Checkpoint inhibitors plus chemotherapy for first-line treatment of advanced non-small cell lung cancer: a systematic review and meta-analysis of randomized controlled trials

Background: We conducted a meta-analysis to evaluate the efficacy and safety of upfront add-on immunotherapy for advanced non-small cell lung cancers (NSCLC). Methods: We performed a literature search on first-line chemotherapy ± immunotherapy in NSCLC. We utilized Revman version 5.3 to calculate the estimated pooled hazard ratio for overall survival (OS) and progression-free survival (PFS) and pooled risk ratio for objective response rate (ORR), all-grade and high-grade adverse events with 95% CI. Results: We analyzed 4322 patients. The pooled hazard ratios for OS, PFS and ORR were 0.74 (95% CI: 0.62–0.88; p = 0.0007), 0.62 (95% CI: 0.57–0.68; p = 0.00001) and 1.51 (95% CI: 1.3–1.74; p = 0.00001), respectively. The pooled risk ratios for all-grade and high-grade adverse events were 1.01 (95% CI: 0.99–1.03; p = 0.27) and 1.17 (95% CI: 1.07–1.28; p = 0.0006), respectively. Conclusion: Add-on immunotherapy significantly improves PFS, OS and ORR for the first-line treatment of advanced NSCLC with a reasonable safety profile.

ligand 1 (PD-L1) and cytotoxic T-cell lymphocyte antigen-4 (CTLA-4) monoclonal antibodies, which promote T-cell activation with subsequent formation of anti-tumor effect, resulting in durable response and improvement in outcome. Single-agent immunotherapy (nivolumab, pembrolizumab and atezolizumab) demonstrated superior overall survival (OS) and better safety profiles compared with chemotherapy in the subsequent management of both squamous and nonsquamous NSCLC [4][5][6][7]. However, single-agent immunotherapy in the first-line setting is limited to the small subset of NSCLC patients whose tumors have a PD-L1 tumor proportion score of 50% or more 'without EGFR or ALK genomic tumor aberrations', for which pembrolizumab is proven to be superior in terms of efficacy and safety profiles [8]. Increasing evidence suggests that combined chemoimmunotherapy can have synergistic anticancer activities through the immunomodulatory effect of checkpoint inhibitors and the immunogenic effect of chemotherapy, such as lowering regulatory T-cell activity and enhancing cross-presentation of tumor antigens [9,10].
Several randomized controlled trials (RCTs) have shown that the addition of immunotherapy to standard chemotherapy improves survival with manageable toxicity profiles. RCTs for the CTLA-4 inhibitor ipilimumab were not included in the analysis owing to separate mechanism of action, lack of OS benefit and different toxicity profiles [11,12]. Therefore, we conducted the meta-analysis of RCTs on PD-1 and PD-L1 inhibitors to evaluate the efficacy and safety of immune checkpoint inhibitors in combination with chemotherapy for the first-line treatment of advanced, metastatic NSCLC.

Methods
We conducted this systematic review according to the Cochrane Handbook for Systematic Reviews [13] and reported per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.

Search methods
We conducted literature search in PubMed, EMBASE and SCOPUS databases using the terms 'immune checkpoint inhibitors and NSCLC', 'nivolumab and NSCLC', 'pembrolizumab and NSCLC', 'atezolizumab and NSCLC', 'avelumab and NSCLC' and 'durvalumab and NSCLC'. A further search was performed on major oncology conferences throughout January 2019, including those of the American Society of Clinical Oncology, European Society of Medical Oncology and International Association for the Study of Lung Cancer. Clinical trials in English were retrieved and filtered, as mentioned in eligibility criteria.

Inclusion & exclusion criteria
The search results were narrowed to the following article types: clinical trial, Phase II; clinical trial, Phase III; checkpoint inhibitor plus chemotherapy versus chemotherapy plus placebo; studies that showed survival data and studies conducted for first-line treatment of advanced or metastatic NSCLC.
The exclusion criteria were as follows: review articles, systematic reviews, letter to editor and case reports; preclinical trials, Phase I trials, or nonrandomized trials; immune checkpoint inhibitors in second-line setting; immune checkpoint inhibitors in adjuvant or neoadjuvant settings; duplicates of previous publications on the same population and study of CTLA-4 inhibitor ipilimumab due to lack of OS benefit in combination with standard chemotherapy in the first-line setting [11,12]. model was applied for all analyses due to heterogeneity among studies. We utilized the inverse variance method to analyze PFS and OS data and reported the outcomes as pooled HRs. Analysis of dichotomous outcomes, such as ORR and AEs, were done by the Mantel-Haenszel method and were reported as risk ratios (RRs) with 95% CIs. Subgroup analyses HRs for PFS and OS were conducted based on the degree of PD-L1 expressions (PD-L1 negative; low-PD-L1 expression: PD-L1 tumor proportion score of 1-49% for pembrolizumab trials or PD-L1 expression on 1-49% of tumor cells (TCs) or 1-9% of tumor-infiltrating immune cells (ICs) for atezolizumab trials; and high-PD-L1 expression: PD-L1 of 50% or greater for pembrolizumab trials or PD-L1 expression of 50% or greater on TCs or 10% or higher ICs in atezolizumab trials). We also performed subgroup analyses of PFS based on age (<65 vs ≥65 years), sex (male vs female) and Eastern Cooperative Oncology Group (ECOG) performance status (0 vs 1), and smoking status (current or former vs never). Publication bias was assessed by funnel plots. We did not perform sensitivity analysis since no study notably influences the results.

Study selection
We retrieved 8409 potential references, and 5044 duplicates were removed. After application of exclusion criteria as mentioned above, seven RCTs were reviewed for the final analysis. We excluded Keynote 021 trial, which is a Phase II trial with expansion cohort from Phase I trial [16]. The data from IMpower-131, -132 and Checkmate-227 trials were extracted from conference abstracts and presentations. (Figure 1) We incorporated additional data from IMpower-130 trial in the analysis following its publication in July 2019 [17]. Figure 1 demonstrates study selection in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.

Study characteristics
The characteristics of the included studies are summarized in Table 1. A total of 4322 patients with advanced NSCLC were included in the meta-analysis. Overall, 2991 patients (69%) had nonsquamous NSCLC, among which 152 patients (5%) had EGFR or ALK alterations. A total of 1331 patients (31%) had squamous NSCLC. The median age of patients ranged from 63 to 65 years; all patients had ECOG performance status score of 0-1 with adequate organ function. All studies were done for the first-line treatment of advanced or metastatic NSCLC utilizing a platinum-based regimen with or without immunotherapy (nivolumab for Checkmate trial, atezolizumab for IMpower trials and pembrolizumab for Keynote trials). Data on the atezolizumab, carboplatin and paclitaxel study arms of IMpower-131 trial (three-arm trial) were not available and not included in the analysis. Another three-arm trial, IMpower-150, had two experimental arms: carboplatin, paclitaxel and atezolizumab (arm A) and carboplatin, paclitaxel, bevacizumab and atezolizumab (arm B) versus carboplatin, paclitaxel and bevacizumab (arm C). We did not include the data comparing between arm A and C because the HR may not reflect the actual effect of add-on immunotherapy (atezolizumab plus chemotherapy vs bevacizumab plus chemotherapy).

Study quality, risk bias & publication bias
Risk of bias for each study was evaluated by RevMan 5.3 software (Cochrane) and is illustrated in Figure 2. IMpower-130, -131, -132, -150 and Checkmate-227 were open-label studies that lacked blinding between investigators and participants. Detection bias was unclear for IMpower-130,-131,-132,-150, and Checkmate-227 trials due to lack of blinding. Moreover, all the studies are sponsored by pharmaceutical companies so other biases remained uncertain. Publication bias was not identified in this study.

Discussion
Multiple randomized clinical trials have been conducted to identify the optimal chemoimmunotherapy treatment strategy for advanced NSCLC patients. Chen et al. conducted pooled efficacy and safety analyses of immune checkpoint inhibitors in NSCLC patients as the first-line treatment option [27]. The study showed statistically improved PFS, OS but not ORR. It is important to note that the analysis also included studies that compare immune checkpoint inhibitors against chemotherapy. We excluded these studies in our review [27]. The metaanalysis done by Xu et al. showed improvement in PFS but not OS in first-line treatment of NSCLC [28]. The study mainly includes Phase I trials [28]. Analysis by Shen et al. showed improvement in PFS, OS and ORR [29]. A recent meta-analysis by Addeo et al. [30] incorporated additional Phase III studies (IMpower-130, -131, -132 and Checkmate-227 studies) that showed significantly prolonged PFS and OS with the addition of immune checkpoint inhibitor [30]. In addition to PFS and OS analyses of EGFR and ALK wild-type patient population, we analyzed pooled HR for PFS in patients with EGFR and ALK alterations ( Figure 4I) that showed statistically significant PFS benefit. However, this benefit is mainly driven by the IMpower-150 trial, which utilized bevacizumab plus chemotherapy as a backbone regimen. We also performed a comprehensive review of safety profiles. Our study met the primary end point of significantly improved OS. The study also revealed statistically significant improvement in PFS and ORR. Pooled analysis for OS based on histologic subtypes favored both histologic subtypes (both squamous NSCLC and nonsquamous NSCLC) but was not statistically significant for the squamous NSCLC subtype. A longer follow-up of the IMpower-131 trial may change this result. Besides, we identified substantial heterogeneity with I 2 of 82% among squamous NSCLC studies with the favorable outcome being driven by the pembrolizumab study ( Figure 3B). The pooled HRs for PFS showed statistically significant benefits for both squamous and nonsquamous NSCLC. Moreover, statistically significant OS and PFS benefits were seen on separate analyses of PD-1 and PD-L1 monoclonal antibodies.
Subgroup analyses of both OS and PFS based on the degree of PD-L1 expressions yielded the statistically significant OS and PFS benefits across different levels of PD-L1 expressions, except pooled HR for OS in patients with low PD-L1 expression (HR: 0.77; 95% CI: 0.55-1.08; p = 0.12). We noted substantial heterogeneity in this analysis. The HRs from the Keynote trials contributed more to the survival benefit in low PD-L1 subgroup. In addition, studies utilized different PD-L1 assay methods, further complicating the picture. Nonetheless, subgroup analyses should be interpreted with caution since they are observational by nature and are not based on randomized comparisons. There are significant false positive and false negative findings which could be misleading [13].
Immune checkpoint inhibitors showed relatively tolerable safety profiles. There was no statistically significant increase in rates of all-grade AEs, but a slight increase in high-grade (grade 3 or higher) AEs with the addition of immunotherapy.
There are several limitations to our review. The PD-L1 assay methods are not consistent across different studies; in the atezolizumab studies, PD-L1 immunohistochemistry is read on both TCs and tumor-infiltrating ICs [20,21,23]. However, trials of nivolumab and pembrolizumab applied PD-L1 expression only on TCs [18,19,25]. Moreover, we require longer follow-up and more mature data from the trials, which may change the overall efficacy and safety in the future. At last, our analysis is not designed to identify the optimal combination strategy, but to prove the impact of add-on immunotherapy to chemotherapy in patients with advanced NSCLC.

Conclusion
Overall, our meta-analysis suggests that the addition of immune checkpoint inhibitor to standard chemotherapy benefited OS and PFS, including tumors with EGFR and ALK alterations in first-line, advanced, metastatic NSCLC. The analysis also showed statistically significant improvement in ORR in the overall patient population with the addition of immune checkpoint inhibitor. The combined regimen had a slight increase in high-grade AEs without a statistically significant increase in rates of all-grade AEs.

Future perspective
It is clear that checkpoint blockade, along with cytotoxic chemotherapy, provides additional therapeutic benefit. Despite successful incorporation of PD-1 and PD-L1 inhibitors into the management of metastatic lung cancer, achieving durable remission remains a challenge. Understanding of complex tumor microenvironments and tumor-IC interaction plays a crucial role in the development of novel therapeutic strategies. Combination strategies utilizing multiple immune-checkpoint blockades is an attractive option and have been implemented in the Checkmate-227 trial [26]. Further strategies may include PD-1 blockade in combination with concurrent activation of the immune system, such as vaccination and use of stimulating antibodies. A greater understanding of molecular medicine along with molecular subclassification of this heterogeneous disease and identification of reliable biomarkers to tailor an optimal treatment strategy is crucial in the era of precision oncology. Additional basic science and clinical research will likely help us to understand more about molecular biology, develop newer biomarkers and evolve new therapeutic approaches which will ultimately improve long-term outcome.

Summary points
• Immune checkpoint blockade plus chemotherapy is proven to be effective in randomized controlled trials (RCTs) for the treatment of lung cancer. • We conducted a literature search for RCTs in advanced non-small-cell lung cancer (NSCLC); seven RCTs with a total of 4322 patients were included in the study. • The study population comprises nonsquamous NSCLC, 69% (5% had EGFR or ALK alterations) and squamous NSCLC, 31%. • The analysis showed that the addition of an immune checkpoint inhibitor lead to improvement in overall survival, progression-free survival and objective response rate. • Addition of atezolizumab to standard therapeutic regimen improved progression-free survival in cohorts of patients with EGFR and ALK altered nonsquamous NSCLC. • Subgroup analysis failed to show OS benefit from add-on immunotherapy in patients with squamous NSCLC that expresses low PD-L1. • The study showed slightly higher rates of high-grade (grade 3 or higher) AEs, but no significant difference in rates of all-grade AEs with addition of immune checkpoint inhibitor. • We identified substantial heterogeneity among studies. The programmed death ligand 1 assay methods are not consistent across different studies. • Long-term follow-up is required for mature data.
• Increasing understanding of molecular biology and the development of newer biomarkers and novel therapeutic approaches will improve outcome in this patient population.