Corey J. Langer, MD, FACP: Gandhi and colleagues presented the final results of the prospective randomized phase III trial. This was 2:1 randomization for pembrolizumab/carboplatin or pembrolizumab/cisplatin; it did include the option for either cisplatin or carboplatin. Again, pembrolizumab was grafted. The control arm was pemetrexed and platinum alone for 4 cycles, followed by pemetrexed maintenance. Importantly, the control group was allowed to cross over to pembrolizumab at the time of disease progression.

One of the criticisms of these trials is that they didn’t get a chance to actually get the active agent. Here, it was built into the study. In fact, in the KEYNOTE-021 cohort G trial, it was well over 70%. In KEYNOTE-189, I think it was in the 50% or 60% range, but that’s just going to continue to climb as more and more people in the control arm progress.

Frankly, the findings were even better than the randomized phase II trial if we look at the hazard ratios: 0.49 for survival. In my career, I do not believe I’ve ever seen a randomized trial in advanced non—small cell lung cancer with a hazard ratio that good in a relatively unselected population, nonsquamous with the absence of ALK translocations or EGFR mutations. But it’s a significant percentage of the patients that we see, and this benefit was seen across the board.

Mark A. Socinski, MD: Yes. Tell us about PD-L1 status.

Corey J. Langer, MD, FACP: The greatest benefit was in the 50%-or-higher group. There the hazard ratio was about 0.42. It was a bit less substantial in the 1%-to-49% group (at 0.50 or so) and then a bit higher for the 0% group (about 0.58 or 0.59). But it was statistically significant right across the board. The progression-free survival benefit was seen across the board, as well, with one minor exception. For those with less than 1% expression, the confidence intervals overlapped. But the trends were clearly there, and the response rates were higher. In fact, for the 50%-or-higher group, the overall response rate was about 65%, 68%. This was about 20% higher than we had seen for single-agent pembrolizumab.

Mark A. Socinski, MD: In that group?

Corey J. Langer, MD, FACP: In the 50%-or-higher group. The big debate is, of course, for those of us who have now adopted this regimen for those with 50% or higher expression, should we go with single-agent pembrolizumab? It is certainly less expensive, less toxic, and easier to give. And the survival in the original KEYNOTE-024 trial was 30 months (median). Again, this is unprecedented. Or should we give the triplet regimen?

Transcript Edited for Clarity