Case 2: Biomarkers and Treatment Decisions in Stage IV NSCLC

EXPERT PERSPECTIVE VIRTUAL TUMOR BOARDPaul K. Paik, MD:My first question, Greg, is to you. Have there been any new discoveries inKRAS-mutant biology, over the past 5 years, that might come into play in this particular case?

Gregory Riedlinger, MD, PhD:This patient was tested on a relatively limited panel. There’s been more recent data suggesting thatKRAS/STK11co-mutated lung adenocarcinomas are not as responsive to immunotherapy asKRAS/TP53co-mutated lung adenocarcinomas. And, potentially, theKRAS/STK11/TP53triple-mutated cancers may actually retain theTP53benefit. The other thing with this case, on this limited panel, is the issue that we run into many times—this limited amount of material. We do some upfront testing for a small panel of genes, but then we want additional new testing that can do tumor mutational burden (TMB), or can pick up on theSTK11, and things like that. A lot of academic medical centers have instituted policies within their pathology laboratories. They have procedures in place to try to conserve tissue, whether that’s through splitting some of the core biopsies into separate cassettes or doing a lot of unstained up front.

Paul K. Paik, MD:I think that highlights an important point. We have a fair bit of biomarker data in this case. The patient’s PD-L1 [programmed-cell death ligand 1] was negative—0%. We don’t know if there’s aKRASmutation that’s generally, mutually exclusive of other drivers. Would we want to send out the tumor material for more extensive sequencing to find some of these co-alterations, or, importantly, to find TMB in this specific case?

Benjamin P. Levy, MD:As the case is now, I’m happy with the information we have. We have aKRASmutation, which is one of the most common mutations we see in smokers. We’ve got the PD-L1. I am not bought in to the whole TMB story yet, and perhaps that’s because we don’t have overall survival yet from the data that was published in theNew England Journal of Medicine. We have a PFS [progression-free survival] advantage comparing dual checkpoint blockade versus chemotherapy in the high TMB patients. But, TMB is not an easy biomarker for me. It does take a lot more tissue to run. It takes more time, and it’s not driving my treatment decision at this time. Anne, I don’t know if you have any other thoughts about TMB?

Anne S. Tsao, MD:I tend to agree with you, Ben. TMB is not quite ready for prime-time across the country. Part of that is because the only way to get an accurate TMB is through whole exome sequencing, which is not practical, from a clinical perspective. And so, we use a substitute—next-generation sequencing. On CheckMate 227, we used the FoundationOne panel. In this particular case, an Oncomine platform, which is a limited set of genes, was used. You don’t even get a real TMB from this. And so, I don’t think that can be used in this particular case. I agree with Ben—TMB is a complicated biomarker. It’s not a very quick yes/no. You have to know which assay is being used and what genes are being tested to really get an accurate assessment. I know that there have been some publications that have suggested that whole exome sequencing is pretty much replicated by a next-generation sequencing platform. For prime-time and for standard of care, we need to think about this a little more.

Benjamin P. Levy, MD:I think TMB would have had a little more traction. It may not have complete traction, but maybe a little more traction if we didn’t have the KEYNOTE-189 data showing a benefit across the board. It’s a very easy decision, at this point. Given the benefit across all PD-L1 expression levels, I think we may have seen a little more traction. If we didn’t have that data, which showed a hazard ratio that’s unheard of for overall survival in an unselected group of patients…I think we have to wait and see where TMB heads. I have been encouraged by some of these genotypes that may or may not predict efficacy to immunotherapy, likeKRASco-occurring withTP53, or the lack of benefit withKRASandSTK11. I think there’s more and more data emerging about specific genotypes and whether or not they predict efficacy to immunotherapy. Let’s wait and see what the overall survival advantage is, if there is one with CheckMate 227.

Paul K. Paik, MD:From a practical standpoint, we’re all from different institutions. Are your pathology labs doing, or do they have access to routine testing—at least targeted exon sequencing, or something comparable to Foundation Medicine? And what is the turnaround time for your testing at this point?

Anne S. Tsao, MD:We don’t have a full tissue panel, but we do have a Guardant360 panel for plasma. You can actually extrapolate TMB from that. Our turnaround time for the blood is about 2 weeks. For tumor tissue, it can be longer. As all of you know, tissue is the issue. You send it off, you get DNA extraction, and then you get a call saying, “We don’t have enough to do the XYZ assay.”

Benjamin P. Levy, MD:Right. We did have a 50-gene Ion Torrent panel up until about 3 months ago. We moved to a next-generation sequencing panel. This is a 300-gene panel that we’re trying to validate. We’re working on validating the TMB right now, as it relates to Foundation Medicine. Again, this is not whole-exome sequencing. The turnaround time is anywhere from 10 to 15 business days. At my former institution, it was much longer. That’s relevant. I think this patient is symptomatic and needs to get started on treatment. She’s ill. She’s tachypneic. If she were a never-smoker, or if we were thinking that this patient had a target, it would be nice to know that information a little quicker. The turnaround times are anywhere from 2 to 3 weeks. We’re not routinely doing TMB yet, and that may change as the data evolves.

Anne S. Tsao, MD:It’s important to remember that we’re in academic centers, right? The majority of patients are treated out in the community—80% of all patients. It’s a challenge for a community oncologist to go to a pathology group that they may not be affiliated with, get the tissue, figure out how to send it off to FoundationOne, and get that information back. They have to have the staff and resources to do that. So, I think TMB is going to be challenging to bring out, for the majority of the clinical practices in the United States.

Benjamin P. Levy, MD:Right. The other thing that I want to say, and this relates to the case, is that this patient had a transbronchial biopsy. Whether we have enough tissue for all of this is sometimes a challenge at our institution. Still, we want to procure as much tissue as possible. Often, there is not enough tissue to do the next-generation sequencing, the TMB that’s involved in that, and the PD-L1. PD-L1 is not that much tissue. It’s 2 or 2 slides in IHC [immunohistochemistry]. But, there are still challenges from time to time, and there is a moment, an educational moment, to let the person who is performing the bronchoscopy know that he or she should get as much tissue as possible.

Paul K. Paik, MD:Greg, what about Rutgers? Do you have a next-generation sequencing platform? What’s the turnaround time for that?

Gregory Riedlinger, MD, PhD:We have a couple of smaller panels—a 15-gene panel and a 50-gene panel—that are really only doing single nucleotide variant (SNV) detection and smaller indels. We had a collaboration with our genomic tumor board and Foundation Medicine going back to 2013, and our oncologists are pretty comfortable with that. A significant number of cases actually get sent for a larger panel—comprehensive sequencing where you’re not just detecting SNVs and indels, but also copy number changes and rearrangements. I think the turnaround time for that has been similar to what you have described—somewhere around the order of 2 to 3 weeks.

Transcript edited for clarity.