Programmed death-ligand 1 testing in patients with non-small cell lung cancer: Ready for prime time?

Abstract

Lung cancer is the leading cause of cancer-related death in the United States, claiming >155,000 lives per year.1 Historically, patients with advanced non–small cell lung cancer (NSCLC) were treated with cytotoxic chemotherapy. Despite modest improvements in survival with chemotherapy, such therapy is not curative, and the prognosis for individuals with advanced NSCLC remains poor. In recent years, the management of patients with NSCLC has been transformed by the emergence of 2 new treatment paradigms: 1) targeted therapy and 2) immunotherapy. Targeted therapy generally involves the use of small molecule inhibitors directed against cancers harboring activating mutations in oncogenes, such as the epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK). Conversely, immunotherapy aims to stimulate antitumor immune responses as a means with which to control cancer growth. A central tenet of cancer immunotherapy is that cancer cells often possess numerous genetic alterations, which may in turn generate “neoantigens” that are capable of being recognized by the host immune system.2 Early attempts to harness this potential therapeutically focused on the development of cancer vaccines and relatively nonspecific cytokine-based therapies (eg, interleukin 2), which produced limited successes. However, more recently, the field of cancer immunotherapy has made significant strides through the development of monoclonal antibodies targeting negative regulators of the immune response known as “immune checkpoints.” In particular, monoclonal antibodies against the programmed cell death protein 1 (PD-1) and its ligand, programmed death-ligand 1 (PD-L1), have demonstrated significant antitumor activity in a broad spectrum of malignancies, ranging from Hodgkin lymphoma to melanoma and NSCLC. In patients with NSCLC, 2 PD-1 inhibitors (nivolumab and pembrolizumab) and a PD-L1 inhibitor (atezolizumab) recently gained regulatory approval for the management of previously treated patients with NSCLC in the United States. These initial approvals were based in part on a series of randomized controlled trials comparing PD-1/PD-L1 inhibitors with standard single-agent cytotoxic chemotherapy delivered in the second-line setting.3-6 In each study, PD-1/PD-L1 inhibitors demonstrated significant improvements in overall survival compared with single-agent chemotherapy. Furthermore, with the exception of the CheckMate 017 study,4 these studies found that the presence of PD-L1 expression in tumor specimens was associated with a greater benefit to PD-1/PD-L1 inhibition.3, 5, 6 It is important to note that treatment-related severe toxicities also were less common among patients receiving PD-1/PD-L1 inhibitors compared with those receiving chemotherapy. As a result, PD-1/PD-L1 inhibitors have become the new standard of care for the management of previously treated patients with NSCLC, even in the absence of PD-L1 expression. Given the success of using immune checkpoint inhibitors in patients with previously treated NSCLC, there has been considerable interest in moving these agents to the front-line setting. However, because only a subset of patients with NSCLC respond to PD-1/PD-L1 inhibitors (approximately 20% of unselected patient populations),3, 4 first-line trials generally have relied on biomarker-enrichment strategies using tumor PD-L1 expression as an entry criterion. For example, in the randomized phase 3 KEYNOTE-024 trial, patients with previously untreated NSCLC and PD-L1 expression in ≥ 50% of tumor cells were randomized to receive either pembrolizumab (a PD-1 inhibitor) or platinum-based chemotherapy, which is the historical gold standard.7 In this study, pembrolizumab produced significant improvements in response rate, duration of response, progression-free survival, and overall survival. Moreover, treatment-related adverse events were less common with pembrolizumab. Importantly, in a contemporaneous phase 3 study, CheckMate 026, previously untreated patients with NSCLC with PD-L1-expressing tumors were randomized to receive either the PD-1 inhibitor nivolumab or platinum-based chemotherapy.8 It is interesting to note that this study, which used a ≥5% cutoff value for tumor PD-L1 expression in the final analysis, failed to meet its primary endpoint of superior progression-free survival compared with chemotherapy. Taken together, these studies have informed a new standard of care for patients with newly diagnosed, advanced NSCLC. Specifically, for those patients without targetable alterations in oncogenic drivers (eg, EGFR or ALK), patients with high PD-L1 expression (≥50% of tumor cells) should now be treated with PD-1 inhibition, whereas those with low PD-L1 expression (<50% of tumor cells) should continue to receive platinum-based chemotherapy. Despite the growing importance of PD-L1 immunohistochemistry (IHC) in patients with NSCLC, testing raises several challenges for clinicians and pathologists alike, particularly when viewed within the context of other predictive biomarkers in NSCLC. First, although genetic biomarkers (eg, EGFR and ALK mutation testing) generally produce binary results, PD-L1 protein expression exists on a biological continuum (ie, a continuous variable), and thereby necessitates the development of cutoff values to define PD-L1 positivity. Second, PD-L1 expression can be heterogeneous and inducible in response to therapies. Third, although EGFR mutations and ALK rearrangements are highly predictive of responsiveness to targeted therapies in patients with NSCLC, PD-L1 expression is less discriminating as a biomarker. Indeed, responses to PD-1/PD-L1 inhibitors have been observed among both PD-L1-positive and PD-L1-negative patients, although response rates generally are higher among PD-L1-positive patients.3, 6 Thus, PD-L1 expression is best viewed as an important enrichment biomarker at this point in time. From a practical standpoint, PD-L1 testing has been complicated by the emergence of multiple different diagnostic assays with which to evaluate PD-L1 expression. Indeed, 4 separate IHC assays have been developed alongside PD-1/PD-L1 inhibitors in the clinic, and 3 of these tests recently gained regulatory approval in the United States as either companion (PD-L1 IHC 22C3 assay [Dako/Agilent, Carpinteria, Calif], pembrolizumab) or complementary (PD-L1 IHC 28-8 assay [Dako/Agilent], nivolumab; and Ventana PD-L1 SP142 assay [Ventana Medical Systems Inc, Tucson, Ariz], atezolizumab) diagnostics. Although companion diagnostics are necessary for the safe and effective use of a drug, complementary diagnostics are not required for use. Instead, complementary diagnostics may provide additional information regarding the benefits of a given therapy.2 Details regarding each PD-L1 IHC assay are summarized in Table 1. It is interesting to note that each test uses a different antibody, along with different scoring criteria and cutoff values to define PD-L1 positivity. Indeed, even when evaluating the same assay, different scoring cutoff values have been used depending on the disease currently being study and/or the line of therapy. For example, pembrolizumab has different PD-L1 expression cutoff values for untreated (≥50%) versus previously treated (≥1%) patients with NSCLC.5, 7 Currently, another limitation of PD-L1 testing is that these assays have not been clinically validated in cytology specimens, which are a common source of diagnostic material for patients with NSCLC. Furthermore, with 3 different PD-1/PD-L1 inhibitors in the clinic for NSCLC and a choice of 3 different assays, pathologists are faced with the challenge of selecting a diagnostic test while not necessarily knowing which corresponding therapeutic agent will be preferred by the treating oncologist. Despite the challenges of PD-L1 testing, this biomarker is playing an increasingly important role in patient stratification and treatment decisions in individuals with NSCLC. Nonetheless, given some of the practical issues surrounding testing raised above, several harmonization efforts currently are underway to investigate the concordance between PD-L1 IHC assays.9, 10 In what to the best of my knowledge is the largest prospective study to date, 13 pathologists evaluated PD-L1 expression in 90 NSCLC specimens using 4 different antibodies (22C3, 28-8, SP142, and 1 laboratory-developed test [E1L3N]), and found that 3 of the 4 assays (22C3, 28-8, and E1L3N) were highly concordant, whereas the SP142 antibody detected significantly less PD-L1 expression in tumor and immune cells.10 It is important to note that in this study, PD-L1 expression scores were highly concordant when evaluating tumor cells, but concordance was poor when scoring immune cells. In the future, additional cross-test comparisons and clinical validation may help to facilitate cross-use of PD-L1 assays. In parallel with these efforts, research into PD-L1 staining in cytology specimens and the identification of additional predictive biomarkers for PD-1/PD-L1 inhibitors are needed. Collectively, this may inform the best strategies for the use of PD-1/PD-L1 inhibitors in the clinic. No specific funding was disclosed. Justin F. Gainor has received personal fees from Bristol-Myers Squibb, Merck, Genentech/Roche, Novartis, Boehringer Ingelheim, Loxo, Theravance Biopharma, Ariad Pharmaceuticals, Kyowa Hakko Kirin, and Clovis and travel expenses from Affymetrix for work performed outside of the current study. Justin F. Gainor, MD, is a medical oncologist specializing in the care of patients with thoracic malignancies. He is an attending physician in the Center for Thoracic Cancers at the Massachusetts General Hospital and an Instructor of Medicine at Harvard Medical School. In addition to his role as a clinical investigator focused on the development of novel immunotherapies, Dr. Gainor is a translational researcher with an interest in identifying determinants of response and resistance to immunotherapies.