A Network Meta-Analysis of Cancer Immunotherapies <i>Versus</i> Chemotherapy for First-Line Treatment of Patients With Non-Small Cell Lung Cancer and High Programmed Death-Ligand 1 Expression.

Overview

abstract

In the absence of head-to-head trials of first-line treatments for metastatic non-small cell lung cancer (NSCLC), synthesis of available evidence is needed. We conducted a systematic literature review and network meta-analysis of randomized controlled trials in patients with stage IV NSCLC and high programmed death-ligand 1 (PD-L1) expression. Patients with other-stage NSCLC or without PD-L1 expression and populations with < 80% stage IV NSCLC were excluded. Outcomes included overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and treatment-related adverse events. English records from MEDLINE and Embase published through October 2020 were eligible, supplemented by hand searches of other sources. Three evidence networks were constructed based on histology (mixed, squamous, non-squamous). OS and PFS results were analyzed applying Bayesian fractional polynomial random-effects models. Hazard ratios over time with 95% credible intervals (CrIs) and expected differences in OS and PFS between each cancer immunotherapy regimen and the chemotherapy common comparator were generated. Seventeen clinical trials were included after screening 32,527 records. Heterogeneity and risk of bias were generally low across trials. In the mixed-histology network of PD-L1-high patients, expected OS was significantly longer with atezolizumab (estimated difference: 10.4 months [95% CrI: 1.9, 18.2]), pembrolizumab (7.2 [2.2, 12.3]), and cemiplimab (13.0 [4.2, 21.0]) versus chemotherapy but not with nivolumab (3.5 [-2.5, 10.6]) or nivolumab plus ipilimumab (6.7 [-0.5, 14.2]) versus chemotherapy. OS improvements were not significant compared with chemotherapy for any regimen in the squamous and non-squamous networks, except pembrolizumab plus chemotherapy in the non-squamous network. All regimens showed significantly longer expected PFS versus chemotherapy in the non-squamous network, whereas the increases were not significant in the mixed or squamous networks. ORR was significantly higher with pembrolizumab and cemiplimab versus chemotherapy in the mixed-histology network, with sintilimab in the non-squamous network, and with combination regimens, including pembrolizumab or atezolizumab, in the squamous and non-squamous networks, except with atezolizumab plus carboplatin, paclitaxel, and bevacizumab. Survival and safety versus chemotherapy were generally similar across cancer immunotherapies and histology networks. These findings may support treatment decisions for patients with high PD-L1 status receiving first-line treatment for NSCLC.