Pembrolizumab alone or combined with chemotherapy versus chemotherapy for the treatment of metastatic cancer: A meta-analysis of randomized clinical trials Authors: Yumin Yue, Quan Wang, Mingtian Wei, Fanghui Ding, Jiang Li, Bobo Zheng Published: 2024 ABSTRACT Background:Whether pembrolizumab alone or in combination with chemotherapy is superior to chemotherapy in metastatic cancer remains controversial. The study aims to give the effectiveness and safety of pembrolizumab-related interventions compared to chemotherapy in metastatic cancer.Methods:Electronic databases were systematically searched until November 20, 2023, for all randomized controlled trials comparing Pembrolizumab alone or in combination with chemotherapy versus chemotherapy for metastatic cancer. The primary outcomes were overall survival (OS) and progression-free survival (PFS). Hazard ratios (HRs) and odds ratios with 95% confidence intervals (CI) were calculated for OS, PFS, overall response rate, and overall adverse events (AEs) by random effects models.Results:16 Randomized controlled trials with 9148 patients were included. Compared with chemotherapy, pembrolizumab was associated with longer OS (HR 0.82; 95% CI 0.73–0.91, P = .0004), more immune-mediated AEs, fewer overall AEs, and grade 3 or 4 AEs, however, no significant difference was found in PFS, overall response rate, and events leading to death. Pembrolizumab with chemotherapy was associated with longer OS (HR 0.74; 95% CI 0.61–0.90, P = .002) and PFS (HR 0.63; 95% CI 0.50–0.79, P < .0001), higher overall response rate, and more immune-mediated AEs comparing to chemotherapy alone, however, no significant advantages were observed in disease control rates, overall AEs, grade 3 or 4 AEs and events leading to death. The patients with programmed cell death ligand 1 tumor proportion scores of at least 50% or combined positive scores (CPS) of at least 10 could derive significantly better OS and PFS benefits from pembrolizumab alone or combined with chemotherapy. Similar OS results were found for first-line treatment and lung cancer subgroup analysis.Conclusions:Pembrolizumab alone or combined with chemotherapy indicates an effective and safe treatment for metastatic cancer. Pembrolizumab alone or combined with chemotherapy provides a better survival advantage under first-line treatment or programmed cell death ligand 1 combined positive scores of at least 10 or programmed cell death ligand 1 tumor proportion scores of at least 50%. However, we found that the specific efficacy of pembrolizumab in unused tumor types could not be effectively evaluated. 1. Introduction Globally, cancer remains the leading cause of death and a serious impediment to increasing life expectancy.[1] Progression to metastasis is a major challenge in cancer treatment, accounting for >90% of total cancer mortality.[2] The programmed cell death protein 1 (PD-1) pathway has received extensive attention for its role in triggering t-cell immune checkpoint responses, leading to tumor cells that can evade immune surveillance and become highly resistant to conventional chemotherapy.[3] Pembrolizumab is the first U.S. Food and Drug Administration-approved anti-PD-1 antibody for treating patients with unresectable or metastatic melanoma.[4] In recent years, it has been applied in treating various solid tumors. A recent meta-analysis consisting of 10,536 patients from 19 studies aimed to compare the safety and efficacy of immune checkpoint inhibitors combinations and chemotherapy alone for cancer treatment. The results showed that the immune checkpoint inhibitors combinations group was significantly better than the chemotherapy group in terms of progression-free survival (PFS), overall survival (OS), and overall response rate (ORR).[5] However, a phase 3 multicenter randomized controlled trial (RCT)[6] enrolled 1010 patients in pembrolizumab plus chemotherapy (n = 351), pembrolizumab monotherapy (n = 307), or chemotherapy alone (n = 352); Results showed no significant difference in OS between patients treated with pembrolizumab plus chemotherapy and those treated with chemotherapy alone (hazard ratios [HR]: 0.86, 95% confidence intervals [CI]: 0.72–1.02), and pembrolizumab was similar to chemotherapy alone(HR: 0.92, 95% CI:0.77, 1.10). Furthermore, KEYNOTE-021,[7] KEYNOTE-604,[8] and KEYNOTE-119[9] showed similar results. Taken together, it remains inconclusive whether pembrolizumab alone or in combination with chemotherapy is superior to chemotherapy in terms of OS in metastatic cancer. Therefore, the primary objective of the present meta-analysis was to evaluate the difference in terms of OS and PFS in metastatic cancer by comparing pembrolizumab alone or in combination with chemotherapy, with chemotherapy. Furthermore, to strengthen this analysis, we not only analyzed the influence of programmed cell death ligand 1 (PD-L1) tumor status on survival, but also analyzed ORR, and adverse events (AEs). 2. Material and methods 2.1. Search strategy The present meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.[10] The study protocol was registered with PROSPERO (CRD42022298112). The meta-analysis was not registered. We systematically searched PubMed, the Cochrane Library, Web of Science, and Google Scholar for all RCTs of pembrolizumab alone or combined with chemotherapy versus chemotherapy in treating metastatic solid tumors. The main search terms and their combinations were pembrolizumab, chemotherapy, metastatic, advanced, cancer, carcinoma, randomized. The original search was conducted on July 30, 2021, and the last updated search was performed on November 20, 2023. There were no language restrictions for each literature search. 2.2. Eligibility criteria All RCTs included in this meta-analysis met the following criteria: (1) pembrolizumab alone or in combination with chemotherapy versus chemotherapy, (2) metastatic solid tumors (breast cancer, urothelial carcinoma, lung cancer, gastric and gastroesophageal junction cancer, renal cell carcinoma, head and neck squamous cell carcinoma, colorectal cancer, melanoma and gynecological tumors). Exclusion criteria were as follows: (1) non-RCT; (2) combined or sequential use of immune checkpoint inhibitors in the chemotherapy group; (3) HR values of OS and PFS were not reported, or survival curves were not provided; and (4) complete response (CR), partial response, stable disease, AEs, and other data that need to be merged were not reported. Two researchers independently screened literature titles and abstracts according to the inclusion and exclusion criteria, downloaded the full text of the literature that initially met the criteria, and further screened. In case of disagreement, a third researcher was consulted. 2.3. Clinical endpoints and data extraction The main focus of this study was to explore the OS and PFS rates in patients with different levels of PD-L1 combined positive score (CPS) and tumor proportion scores (TPS). The study analyzed patients with PD-L1 CPS of at least 1, PD-L1 CPS of 10 or higher, PD-L1 TPS of at least 1%, PD-L1 TPS of 50% or higher, as well as all patients included in the meta-analysis. Secondary outcomes were ORR (CR + partial response), disease control rate (DCR, CR + partial response + stable disease for 24 weeks or more), overall AEs, grade 3 or 4 AEs, immune-mediated AEs, and events leading to death. The 2 authors extracted data independently according to the predesigned data extraction table and checked the data individually. In case of inconsistent data retrieval, they had to discuss and form a joint decision. 2.4. Risk of bias assessment Two authors independently assessed the risk of bias using the Cochrane Collaboration risk of bias assessment tool.[11] Quality of evidence was assessed using the grades of recommendation, assessment, development, and evaluation in the present study.[12] 2.5. Statistical analysis Heterogeneity between studies was identified by the chi-square and quantified using the I2 statistic.[13] Due to the potential clinical heterogeneity between RCTs and among tumors, we used a random-effects model to calculate the pooled effect which was presented as HR with their 95% CIs for OS and PFS, odds ratios with their 95% CIs for dichotomous outcomes. We used Stata version 14.0 (StataCorp, College Station, TX) to conduct Egger and Begg tests to explore publication bias and meta-regression to the source of heterogeneity. Subgroup analyses were performed by lines of therapy and primary tumor type for all outcomes. Statistical analysis was performed with Review Manager Version 5.4.1. 3. Results 3.1. Literature search results and characteristics A total of 227 papers were retrieved through the database according to the retrieval strategy, and 163 papers remained after deduplication. By evaluating abstracts or full texts, 16 RCTs (20 papers)[6–9,14–29] with 9148 patients were included for the final analysis (Fig. 1). (PRISMA 2020 flow diagram): 16 RCTs (20 papers) with 9148 patients were included for the final analysis. The baseline characteristics included in the study are presented in Table 1. The tumor types included were shown below: breast cancer (3 RCTs[9,14,18]), urothelial carcinoma (2 RCTs[6,23]), lung cancer (7 RCTs[7,8,16,17,19,21,22]), gastric and gastroesophageal junction cancer (2 RCTs[19,25]), colorectal cancer (1 RCT[20]), head and neck squamous cell carcinoma (1 RCT[24]). 8 studies[9,16,20–29] evaluated the efficacy of pembrolizumab versus chemotherapy, 6 studies[7,8,14,17–19] evaluated the efficacy of pembrolizumab combined with chemotherapy versus chemotherapy, and 2 studies[6,15] compared pembrolizumab combined with chemotherapy or pembrolizumab versus chemotherapy. 6 RCTs[6–8,15,17,18] evaluated the pembrolizumab combined with chemotherapy as a first-line setting, while 4 RCTs[6,20–22] assessed pembrolizumab as first-line therapy for metastatic cancer. 6 studies[6,9,14,15,24,25] provided CPS-related survival data and 6 RCTs[7,16,17,21,22,24] had available TPS-elated survival data. All studies were multicenter RCTs except 2.[14,19] Table 1Characteristics of included randomized clinical trials.Study/Publication timeTrial registration NoNo. of participating countries and centersSettingTreatment protocolDuration of studyPrimary endpointNo. of patientsPembrolizumab/ P + CChemotherapyP + CPCKEYNOTE-119, 2021NCT0255565731/1502–3 line, TNMBCPembrolizumabInvestigator choice chemotherapy (capecitabine, eribulin, gemcitabine, or vinorelbineNov 25, 2015 and April 11, 2017OS–312310KEYNOTE-361 2021NCT0285330521/2011 line, MUCPembrolizumabGemcitabine, and either cisplatin or carboplatinOct 19, 2016 and June 29, 2018PFS, OS351307352KEYNOTE-021, 2021NCT0203967426/21 line, stage IIIB or IV NSCLCPembrolizumab plus Pemetrexed and carboplatin, followed by pembrolizumab for 24 months and optional indefinite pemetrexed maintenance therapyPemetrexed and carboplatin, pemetrexed maintenance therapyNov 25, 2014, and Jan 25, 2016ORR,60–63Tolaney SM 2020NCT030516591/40–2 line, HoR(+) ER(−) MBCEribulin mesylate plus pembrolizumabEribulin mesylateApril 6, 2017, and August 28, 2018PFS,44–44KEYNOTE-062 2020NCT0249458329/2001 line, advanced G/GEJ cancer, CPS≥1Pembrolizumab plus Cisplatin and fluorouracil or capecitabineCisplatin plus fluorouracil or capecitabineSeptember 18, 2015, and May 26, 2017OS, PFS(CPS > 1, CPS > 10)257256250KEYNOTE-604 2020NCT0306677818/1401 line, MSCLCPembrolizumab plus etoposide and cisplatinSaline placebo plus etoposide and cisplatinMay 15, 2017, and July 30, 2018PFS, OS228–225KEYNOTE-010 2020NCT0190565724/202≥2 line, MSCLCPembrolizumabDocetaxelAug 28, 2013, and Feb 27, 2015PFS, OS344 (2 mg/kg)/346 (10 mg/kg343KEYNOTE-189 2020NCT0257868016/1261 line, metastatic nonsquamous NSCLCPembrolizumab plus pemetrexed and platinumPlacebo plus pemetrexed and platinumFebruary 26, 2016, and March 6, 2017PFS, OS410–-206KEYNOTE-355 2020NCT02819518,29/2091 line, TNMBCPembrolizumab plus taxane, gemcitabine, or platinum agentsTaxane, gemcitabine, or platinum agentsJan 9, 2017, and June 12, 2018PFS, OS566–281PROLUNG 2020NCT025745981/1≥2 line, MSCLCPembrolizumab Plus DocetaxelDocetaxelDecember 2016 through May 2019ORR40–38KEYNOTE-177 2020NCT0256300223/1921 line, MSI-H–dMMR stage IV colorectal cancerPembrolizumabmFOLFOX6/ FOLFIRI or plus bevacizumab/cetuximabPFS, OS–153154KEYNOTE-024 2019NCT0214273816/1421 line, stage IV NSCLC with TPS > 50%PembrolizumabCarboplatin plus pemetrexed, cisplatin plus pemetrexed, carboplatin plus gemcitabine, cisplatin plus gemcitabine, or carboplatin plus paclitaxelSeptember 19, 2014, and October 29, 2015PFS–154151KEYNOTE-042 2019NCT0222089432/2131 line, stage IV NSCLC with TPS > 1%PembrolizumabCarboplatin plus paclitaxel or pemetrexed2014, to March 6, 2017OS(TPS > 50%)–637637KEYNOTE-045 2019NCT0225643629/120≥2 line, recurrent MUCPembrolizumabPaclitaxel, docetaxel, or vinflunineNovember 5, 2014, and November 13, 2015PFS, OS–270272KEYNOTE-040 2018NCT0225204220/972–3 Line, recurrent or metastatic HNSCCPembrolizumabMethotrexate, docetaxel, or cetuximabDec 24, 2014, and May 13, 2016OS–247248KEYNOTE-061 2018NCT0237049830/1482 Line, advanced G/GEJC(CPS > 1)PembrolizumabPaclitaxelJune 4, 2015, and July 26, 2016,PFS, OS–296296ARCC = advanced renal cell carcinoma, C = Chemotherapy, ER(−) = ERBB2-negative, G/GEJ = gastric and gastroesophageal junction cancer, HNSCC = head and neck squamous cell carcinoma, HoR(+) = hormone receptor-positive, MUC = metastatic urothelial carcinoma, NSCLC = non-small-cell lung cancer, ORR = objective response rate (CR+PR), OS = overall survival, P = Pembrolizumab, PFS = progression-free survival, TNMBC = triple-negative metastatic breast cancer. Characteristics of included randomized clinical trials. ARCC = advanced renal cell carcinoma, C = Chemotherapy, ER(−) = ERBB2-negative, G/GEJ = gastric and gastroesophageal junction cancer, HNSCC = head and neck squamous cell carcinoma, HoR(+) = hormone receptor-positive, MUC = metastatic urothelial carcinoma, NSCLC = non-small-cell lung cancer, ORR = objective response rate (CR+PR), OS = overall survival, P = Pembrolizumab, PFS = progression-free survival, TNMBC = triple-negative metastatic breast cancer. 3.2. Risk of bias assessment The risk of bias assessment for included studies is shown in eFigure 1, Supplemental Digital Content, http://links.lww.com/MD/O159. All studies described appropriate randomization methods. 3 RCTs[8,15,18] were double-blind and the rest were open-label. All studies reported that the subjective outcomes were reviewed by a blinded, independent radiologist except 2 studies.[14,19] 3.3. OS and PFS The OS results favored pembrolizumab alone or combined with chemotherapy (Fig. 2). Compared to chemotherapy, the HR for OS was improved in patients treated with either pembrolizumab alone (HR 0.82; 95% CI 0.73–0.91, P = .0004) or in combination with chemotherapy (HR 0.74; 95% CI 0.61–0.90, P = .002). (Overall survival): Compared to chemotherapy, the HR for OS was improved in patients treated with either pembrolizumab alone (HR 0.82; 95% CI 0.73–0.91, P = .0004) or in combination with chemotherapy (HR 0.74; 95% CI 0.61–0.90, P = .002). CI = confidence intervals, HR = hazard ratios, OS = overall survival. As shown in Figure 3, pembrolizumab combined with chemotherapy was significantly superior to chemotherapy in PFS (HR 0.63; 95% CI 0.50–0.79, P < .0001). However, the pooled results showed no significant difference in PFS between pembrolizumab and chemotherapy groups (HR 0.9, 95% CI 0.68–1.19, P = .46). (Progression-free survival): Pembrolizumab combined with chemotherapy was significantly superior to chemotherapy in PFS (HR 0.63; 95% CI 0.50–0.79, P < .0001). However, the pooled results showed no significant difference in PFS between pembrolizumab and chemotherapy groups (HR 0.9, 95% CI 0.68–1.19, P = .46). CI = confidence intervals, HR = hazard ratios, PFS = progression-free survival. 3.4. The association between PD-L1 expression status and survival Patients with at least 50% TPS could obtain significantly more OS and PFS benefits from pembrolizumab alone or combined with chemotherapy compared to just chemotherapy (eTable 1, Supplemental Digital Content, http://links.lww.com/MD/O160). Only 1 study[7] reported survival data of pembrolizumab plus chemotherapy versus chemotherapy in patients with TPS >1%, and the results showed no significant difference in OS and PFS. Compared to chemotherapy in patients with TPS of at least 1%, the pooled results showed that pembrolizumab had a significant advantage in terms of OS (HR 0.75, 95% CI 0.68–0.83), but there was no statistical difference in PFS. In patients with CPS of at least 10, OS was significantly longer with pembrolizumab alone or plus chemotherapy compared with chemotherapy alone (HR 0.80 95% CI 0·65–0·98, HR 0·69 95% CI 0·49–0·97, respectively). In patients with CPS of at least 1, the OS of pembrolizumab was significantly longer than that of chemotherapy (HR 0.84, 95% CI 0.75–0.94), however, pembrolizumab plus chemotherapy was not significantly better than chemotherapy (HR 0.85, 95% CI 0.72–1.00). Pembrolizumab plus chemotherapy has a longer PFS than chemotherapy, but this advantage is not found in pembrolizumab alone regardless of whether CPS is ≥1 or 10. 3.5. ORR and DCR Compared with chemotherapy, Pembrolizumab combined with chemotherapy significantly increased the objective response rate (odds ratios 1.92, 95% CI 1.33–2.79, P = .0006), however, this advantage was not maintained with pembrolizumab alone (eFig. 2, Supplemental Digital Content, http://links.lww.com/MD/O159). Pembrolizumab alone or combined with chemotherapy versus chemotherapy had no significant differences in terms of DCR (eFig. 3, Supplemental Digital Content, http://links.lww.com/MD/O159). 3.6. Adverse events Patients who received the pembrolizumab had fewer overall AEs than those who received chemotherapy (odds ratios 0.26, 95% CI 0.18–0.36, P < .00001, eFig. 4, Supplemental Digital Content, http://links.lww.com/MD/O159). However no difference was found between pembrolizumab plus chemotherapy and chemotherapy alone. Similar results were found for grade 3 or 4 AEs (eFig. 5, Supplemental Digital Content, http://links.lww.com/MD/O159). Pembrolizumab alone or combined with chemotherapy lead to more immune-mediated AEs than chemotherapy (eFig. 6, Supplemental Digital Content, http://links.lww.com/MD/O159). Pembrolizumab alone or combined with chemotherapy versus chemotherapy had no significant differences in terms of events leading to death (eFig. 7, Supplemental Digital Content, http://links.lww.com/MD/O159). 3.7. Publication bias and subgroup analyses Egger tests revealed publication bias in AEs (P = .041, eFig. 8, Supplemental Digital Content, http://links.lww.com/MD/O159), However, Begg tests showed no publication bias in all outcomes (eFig. 9, Supplemental Digital Content, http://links.lww.com/MD/O159). In the subgroup analysis of OS (eTable 2, Supplemental Digital Content, http://links.lww.com/MD/O160), similar results to the main outcomes were found for first-line treatment and lung cancer. Only 1 study[14] reported data on Pembrolizumab plus chemotherapy versus chemotherapy in 2 or more line treatments, showing no statistical difference between the 2 groups. There was a significant difference in other cancers between pembrolizumab plus chemotherapy and chemotherapy. In the subgroup analysis of PFS and ORR (eTables 3 and 4, Supplemental Digital Content, http://links.lww.com/MD/O160), the inconsistent outcomes compared to the main analysis were found in 2 or more line treatments. No difference was found between pembrolizumab plus chemotherapy and chemotherapy. In the subgroup analysis of DCR (eTable 5, Supplemental Digital Content, http://links.lww.com/MD/O160), all subgroup results were consistent with the main outcome, except for first-line treatment. Pembrolizumab plus chemotherapy obtained more DCR than chemotherapy alone, but only 1 study reported that pembrolizumab had a significantly less pathological CR than chemotherapy in first-line treatment. In the subgroup analysis of AE (eTable 6, Supplemental Digital Content, http://links.lww.com/MD/O160), no inconsistent results were found in all subgroups. 3.8. Grades of recommendation, assessment, development, and evaluation of the outcomes The grades of recommendation, assessment, development, and evaluation approach was used to assess the evidence for the outcomes (eTable 7, Supplemental Digital Content, http://links.lww.com/MD/O160). The level of evidence is high in all outcomes. 4. Discussion To our knowledge, this is the first and largest pooling analysis involving only RCTs to evaluate the efficacy and safety of pembrolizumab alone or combined with chemotherapy versus chemotherapy for the treatment of metastatic cancer. The present study found that pembrolizumab plus chemotherapy was associated with longer OS and PFS, higher ORR, and more immune-mediated AEs. A series of clinical studies in recent years have confirmed the impressive efficacy of immunotherapy in dMMR/MSI-H tumors, especially the KEYNOTE-177 study, which directly pushed pembrolizumab into the first-line treatment of dMMR/MSI-H advanced colorectal cancer. However, no difference was found in DCR, overall AEs, grade 3 or 4 AEs, and events leading to death between pembrolizumab plus chemotherapy and chemotherapy alone. Compared with chemotherapy, Pembrolizumab was associated with longer OS, more immune-mediated AEs, fewer overall AEs, and grade 3 or 4 AEs. Similar results were found in PFS, ORR, DCR, and events leading to death between pembrolizumab and chemotherapy. Pembrolizumab is a typical PD-L1 antibody, and it seems reasonable that the higher the level of PD-L1 expression, the more pembrolizumab would benefit. The present pooling analysis showed that patients with metastasis cancer could derive significantly more OS benefit from pembrolizumab alone or combined with chemotherapy compared to chemotherapy at PD-L1 CPS of at least 10 and TPS of at least 50%, respectively. For patients with PD-L1 CPS of at least 1 and TPS of at least 1%, the observed OS benefit was also found in pembrolizumab versus chemotherapy. However, no significant difference was noted between pembrolizumab combined with chemotherapy and chemotherapy alone. The most likely reason for this phenomenon is that very little literature provides relevant data, which may affect the reliability of the results. A previous meta-analysis[30] involving 8 RCTs reported that PD-1 or PD-L1 inhibitors also benefited patients with PD-L1 negative metastatic cancer, so the clinical benefit of PD-L1 testing had declined. However, this study has the following risks of bias. Firstly, only 4174 patients were included in the study, which may have certain selectivity biases. Secondly, this study included 2 anti-PD-1(nivolumab and pembrolizumab) and 1 PD-1(nivolumab) agents, with fewer studies but more immunotherapeutic agents, which reduced the reliability of the study. A recent meta-analysis[31] examined the relationship between PD-L1 expression status and the efficacy of PD-1/PD-L1 inhibitors and OS in solid tumors. The study found that PD-1/PD-L1 inhibitors significantly improved OS in both PD-L1 positive and PD-L1 negative patients compared to in controls, but the extent of the benefit was clinically PD-L1 dependent. However, the study also included 5 anti-PD-1/PD-L1 agents (nivolumab, pembrolizumab, atezolizumab, avelumab, and durvalumab) that affect the reliability of the results to a certain extent. In order to minimize the influence of confounding factors, the present pooling analysis includes 16 RCTs with 9148 patients to study pembrolizumab alone, or combined with chemotherapy, versus chemotherapy for the treatment of metastasis cancer and showed that the degree of OS benefit was dependent on the level of clinical PD-L1 expression, and that the higher the expression of PD-L1, the greater the likelihood of survival benefit. The present pooling analysis showed that the PFS results favored pembrolizumab combined with chemotherapy, but no significant difference was found between pembrolizumab and chemotherapy alone. The results of all subgroup analyses also supported this result, except in terms of TPS ≥ 1% and 2 or more lines. A previous meta-analysis of the correlation between OS and PFS and the efficacy of PD-1 inhibitors in cancer showed that OS was not associated with PFS measured by RECIST standards.[32] The results of our study also showed that the advantage of pembrolizumab was found in OS, but not in PFS, and this is not accidental. Therefore, OS is the most reliable indicator to evaluate efficacy. The ORR was comparable between the pembrolizumab and chemotherapy group, however, the ORR of pembrolizumab plus chemotherapy was superior to the chemotherapy alone in the study. This may indicate that pembrolizumab did not benefit significantly from PFS. We found that pembrolizumab was associated with fewer AEs than chemotherapy. Magee DE also reported similar results.[33] This study has several limitations. First, this study is based on inter-study analysis, and the results still need to be combined with the actual clinical situation before being applied to clinical cases. Second, multiple solid tumors were included in this study. Due to the limitation of the number of studies, only subgroup analysis was conducted for lung cancer and other tumors, lacking subgroup analysis for each tumor. Although the results of subgroup analysis were basically consistent with the overall results, there was still a risk of bias. Finally, the lack of detailed restrictions on maintenance therapy for metastatic tumors may have an impact on OS. Third, due to the differences in tumor types, as well as the dosage and duration of medication for each tumor, the specific efficacy of pembrolizumab could not be effectively assessed, and large sample sizes of randomized controlled studies are needed to further evaluate it. 5. Conclusion In summary, pembrolizumab was associated with longer OS, and fewer overall AEs in patients with metastatic cancer, while pembrolizumab plus chemotherapy significantly improved OS, PFS, and ORR, and had a comparable safety profile versus chemotherapy alone. The results of subgroup analysis showed that the survival advantage of pembrolizumab alone or combined with chemotherapy was more obvious under the background of first-line treatment or PD-L1 CPS of at least 10 or PD-L1 TPS of at least 50%. Author contributions Conceptualization: Yumin Yue, Fanghui Ding, Jiang Li, Bobo Zheng. Formal analysis: Yumin Yue, Mingtian Wei, Jiang Li. Funding acquisition: Bobo Zheng. Investigation: Quan Wang, Mingtian Wei. Methodology: Mingtian Wei, Fanghui Ding. Project administration: Fanghui Ding, Jiang Li. Resources: Mingtian Wei. Software: Yumin Yue, Bobo Zheng. Validation: Quan Wang, Mingtian Wei. Visualization: Fanghui Ding. Writing – original draft: Yumin Yue. Writing – review & editing: Quan Wang, Fanghui Ding, Jiang Li, Bobo Zheng. Supplementary Material