Over the recent decade, we witnessed important progress in the treatment of patients with advanced NSCLC in three domains. First, cytotoxic chemotherapy, where histology‐directed chemotherapy with cisplatin‐pemetrexed, followed by pemetrexed maintenance therapy in appropriate candidates, has resulted in a median overall survival (OS) of 16.9 months in adenocarcinoma.1 Second, the use of tyrosine kinase inhibitors (TKIs) in tumors driven by specific molecular pathways, such as EGFR, ALK and others, has largely improved progression‐ free survival (PFS) compared to the one with chemotherapy in randomized studies, and had led to OS times of several years in many of these patients.2 Third, immunotherapy with immune checkpoint inhibitors (ICI) directed against the immunosuppressive molecules programmed cell death 1 (PD‐1) and programmed cell death ligand 1 (PD‐L1) has been in clinical trials since 2009. At present, the anti‐PD‐1 antibodies nivolumab (a fully human IgG4 antibody) and pembrolizumab (an engineered humanized IgG4 antibody) have been approved for NSCLC by different regulatory agencies worldwide. EMA approved nivolumab for advanced NSCLC after prior chemotherapy, and pembrolizumab for advanced NSCLC in adults whose tumors express PD‐L1 and who have received at least one prior chemotherapy regimen. At the time of writing of this contribution, there were no public randomized study data on the use of these agents in 1st line therapy, we will therefore concentrate on the relapse therapy setting. Despite the real progress made by ICI therapy, we must realize that at present only about 20% of the patients respond to single‐agent ICI treatment, while 50% have early progression (Checkmate 0173; Checkmate 0574). Moreover, the cost of these drugs is considerable. Hence it is important to define optimal candidates in clinical practice. Elements in this decision are a) clinicopathological factors; b) possible predictive biomarkers; and c) the availability of other treatment choices. As for (a) clinicopathological factors, there is no evidence that age, gender or ethnicity determine activity of ICIs. Smoking history, on the other hand, is strongly associated with better response rate to ICI therapy. Although EGFR oncogene pathway activation has been linked to upregulation of PD‐L1 in tumor cells, response rates to ICIs in these patients are generally reported to be lower. In e.g. Checkmate 057, the overall response rate was 19%. It was 22% in smokers vs. 9% in non‐smokers, 18% in EGFR‐wild type vs. 11% in EGFR‐mutant tumors. Further understanding and refinement of the use of ICIs in tumor with an oncogene driver is needed. (b) A multitude of potential predictive biomarkers of response to PD‐1/ PD‐L1 pathway inhibitors have been reported. In particular, PD‐L1 expression in tumor and/or immune cells, the presence of TILs (tumor‐infiltrating lymphocytes, CD8+ T‐cells in particular), and the overall mutational load in the tumor cells have been linked to activity of ICIs.5 PD‐L1 expression on tumor cells, determined by immunohistochemistry (IHC) staining is by far the one most close to clinical practice for selecting optimal candidates for immunotherapy. This biomarker is quite distinct and less powerful than e.g. EGFR mutation as predictor of efficacy of EGFR TKIs. EGFR mutation is limited to the tumor, it is located in a distinct pathway, and is a yes/no phenomenon. PD‐L1 IHC, on the other hand, relates to the tumor and its micro‐environment, is only one of the many checkpoints in a complex interaction, and is a gradual phenomenon. Nonetheless, as can be noted from the figure, in most datasets of phase III studies‐except Checkmate 017‐PD‐L1 IHC predicts efficacy of ICI therapy. We added the large phase I study Keynote 001 to the figure as a very illustrative example: Over the categories of PD‐L1 expression, response rate increased from 8% in the lowest to 45% in the highest category.6 In the Keynote 010 phase III study, the hazard ratio of OS versus chemotherapy was 0.76 in t e low‐, but 0.54 in the highexpression group.7 In the Checkmate 057, the OS hazard ratio even was 1.00 in all patients with tumors having PDL1 <10%. Thus, except for the Checkmate 017 dataset, PD‐L1 IHC enriches the response rate and differential OS benefit vs. chemotherapy, and can be used to designate these expensive agents to the optimal candidates. (c) Docetaxel single‐agent chemotherapy was the comparator in the phase III studies on relapse therapy. In the meanwhile, progress has been made in conventional relapse therapy as well. In the LUME‐Lung 1 trial, the combination of docetaxel and the triple angiogenesis inhibitor nintedanib resulted in a significantly better OS than docetaxel alone‐with a difference in median OS of 2.3 months‐in patients with adenocarcinoma.8 In the REVEL study, patients treated with docetaxel plus ramucirumab, a VEGF receptor 2 inhibiting monoclonal antibody, had significantly better OS than those treated with docetaxel alone across all NSCLC histologies.9 In conclusion, the choice of ICI therapy for relapsing NSCLC needs to be considered in the available treatment options for these patients, and this can be based on clinicopathological factors, predictive biomarkers, and comparison of efficacy of various treatments in specific subgroups. While PD‐L1 is not a biomarker with the strength such as e.g. EGFR mutation, it helps to optimize the response rate and differential OS benefit of ICI therapy vs. chemotherapy, and to and can be used to designate these expensive agents to the optimal candidates. (Table Presented) .
Vansteenkiste, J., & Wauters, E. (2017). SC22.01 How Do I Define Optimal Candidates for Immunotherapy in My Practice? Journal of Thoracic Oncology, 12(1), S125–S127. https://doi.org/10.1016/j.jtho.2016.11.113