Clinical development of immune checkpoint blockade has dramatically changed the treatment paradigm and prognosis for patients with non–small cell lung cancer. Immune checkpoint blockade with PD-1 and PD-L1 antibodies generates clinically significant, durable responses in patients with advanced non–small cell lung cancer. These agents are approved for first- and second-line treatment, either as single agents or in combination with chemotherapy and angiogenesis inhibitors. Although the toxicity profile of these treatments is favorable, a unique set of immune-mediated adverse events, such as pneumonitis, has been observed. Broader use of these agents is improving survival for patients with advanced lung cancer.
Immune checkpoint blockade with anti-PD-1 or anti-PD-L1 antibodies can generate durable responses in a subset of patients with advanced non–small cell lung cancer.
Combinations with chemotherapy and angiogenesis inhibitors improve response rates and overall survival.
PD-L1 expression is a biomarker for patient selection.
Novel biomarkers, such as tumor mutation burden and gene expression profiling, are in development.
The clinical development of immune checkpoint blockade has dramatically changed the treatment paradigm and prognosis for patients with non–small cell lung cancer (NSCLC) ( Box 1 ). Therapeutic antibodies targeting the T-cell receptor programmed cell death-1 (PD-1) and its ligand (PD-L1) are now approved as monotherapy and in combination with chemotherapy for advanced NSCLC. Here we review the mechanism of action, biomarkers, clinical data, and unique side effect profile of these agents.
|PD-1||Programmed cell death-1|
|PD-L1||Programmed cell death ligand 1|
|CTLA-4||Cytotoxic T-lymphocyte antigen 4|
|TPS||Tumor proportion score|
|irAE||Immune-related adverse event|
|Pembrolizumab, nivolumab||Anti-PD-1 antibodies|
|Atezolizumab, durvalumab||Anti-PD-L1 antibodies|
Mechanism of action
Recent genomic and transcriptomic profiling of early stage lung cancers suggests that evasion of the adaptive immune system is an important feature of lung cancer evolution. , Although lung cancers have a myriad of strategies to evade antitumor immune responses, T-cell dysfunction induced by engagement of inhibitory receptors has garnered the most attention given the success of blocking these interactions in the clinic. Although we focus on two pathways, note that T cells express several activating and inhibitory receptors that are under active clinical investigation as therapeutic targets. Moreover, the tumor microenvironment contains other immune subsets, such as regulatory T cells, myeloid-derived suppressor cells, macrophages, and B cells, that play important roles in modulating antitumor immune responses. ,
Cytotoxic T-Lymphocyte Antigen 4
T-cell activation requires antigen presentation by the major histocompatibility complex to the T-cell receptor and a second signal, provided by interaction between the costimulatory receptor CD28 and B7 family ligands (CD80/86) on antigen presenting cells ( Fig. 1 ). Cytotoxic T-lymphocyte antigen 4 (CTLA-4) is an inhibitory receptor found on effector and regulatory T cells, which competes for binding with CD28 on T cells to constrain the initial stage of immune activation. , Ipilimumab and tremelimumab are therapeutic antibodies targeting CTLA-4. Ipilimumab is a fully human IgG1 monoclonal antibody. Blockade of CTLA-4 with ipilimumab led to improved survival in patients with metastatic melanoma and was the first immune checkpoint antibody approved for patients. Tremelimumab is also a fully human monoclonal antibody (IgG2 subtype), which is currently in clinical development.
Programmed Cell Death-1/Programmed Cell Death Ligand-1
PD-1 is an inhibitory receptor that is upregulated on T cells after antigen exposure. , Engagement of PD-1 by its ligands PD-L1 (B7-H1 or CD274) and PD-L2 (B7-DC or CD273) leads to decreased T-cell effector function. Human tumors harbor PD-1 expressing tumor-infiltrating lymphocytes. Importantly, work from multiple mouse models of chronic infection and cancer demonstrated that blocking the interaction between PD-1 and its ligands partially reverses T-cell dysfunction improving viral clearance and constraining tumor growth. , T and B cells, macrophages, endothelial cells, and malignant cells can all express PD-L1 (see Fig. 1 ). PD-L1 expression is upregulated by several proinflammatory cytokines, including interferon-γ, in a mechanism termed “adaptive immune resistance.” In contrast, PD-L2 has a more restricted expression on macrophages and tumor cells. PD-L1 can also be induced intrinsically in tumor cells, through activation of oncogenes or loss of tumor suppressor pathways. Multiple PD-1 (pembrolizumab, nivolumab) and PD-L1 (atezolizumab, durvalumab) blocking antibodies are approved for the treatment of NSCLC. The clinical activity of anti-PD-1 and anti-PD-L1 antibodies seems similar (as discussed later), suggesting that PD-L1 is the dominant ligand for PD-1 in the tumor microenvironment. However, these agents have not been directly compared or tested in combination to determine if they have additive or synergistic activity.
Clinical activity of immune checkpoint blockade
Brahmer and colleagues initially reported on the clinical activity of PD-1 blockade in patients with advanced solid tumors in 2010. This phase I study demonstrated that anti-PD-1 antibody therapy had a tolerable safety profile and provided the first hint of clinical activity in NSCLC, which had previously been considered an immunologically inactive malignancy. Subsequently, Topalian and colleagues and Brahmer and colleagues reported larger clinical trials of anti-PD-1 and PD-L1 antibodies, respectively. These studies demonstrated that PD-1/PD-L1 blockade had durable antitumor activity in squamous and nonsquamous NSCLC, and suggested PD-L1 expression on tumor cells could serve as a biomarker for activity. Across multiple tumor types, patients with PD-L1-negative tumors did not respond, whereas 36% (9/25) of patients with PD-L1-expressing tumors demonstrated an objective response.
Before the development of immune checkpoint blockade, the standard first-line therapy for NSCLC that lacked a driver mutation was cytotoxic chemotherapy with a platinum-based doublet. Docetaxel was commonly used as second-line therapy, but has limited response rates with significant toxicity. Multiple randomized controlled trials have now demonstrated superior safety and activity for anti-PD-1 (nivolumab, pembrolizumab) and anti-PD-L1 (atezolizumab) therapy compared with docetaxel ( Table 1 ). Nivolumab is a fully human IgG4 antibody against PD-1. CheckMate 017 was a randomized, international, phase III study that compared nivolumab with docetaxel in patients with squamous histology, and demonstrated that nivolumab improved overall survival (OS) (9.2 vs 6.0 months; hazard ratio [HR], 0.59) and progression-free survival (PFS) (3.5 vs 2.8 months; HR, 0.62). In this study, PD-L1 expression was determined with the 28-8 antihuman PD-L1 antibody and was not predictive of response. Similarly, in nonsquamous NSCLC (CheckMate 057) OS was longer for patients treated with nivolumab (12.2 vs 9.4 months; HR, 0.73) compared with docetaxel with an association noted between PD-L1 expression and benefit. Nivolumab is approved for the second-line treatment of squamous and nonsquamous NSCLC, irrespective of PD-L1 expression.
|Study||Population||Treatment Groups||Key Findings|
|CheckMate 017||Advanced squamous NSCLC, post-platinum doublet||Nivolumab vs docetaxel||mOS, 9.2 mo vs 6.0 mo |
|CheckMate 057||Advanced nonsquamous NSCLC, post-platinum doublet||Nivolumab vs docetaxel||mOS, 12.2 mo vs 9.4 mo |
|KEYNOTE-010||Advanced NSCLC, post-platinum doublet |
PD-L1 TPS ≥1%
|Pembrolizumab 2 mg/kg vs |
Pembrolizumab 10 mg/kg vs
|mOS, 10.4 mo vs 12.7 mo vs 8.5 mo |
HR, 0.71 (2 mg/kg)
HR, 0.61 (10 mg/kg)
|OAK||Advanced NSCLC, post 1–2 prior lines of therapy||Atezolizumab Docetaxel||mOS, 13.8 mo vs 9.6 mo |
Pembrolizumab is a selective, humanized IgG4 monoclonal antibody against PD-1. In the large, multicenter, phase I KEYNOTE-001 study, previously treated and untreated NSCLC patients were assigned to multiple pembrolizumab dosing regimens. To identify a biomarker cutoff to select patients more likely to benefit, tumor biopsies were stained with the anti-PD-L1 antibody clone 22C3. Patients with PD-L1 expression on greater than 50% of tumor cells were more likely to benefit from pembrolizumab. The KEYNOTE-010 study randomized patients with PD-L1-positive NSCLC (PD-L1 expression on ≥1% of tumor cells) to pembrolizumab 2 mg/kg or 10 mg/kg versus docetaxel (see Table 1 ). Median OS was prolonged in the 2 mg/kg (10.4 months; HR, 0.71) and 10 mg/kg (12.7 months; HR, 0.61) pembrolizumab groups relative to docetaxel (8.5 months), leading to approval in the second-line setting for patients with PD-L1-positive tumors.
Atezolizumab is a humanized IgG1 monoclonal anti-PD-L1 antibody that also improved OS (13.8 months) compared with docetaxel (9.6 months; HR, 0.73) in patients with squamous or nonsquamous NSCLC. In the phase I study of atezolizumab, PD-L1 expression was examined in tumor and immune cells using the Ventana SP142 IHC assay, with a stronger association noted between PD-L1 expression on tumor-infiltrating immune cells and treatment response. The larger, randomized, phase III OAK study confirmed that patients with higher PD-L1 derived greater clinical benefit, but improved OS was also seen in patients who did not express PD-L1.
Given the safety profile and durable responses seen in early phase studies, PD-1/PD-L1 blockade was quickly evaluated as first-line treatment ( Table 2 ). Previously untreated patients with PD-L1 expression greater than or equal to 50% had a response rate of 50% in the KEYNOTE-001 study. The KEYNOTE-024 study went on to directly compare pembrolizumab monotherapy with chemotherapy in untreated advanced NSCLC patients with PD-L1 expression on at least 50% of tumor cells. Pembrolizumab was associated with higher response rates (44.8% vs 27.8%), PFS (10.3 vs 6.7 months), and OS compared with platinum-based chemotherapy, leading to Food and Drug Administration approval for patients with squamous and nonsquamous NSCLC in the first-line setting. In recently updated analysis, the median OS for pembrolizumab was 30 months compared with 14.2 months for chemotherapy (HR, 0.63). , To determine the role of immune checkpoint blockade in a population less restricted for PD-L1, KEYNOTE-042 randomized patients with previously untreated, locally advanced, or metastatic NSCLC and a PD-L1 expression of 1% or greater to pembrolizumab or platinum-based chemotherapy. Patients treated with pembrolizumab had a median OS of 16.7 months, compared with 12.1 months for (HR, 0.81) for chemotherapy. In an exploratory analysis of the PD-L1 1% to 49% population, the OS was similar for pembrolizumab and chemotherapy. Significantly fewer grade 3 or higher treatment-related adverse events were noted for pembrolizumab (18%) compared with chemotherapy (41%). The phase 3 CheckMate 026 study compared first-line nivolumab with platinum-based chemotherapy in patients with PD-L1 expression on greater than 5% of tumor cells. The median PFS in the nivolumab group was 4.2 months compared with 5.9 months with chemotherapy, with no difference in OS (14.4 vs 13.2 months). In exploratory analyses, there was also no difference in survival in the subgroup of patients with greater than 50% PD-L1 expression. However, patients with a high tumor mutation burden determined by whole exome sequencing had higher response rates (47% nivolumab vs 28% chemotherapy) and PFS (9.7 vs 5.8 months). There are several potential reasons for the disparate outcomes of the CheckMate and KEYNOTE studies as noted by the authors, including differences in the immunohistochemistry (IHC) antibody used to select patients, imbalances in PD-L1 or tumor mutational burden high groups between arms, and a higher percentage of never smokers in the CheckMate study. Importantly, the KEYNOTE-042 study did not allow crossover and few patients randomized to the chemotherapy arm went on to receive pembrolizumab (20%) compared with the percentage of patients on the CheckMate 026 trial who went on to receive nivolumab (60%). ,
|Study||Population||Treatment Groups||Key Findings|
|KEYNOTE-024 ,||Previously untreated advanced NSCLC (no EGFR or ALK mutation) |
PD-L1 TPS ≥50%
|mOS, 30.0 mo vs 14.2 mo |
|KEYNOTE-042||Untreated locally advanced or metastatic NSCLC |
PD-L1 TPS ≥1%
|mOS, 16.7 mo vs 12.1 mo |
|KEYNOTE-189||Metastatic nonsquamous NSCLC (no EGFR or ALK mutation)||Carboplatin/cisplatin + pemetrexed + pembrolizumab |
Carboplatin/cisplatin + pemetrexed
|12-mo OS, 69.2% vs 49.4% |
|KEYNOTE-407||Metastatic squamous NSCLC||Carboplatin/paclitaxel or nab-paclitaxel + pembrolizumab |
Carboplatin/paclitaxel or nab-paclitaxel
|mOS, 15.9 mo vs 11.3 mo |
|Checkmate 026||Stage IV or recurrent NSCLC |
|mOS, 14.4 mo vs 13.2 mo |
|Checkmate 227 ,||Stage IV or recurrent NSCLC |
|Nivolumab + ipilimumab |
|mOS, 17.1 mo vs 14.9 mo |
|IMpower 150||Metastatic nonsquamous NSCLC||Atezolizumab + carboplatin/paclitaxel + bevacizumab |
carboplatin/paclitaxel + bevacizumab
|mOS, 19.2 mo vs 14.7 mo |