicate papers, abstracts, and reviews. After reading the full texts, studies with issues such as duplicate reports and no specified data were excluded, and seven papers were finally included. e study flow diagram is shown in Figure 1, and the basic information of the literature is shown in Table 1.
e results of the Cochrane risk bias assessment are shown in Figure 2. All seven included studies were of high quality and mostly at low risk of bias (green section in the figure). Four studies had performance bias, and one study had detection bias due to the failure of the study design to meet double-blind requirements for informed consent (red section in the figure).
Of the 7 included studies, 5 included all-subject PFS outcomes, and 5 included all-subject OS outcomes. As shown in Figure 3

Discussion
Currently, chemotherapy remains the basic treatment for TNBC. Notwithstanding the progress of chemotherapy combined with antiangiogenic drugs and immune-targeted drugs, the efficacy is still unsatisfactory. Research has revealed higher levels of PD-L1 in the tumor microenvironment of TNBC compared to other breast cancers, for which immunosuppressive therapy targeting PD-1/PD-L1 may provide a novel therapeutic insight [22]. T lymphocytes recognize and kill tumor cells. PD-1 is a T-lymphocyte regulatory protein that reduces the tumor-killing effect of T lymphocytes, which may contribute to tumorigenesis. PD-L1 is a ligand protein produced by tumor cells that binds PD-1 to attenuate T-cell-mediated immunosurveillance and provide an immune escape from cancer cells [23]. PD-1/PD-L1 inhibitors can specifically bind PD-1/PD-L1 and restore the normal immune function of T lymphocytes to kill tumor cells. In recent years, the application of PD-1/PD-L1 inhibitors in TNBC has achieved considerable progress, but the treatment efficiency remains unfavorable and circumscribed.
In this study, a meta-analysis was performed on seven RCT studies in which four drugs were employed, including atezolizumab used in two studies [18,19], durvalumab in one [15], cobimetinib in one [16], and pembrolizumab in three [17,20,21]. Atezolizumab was the first PD-L1 inhibitor approved by the FDA and the first immunotherapy regimen for breast cancer. In the IMpassion130 study [18], first-line treatment of TNBC with atezolizumab in combination with paclitaxel achieved favorable PFS results in all patients, with a more pronounced advantage in the PD-L1-positive patient group, but the difference in OS between the two groups was not statistically different. However, the advantages of combining atezolizumab with paclitaxel for the treatment of metastatic TNBC were absent from the results of IMpassion131 [19] compared to the previously published trial results of the IMpassion130 study, and the reasons for the absence of differences remain elusive. Durvalumab is a PD-L1 antibody with promising efficacy in bladder and lung cancers. In the SAFIR02-BREAST IMMUNO study [15], durvalumab improved OS in TNBC patients but showed no significant difference in the subgroup analyses. Furthermore, an exploratory analysis of TNBC patients showed an HR of 0.18 for durvalumab efficacy in patients with CD274 gene gain/amplification and 1.12 in patients with normal/lost CD274 gene, which provides a new reference for the selection of the durvalumab treatment population.  e results of this study showed no significant effect of PD-1/PD-L1 inhibitors on PFS and OS in patients without PD-L1 expression classification but significantly improved the PFS and OS of patients in the PD-L1-positive subgroup, which confirms the close correlation between the effect of PD-1/PD-L1 inhibitors and PD-L1 expression. Differences between this result and the results of separate studies may be attributed to the following reasons. e first is the difference in the determination of PD-L1 expression. PD-L1 expression is determined through immunohistochemistry (IHC) and is found on both tumor cells and immune cells. However, PD-L1 is a versatile but not yet impeccable biomarker for predicting anti-PD-1 or anti-PD-L1 antibodies in patients with various tumors [24]. e second is the difference in the method of drug administration, whose application methods include combination chemotherapy and preoperative neoadjuvant chemotherapy, with certain variations in efficacy. In addition, a meta-analysis of AEs in this study showed that PD-1/PD-L1 inhibitors increase the risk of high-grade AEs and SAEs. Nonetheless, the source of adverse events, from medication issues or the expression of PD-L1, remains unconfirmed as the type of intervention is determined in the subjects based on PD-L1 expression levels.
In conclusion, PD-1/PD-L1 inhibitors are effective in the treatment of TNBC, which is strongly correlated with the expression of PD-L1, and patient selection and clinical application require further investigation and verification.

Conflicts of Interest
e author declares that there are no conflicts of interest.