Roy S. Herbst, MD, PhD: In someone who has failed frontline therapy but continues to have a reasonable performance status, second-line therapy has been shown to be of benefit. Depending on what the patient got in frontline, that’s how you would determine second-line treatment. Nowadays, the second-line treatment of choice is usually an immunotherapy. If someone has positive PD-L1—at the 1% level, meaning at least some staining for PD-L1—one can use pembrolizumab, the PD-1 inhibitor. If someone has no PD-L1 staining, or even if they have PD-L1 staining, the other drugs, of course, are nivolumab and atezolizumab. Neither of them has a label.
For nivolumab, I think the data—again, they’re not on the label—would suggest that it’s very similar to pembrolizumab in the patients who are PS of 0. There isn’t much activity. Interestingly, for atezolizumab, the most recently approved drug, it does appear that even in the negative group there was a significant benefit for the PD-L1 inhibitor versus docetaxel. Does that mean the drug is somewhat different in targeting PD-L1? Perhaps it’s able to prime and activate more tumors that are negative. That’s a hypothesis that I think needs further testing, but those would be the agents we would use in the patients who were negative.
For second-line therapy, you would want to look at the tissue type, the prior therapies, and the performance status. Aggressiveness of disease is also important. So, if someone has a disease that’s quite symptomatic, you might be a little worried about giving a PD-1 or a PD-L1 inhibitor alone, knowing that the chance of benefit is only 20%—still not bad, but it might take some time to work. That’s where the whole idea of combinations is coming in.
An attractive combination will be to give chemotherapy alone followed by the PD-1/PD-L1 therapy. We see that being used. That’s a reasonable approved mechanism of action, or one could use chemotherapy plus immunotherapy. That would need to be on protocol. It’s not approved, though there was a very well-known trial, KEYNOTE-021, which actually looked at carboplatin and pemetrexed with pembrolizumab. It seemed to show that there were some high response rates, both in the PD-L1-negative and the PD-L1-positive groups. So, we’re waiting for more data on those groups.