Mark A. Socinski, MD: I’m going to transition to the next paradigm that’s on the horizon regarding I-O/I-O combinations. We saw some data, between AACR [American Association for Cancer Research] and ASCO [American Association of Clinical Oncology], looking at CheckMate-227 and a favorite anti—PD-1 [programmed death-ligand 1] and anti-CTLA [cytotoxic T-lymphocyte–associated protein 4) approach—in this case nivolumab [Opdivo]/ipilimumab [Yervoy]. What are your thoughts on CheckMate-227 and where it is? This piqued everyone’s interest based on a second biomarker beyond PD-L1, which is tumor mutation burden [TMB].

Leora Horn, MD, MSc, FRCPC: We saw from CheckMate-227 and -568, which was the single arm study with nivolumab/ipilimumab, that the PD-L1 and TMB are not necessarily co-expressed. You can be PD-L1 high and TMB low, but you can be PD-L1 low and TMB high. One doesn’t substitute for the other.

Mark A. Socinski, MD: They don’t correlate at all.

Leora Horn, MD, MSc, FRCPC: Correct. The CheckMate-227 is a huge trial: It’s 2000 patients, and the biggest problem is we’re seeing little bits of data at a time. We’re seeing data on 200 patients, and then we’re seeing data on 300 patients. We could make more sense of the study if we saw the entire trial. Many of the studies we’ve seen are chemo plus or minus I-O, and there’s no I-O-exclusive arm. CheckMate-227 does have that third arm, which will help us figure out who should and shouldn’t receive chemotherapy or I-O.

We’ve seen, at the IASLC [International Association for the Study of Lung Cancer] World Lung Conference, as well as at ASCO, that TMB did predict response with nivolumab/ipilimumab. Patients with high TMB had an improved progression-free survival with this combination therapy. Nivolumab compared to chemotherapy has shown unprecedented results, with an OS [overall survival] benefit in the patients with high TMB. Again, they’re using the cutoff with greater than 10. We have to keep in mind which cut point we’re looking at with the different trials. Once we see the OS data, we may determine that this is an option. If you have a patient who’s PD-L1 is less than 50%, you can send TMB and determine, perhaps, that they don’t need pembrolizumab single agent and instead receive nivolumab/ipilimumab. It’s going to become more confusing, I think. The biggest issue is, who do you test with? Where do you send your panel? If you’re sending it out, is it enough to wait a month to start those patients on therapy? That’s going to be the biggest challenge, once these data are available and the agents get approved.

Thomas E. Stinchcombe, MD: I have a lot of enthusiasm after AACR; those progression-free survival curves separate out nicely. We really need to see the full dataset because, like Leora said, we’ve seen bits and pieces of this.

Mark A. Socinski, MD: It’s fair to say that I don’t remember the exact number, but it was only a minority of patients who actually had TMB analyzed, which were available for TMB, so it’s a subset of that.

Thomas E. Stinchcombe, MD: I think 45% or 50% were TMB evaluable; of those, about half were high. And then when you try and counterbalance for prognostic factors, it’s hard to interpret.

Maximilian Hochmair, MD: The CheckMate-200 and -227 are TMB positive, PD-L1 negative, with long durations of response. Those patients receive these 2 I-O combinations. In our country, we get TMB from FoundationOne. We don’t have our own test. We have to implement that in our pathological report. As soon as we have FDA [US Food and Drug Administration] approval or EMA [European Medicines Agency] approval, we will start to test TMB locally also.

Mark A. Socinski, MD: There are lots of issues concerning the measurement of TMB. What’s the appropriate cutoff? Is it different from monotherapy? Is it different for I-O combinations or with chemotherapy? The other issue is plasma-based TMB. Is this an option? We’ve become accustomed to using molecular testing based on blood and plasma that we’re hopeful that plasma-based TMB approaches will be helpful.

Thomas E. Stinchcombe, MD: Yeah.

Mark A. Socinski, MD: In many ways, I agree with your initial comments about CheckMate-227. It kind of follows the paradigm that if you can’t convince them, confuse them. It’s been kind of, as you said, bits and pieces in different subsets—PD-L1-negative. It’s really hard to receive clear messages from 2000 patients from across these studies. We’re not getting a clear message from this. It’s hard to know in routine clinical practice, is this something that, say, a community oncologist in Tennessee should be sending out and waiting for a TMB measure at this point? Or do you think it’s too early?

Leora Horn, MD, MSc, FRCPC: I think it’s too early. We have to sort out the TMB, how you’re testing. Interestingly, we’ve started seeing more patients who are getting tested, have some sort of TMB platform on their testing, be it FoundationOne or Caris or wherever testing is done. At centers like ours, we do our own in-house panel. We don’t have TMB there. Is there even enough tissue to send out for TMB? And who’s going to pay for 2 different panels? It’s early to adopt that regimen outright.

Mark A. Socinski, MD: Do you agree?

Thomas E. Stinchcombe, MD: Yes. We had one of the trials, and logistically, you have to plan ahead. I’ve put people on trials with brain metastases where they’re getting the radiation treatment that I can send out for. Logistically, there’s going to be a lot of challenges, as you’re trying to implement it if this trial is positive. We’d like to see harmonization, too, across the platforms to make sure high TMB is really high TMB.

Mark A. Socinski, MD: We have a blueprint like we had for PD-L1. We have that going on at IASLC right now. We’ll get some clarity on what’s the best platform for tumor mutation burden analysis. I take, from the comments here, that I-O/I-O combinations are exciting but maybe not quite ready for prime time. The TMB is a little early for routine clinical practice. Essentially, as I look at the past 6 months, what we’ve essentially done is take traditional first-line chemotherapy that we’ve used in nondriver populations. Prior to 6 months ago, we would use immunotherapy second line. What we’ve done is take first- and second-line regiments and combined them with these complicated chemotherapy/I-O regimens. We talked a bit about the optimal strategy based on histology.

Transcript Edited for Clarity