First-line PD-1/PD-L1 inhibitor plus chemotherapy vs chemotherapy alone for negative or < 1% PD-L1-expressing metastatic non-small-cell lung cancers

Abstract

Aims

Single-agent anti-PD-1/PD-L1 clinical efficacy against < 1% PD-L1-expressing non-small-cell lung cancers (NSCLCs) is controversial.

Methods

This meta-analysis examined randomized-trial data comparing first-line PD-1/PD-L1-inhibitor + chemotherapy (CT) vs CT alone for advanced < 1% PD-L1 NSCLCs. Outcome measures included overall survival (OS), progression-free survival (PFS) and objective response rate (ORR).

Results

IMpower (atezolizumab + CT), Keynote (pembrolizumab + CT) and CheckMate (nivolumab + CT) trials included 2037 NSCLCs (1246 PD-L1–negative; 791 < 1% PD-L1 expression). Anti-PD-1/PD-L1 + CT was significantly associated (hazard ratio [95% confidence interval]) with prolonged OS (0.75 [0.63–0.89]; p = 0.0008) and PFS (0.72 [0.65–0.80]; p < 0.0001), and higher ORR (odds ratio 2.06 [1.50–2.83]; p < 0.0001).

Conclusions

First-line anti-PD-1/PD-L1 + CT combination appears superior to CT alone for advanced, < 1% PD-L1-expressing NSCLCs for OS, PFS and ORR.

Introduction

Almost half of non-small-cell lung cancers (NSCLCs) are diagnosed at a metastatic or advanced stage. Over the past several decades, in the absence of oncogenic addiction, treatment of such disease has relied on a platinum-based chemotherapy doublet (Sandler et al. 2006). But the survival results under such treatment have remained disappointing. The relatively recent arrival, especially for lung cancers, of immune-checkpoint inhibitors (ICIs), i.e., directed against programmed cell-death protein (PD-1) and its ligand (PD-L1), has transformed their management. As second-line therapy, three anti-PD1/PD-L1 molecules have been approved. Compared to docetaxel, the reference second-line chemotherapy, nivolumab (anti-PD-1) and atezolizumab (anti-PD-L1) prolonged overall survival (OS) (Borghaei et al. 2015; Brahmer et al. 2015; Herbst et al. 2016; Rittmeyer et al. 2017), independently of cancer-cell PD-L1 expression, and pembrolizumab did so for patients whose NSCLCs expressed > 1% PD-L1.

More recently, as first-line therapy, the results of several trials showed the benefit of using anti-PD-1/PD-L1, either alone or in combination with chemotherapy. In the phase-3 Keynote-024 trial, pembrolizumab monotherapy was superior to chemotherapy in terms of response and OS for patients with ≥ 50% of their tumor cells expressing PD-L1 (Reck et al. 2016). In the more recent Keynote-042 trial, on patients with NSCLCs expressing ≥ 1% PD-L1, Pembrolizumab also obtained improved efficacy, but with a more modest clinical benefit (Mok et al. 2019). In contrast, a phase-3 trial comparing nivolumab alone to chemotherapy did not find any OS benefit for patients with advanced NSCLCs expressing > 5% PD-L1 (Carbone et al. 2017). In this context, several phase-3 trials have evaluated the potential of combining ICIs and platinum-based chemotherapy, regardless of the PD-L1-expression level, for first-line therapy. The outcomes were positive, for the most part, concerning the principal criterion [progression-free survival (PFS) or OS], with the benefit being apparently even greater for patients whose tumors strongly expressed PD-L1 and uncertainty about a real benefit for those with low (< 1%) expression.

The objective of this study was to determine, through meta-analysis of published study results, whether such anti-PD-1/PD-L1+ chemotherapy combination therapy has a real benefit for patients with NSCLCs expressing < 1% PD-L1.

Methods

This meta-analysis of reported data concerned first-line therapy of advanced NSCLCs without oncogenic addiction. The goal was to compare an ICI–chemotherapy combination to chemotherapy alone for patients whose tumors expressed < 1% PD-L1.

Research strategy

We searched the PubMed and Cochrane databases on 30 April 2019. We also manually searched the abstracts on our subject accepted by the AACR, ASCO, ESMO, WLCC and ELCC congresses until April 2019. The following search terms were used: “immune checkpoint inhibitor or immunotherapy”, “nivolumab or pembrolizumab or atezolizumab or avelumab or durvalumab”, “advanced or metastatic”, “non-small-cell lung cancer or NSCLC”, “PD-1 or PD-L1”, and “randomized controlled trial.” Only phase-2 or -3 randomized trials comparing first-line, anti-PD-1/PD-L1+ chemotherapy to chemotherapy alone were retained.

Data extraction

All the study data were extracted independently by two readers using a predefined protocol; discrepancies were resolved by discussion with a third reader. The following information was collected: each patient’s characteristics [e.g., age, sex, description and doses of treatment(s) administered, tumor histology], tumor stage, PD-L1-expression level.

The therapeutic efficacy criteria analyzed were OS, PFS, objective response rate (ORR) and median duration of response (DOR). We also analyzed the same outcomes in a second meta-analysis comparing chemotherapy alone vs chemotherapy + anti-PD-1 or chemotherapy + anti-PD-L1.

Statistical analyses

The Cochrane collaboration method for meta-analysis, using the Review Manager software (RevMan version 5.3; Oxford, UK), was applied to compute the analyses. Statistical heterogeneity was assessed with χ² tests and I² statistics, with a χ² test p < 0.10 indicating heterogeneity and I² values of 30–60% corresponding to moderate heterogeneity.

A fixed effect model was used to calculate the cumulative hazard ratio (HR) when heterogeneity among studies was weak and a random model was applied when that heterogeneity was marked. Subgroup analyses were run each time for the different anti-PD-1/PD-L1 ICIs for patients with < 1% PD-L1 expression. Meta-analysis results with 95% confidence interval (CI) are reported with odds ratios (ORs) for the ORRs, and HRs for OS and PFS. All tests were two sided and p < 0.05 defined significance. Median DOR differences were compared with “t Student’s” test. A Begg’s funnel plot was used to analyze between-population heterogeneity.

Results

The electronic search identified 273 references and the manual search of congress abstracts added 6 more (Fig. 1). Ten studies meeting inclusion criteria were retained for this analysis but two were finally excluded: the POSEIDON trial (Mok et al. 2017), whose results are not yet available, and Checkmate-227 part 1, whose participants’ NSCLCs had > 1% PD-L1 expression (Hellmann et al. 2018).

Fig. 1.

Flow diagram showing inclusion and exclusion of articles

Thus, this meta-analysis included eight studies, seven phase-3 and one phase-2 randomized; three had evaluated combinations with pembrolizumab (Papadimitrakopoulou et al. 2017; Gandhi et al. 2018; Paz-Ares et al. 2018), four with atezolizumab (Socinski et al. 2018; Jotte et al. 2018; Papadimitrakopoulou et al. 2018; Cappuzzo et al. 2018) and one with nivolumab (Borghaei et al. 2018) (Table 1). Concerning histological type, five studies included only patients with non-squamous-cell tumors, two those with squamous-cell cancers, and both histological types were included in one. These 8 studies included 2037 patients, 1246 with negative PD-L1 expression and 791 with < 1% PD-L1 expression. They were predominantly men, smokers, with median age 64 years; 1423 (69.8%) had non-squamous-cell NSCLCs, the remaining 614 (30%) were squamous-cell NSCLCs. Characteristics of the patients enrolled in the eight trials are reported in Table 2. The Begg’s funnel plot (Fig. 2) indicated heterogeneity among populations included.

Table 1.

Published studies included in this meta-analysis

Trial (ref)	Phase	Drug combinations: arm 1 vs arm 2	Randomized	Patients enrolled, n	NSCLC histology	Endpoint (s)
Keynote-021G (Papadimitrakopoulou et al. 2017)	II-R	Pembrolizumab + ChT vs ChT (carboplatin–pemetrexed)	1:1	44 (< 1% PD-L1+)	Nsq	ORR
Keynote-407 (Paz-Ares et al. 2018)	III	Pembrolizumab + ChT vs ChT (carboplatin–paclitaxel)	2:1	194 (< 1% PD-L1+)	Sq	OS and PFS
Keynote-189 (Gandhi et al. 2018)	III	Pembrolizumab + ChT vs ChT (platin–pemetrexed)	2:1	190 (< 1% PD-L1)	Nsq	OS and PFS
Checkmate-227 (Borghaei et al. 2018)	III	Nivolumab + ChT vs ChT (platin–pemetrexed) or (platin–gemcitabine)	1:1:1	363 (< 1% PD-L1+)	Nsq and Sq: < 1% PD-L1+	OS
IMpower-130 (Cappuzzo et al. 2018)	III	Atezolizumab ChT vs ChT (carboplatin–n-paclitaxel)	2:1	356 (PD-L1−)	Nsq	PFS and OS
IMpower-131 (Jotte et al. 2018)	III	Atezolizumab + ChT vs ChT (carboplatin–n-paclitaxel)	1:1:1	331 (PD-L1−)	Sq	PFS and OS
IMpower-132 (Papadimitrakopoulou et al. 2018)	III	Atezolizumab + ChT vs ChT (platin–pemetrexed)	1:1	163 (PD-L1−)	Nsq	PFS and OS
IMpower-150 (Socinski et al. 2018)	III	Atezolizumab + ChT + Bev vs ChT + Bev (carboplatin–paclitaxel)	1:1:1	396 (PD-L1−)	Nsq	PFS and OS

NSCLC non-small-cell lung cancer, ChT chemotherapy, PD-L1 programmed cell-death protein-1 ligand, Nsq non-squamous cell, ORR objective response rate, OS overall survival, PFS progression-free survival, Bev bevacizumab

Table 2.

Characteristics of patients from the eligible clinical studies

Trial	n	Median age (years)	Males, n (%)	Ever smoker, n (%)
CheckMate-227	363	64	225 (62)	NA
IMpower-130	679	65	400 (59)	614 (90)
IMpower-131	683	65	557 (82)	627 (92)
IMpower-132	578	64	384 (66)	511 (88)
IMpower-150	692	63	425 (61)	584 (84)
Keynote-189	616	65	363 (59)	543 (88)
Keynote-407	559	65	455 (81)	NA
Keynote-021	123	64	48 (39)	99 (80)

NA not applicable

Fig. 2.

Funnel-plot-assessed between-population heterogeneity

The meta-analysis of the four studies with survival data showed a significant OS benefit for the ICI–chemotherapy combination, compared to chemotherapy alone (HR 0.75 [95% CI 0.63–0.89]; p = 0.0008). That benefit was also found for chemotherapy combinations with pembrolizumab or atezolizumab (Fig. 3).

Fig. 3.

Meta-analysis of overall survival

ICI–chemotherapy, compared to chemotherapy alone, for patients whose NSCLCs had < 1% PD-L1 expression, also obtained a significant PFS benefit (HR 0.72 [95% CI 0.65–0.80]; p < 0.00001), regardless of the ICI molecule combined with chemotherapy (Fig. 4).

Fig. 4.

Meta-analysis of progression-free survival

The ORR for combination ICI–chemotherapy was statistically superior to that of chemotherapy alone in six of the seven studies for which results have been published (OR 2.06 [1.50–2.83]; p < 0.00001). Because of heterogeneity between studies (χ² avec p < 0.10 and I² = 46%), a random model was used to determine ORs. The benefit was obtained, regardless of the nivolumab or pembrolizumab combination administered. Significant benefits were observed with atezolizumab–bevacizumab–chemotherapy, (IMpower-150) or atezolizumab–platinum–pemetrexed (IMpower-132) (Fig. 5).

Fig. 5.

Meta-analysis of the objective response rate

The median DOR was reported in only four of the eight publications for patients with negative or < 1% PD-L1-expressing NSCLCs; it was significantly longer for patients given the combination therapy: 8.1 vs 4.9 months (p < 0.0008) (Table 3). The second meta-analysis comparing chemotherapy alone or combined with anti-PD-1 or anti-PD-L1 revealed no significant PFS (Fig. 6) or ORR (Fig. 7) differences according to the ICI used.

Table 3.

Median duration of response (DOR) for patients with negative or < 1% PD-L1-expressing NSCLCs

Trial	Median DOR months: ICI + ChT vs ChT
IMpower-131	6.9 vs 5.2
IMpower-132	10.1 vs 4.2
IMpower-150	8.2 vs 5.5
CheckMate-227	7.2 vs 4.7

NSCLCs non-small-cell lung cancers, ICI immune-checkpoint inhibitor, ChT chemotherapy

Fig. 6.

Forest plot of the anti-PD-1 or anti-PD-L1 impact on PFS

Fig. 7.

Forest plot of the anti-PD-1 or anti-PD-L1 impact on ORR

Discussion

PD-1/PD-L1 inhibitors have demonstrated their efficacy as first-line treatment for NSCLCs without EGFR mutation or ALK translocation. The results of several preclinical studies showed the immunostimulatory effects of the majority of chemotherapy drugs (Heinhuis et al. 2019), which justified their use in combination with ICIs. All phase-3 trials on these combinations conducted in this context obtained PFS and OS efficacy compared to chemotherapy alone, with an even greater benefit as tumor expression of PD-L1 increased and uncertain effectiveness against negative or < 1% PD-L1-expressing NSCLCs (Landre et al. 2019).

This meta-analysis demonstrated comparable PFS, OS and ORR efficacies of ICI–chemotherapy combinations, compared to chemotherapy alone, for the subgroups of patients with negative or < 1% PD-L1-expressing NSCLCs. No difference was observed when the analysis was restricted to anti-PD-1 or anti-PD-L1. The results of other recently published meta-analyses (Zhou et al. 2018; Wang et al. 2019) are in agreement with our findings, even though they did not specifically address the subgroup of patients with negative or < 1% PD-L1-expressing NSCLCs and they used different inclusion methodologies.

Our study has some limitations. It was a meta-analysis based on published data, not individual patient data; in addition, the OS information in four of the eight papers analyzed did not yet have sufficient hindsight. The populations included were not strictly the same: patients enrolled in the IMpower-150 trial could have an EGFR-activating mutation or an ALK translocation. Finally, clinical and/or biological tolerance of the different combinations in patients with negative or < 1% PD-L1-expressing NSCLCs could not be assessed, because those specific data were not available. Only the CheckMate-227 trial provided specific toxicity data (Borghaei et al. 2018). According to the meta-analysis of Zhou et al. (2018), adjunction of anti-PD-1/PD-L1 to chemotherapy was significantly associated with an increased risk of grade-3/4 adverse events, immune toxicities and the treatment-discontinuation rate because of adverse events, but without significantly increasing the numbers of treatment-associated deaths or adverse events linked to stopping treatment.

Conclusion

The ICI–chemotherapy combinations, compared to chemotherapy alone, as first-line treatment of patients with negative or < 1% PD-L1-expressing NSCLCs significantly prolonged OS and PFS. Hence, ICI–chemotherapy combinations can now be considered the new standard-of-care regimen for these patients. Future studies are needed to determine the optimal combinations in terms of efficacy and toxicity, as a function of the chemotherapy and ICI molecules prescribed. The costs of these agents and the quality of life they procure should also continue to be evaluated in clinical practice.

Executive summary

• Clinical efficacy of anti-PD-1/PD-L1 agents in patients with negative or < 1% PD-L1-expressing non-small-cell lung cancers is controversial.

• This meta-analysis examined data from randomized trials comparing first-line PD-1/PD-L1-inhibitor + chemotherapy vs chemotherapy alone to treat advanced NSCLCs.

• This meta-analysis was based on 2037 patients (1246 PD-L1-negative and 791 with < 1% PD-L1 expression), 69.9% non-squamous-cell cancers, median age of 64 years.

• For patients with untreated, advanced, negative or < 1% PD-L1-expressing NSCLCs, the anti-PD-1/PD-L1+ chemotherapy combination, compared to chemotherapy alone, was associated with significantly prolonged OS and PFS, and improved ORR.

Abbreviations

ALK

Anaplastic lymphoma kinase

ChT

Chemotherapy

CI

Confidence interval

DOR

Duration of response

EGFR

Epidermal growth factor receptor

HR

Hazard ratio

ICI

Immune-checkpoint inhibitor

NSCLC

Non-small-cell lung cancer

Nsq

Non-squamous cell

OR

Odds ratio

ORR

Objective response rate

OS

Overall survival

PD-1

Programmed cell-death protein-1

PD-L1

Programmed cell-death protein-1 ligand

PFS

Progression-free survival

Sq

Squamous cell

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.