Numerically, there were more deaths in the Opdivo (nivolumab) arm of the CheckMate-057 trial compared to the docetaxel arm within three months of initiation of therapy, although the trial did show that Opdivo significantly improved overall survival (OS) in patients with advanced non-small cell lung cancer (NSCLC), according to findings presented at the IASLC 17th World Conference on Lung Cancer in Vienna.
The absolute difference in death events between arms within the first three months was 15.
“Although nivolumab significantly improved overall survival versus docetaxel in patients with previously treated advanced nonsquamous NSCLC in the CheckMate-057 trial, the Kaplan−Meier overall survival curves for nivolumab and docetaxel crossed at approximately seven months, suggesting non-proportional hazards between arms,” said Solange Peters, M.D., Ph.D., head of the Thoracic Malignancies Program in the Department of Oncology at the University of Lausanne in Switzerland.
This led Peters and colleagues to investigate the relationship between patient baseline disease characteristics — including PD-L1 expression — and death within the first three months of treatment in a post-hoc analysis of data from 582 patients participating in CheckMate-057. All patients had nonsquamous NSCLC and experienced progressive disease (PD) during or after platinum-based doublet chemotherapy; 292 patients were randomly assigned to receive intravenous Opdivo at 3 mg/kg every two weeks, and 290 patients were assigned to intravenous docetaxel at 75 mg/m2 every three weeks. Opdivo patients were allowed to continue this treatment beyond PD.
CheckMate-057 met the primary endpoint of OS; median OS was 12.2 months with Opdivo versus 9.5 months in the docetaxel arm. The one-year OS rate was 51 percent versus 39 percent with Opdivo versus docetaxel, respectively, and the 2-year OS rate continued to favor Opdivo.
However, a numerically higher risk of death was observed with Opdivo compared with docetaxel, with 59 deaths occurring in Opdivo -treated patients versus 44 deaths in the docetaxel arm. At three months, 80 percent of Opdivo -treated patients versus 85 percent of docetaxel-treated patients were alive, although after this time-point among patients alive longer than 3 months, OS favored Opdivo over docetaxel.
The investigators observed that these early deaths were not treatment-related and were attributed to disease progression. Therefore, they evaluated the association between baseline characteristics and early death. Patients were stratified according to prior therapies received, location of disease sites, smoking status, ECOG performance status (PS), PD-L1 expression and the presence of EGFR mutation.
No single baseline factor reliably characterized the OS subgroups.
However, by multivariate analysis, factors evaluated in concert were found to associate with higher risk of death during the first three months with Opdivo versus docetaxel. The analysis revealed that Opdivo-treated patients with a poorer prognosis (ECOG PS 1, and/or progressive disease) at baseline combined with having lower or no PD-L1 expression may be at higher risk of death within the first three months of treatment. “However, the majority of nivolumab-treated patients with these factors were alive and did not die within the first three months, and experienced substantial subsequent benefit,” Peters commented.
She described baseline PD-L1 expression as “a continuum, ranging from 1 percent to 100 percent, with increasing expression associated with enhanced overall response and OS benefit from nivolumab; this continuum had no effect on response rates in docetaxel-treated patients.”
“Nivolumab is the standard of care as second-line treatment for advance squamous and nonsquamous NSCLC. While enhanced clinical activity with nivolumab correlates with increasing tumor PD-L1 expression, patients with low or no PD-L1 expression have a comparable probability of response,” she added.
Peters also noted that deep and durable Opdivo responses were observed with Opdivo irrespective of PD-L1 expression levels in the tumor. The favorable safety profile demonstrated with Opdivo was also irrespective of PD-L1 status.
A landmark three-month evaluation of OS by PD-L1 status in patients alive at three months showed that median OS in patients with PD-L1 expression less than 1 percent was 14.7 months with Opdivo versus 11.4 months with docetaxel, whereas median OS in patients with PD-L1 expression 1 percent or greater was 17.4 versus 11.3 months with Opdivo versus docetaxel.
“PD-L1 status alone is not considered an appropriate biomarker for nivolumab treatment selection in pretreated advanced NSCLC, but rather should be considered in the context of other patient/disease characteristics,” Peters advised. “Additional biomarker research is ongoing to help identify patients with advanced NSCLC who may derive the most clinical benefit from nivolumab.”