Current Biomarkers in Lung Cancer

Roy S. Herbst, MD, PhD: Over the last several years, we’ve evolved to the point, at the Yale Cancer Center Smilow Cancer Hospital, to test pretty much all of our new patients for mutational profiles. Certainly, you’re more likely to find actionable mutations in nonsquamous lung cancer, but we tend to do it for all non—small cell lung cancer, because you can find those mutations in patients with squamous cell cancer as well. The most important tests to have the day you see the patient with newly diagnosed disease would be for EGFR mutation, epidermal growth factor receptor; ALK mutation, anaplastic lymphoma kinase; and then ROS1 mutation. Those 3 I want to have, because, if a patient had one of those mutations, that would lead me to give them an oral targeted therapy.

In the last 6 months, we have been very focused on PD-L1, and we’ve actually put in place at Yale New Haven Hospital the ability to have PD-L1 testing done before the patient comes to see us. So, it’s done as a reflex. PD-L1 is important because, for those 20% who have high—meaning 50% or more—membranous staining for PD-L1, they can get pembrolizumab. So, that’s EGFR, ALK, and ROS1 mutations, and PD-L1. Certainly, we have the ability to look for another 50 or 60 genes in a small panel and a larger panel of 409 genes, but we usually save that for patients who are looking at next steps or looking for early phase trials.

A driver mutation is found in the United States in about 15% to 20% of patients, probably closer to 15%. So, most patients with non—small cell lung cancer will not have a driver mutation. Most patients are smoking-related lung cancer patients, and 20% to 30% will have a RAS mutation. That is a driver mutation; that’s one that we don’t have much that we can do about. So, we have about 80% or more of patients who have genotypes that are not actionable with any new targeted drugs. That’s a group that really is important to identify, and, of course, immunotherapy is now offering new hope to this group.

PD-L1 is not a perfect marker. We know there’s heterogeneity. We know there could be differences depending on how you score it. But, that said, it has panned out, especially in the studies with pembrolizumab—the KEYNOTE-001 trial; the KEYNOTE-010 trial, which I led; and the KEYNOTE-024 trial—that it is a positive predictor for activity for the PD-1 inhibitor. Now given that frontline therapy is used based on positive PD-L1 status, it is certainly measured at our hospital and I would make a case at every hospital—and every patient should ask for that. Because, if you’re one of those 20% or 25% where you can get PD-L1 testing, you have a positive result, and you’re one of those patients who can get pembrolizumab alone versus chemotherapy, I would say that you would want that—a patient would want that.