Background: Real-world data have suggested a detrimental effect of steroid use in patients with advanced non-small-cell lung cancer (NSCLC) receiving immunotherapy. However, previous studies included heterogeneous cohorts of patients receiving different lines of treatment with several immuno-oncology agents and various combinations of chemotherapy and immuno-oncology agents.
Patients and methods: A comprehensive clinicopathologic database of patients with NSCLC and programmed cell death ligand 1 >50% treated with frontline pembrolizumab monotherapy was constructed in 14 centers in Italy, Spain, Greece, and Switzerland. A multivariate analysis adjusting for the established prognostic factors was performed using a Cox regression model.
Results: For the 265 eligible patients, the median age at diagnosis was 67 years, 66% were male, 90% were current or former smokers, 18% had had an Eastern Cooperative Oncology Group performance status of 2 or 3. Of the NSCLC subtypes, 64% were adenocarcinoma and 25% were squamous cell. Of the patients, 18% had had brain metastases at diagnosis and 24% had received steroids before or during pembrolizumab treatment. The median time to progression was 4.4 months with and 13.7 months without steroid use (hazard ratio [HR], 2.55; 95% confidence interval [CI], 1.69-3.85; log-rank P < .001). The median survival was 22.5 months for the whole cohort, 7.7 months for the steroid group, and not reached for the non-steroid group (HR, 3.64; 95% CI, 2.34-5.68; log-rank P < .001). On multivariate analysis accounting for all established prognostic variables, steroid use was still independently associated with a high risk of progression (HR, 1.864; 95% CI, 1.179-2.949; P = .008) and death (HR, 2.292; 95% CI, 1.441-3.644; P < .001) CONCLUSIONS: In patients with advanced NSCLC and programmed cell death ligand 1 expression > 50% receiving frontline pembrolizumab monotherapy, any use of steroids before or during treatment was associated with an 86% increase in the risk of progression and a 2.3-fold increase in the risk of death, even accounting for palliative indication-related bias, including the presence of central nervous system metastasis. The use of steroids for palliative indications should be restricted to absolutely necessary for patients receiving immuno-oncology monotherapy.
Keywords: Brain metastases; Immunotherapy; Non–small-cell lung cancer; Pembrolizumab; Steroids.
Copyright © 2020 Elsevier Inc. All rights reserved.