Programmed Cell Death 1 (PD-1) Ligand (PD-L1) Expression in Solid Tumors As a Predictive Biomarker of Benefit From PD-1/PD-L1 Axis Inhibitors: A Systematic Review and Meta-Analysis

Abstract

Purpose Drugs targeting the programmed cell death 1 (PD-1)/programmed cell death ligand 1 (PD-L1) pathway show significant clinical activity across several tumor types. However, a majority of patients do not respond to these agents. Use of biomarker assays to predict response to these agents is an active area of research; however, the predictive value of PD-L1 immunohistochemistry (IHC) assays is largely inconsistent across clinical trials. In this meta-analysis of clinical trials of PD-1/PD-L1–targeted agents, we evaluate the predictive value of a tumor and tumor-infiltrating immune cell PD-L1 IHC assay as a biomarker for objective response to PD-1/PD-L1 inhibitors. Methods We searched databases (PubMed, Medline, ASCO abstracts, European Society for Medical Oncology abstracts, and Scopus) up until December 2016 for clinical trials using PD-1/PD-L1 inhibitors with reported PD-L1 biomarker data. Objective response rates (primary end point) from all phase I to III trials investigating nivolumab, pembrolizumab, atezolizumab, durvalumab, and avelumab in advanced solid tumors were collected. Odds ratios (ORs) for response in PD-L1–positive patients compared with PD-L1–negative patients were calculated using the DerSimonian-Laird random effects model to combine trials. We performed meta-analysis as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Results Forty-one distinct trials with 6,664 patients were identified. PD-L1 expression was predictive of favorable response across all tumor types (OR, 2.26; 95% CI, 1.85 to 2.75; P < .001), with the significantly largest effect observed in non–small-cell lung cancer (OR, 2.51; 95% CI, 1.99 to 3.17; P < .001). A subgroup analysis across all non–small-cell lung cancer trials using nivolumab and Dako clone 28-8 (Dako, Carpinteria, CA) IHC antibody assay yielded a significantly higher objective response rate in patients with tumor PD-L1 expression even at the minimum cutoff value of 1% (OR, 2.17; 95% CI, 1.03 to 4.57). Conclusion Our meta-analysis shows that tumor and tumor-infiltrating immune cell PD-L1 overexpression based on IHC is associated with significantly higher response rates to PD-1/PD-L1 axis inhibitors across a range of malignant solid tumors.