Insights From: Geoffrey R. Oxnard, MD, Harvard Medical School; Sandip Patel, MD, UC San Diego Moores Cancer Center
Sandip Patel, MD: The liquid biopsies vary amongst their sensitivity, specificity, and turnaround time. Typically, the cell-free DNA next-generation sequencing approaches sequence the widest number of genes, often over 70 genes, and will capture the entire exon. These are my preferred assays for the majority of my patients, especially at diagnosis, because they cover the widest swath of cancer DNA. We’re less likely to miss a pertinent genomic aberration. Typically, though, these panels take a little longer to run—2, sometimes 3 weeks—than some of the more focused panels. These are the digital droplet PCR, ARMS/Scorpion-based techniques, as well as PCR of single specific genes. These often can have turnaround times in the range of 48 to 72 hours. But you have to know the specific gene you’re looking for ahead of time. And so, especially at diagnosis, I tend to use the broad cell-free DNA-based approaches. But, for resistance monitoring, a more focused approach—or a similar broad approach, so that we can maintain consistency and also look at potential other concomitant aberrations—is typically what I utilize in my practice.
When looking at a specific gene, especially EGFR exon 19 or L858R, the majority of these assays are roughly comparable. However, if you don’t know for a given patient what their molecular aberration at baseline will be, a broader panel is likely to be more informative for the clinician and beneficial to the patient in terms of catching mutations that otherwise may be missed on smaller panels. However, the patient must be at a point in their therapy where they can wait the 2 to sometimes 3 weeks required for a broader panel, as opposed to the 2 or 3 days that may be required for a very focused panel.
Geoffrey R. Oxnard, MD: The NGS assays are the newest assays, and so their validation is not as well established. We’ve done work comparing NGS assays, for example, to ddPCR, and we find very clean linearity. Their ability to measure allele fractions is about the same. I think I really trust these assays, again, above that 0.5% threshold. As you get down to below 0.5%, the low-level positives in the 0.1% to 0.2% range, I’m worried that sometimes you can see low-level noise. Some of that noise is mutations you’re finding in the cells. There are actually low levels of DNA floating around that are not from the tumor that you’re finding in these patients. And so, it can confuse you. We know that the results we get out of a blood test do not always align with the results we get out of a tumor test. I think we should send them. We take some things with a grain of salt. If they’re a high enough level, I would say trust them and go with it. If they’re a low level, be a little cautious with what you interpret. I still think that there’s a need for prospective clinical trials that clarify all of this with these new NGS assays and can tell us which of the types of results we can trust most confidently, and which are the types of results that we need to be a little more concerned about as we try to pick targeted therapies.
How am I using these NGS results when I get them back? If I find something targetable, if I get that MET mutation that comes back, I’m reaching for a targeted therapy. I have patients who are sick. They can’t wait for tumor testing, they can’t wait for a biopsy, and we’re trying to find something that will work, right? For those sick patients, you want to have a result in-hand, and I have found that an NGS panel gives me that result so I can start a targeted therapy when I need it.
Sometimes, we can use these tests to get in clinical trials. Most clinical trials today are using tumor-based biomarkers for enrollment, not liquid biopsy-based biomarkers for enrollment. But, increasingly, clinical trials are recognizing how these tests are being incorporated into our routine practice and are, therefore, making their eligibility criteria more flexible and often teaming up with liquid biopsy companies to allow certain results to be adequate for enrollment. And so, we sometimes can find both clinical trials and available targeted therapies. I will say it’s not that uncommon that you find a result with liquid biopsy and you need to get a tumor biopsy to confirm it before you’ll be eligible for a trial, and so we need to recognize that if what your patient wants is trials, a biopsy is the better way of getting there.
Sometimes, I get a result back and I see high levels of P53 mutations or high levels of LKB1. I’m seeing a result that tells me there’s a lot of tumor in here, but there’s no mutation of interest. There’s just no HER2, there’s no MET, there’s no BRAF, and there’s no rearrangement. And for that patient, I tell them, “Look, we tried the genomics. It’s not working out, and what you need is chemotherapy, or what you need is immune therapy, but this is not going to be your approach.”
That, I think, is a compelling use of these tests that we aren’t talking about as much. You see that patient who is maybe a heavy smoker, maybe is EGFR- and ALK-negative, and he’s saying to you, “Wait, what about BRAF?” You’re looking for a way of avoiding an extra biopsy and extra testing for a sick lung cancer patient. A liquid biopsy could be, if the results are believable and show a lot of tumor content, a way of avoiding the biopsy, moving past that discussion into targeted therapies and immune therapies.