Stephen Liu, MD: KEYNOTE-042 was a randomized trial looking at pembrolizumab monotherapy in a select patient population, and it has influenced care. This was a large study of over 1200 patients with advanced non–small cell lung cancer, excluding patients with sensitizing EGFR mutations or ALK fusions. All patients did have PD-L1 expression defined in that study by at least 1% of cells positive using the [Agilent Technologies, Inc] Dako 22C3 clone. They’re randomized to pembrolizumab monotherapy or histology-specific chemotherapy.
We saw an improvement in overall survival with this approach leading to its FDA approval for patients with PD-L1 of 1% or greater. This was driven, as expected, by the PD-L1–high cohort that we saw in KEYNOTE-024. The population of interest was PD-L1 low. Are we really providing a better option for more patients? If we analyze patients with PD-L1 expression between 1% and 49%, where the use of pembrolizumab wasn’t clear prior to this trial, there was not a clear benefit over chemotherapy. It didn’t appear significantly worse, but it didn’t appear significantly better. The study wasn’t designed or powered to look at that subset, although clinically that is the subset of interest. It is an option for patients, particularly if patients aren’t eligible for other therapies. But in this day and age we have other options that I would say offer a clearly superior benefit to chemotherapy. Strategies including dual-checkpoint blockade or combinations of chemotherapy-immunotherapy. In a world where those options exist, I really think those have a leg up over pembrolizumab monotherapy for PD-L1 low, but that is an option for patients under certain circumstances.
IMpower110 was a randomized phase 3 trial looking at single-agent atezolizumab, a PD-L1 antibody, versus histology-specific chemotherapy. This was a relatively large study. Patients were randomized 1:1 to receive atezolizumab at a dosage of 1200 milligrams every 3 weeks or histology-specific chemotherapy. If we look at the PD-L1–high group looking at PD-L1 expression on tumor cells or on neighboring immune cells, which was about 35% to 40% of the entire population, there was a clear and overwhelming survival benefit. The median overall survival was 20.2 months compared with 13.1. That’s an impressive hazard ratio of 0.59 for the PD-L1–high group. One-year survival rates of 65% versus 50.6%. Response rates were higher, over 38%. Outcomes favoring the use of atezolizumab monotherapy in that PD-L1 high.
They also explored outcomes using the Dako 22C3 antibody, and the results were very consistent. In patients with PD-L1–high disease, atezolizumab is a reasonable option. It was FDA approved, and it is another option for patients. It’s not wise to compare it directly with other agents in this setting, such as pembrolizumab. Those outcomes are very comparable. Although there may not be a compelling reason to change—if there were, for example, nonclinical reasons, formulary restrictions that guided 1 over the other—I feel comfortable with either approach.
Transcript Edited for Clarity