PS02.09 Use of Systemic Corticosteroids During the First Month of Nivolumab Therapy in Patients with Advanced Non-Small Cell Lung Cancer: Topic: Medical Oncology

Checkpoint inhibitors enhance inherent anti-tumor immune responses and are approved for advanced non-small cell lung cancer (NSCLC), while corticosteroids are also frequently used in these patients to quell inflammation. Based on this mechanistic opposition, it is presumed that immunosuppression by corticosteroids will hinder the efficacy of PD-1 inhibitors, but the true nature and magnitude of their interaction remains unknown. We sought to describe the use of corticosteroids in NSCLC patients receiving the PD-1 inhibitor nivolumab. A single-center retrospective review included patients with advanced NSCLC treated with nivolumab starting between April 2015 and April 2016. The use of immunosuppressive doses of systemic corticosteroids (> 10 mg prednisone per day or equivalent) during nivolumab therapy was recorded and associated with objective outcomes of number of completed nivolumab cycles and median overall survival (OS) from start of nivolumab to death. Continuous variables were compared by the Wilcoxon rank-sum test, and median OS was estimated by the Kaplan-Meier method and compared using the log-rank test. In total, 210 patients with NSCLC were treated with at least one complete cycle (two treatments) of nivolumab. 31% (n=66) received immunosuppressive doses of systemic corticosteroids. The most common indications for concurrent corticosteroids included sequelae from active or treated brain metastases (27%), COPD or other respiratory disease (21%), and disease-related pain and constitutional symptoms such as fatigue and lack of appetite (18%). For patients on immunosuppressive doses of steroids at initiation or within the first 30 days of nivolumab therapy (n=25), the median number of nivolumab cycles was two, compared to five cycles in patients not exposed to corticosteroids (p=0.002). Similarly, the median OS for patients on steroids during their first cycle was 4.3 months (95% CI 2.6 - 9.7), compared to 11 months for patients not on steroids (95% CI 9.6 - 13.8). Eleven patients were treated for potential immune-related adverse events (IRAE) not requiring discontinuation of therapy, while 20 patients (9.5%) discontinued therapy following steroid initiation for IRAE. Nearly one-third of NSCLC patients treated with nivolumab were prescribed concurrent systemic corticosteroids during the course of nivolumab therapy. Patients requiring steroids during initiation of nivolumab therapy had shorter nivolumab treatment, suggesting a lack of clinical benefit, and worse overall survival. This finding requires validation in larger cohorts of patients receiving immune checkpoint inhibitors.