Liquid Biopsies, NGS Expected to Become Standard in Lung Cancer

Maximilian Diehn, MD, PhD, discussed the future of NGS platforms and liquid biopsies and his vision for them eventually becoming standard practice for patients with lung cancer.

Maximilian Diehn, MD, PhD

The use of next-generation sequencing (NGS) platforms and liquid biopsies in lung cancer could evolve to where these procedures are part of standard care, experts say.

“ctDNA assays are now in the clinic; they are here, they are not just in the realm of research anymore,” said Maximilian Diehn, MD, PhD. “For patients with lung adenocarcinoma [for whom you’re] looking at active EGFR mutations or the T790M mutation, these are assays routinely being used in many centers to address patients who either can’t have biopsies in the frontline setting or try to minimize the number of biopsies you have to do. That is a very important thing we should be offering to our patients in the community.”

OncLive: Your presentation covered NGS platforms and liquid biopsies. Can you provide an overview?

During the 2017 OncLive® State of the Science SummitTM on Advanced Non—Small Cell Lung Cancer, Diehn, an assistant professor in the Department of Radiation Oncology at Stanford Medicine, discussed the future of NGS platforms and liquid biopsies and his vision for them eventually becoming standard practice for patients with lung cancer.Diehn: I focused my talk on a particular application of NGS related to the detection of ctDNA, which refers to fragmented DNA that can be found in the circulation of patients with cancer, and are becoming increasingly relevant in the research and management of patients with lung cancer.

I spoke about how we are currently using ctDNA in the clinic. There is a regulatory approved assay, particularly for EGFR mutations related to patients with lung adenocarcinoma—where we can detect those now in the plasma or blood. We are using those in patients where we can’t perform a biopsy. For example, it’s too dangerous for the patient to have a biopsy, or they had a biopsy and there wasn’t enough tissue. So, we can consider using this test to determine whether they have activating mutations at the time of diagnosis and, if they are detectable, then the patient can get the EGFR-targeted drugs.

Also, more frequently, we are using these tests at the time of progression on EGFR-targeted therapy because there is an FDA-approved drug osimertinib (Tagrisso) that targets the T790M mutation. If we can show that a patient developed T790M in the blood, then that is good [reason] to get the patient the drug. Then, we can avoid the need for biopsy at the time of progression by doing this blood test first.

In the second half, I talked about some of the emerging work that a number of groups are doing with applying NGS to detect ctDNA. The benefits of that approach are that you can detect and measure mutations in many genes at one time than just EGFR. That opens up many new applications, including a very sensitive response assessment to see how patients are responding to treatments. It can potentially be used for determining if patients are cured or not after they have had surgery or local treatment. We can potentially use it some day in lieu of imaging as a way to follow our patients—just following up a blood test with patients rather than having to repeat scans all the time.

We have seen a lot of data with liquid biopsies in recent years. What next steps would you like to see?

For the same kinds of patients for whom we’re using the T790M test clinically already, if we apply the NGS approaches like the one my lab developed, then we can show that many patients have multiple resistance mechanisms. That matters to how they perform when you treat them with a next line of therapy. In the future, we should be personalizing therapy potentially based on what mutations are appearing in a patient’s blood and selecting drugs based on that. Such studies are starting to be done. We need to do clinical studies that prove utility of applying these assays, so they can become part of the standard of care. One of the exciting, but also challenging, aspects of this field is that there are so many potential applications. But, if we want any one of those to be the standard of care, we have to show that it makes a difference to patient outcomes. Just because you can do it doesn’t mean it’s actually useful.

We really need to do many well-designed trials that address particular questions. For example, one thing we are very interested in is the detection of minimal residual disease (MRD). For patients who have had surgery for early-stage lung cancer, we know that some fraction [of them] will eventually recur, even though their scans are totally clear after surgery. This means that they must have microscopic disease left in their bodies—known as MRD. We have shown that, with our assay, we can detect which patients have these microscopic cells left. If you can still detect the ctDNA after surgery, those patients are all guaranteed to recur. Whereas, if you don’t detect any disease after surgery, the vast majority of patients will not develop recurrence—at least in our initial cohort.

That suggests that we might use this to determine which patients might need additional systemic treatment after surgery. To prove that we need prospective trials to validate the result that we have, as well as to actually randomize patients to treatment or no treatment, to show that for the patients who have residual ctDNA, you will cure a larger fraction and improve the outcomes.

What challenges still exist with NGS?

We need to do similar clinical utility studies in all of the different areas that we want to apply ctDNA, and they are starting to be done. I am sure we’ll hear a lot more in 5 to 10 years. Ultimately, we can get more of these assays into the standard of care. The NGS approaches are very powerful, but there are a number of challenges. One important challenge is sensitivity. Most of the NGS assays are not as sensitive as the more focused PCR-based assays. That has changed, and we and other groups have developed very sensitive methods that rival the PCR methods, at least some of the approaches have overcome that technical hurdle.

Do they have the ability to detect other mechanisms of resistance?

Another major hurdle is the cost, and NGS platforms are still more expensive because of the sequencing cost. To really leverage the full power of that approach, we need cheaper sequencing costs, because that would allow us to make the assays more sensitive and see them potentially become part of the standard of care. The resistance mechanisms depend on the patient cohort, but particularly in the EGFR- activating mutation space, after you treat with first-line EGFR inhibitors, the most common resistance mutation is T790M. About 50% to 60% of patients will get that. We showed, in a study we published last year, that about half of the patients who have T790M as shown on a tumor biopsy also have at least 1 additional resistance mechanism that we can detect in the blood. It is not seen in the tumor, but if we look in the blood, we see it.

There are a number of different mechanisms; the most common we saw was MET amplification, However, there is HER2 amplification, PIK3CA mutations, and a number of other point mutations that we and others have seen. There are drugs that target MET and HER2, so one could envision that patients who have both T790M, as well as MET, might be best treated with a combination of a T790M inhibitor and a MET inhibitor, and some sequencing or concurrent strategy. Those kinds of studies are being done now to see whether those patients do better. We’ve seen that, in those patients with T790M and MET, by the time they get the T790M inhibitor, they have worse outcomes. That makes sense because they’re only inhibiting 1 of the 2 resistance mechanisms.