Efficacy and safety of combination PD‐1/PD‐L1 checkpoint inhibitors for malignant solid tumours: A systematic review

Abstract Treatment of multiple malignant solid tumours with programmed death (PD)‐1/PD ligand (PD‐L) 1 inhibitors has been reported. However, the efficacy and immune adverse effects of combination therapies are controversial. This meta‐analysis was performed with PubMed, Web of Science, Medline, EMBASE and Cochrane Library from their inception until January 2020. Random‐effect model was adopted because of relatively high heterogeneity. We also calculated hazard ratio (HR) of progression‐free survival (PFS), overall survival (OS) and risk ratio (RR) of adverse events (AEs), the incidence of grade 3‐5 AEs by tumour subgroup, therapeutic schedules and therapy lines. Nineteen articles were selected using the search strategy for meta‐analysis. Combined PD‐1/PD‐L1 inhibitors prolonged OS and PFS (HR 0.72, P < 0.001) and (HR 0.66, P < 0.001). In addition, incidence of all‐grade and grade 3‐5 AEs was not significant in the two subgroup analyses (HR 1.01, P = 0.31) and (HR 1.10, P = 0.07), respectively. Our meta‐analysis indicated that combination therapy with PD‐1/PD‐L1 inhibitors had greater clinical benefits and adverse events were not increased significantly.

with PD-1/PD-L1 inhibitors improved efficacy. 4,[9][10][11] The efficacy of combination of PD-1/PD-L1 inhibitors with ipilimumab is also encouraging in melanoma. 12 Besides, combination of PD-1/PD-L1 inhibitors with nab-paclitaxel in breast cancer 13 and with dabrafenib and trametinib in melanoma 14 has shown similar efficacy. There are now >100 ongoing clinical trials of PD-1/PD-L1 inhibitors as monotherapy or in combination with other agents in different tumour types. 15 Nevertheless, the use of these agents can be limited by adverse events (AEs), such as nausea, fatigue, decreased appetite, diarrhoea and vomiting. 16 The clinical benefit associated with combination PD-1/PD-L1 inhibitors should be balanced against associated toxicity.
Addition of PD-1/PD-L1 inhibitors to treatment remains controversial, and individual studies are not sufficient to clarify this.
Whether PD-1/PD-L1 checkpoint inhibitors will achieve significant efficacy for all tumour types or different therapeutic schedules is still up for question. Therefore, we performed a meta-analysis of phase II/III randomized controlled trials to compare the efficacy and safety of combination PD-1/PD-L1 checkpoint inhibitors for malignant solid tumours. It is important for clinical policymakers to explore the degree of efficacy in different tumour types, therapeutic schedules and therapy lines. Additionally, the incidence of AEs may provide clinicians with important and clinically useful information.

| Search strategy
This meta-analysis was performed with PubMed, Web of Science, Medline, EMBASE and Cochrane Library from their inception until January 2020 to identify relevant studies. A combination of freetext terms and medical subject headings terms was used for the subject search. Search terms included "nivolumab" OR "BMS 936558" OR "BMS 936559" OR "MDX 1105" OR "pembrolizumab" OR "lambrolizumab" OR "MK 3475" OR "pidilizumab" OR "CT 011" OR "durvalumab" OR "MEDI 4736" OR "atezolizumab" OR "MPDL 3280a" OR "avelumab" OR "AMP 224" OR "PD-1" OR "PD-L1" OR "programmed death 1" OR "programmed death ligand 1" OR "programmed cell death ligand 1" OR "programmed death ligand 1" OR "B7-H1" OR "CD274" AND "tumor" OR "cancer" OR "carcinoma" OR "neoplasm" OR "malignancy" OR "sarcoma". We also had two researchers independently screen the titles and abstracts of the retrieved articles.

| Study selection
Studies were included if they met the following criteria. (a) Literature type: phase II/III randomized controlled trials. (b) The experimental intervention group was treated with combination PD-1/PD-L1 checkpoint inhibitors with other therapies (immunotherapy, chemotherapy, targeted therapy and radiotherapy), whereas the control group received other therapies without PD-1/PD-L1 inhibitors. (c) Efficacy and safety data were available. Exclusion criteria were as follows: (a) studies with post-operative adjuvant therapy and neoadjuvant therapy; (b) not in English; and (c) multiple articles that analysed the same trials. In the latter case, we analysed the latest data.

| Data extraction and quality assessment
Data from each study were extracted by two researchers independently. A third researcher was consulted to reach a majority decision.
The following information was used: (a) authors' names, year of publication, tumour type, therapy lines, sample size and interventions; and (b) the primary efficacy outcomes were OS and PFS, and the secondary outcome was AEs. The meta-analysis was conducted in accordance with the guidelines of the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) 2015 statement. 17

| Statistical analysis
We calculated the hazard ratio (HR) and 95% confidence interval (CI) for OS and PFS and the risk ratio (RR) and 95% CI for AEs. We also performed subgroup analyses of OS, PFS and incidence of grade 3-5 AEs for patients with different tumour types, therapeutic schedules and therapy lines. Revman version 5.3 (The Cochrane Collaboration) was used to perform the meta-analysis. Heterogeneity between studies was evaluated using the chi-squared test and I 2 statistics.
Because of the complexity of the control conditions and the variety of solid tumours, a random-effect model was used to enhance the credibility of the results. We used Begg's and Egger's tests with Stata SE version 12 (Stata Corporation), with significance set at P < 0.1, to evaluate publication bias. All the statistical tests were two-sided, and P < 0.05 was considered statistically significant.

| Eligible studies and characteristics
The search strategy generated 26 502 relevant clinical records from the five databases. After screening and eligibility assessment, 19 eligible 5,14,18-34 phase II/III randomized controlled trials were selected for meta-analysis, including 10 178 patients. The detailed search and study selection process is shown in Figure 1. In addition, RCTs was evaluated with the Cochrane Collaboration tool, which demonstrated relatively high methodological quality ( Figures S1 and S2).

| Incidence of grade 3-5 AEs
The incidence of grade 3-5 AEs was examined in 5568 patients in the experimental groups and 4416 patients in the control groups. We and Figure S5).

| Incidence of all-grade and grade 3/4 AEs
The incidence of all-grade and grade 3/4 AEs was examined in 5315 patients in the experimental groups and 4258 patients in the control groups. A total of 9573 patients experienced AEs of any grade.

| D ISCUSS I ON
In the past 10 years, >10 cancers have been recommended for treatment with PD-1/PD-L1 checkpoint inhibitors, with objective response rates of 10%-30% and good toxicity. 35 Compared with traditional therapies, PD-1/PD-L1 inhibitors can prolong survival because of the memory of the adaptive immune system. 36 Nevertheless, we have to acknowledge that many patients do not benefit from the treatment or relapse after a period of response, especially in breast and colon cancers. 20,33,37 Tumour-mediated mechanisms of immunotherapy resistance are improved by synergism with targeted therapies or chemotherapy. 38 Many studies have demonstrated that combination with chemotherapy, moleculartargeted therapy and immunotherapy has good curative effect and adequate safety. 5,22,39 In the presence of efficacy based on therapeutic schedules, we found that adding PD-1/PD-L1 inhibitors to various therapeutic schedules achieved different degrees of clinical benefit. In 8 chemotherapy groups, combined chemotherapy with PD-1/PD-L1 inhibitors achieved the impressive efficacy, which was consistent with recent clinical trials. 19,33 A pre-clinical trial 40 showed that chemotherapy induces PD-L1 overexpression via nuclear factor-κB, which aggravates immunosuppression in ovarian cancer. The mechanisms of action of chemotherapeutic agents include the death of tumour cells with immunogenicity, reduced immunosuppressive effect and sensitization of tumour cells to immune effector cells. 40 When it comes to adding PD-1/PD-L1 inhibitors, many studies have investigated the mechanism. Firstly, combination therapies can increase cross-presentation of tumour antigens and up-regulation of major histocompatibility complex (MHC) class I antigens. 41 Secondly, in the presence of interleukin (IL)-2, IL-5 and other cytokines, combination therapies enhance CD8 T-cell activation and their ability to attack tumour cells. 42 Our research indicated that the addition of PD-1/PD-L1 inhibitors prolonged OS and PFS notably in molecular-targeted treatment. There has been an increase in the use of anti-vascular endothelial growth factor (VEGF) agents for molecular-targeted therapy. 43 (1) and Rini (2) in order to make a better distinction.
down-regulate anticancer immunity of cytotoxic T lymphocytes. 45 Anti-VEGF agents have been shown to have multiple mechanisms of action. 43,46 Some studies 47,48 have reported that anti-VEGF agents up-regulate PD-L1 on endothelial cells and tumour cells and cause abnormal vascularization in mouse models, which aggravates immunosuppression. It has been suggested that treatment with PD-1/PD-L1 inhibitors ameliorates immune escape and promotes normalization of tumour vasculature. 44,49 Only one included article mentioned that combined PD-1/PD-L1 inhibitors with radiotherapy improved the curative effect. When radiotherapy is combined with PD-1/PD-L1 inhibitors, it can increase inflammatory processes, restrain leucocyte adhesion to ECs, promote apoptosis and reduce oxidative burst in macrophages. 50 In NSCLC, radiotherapy can up-regulate tumour cell PD-L1 expression. 51 Besides, the greatest benefit was observed with immunotherapy (ipilimumab) when plus PD-1/PD-L1 inhibitors for malignant The OS and PFS of first-line treatment were significantly higher than those of second-line or beyond treatment.

F I G U R E 6 Forest Plot of Hazard ratio of grade 3-5 AEs rates based on therapeutic schedules in total population
Our meta-analysis demonstrated that combination treatment with PD-1/PD-L1 checkpoint inhibitors did not significantly increase incidence of all-grade AEs. Nearly 95% of patients experienced at least 1 AE, which is consistent with Hoffner. 16 Second, when immunotherapy (ipilimumab) plus PD-1/PD-L1 or combination PD-1/PD-L1 inhibitors used in melanoma, the rate of grade 3-5AEs showed AEs increased significantly, which is also consistent with previous results. 57,58 The most common AEs were fatigue, nausea and diarrhoea. The incidence of rash was raised rapidly, which might be attributed to the use of PD-1/PD-L1 inhibitors. It has been shown that PD-1 blockade increases the risk of immune-mediated AEs when combined with chemotherapy. 59 We think that the decline of anaemia could be due to the addition of PD-1/PD-L1 inhibitors 3267 patients of included studies receiving lower dose chemotherapy in experiment group than 2470 patients in control group. 5,19,21,23,26,29,33,34 As far as we known, the present study is the first to analyse comprehensively the efficacy and safety of combination treatment with

| CON CLUS IONS
For malignant solid tumours, patients treated with first-or second-line combination therapy with PD-1/PD-L1 inhibitors had significantly prolonged PFS and OS, with only a small increase in the incidence of AEs.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest. writing-review and editing (lead).

E TH I C A L A PPROVA L
This article does not contain any studies with human participants or animals performed by any of the authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are included within the article.