Abbreviations: ICI, immune checkpoint inhibitor; NSCLC, non-small-cell lung cancer; SCLC, small-cell lung cancer; RCT, randomized controlled trial; OS, overall survival; PFS, progression-free survival; ORR, objective response rate; AE, adverse event; HR, hazard ratio; OR, odds ratio; PD-1, programmed cell death protein-1; PD-L1, programmed cell death 1 ligand 1; CTLA-4, cytotoxic T lymphocyte antigen 4.

Background

Lung cancer is the leading cause of cancer-related deaths worldwide, and almost half of patients are diagnosed with advanced or metastatic disease when early symptoms appear [1, 2]. Over the past several decades, platinum-based chemotherapy was regarded as the first-line standard treatment for advanced non-small-cell lung cancer (NSCLC) patients without epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinse (ALK) mutations, with a response rate of 12.5–37%, median progression-free survival (PFS) of 4 to 8 months, and median overall survival (OS) of 8 to 13 months [3]. Due to the poor prognosis, novel and effective treatment strategies are urgently warranted for patients with advanced NSCLC.

Currently, immune checkpoint inhibitors (ICIs) including programmed cell death protein 1 (PD-1), programmed cell death 1 ligand 1 (PD-L1) and cytotoxic T ymphocyte antigen 4 (CTLA-4) inhibitors were widely used in advanced cancers. Several ICIs, such as pembrolizumab (lambrolizumab or MK-3475, a highly IgG4 monoclonal anti-PD-1 antibody), nivolumab (a fully human IgG4 anti-PD-1 antibody), atezolizumab (MPDL3280A, an IgG anti-PD-L1 antibody), and ipilimumab (a fully human monoclonal IgG1κ anti-CTLA-4 antibody) have shown promising anti-tumour activity and safety in advanced NSCLC. However, it is estimated that only 50% of patients could benefit from single-agent ICI [4], and several risk factors, such as PD-L1 expression level, EGFR and ALK genetic alteration status, are important factors affecting the curative effect. Carbone et al. [5] reported that single-agent nivolumab could not improve survival compared with chemotherapy in stage IV or recurrent NSCLC with a PD-L1 expression level of 5% or more. In the phase 3 KEYNOTE 042 trial [6], single-agent pembrolizumab resulted in prolonged PFS and OS than chemotherapy only in the PD-L1 expression level of ≥ 50% group. However, no survival benefit could be seen in the PD-L1 expression leve of ≥ 1% or ≥ 20% groups. A another study by Lisberg et al. [7] reported that pembrolizumab monotherapy showed no survival benefit in PD-L1-positive, tyrosine kinase inhibitor (TKI) naïve, and EGFR-mutant patients with advanced NSCLC.

Due to the limitations of single-agent ICI, several randomized controlled trials [8,9,10,11,12,13,14,15] (RCTs) have evaluated the efficacy and safety of ICI and chemotherapy in the first-line treatment for advanced NSCLC. In order to evaluate the efficacy and safety of ICI + CT versus CT alone in advanced NSCLC, we performed this meta-analysis and examined the tumor response, survival of patients and treatment-related AEs in patients with advanced NSCLC.

Methods

Strategy of study screening

We identified original articles by searching databases including PUBMED, EMBASE and Cochrane Library from inception until December 2021. As for the literature search, we used any of the following key words: “immune checkpoint blockade OR immune checkpoint inhibitor OR immune therapy OR immunotherapy OR PD-1 OR PD-L1 OR pembrolizumab OR nivolumab OR atezolizumab OR tremelimumab OR avelumab OR durvalumab OR ipilimumab” AND “Non-small Cell Lung Cancer OR NSCLC” AND “chemotherapy”. To avoid missing relevant studies, we also searched manually through relevant references to identify other relevant clinical trials. Only randomized controlled studies (RCTs) that investigated the efficacy and safety of ICI + CT in the first-line treatment for advanced NSCLC were eligible for inclusion in the meta-analysis. Other inclusion criteria were articles published in English and presentation of data for any of the efficacy and safety outcomes of interest that were OS, PFS, objective response rate (ORR) and treatment related adverse events (AEs). Papers of non randomized trials, reviews, meta-analysis, letters, and case reports were excluded. The trials identified through the search were independently screened by two authors (L.F. M and J.F. H) for inclusion. Any disagreements were arbitrated by a third author (P.H. L).

Data extraction and Quality assessment

Two authors (P.H. L and S.X. H) independently extracted data concerning author details, year, study design, phase, number of patients, age, sex, and treatment regimens, and PD-L1 status according to a predefined data extraction form. Clinical outcomes including ORR, PFS, OS, grade 3–5 treatment related AEs, and treatment related deaths were recorded for further analysis. Data of OS for patients with PD-L1-negative (< 1%), PD-L1-positive (≥ 1%), PD-L1-low (1–49%) and PD-L1-high (≥ 50%) tumors was also recorded in detailed. When multiple papers of the same trial were identified, data was extracted and recorded as a single trial. If any discrepancy occured, problems were resolved by discussion and consensus. Two authors (P.H. L and H.L. W) used the Cochrane Collaboration risk of bias assessment tool to assess the risk of bias of the included RCTs [18].

Statistical analysis

Stata SE 15.0 (Stata Corporation, College Station, TX, USA) was used to conduct meta-analysis in the study. We calculated the pooled hazard ratio and 95% CI for OS and PFS and the pooled odds ratio and 95% CI for ORR and the incidence of grade 3–5 treatment related AEs. Between-study heterogeneity was analysed through I-squared (I2) tests in the meta-analysis. The heterogeneity was considered as high (either I2 > 50% or p < 0.1), then the randomized-effects model was applied; otherwise, the fixed-effects model was used. P value < 0.05 would be treated as statistically significant.

Results

Search results and study characteristics

The fowchart of the selection process and detailed identifcation are shown in Fig. 1. After the duplicate removal, eight RCTs with a total of 4227 patients were included [8,9,10,11,12,13,14,15]. Among the eight global, multi-center RCTs, six [8, 9, 11, 12, 14, 15] were phase 3 studies and two [10, 13] were phase 2 studies. All patients were diagnosed with NSCLC by pathology and were adults with advanced or metastatic disease, and received ICI + CT in the first-line treatment. Across these eight trials, six trials reported OS for patients with PD-L1-negative (< 1%) tumors, and five trials reported OS for patients with PD-L1-low (1–49%) and PD-L1-high (≥ 50%) tumors, while only 4 trials reported OS for patients with PD-L1-positive (≥ 1%) tumors. All articles were published between 2012 and 2021. All the eight trials were identified in the systematic evaluation, including three pembrolizumab plus chemotherapy [8,9,10], two atezolizumab plus chemotherapy [11, 12], two ipilimumab plus chemotherapy [13, 14], and one nivolumab + ipilimumab + chemotherapy [15]. The main characteristics of the included studies are shown in Table 1.

Fig. 1
figure 1

Flow diagram representing the selection process

Full size image
Table 1 The main characteristics of included studies
Full size table

Efficacy

Data of OS, PFS, and ORR was reported in all the eight included trials. Randomized-effects model was used in these outcome measurements because of the significant heterogeneity (I2 > 50%). The forest plot of these outcomes are showed in Fig. 2.

Fig. 2
figure 2

Forest plot of OS, PFS and ORR. A, Forest plot of OS. B, Forest plot of PFS. C, Forest plot of ORR

Full size image

The meta-analysis indicated that ICI + CT significantly improved OS (HR = 0.74, 95% CI: 0.62–0.85, p < 0.001) (Fig. 2 A), PFS (HR = 0.66, 95% CI: 0.57 − 0.75, p < 0.001) (Fig. 2B), and (OR = 1.89; 95% CI, 1.43–2.49, p < 0.001) (Fig. 2 C) in comparison to CT alone. Subgroup analysis showed that improved OS for ICI + CT was observed in either pembrolizumab plus chemotherapy, atezolizumab plus chemotherapy, ipilimumab plus chemotherapy or nivolumab plus ipilimumab plus chemotherapy group (HR = 0.57, 95% CI: 0.45 − 0.69, p < 0.001; HR = 0.82, 95% CI: 0.70 − 0.95, p < 0.001; HR = 0.90, 95% CI: 0.77 − 1.04, p < 0.001, and HR = 0.70, 95% CI: 0.52 − 0.88, p < 0.001, respectively) (Fig. 3).

Fig. 3
figure 3

Subgroup analysis of OS in patients treated with different combination regimens. pemb, pembrolizumab; atez, atezolizumab; ipil, ipilimumab; nivo, nivolumab; CT, chemotherapy

Full size image

In the PD-L1 subgroups, ICI + CT was associated with significantly longer OS than CT in either PD-L1-negative, PD-L1-positive, PD-L1-low and PD-L1-high group (HR = 0.59, 95% CI: 0.49 − 0.68, p < 0.001; HR = 0.58, 95% CI: 0.48 − 0.68, p < 0.001; HR = 0.63, 95% CI: 0.52 − 0.75, p < 0.001, and HR = 0.55, 95% CI: 0.42 − 0.68, p < 0.001, respectively). (Fig. 4).

Fig. 4
figure 4

Subgroup analysis of OS in patients with different PD-L1 expression levels

Full size image

Safety

The incidence of grade 3–5 treatment-related AEs were reported in all the eight publications. Randomized-effects model was used because of the high heterogeneity (I2 > 50%). As shown in Fig. 5 A, ICI + CT significantly increased the incidence of grade 3–5 treatment-related AEs (OR = 1.46, 95% CI: 1.19–1.79, p < 0.001) compared with CT alone.

Fig. 5
figure 5

 A, Forest plot of grade 3–5 treatment-related AEs. B, Forest plot of treatment-related deaths

Full size image

Among the eight trials, treatment-related deaths were reported in seven trials. Fixed-effects model was used because of the low heterogeneity (I2 < 50%). As shown in Fig. 5B, there was no statistical difference in the incidence of treatment-related deaths between the ICI + CT and CT groups (OR = 1.94, 95% CI: 0.97–3.88, p = 0.061).

Quality of the included studies

The risks of bias of the included studies in this meta-analysis are summarized in Fig. 6. The methodological quality was assessed as high in all the eight RCTs.

Fig. 6
figure 6

The assessment of bias of included studies using the Cochrane Collaboration Risk of Bias Assessment Tool

Full size image

Discussion

Immune checkpoint inhibitors have played an important role in the treatment of advanced NSCLC nowadays [19, 20]. In recent years, many clinical trials showed that ICI combination therapies offered a better survival benefit than monotherapies in advanced NSCLC [17, 21–22]. Mo et al. [23] reported that ICI combination therapies including ICI + CT, double-agent ICIs (nivolumab plus ipilimumab) and ICIs plus targeted therapy plus chemotherapy could significantly improve OS and PFS over monotherapies in patients with advanced NSCLC. However, despite the advent of novel ICI combination therapies, the optimal first-line setting for advanced NSCLC has not been established, and the combination of ICI and CT has become one of the most promising approaches in the treatment of advanced NSCLC. Due to the insufficient evidence regarding the efficacy and safety of ICI + CT versus CT alone in advanced NSCLC and the controversial role of PD-L1 as a prognostic predictor, a meta-analysis is warranted to provide more evidence for clinical use of this treatment strategy.

Our meta-analysis shows that ICI + CT significantly improved OS (HR = 0.74, 95% CI: 0.62–0.85, p < 0.001), PFS (HR = 0.66, 95% CI: 0.57 − 0.75, p < 0.001), and (OR = 1.89; 95% CI, 1.43–2.49, p < 0.001) compared with CT alone in advanced NSCLC, and significantly longer OS was observed in either pembrolizumab + chemotherapy, atezolizumab + chemotherapy, ipilimumab + chemotherapy, and nivolumab + ipilimumab + chemotherapy subgroup (all p < 0.001), indicating that ICI + CT is more effective than CT alone in the first-line treatment of advanced NSCLC. Similar results were found in other malignant tumors. In the phase 3 KEYNOTE-355 trial [24], pembrolizumab plus chemotherapy showed improved PFS versus chemotherapy among patients with metastatic triple-negative breast cancer with combined positive score (CPS) of 10 or more. In the CheckMate 649 trial [25], nivolumab in combination with chemotherapy was associated with significantly longer OS and PFS versus chemotherapy alone in previously untreated patients with advanced gastric, gastro-oesophageal junction, or oesophageal adenocarcinoma. In the IMpower133 trial [26], atezolizumab + chemotherapy resulted in significantly longer OS and PFS than chemotherapy alone in the first-line treatment of extensive-stage small-cell lung cancer (SCLC). These results suggest combined inhibition of immune checkpoint PD-1/PD-L1/CTLA-4 signaling pathway and chemotherapy resulting in enhanced anti-tumor activity. Preclinical studies suggests that PD-1/PD-L1/CTLA-4 checkpoint inhibitors increase T cells’ responses and reduce the acquired immune system tolerance which is overexpressed by cancer and tumor icroenvironment [27], and chemotherapeutic agents may increase immune-potentiating effects under certain condition, thereby enhancing the anti-tumor immune effects in tumors [28]. In addition, biomarkers for predicting an enhanced benefit for ICI combination therapies remain elusive, and whether PD-L1 can be used as a biomarker to predict outcome is controversial [29, 30]. In this study, compared with CT alone, ICI + CT showed significantly longer OS in either PD-L1-negative, PD-L1-positive, PD-L1-low and PD-L1-high group (all P < 0.001). The results were consistent with those from Landre et al. [31], who reported that PD-1/PD-L1 inhibitor plus chemotherapy showed improved OS, PFS and ORR versus CT alone for negative or < 1% PDL1expressing in the firstline treatment for metastatic NSCLC. These indicate that the addition of ICI to chemotherapeutic agents could benefit patients regardless PD-L1 expression levels, and PD-L1 can not be used as a biomarker to predict outcome for patients treated with ICI + CT in advanced NSCLC.

Regarding toxicities, the safety and tolerability profile of single-agent ICI was well established in cancers [32,33,34]. However, ICI combination therapies were reported to show increased treatment-related AEs over monotherapies in many studies. In the phase 3 CheckMate 649 trial [25], nivolumab plus chemotherapy significantly increased the incidence of grade 3–5 treatment-related AEs (59% vs. 44%) versus chemotherapy alone in patients with advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma. A recent meta-analysis [23] demonstrates that ICI combination therapies, including ICI + CT, were associated with significantly increased grade 3 or higher AEs (OR = 1.30, 95% CI: 1.03–1.57, p = 0.007) compared with monotherapies. In our study, the meta-analysis showed that the incidence of grade 3–5 treatment-related AEs in the ICI + CT group were significantly higher than that in the CT group (OR = 1.46, 95% CI: 1.19 − 1.79, p < 0.001). Treatment-related deaths showed similar between the two groups (OR = 1.94, 95% CI: 0.97–3.88, p = 0.061). Although the general safety profile of ICI + CT was found to be worse than that of CT, the incidence of treatment-related deaths is overall rare (0.7–1.9%) (Table 1), and the toxicities were manageable with appropriate monitoring.

Despite encouraging results, our study has several limitations. First, patients in each trial received different combination regimens (pembrolizumab plus chemotherapy, atezolizumab plus chemotherapy, ipilimumab plus chemotherapy, and nivolumab plus ipilimumab plus chemotherapy), and the anti-tumor mechanisms of ICIs (including PD-1, PD-L1 and CTLA-4 inhibitors) are different, which add heterogeneity to our analysis. Second, the number of included studies is small and only one trial was included in the nivolumab plus ipilimumab plus chemotherapy group, which may lead to a limitation in the evaluation of results in this study. Finally, the follow-up time among each trial is different, and the data of OS from some included trials were not mature enough because of the limited follow-up time.

Conclusions

Compared with CT alone, ICI + CT greatly enhances OS, PFS, and ORR rates in the first-line treatment for advanced NSCLC, with increased grade 3–5 treatment-related AEs. Survival benefit was observed for ICI + CT among all patients regardless PD-L1 expression levels. PD-L1 can not be used as a biomarker for predicting outcome for patients treated with ICI + CT. Due to the limitations in our study, further investigations are required.