Blood-Based TMB as a Predictive Biomarker in NSCLC

Transcript:

David R. Gandara, MD: Well, let’s go back to the ESMO [European Society for Medical Oncology] lung session that occurred just a couple of hours ago and go off on a little bit different topic, something that I certainly have interest in and that was the presentation by Ed Kim [Edward S. Kim, MD]—a proof of principle trial called B-F1RST. And maybe I’ll just introduce it first. So, this was based on our recent publication that you can measure tumor mutational burden [TMB] in blood using a variation of the Foundation Medicine assay that’s used in tissue to measure TMB. And that by using both a test set and a validation set from the prior POPLAR and OAK trials, you can establish a cut point, which is equivalent to that in tissue, which is greater than equal to 16 mutations, and it equates to about 16 mutations per megabase of DNA. That is predictive of response rate and progression-free survival; maybe not quite so solid for overall survival. But this has led to this B-F1RST trial studying this prospectively. So, both Dr Vokes and I just heard the data from this study. What are your thoughts about this and what did they find?

Everett E. Vokes, MD: In many ways, David, you put a lot of your career into that interest in developing a liquid assay for this. I thought it was very intriguing. This cutoff of 16 seemed to identify a group of patients that benefit from single-agent atezolizumab [Tecentriq][for] first-line [treatment] more than those with a lower. And then, as you pointed out during the question, if you set that cut[off] point higher at 20 or 24, I believe, you get even more selection and benefit. But the trade-off is that very few patients are selected out. I thought the discussant made a good case that for that study to have been even more convincing, it would have been nice to have a correlate with tissue. So that was missing. Of course, in the tissue, and I’ll have to ask you if this is the same, the cutoff is 10 mutations per megabase, and here’s it’s 16. How are those different?

David R. Gandara, MD: That’s actually not exactly correct. For our audience, I’ll just say this is a convoluted subject. If I can take just a minute, and the reason I say this is that the test that we’re talking about is a Foundation Medicine assay from FoundationOne that is analytically validated. It’s an FDA [US Food and Drug Administration]-approved test because it has the genomics plus the TMB. However, it was all derived based on a TMB level of high, meaning greater than equal to 16. In one study, which was the CheckMate-227 study, which looked at nivolumab plus ipilimumab, instead of using that cut[off] point, the investigators looked at a retrospective study of about 100 patients and said for nivolumab plus ipilimumab, we think actually the best cut point is greater than equal to 10.

So, this is part of the confusion, and why I think both investigators in lung cancer, treating oncologists, even patients, want to know what are these differences. And I thought the B-F1RST study this morning, although it’s a small proof of principle trial, it showed that at least in this prospective evaluation, that greater than equal to 16 was more predictive of response and progression-free survival than 10. And, as you mentioned, actually if you went to 20, it was even better, but then you really have a very small proportion of the population. So, like was previously done with PD-L1 [programed cell death ligand 1], there is a harmonization process that the Friends of Cancer Research, the FDA, the NCI [National Cancer Institute] are all involved in to try to see if we can standardize some of these things for tumor mutational burden. I would point out though that the level that is best for clinical use for monotherapy from immunotherapy may be different from the level that’s best for an I-O [immuno-oncology] combination. Though I think we have to keep that in mind.

Everett E. Vokes, MD: But the basic test is measuring the same thing in the liquid biopsy versus the tumor.

David R. Gandara, MD: It’s not measuring exactly the same thing.

David R. Gandara, MD: And the reason is the computational algorithm for determining based on what is present in blood is a little bit different than tissue. However, and not that I’m giving an advertisement for Foundation Medicine, they are also harmonizing the blood and tissue test, so that eventually. Right now, there’s not a commercial test for blood TMB. It’s still considered research, but they are moving in that direction, and it will be equivalent with the tissue test. Now, what we heard today was the B-F1RST study and it was first because there is a subsequent trial which is a phase III trial which is ongoing. It’s accruing very well. It’s called B-FAST. This is a phase III prospective study in which patients are randomized to atezolizumab or to platinum chemotherapy, based on a high blood TMB level. And, again, this will have tissue correlates. It’s a phase III trial.

Everett E. Vokes, MD: Yes.

David R. Gandara, MD: What we heard about today you could criticize, but it’s really a proof of principle. It was showing that prospectively, they could do the TMB, and they couldn’t do it in some patients. In other words, in some patients, they did not shed enough tumor DNA into their blood to be able to determine TMB.

Everett E. Vokes, MD: But I thought it was very interesting that the great majority they actually could do it on.

David R. Gandara, MD: That’s true.

Everett E. Vokes, MD: There were some dropouts where it was not measurable, but the great majority was. And then the, and I really appreciate the explanation here because I think TMB will emerge into a test that most patients will have at some point in their evaluation. And the idea that the actual cutoff is somewhat weighing pros and cons against how many people, patients, should be included and where is the optimal cutoff point, and that it may not be the same for all drugs or combinations.

Transcript Edited for Clarity