Everett Vokes, MD: At the current state of affairs, where we have second-line immune therapies that we routinely give, for pembrolizumab, there’s guidelines to look at PD-L1 expression. But for nivolumab and atezolizumab, those recommendations don’t exist. Naiyer, do you test? And if you have a result available, does it influence which drug you choose? How do you use the information?

Naiyer Rizvi, MD: I think it’s hard to really choose a drug for second-line therapy (between the 3). Clearly, for pembrolizumab, the label is for more than 1% PD-L1 expression. I think all 3 are totally reasonable options. A rebiopsy is not going to change your management, so the PD-L1 assay doesn’t really help you. Some of the data that was updated at The European Society for Medical Oncology (ESMO) 2017 Congress was based around the OAK study and the biomarker that was used. I think that there’s some question about how well SP142 performed and how to interpret the OAK data. For patients where tumor blocks were available, which I think was in a little over half, they rescored the patients with the 22C3 assay. And interestingly, for patients who have no PD-L1 expression, by either assay, I think by SP142, the hazard ratio was 0.55, and by 22C3, the hazard ratio was 0.61. This is interesting data because you know that the hazard ratios for the PD-L1 0% patients treated with pembrolizumab and nivolumab have been closer to the middle. So, I think that data were pretty interesting.

Everett Vokes, MD: Very good. But I think you would then advocate that if in doubt, atezolizumab would be the drug you might choose.

Naiyer Rizvi, MD: I didn’t think that before, but maybe I’m getting there. Yes.

Everett Vokes, MD: Ram [Suresh]?

Suresh Ramalingam, MD: Increasingly, now, with the first-line use of checkpoint inhibitors for high expressers, you’re going to be left with the low-expressing population in the second-line setting. And, to me, as Naiyer said, all 3 drugs have comparable efficacy. The difference: nivolumab is given every 2 weeks. The other 2 drugs are given every 3 weeks. From that standpoint, for patient convenience, the every 3-week drug may be better. I don’t think there’s a big difference, clinically speaking, between these drugs at this point.

Everett Vokes, MD: Yes, they’re largely interchangeable. Roy, is there anything you would want to add?

Roy Herbst, MD, PhD: No. I agree, although I was intrigued by the atezolizumab data, as well. When they compared the 2 assays, they were seeing activity in the negatives. So, I think the every-3-week schedule, if you have a negative patient, might give that a little bit of an edge in the future. Certainly, pembrolizumab for the 1% or more population, based on KEYNOTE-010. I think these are all going to end up being backbones for combinations in the future.

Transcript Edited for Clarity