Update on PD-L1 Testing for NSCLC

Transcript:Mark A. Socinski, MD: Ross, we heard from Dr. Weiss and Dr. Kris about PD-L1 testing being a standard of care at their institutions. They’re doing it reflexively. What are your views on that?

Ross Camidge, MD, PhD: I think it has become part of the standard panel that we test for, so it doesn’t exist in isolation. We have to look for all of the driver oncogenes. And in addition, we have to look for PD-L1 immunohistochemistry in those patients. Sometimes, you’re going to have people who are positive on both. You can have an EGFR mutation and a PD-L1 expression. I think we have to discuss what that really means.

Mark A. Socinski, MD: Yes, we will. The issue—part of the confusion, I think—at the level of the community oncologist is that they hear about a lot of different PD-L1 testing platforms and different antibodies. We do have some data. The Blueprint Project—we have an NCCN effort that has looked at this. What are your thoughts on the different antibodies?

Ross Camidge, MD, PhD: Initially, when they said, “We’re going to compare these antibodies,” I didn’t quite see the value of it. But it really has been a herculean effort to say that every drug company with a PD-1 or PD-L1 must develop their own companion diagnostic. And at the end of the day, when you actually compare them side by side, the amazing thing is that 3 out of the 4 diagnostic tests tell you exactly the same thing. That’s a great benefit for the community, because as long as you get something that is referenced to the standard, you can find whatever the best-validated, cheapest assay that you can use is. The one exception is the assay that was associated with atezolizumab, which is a little bit off left field in terms of how it correlates with the others, and I think that has to be viewed somewhat differently from the others.

Mark A. Socinski, MD: Right. And where I’ve struggled with those data is in cross-trial comparisons, because you’re really not talking apples with apples if you’re using different antibodies that we know have different performance characteristics.

Ross Camidge, MD, PhD: The thing I’d really like to jump on there is that you have to think of PD-L1 as this continuous variable. It goes from 0% to 100%, and by setting something that says, “We’re including people in our trial who are above this level,” the patients don’t suddenly all become homogenous above that level. If you say, “I’m only going to look at people who are 50%” in your trial, maybe everyone is 51%. But in Mark’s trial, maybe they’re all 90%. And so, it really is a problem in terms of comparing between studies.

Mark A. Socinski, MD: Is there a better biomarker?

Ross Camidge, MD, PhD: No.

Mark A. Socinski, MD: OK. And to the credit of PD-L1, the data have been pretty consistent across all the studies about how the more you express, the greater the benefit you seem to get. Not that you don’t get benefit if you don’t express, but there is some effect of PD-L1 intensity on the relative benefit of all of these agents.

Ross Camidge, MD, PhD: Yes. I think its strength comes in randomized studies, where whatever those floors are, you hope they’re going to balance out between the arms. But if you have a small phase I study, and you have a response rate of 40%, maybe you just stacked your patient population a little differently from someone else.

Mark A. Socinski, MD: I want to also just touch on for a moment—and maybe ask Dave Spigel about—toxicity management and the spectrum of immune-related toxicities. And what message should we be delivering about that whole issue?

David R. Spigel, MD: I think it’s something that has evolved. When there were no drugs and they were only available in trials, it was new to everybody. It was easy to see that it wasn’t docetaxel toxicity, it wasn’t usual chemotherapy toxicity, and it was tempting to say that these drugs had no toxicity. As we’ve now gotten deeper into this and we have 3 approved drugs, it’s clear that there are unique toxicities to these agents. So, what we try to communicate to our colleagues in our practices—to our nurses, as well—is that you have to be attuned to toxicity. And these are unusual. They’re immune-related toxicities. They’re not the usual things we think of with chemotherapy. Although patients can get cytopenias, fatigue, and other things we’re used to, we have to consider the unusual toxicities—so, things like looking for pneumonitis or inflammation in the lungs, or thinking about looking for inflammation in organs like the pancreas, the liver, and the thyroid.

One that we struggle with a lot, because it can sneak up on you, is colitis, where you get diarrhea that you think is minor. We’ve gotten so used to managing minor diarrhea with our therapies. But this is a colitis that can become quite serious and lead to hospitalization—rarely, more serious things like death. So, the important thing is to recognize these drugs have toxicities. You need to jump on them early and educate the patients and families.

Mark A. Socinski, MD: Jump on them when they’re grade 1 or 2.

David R. Spigel, MD: I think so. And I think that it may mean just holding therapy. It may mean instituting steroids earlier than you’d think. The other lesson I think all of us would agree with is that you just don’t stop steroids early. You continue them longer than you think and taper them down.

Mark A. Socinski, MD: To build on that, I’ve had a couple of oncologists who have talked to me about some reluctance to give steroids, because they thought they were going to blunt the effect of the treatment.

David R. Spigel, MD: Right, sure.

Mark A. Socinski, MD: Can you comment on that?

David R. Spigel, MD: Well, we’ve learned from melanoma with ipilimumab that that doesn’t appear to be the case. We don’t really have solid data from all the lung cancer trials yet, but so far, there has been frequent, common use of steroids in all the pivotal studies, and yet they were positive. So, we don’t really feel that we have a strong story that it’s completely safe in terms of not blunting the efficacy, but it seems to be something that can manage toxicity safely and still allow patients to benefit. But it is hard to imagine that there couldn’t be some negative interaction there at some level. Maybe it’s not important if it’s short-term use.

Mark A. Socinski, MD: But given your points, this is a very different spectrum than we’re used to. Early recognition is the key. Management of toxicity trumps the efficacy issue, in my opinion, because you can get someone really sick if you don’t stop it early. And since we don’t have convincing data that steroids do it—we have the melanoma data to suggest that they don’t—then I think to err on the side of being aggressive in managing the toxicities and keeping them lower grade versus higher grade is a…

David R. Spigel, MD: One thing we joke about sometimes is that diarrhea is not as bad as lung cancer, and diabetes is not as bad as lung cancer, but for folks who are doing well who develop these toxicities, it’s a big deal.

Mark A. Socinski, MD: Yes.

Mark G. Kris, MD: I’d like to just throw in a word for infliximab. I think we have way too much reluctance to use infliximab. I took care of an elderly lady, and we really harmed her with the high-dose steroids. She had colitis, and we learned our lesson with this person. The next time it happened, we gave her infliximab. Within days, she was fine, with none of the steroid-induced side effects. So, I urge people, rather than think of it as a “break glass in case of fire” kind of thing, I think it really should be the first thing. And for an elderly person, somebody with congestive heart failure who has diabetes, infliximab is the way to go.

Mark A. Socinski, MD: Yes.

Ross Camidge, MD, PhD: I can give one small example over the wall—give the counterargument. So, I had somebody who had really bad diarrhea. They were on ipilimumab/nivolumab. Steroids weren’t fixing it. I convinced the patient to try infliximab, and then they turned out to have C. diff (Clostridium difficile colitis). So, you still have to do the medical work.

Mark G. Kris, MD: Absolutely. Nothing says you can stop being a doctor. And I’ve had many, many cases of that. So, by all means, do not assume that it’s colitis related to the drug. The same is true on EGFR TKIs, too. You really have to rule them out. And by the way, its testing now is extraordinary, not just for C. diff, but also all the enteropathogens. You get an answer literally within hours now.

Mark A. Socinski, MD: I think that’s a good point. Assuming that they’re toxicity related, you almost have to exclude all the other diagnoses.

Mark G. Kris, MD: Absolutely.

Transcript Edited for Clarity