Expert Tackles Efficacy of PD-L1 Testing for NSCLC in 2017

Daniel B. Costa, MD, PhD

In 2017, a lung cancer diagnosis must be accompanied by biomarker testing, whether it be for a genomic alteration, such as EGFR, ALK, or ROS1, or the immune marker PD-L1, according to Daniel B. Costa, MD, PhD, MMSc.

Moreover, even with FDA-approved agents designed to target these abnormalities or expression levels, researchers are still hoping to uncover additional markers and match novel therapies to them.

“It needs to be enforced that, if you want to appropriately treat advanced NSCLC in 2017, testing needs to be done,” said Costa, an associate professor of medicine at Harvard Medical School, medical staff, Hematology/Oncology, medical director of the Cancer Clinical Trials Office, and Medical Center Thoracic Oncology Group Leader, of Beth Israel Deaconess Medical Center. “It is almost impossible to talk to a patient who doesn’t already understand that there is tumor heterogeneity and that every patient is treated differently. There is no cookie-cutter way to treat a patient anymore.”

In an interview during the 2017 OncLive^® State of the Science Summit on Advanced Non—Small Cell Lung Cancer, Costa spoke on biomarker testing in NSCLC and the necessary next steps for PD-L1 testing in these patients.

OncLive: What are the immediate next steps with PD-L1 testing for patients with NSCLC?

Costa: If you treat advanced NSCLC now, we cannot think of the disease as a one-size-fits-all approach after diagnosis anymore. It really requires substratification, and we are going beyond just the histological differentiation between the most common types of NSCLC: squamous cell carcinoma and adenocarcinoma. We are going beyond that to genomic markers and biologic markers.

PD-L1 is a very important biomarker that has clinical implications. This protein is expressed both in tumor cells and in the microenvironment of the tumor. It’s been consistently shown now—depending on the way that you try to monitor this—to predict that if you use the correct assay, the greater benefit of drugs that unleash the immune system.

More specifically, pembrolizumab’s (Keytruda) approval is based on the expression of PD-L1, so we know that the higher the expression of PD-L1, the higher the rate of antitumor response and tumor control. Now that this drug is approved based on this biomarker, it is increasingly important that all patients with NSCLC test for it.

What other promising biomarkers are we exploring?

If we are talking about immune-based therapies, PD-L1, in my particular view, is a very clinically useful biomarker because it can be done routinely in most of the samples that we have and it doesn’t have clinical implications. However, it’s not the perfect biomarker. The presence of PD-L1 expression, even at the highest level, doesn’t necessarily predict a 100% response rate. It is clear that there needs to be more refinements if we can really dwindle down on the patients who would or would not respond.

There are other biomarkers that are very important and will likely become part of an assessment of immune response. PD-L1 is just one of many, and others will come through for expression in both tumor cells and the tumor-infiltrating microenvironment.

The other one is the mutation burden of the tumor. It is obvious that a tumor is invading the immune system because something is happening as it’s becoming a tumor or becoming metastatic. This leads to differences within the tumor that would lead the immune system to attack, and that’s why the immune system wants to invade it.

The road of advanced genomics, to look at neoepitopes and tumor burden, will also be important. However, if you look at what is clinically indicated and can be translated to the clinic right now, we are somewhat stuck—at least in the immediate future with PD-L1 expression. However, it’s obvious we want to build on that either with other ancillary or other advanced tests that can supplement PD-L1 expression in the tumor.

Are there ongoing trials in this space?

Yes. The whole community is very excited to see biomarker-driven trials. We are talking now about PD-L1, but there are also genomic markers being used routinely in the clinic such as EGFR mutations, ROS1 gene rearrangements, and ALK. We have drugs approved for them. The community is excited with the idea that you can subdivide cases and you can use these biomarkers to develop drugs or combinations of drugs.

We are very excited with ongoing clinical trials that are moving precision oncology to the most advanced stages to the earliest stages of tumor burden, where, instead of just disease control for a period of time, we are extending survival and perhaps enhancing survival.

For immunology, the use of biomarkers and combination therapy with other approved therapies, such as chemotherapy and radiation therapy, have significant promise to expand the pool of patients who we now use the therapies in and move it forward within the therapeutic spectrum.

What other challenges in this area can we potentially tackle in the near future?

If you actually take it to the ground level, the first challenge is making sure that all patients have the test and that it’s mandated. It is easy to get everybody on board and do it, but we are very cognizant that in the bigger picture around the world and in the US community settings, it’s pretty clear that it’s difficult to keep up with the technologies that exist and also to be able to offer these tests to every patient.

The first step we need is to make sure that appropriate testing is done, even the minimal testing that we have to do it. That is number one. The second part is that because of the fact the biomarkers are so linked with drug development, it becomes complicated when we have so many different vendors and drugs being produced by different pharmaceutical companies.

We really do, as a community, need to come together, including academia, pharmaceutical, and regulatory agencies such as the National Institutes of Health and FDA. We need to have a joint picture on technology that seems promising and that we use it all together to develop drugs instead of just piecemealing it and trying to recognize it later. That is a particular challenge that will hopefully be addressed as we move forward.

What is the current understanding of whether PD-L1—negative patients will benefit from checkpoint inhibitors?

The expression of PD-L1 in tumor cells or the microenvironment is predictive of a higher response. However, it’s pretty clear that in all randomized trials of patients receiving checkpoint inhibitors after failure of traditional chemotherapy compared with second-line docetaxel, all of these patients—irrespective of the lowest or no PD-L1 expression—seem to do better [with anti—PD-1 treatment]. They have fewer toxicities, higher chance of response, and better durations of response than chemotherapy.

For the practicing clinician, a biomarker to select first-line therapy is very adequate in helping to decide. However, once you pass traditional chemotherapy and, even in patients whose tumors don’t have high PD-L1 expression, it’s clear that checkpoint inhibitors are the gold standard therapies. It helps an oncologist understand what to expect.

If you have a patient with the lowest PD-L1 expression, your enthusiasm for a very robust response is not there. However, nobody would offer a patient like that traditional second-line chemotherapy in 2017. Immunotherapy is still the way to go.

Refining the biomarker or identifying other markers that could help select either the best checkpoint inhibitor combination—or even developing less toxic and better active chemotherapy—would be the way to go for those patients.

Will PD-L1 testing continue to have a role in NSCLC or will something else become the standard?

The landmark KEYNOTE-024 trial that led to the FDA approval of pembrolizumab for first-line therapy in patients with PD-L1 expression equal to or higher than 50% was practice changing. It put PD-L1 in the upfront setting of both lung adenocarcinoma and squamous carcinomas. For a while, that will be a very important biomarker, especially as we reconcile the different types of immunohistochemistry. However, if we come up with an idea of which ones to use—a companion diagnostic to pembrolizumab or others—that will be very important.

If there is a combination of PD-L1 plus something else, be it tumor load of mutations or neoepitopes to enhance the response in selection of patients, you can really tease out the patients who will respond to monotherapy with anti—PD-1/PD-L1. Perhaps that will supplement it. The future can be just a year or 2 until practice changes. However, for this first-line selection, now that we have created this new slice of the pie where checkpoint inhibitors can be used instead of chemotherapy, it will be hard to supplement it.