Levy Lends Insight on Immunotherapy Biomarkers in NSCLC

Benjamin P. Levy, MD

Established and emerging biomarkers are demonstrating predictive value with immunotherapy for patients with non—small cell lung cancer (NSCLC), but equally as compelling are those that can predict lack of response, said Benjamin P. Levy, MD.

Biomarkers of response to immunotherapy include PD-L1, tumor mutation burden (TMB), and potentially KRAS alongside TP53, whereas biomarkers that serve as deterrents include EGFR mutations and, most recently, LKB1 or SKT11 mutations.

“These are readily available mutations to interrogate on tumors that may soon help guide decision-making, if not now,” said Levy, adding that physicians must continue to search for better predictive biomarkers both in tissue-based and plasma-based assays.

OncLive: What biomarkers should physicians be aware of?

In an interview during the 2018 OncLive^® State of the Science Summit™ on Advanced Non—Small Cell Lung Cancer, Levy, assistant professor of Oncology, clinical director, Medical Oncology, Johns Hopkins Sidney Kimmel Cancer Center, Sibley Memorial Hospital, discussed established and emerging biomarkers that could have predictive value with immunotherapy.Levy: We have come a long way. We know that PD-L1 remains a biomarker. The cutoff point of 50% is still used to drive decision-making. There are other biomarkers to consider, such as TMB. This is a story that is beginning to be told. I don't think the story is completely done yet. TMB may be another biomarker to look at that may predict efficacy to either single checkpoint inhibitors or dual checkpoint blockade, as we have seen from recent data published in the New England Journal of Medicine.

There are other emerging biomarkers that are consistently showing benefit or predictive value with immunotherapy. One of those is KRAS with TP53. This is a story that's starting to be told. Data were published in July 2018 looking at KRAS mutations and TP53 occurring alongside them. Data showed that if you have those 2 alterations together, there may be a higher likelihood of response to immunotherapy.

Are PD-L1 and TMB routinely tested for at Johns Hopkins Medicine?

Those are important predictive markers for benefit, but just as important are markers that may predict lack of benefit to immunotherapy, [so we know] who not to use immunotherapy in. The ones that have emerged with data have been EGFR mutations, at least for single-agent checkpoint inhibitors. There have been consistent data that have suggested that EGFR-mutant NSCLCs do not respond to immunotherapy. Then, more recently, LKB1 or SKT11 mutations also may predict lack of benefit to immunotherapy.We're still working out TMB in-house. We are still waiting to get it validated, but PD-L1 is of course tested for. EGFR mutations are tested for, and KRAS is also looked at. We do a next-generation panel that gives you the co-alterations with KRAS, as well. LKB1, a marker that may demonstrate lack of benefit to immunotherapy, is also tested in our next-generation panel.

Where does the standardization for TMB need to come from?

How do you approach a patient with a TPS less than 50%?

What are the remaining questions?

Those are the ones that can be routinely interrogated, but there are other biomarkers being looked at in clinical trials. One is longitudinal circulating tumor (ct)DNA assessment, which looks at changes in ctDNA over time as a predictor of efficacy to immunotherapy. [Another approach is] looking at the microbial elements in stool to better understand which microbes may help predict efficacy to immunotherapy versus those that don't.We have a lot of work to do with TMB. Certainly, FoundationOne has a [test for] TMB, and they use 10 mutations/megabase as the cutoff. There are different platforms that are being used. Do we use whole-exome sequencing? Do we do next-generation sequencing? Those are all elements that we still need to work out and validate. Interestingly, every study has different cutoffs. They have similar platforms, but show some nuances. We are still at the ground floor with TMB. It needs further refinement.A patient who has a TPS less than 50% is certainly a challenge, especially if they are close to 50%. I have had patients with a TPS of 40% and they are heavy smokers. We’ve done their FoundationOne analysis and they have a high TMB. I have offered that patient single-agent immunotherapy. For the most part, patients with a TPS score less than 50% who don't have a driver mutation are given chemotherapy with immunotherapy. Those with greater than 50% TPS should get single-agent pembrolizumab (Keytruda), and those with less than 50% without a driver mutation should get carboplatin, pemetrexed, and pembrolizumab.There are a lot of remaining questions. The biomarker search is so important. We are on the ground floor with TMB. We need better predictive biomarkers, both in tissue and in plasma. We also need to define the subsets that won't benefit from immunotherapy. These drugs are costly, and they have side effects that can be challenging at times. We have to define those patients who really won't benefit.

That's where a lot of effort needs to go. We have to look at either analysis in the blood via ctDNA or in tissue analysis through things like LKB1, EGFR, and other mutations that may be immunosuppressive. Through this, we hope to understand which mutations immunotherapy won't work for and then, moreover, how to augment the lack of benefit by adding different drugs to immunotherapy.