Novel Predictive Biomarkers for Non-Small Cell Lung Cancer - Episode 10

NSCLC: Use of TMB in CheckMate-227

Transcript:Gregory J. Riely, MD, PhD: CheckMate-227 was a clinical trial that enrolled patients with non—small cell lung cancer and has a variety of pieces of data that we’ve seen in several presentations. The initial cutoff point for CheckMate-227 was patients who had PD-L1 [programmed-cell death ligand 1] expression greater than 1% or PD-L1 expression less than 1%. The trial’s gone on to divide those patients further by their tumor mutation burden [TMB]. The cutoff point in tumor mutation burden is 10 mutations per megabase.

As the trial has gone forward, there are 3 arms in the trial in terms of treatment. There’s chemotherapy, chemotherapy plus nivolumab, and an arm of patients who got ipilimumab plus nivolumab. Ipilimumab is a CTLA-4 [cytotoxic T-lymphocyte antigen-4] antibody and nivolumab is a PD-1 antibody.

I think over the course of 2 big scientific meetings recently, both AACR [American Association for Cancer Research] and ASCO [American Society of Clinical Oncology], we’ve seen 2 different pieces of data presented. But it seems clear that patients with high tumor mutation burden seemed to benefit most from the combination of ipilimumab and nivolumab. I think regarding the subsets of patients who are PD-L1—negative or who have PD-L1 expression 1% or greater, among both groups it does appear that those patients seemed to benefit from ipilimumab and nivolumab if they had high TMB.

In CheckMate-227, there are certainly multiple arms of patients. There were multiple biomarkers used. But in the group of patients who had high TMB, there was a clear correlation with benefit with ipilimumab and nivolumab. That was significantly better than those patients who received chemotherapy alone. We’ve talked a lot about PD-L1 as a biomarker in patients with immunotherapy, but TMB, or tumor mutational burden, is also an important biomarker.

How these 2 biomarkers interact is, of course, critically important to understand. One might presume that if you’ve heard PD-L1 predicts for patients who are going to respond, then you’ve also heard that patients with high TMB are going to respond. You might predict that those 2 biomarkers would overlap. But it turns out that it doesn’t overlap very cleanly. There are a lot of patients who have no PD-L1 expression but have a high tumor mutation burden. For those patients with high tumor mutation burden, regardless of their PD-L1 status, it does appear that they’re going to benefit from immunotherapy whether it’s the combination of ipilimumab/nivolumab, single agent PD-1, or a combination of chemotherapy and PD-1 inhibitor.

Jared Weiss, MD: TMB is a somewhat controversial biomarker. I think we do have multiple data sets showing that it’s a legitimate predictive biomarker. That’s not controversial to me. What’s controversial is the optimal way to integrate it into clinical practice. The first question is, who do you test? If you’re using next-generation sequencing for nonsquamous carcinoma already, then on most of the panels, by default, you’re getting tumor mutational burden. The question is often brought up, Should I be testing squamous patients with next-generation sequencing to get tumor mutational burden? I don’t think that’s an unreasonable thing to do, but to me, the data and the clinical applicability of the data have not passed the threshold where I’m going to get next-generation sequencing on my squamous patients just to test tumor mutational burden.

The next question, beyond who you’re going to test, is how you’re going to use it. This is a very controversial area. I think what emerged from this study was that for patients with a high TMB, defined as at least 10 mutations per megabase, ipilimumab/nivolumab is a reasonable treatment to consider in the front line. In my practice, I would consider it for a very fit, very motivated patient with no immunotherapy contraindications. When you get into the PD-L1—negative TMB-high patient, it gets a little bit more controversial. This is not a common population. It’s about 16% of non–small cell lung cancer. In my opinion, reserving the right to be very wrong as more data emerge, ipilimumab/nivolumab is actually a fine idea for such patients.

The final finding with regards to TMB from that study was that in patients who were PD-L1—negative and TMB-low, all attempts to integrate immunotherapy failed to beat chemotherapy. Ipilimumab/nivolumab failed to beat chemotherapy alone, and nivolumab/chemotherapy failed to beat chemotherapy. If you look at the at-risk number at the beginning of those curves, you’re talking about small patient numbers. However, that result does fit together with prior data sets showing very similar findings. I think it is likely to be true. I think it would be reasonable, if you had a patient with PD-L1–negative, tumor mutation burden-low disease, to not expose that patient to the risks and expense of immunotherapy in the absence of likely benefit.

Gregory J. Riely, MD, PhD: Since PD-L1 testing and TMB lead to identification of different groups of patients, it seems to me that, going forward, we are really going to want to know both pieces of data about a patient. We’re going to want to know their PD-L1 status and their TMB status. You might identify a patient with high TMB but no PD-L1. That patient should get immunotherapy. In contrast, you might identify a patient with PD-L1 expression 50% or greater but low TMB. The current data suggest that patient should receive single-agent immunotherapy as an initial approach to treatment. I think we need to learn a lot more about how these biomarkers interact and how we can use them together to really predict the best patients to give immunotherapy to.

Transcript Edited for Clarity.