Insights From: Geoffrey R. Oxnard, MD, Harvard Medical School; Sandip Patel, MD, UC San Diego Moores Cancer Center
Sandip Patel, MD: The percent of cell-free DNA that’s reflected in a liquid biopsy report is conceptually equivalent to the mutant allele frequency percentage that’s reported for tumor-based NGS methods. Effectively, these are the percentage of DNA detected that has that specific cancer related molecular aberration. Typically, these percentages range anywhere from 0.1%, or the minimal level of detection for cell-free DNA approaches, all the way to the percentages where we start to consider germline aberrations in certain genes such as BRCA and P53. You may get up to 50% of the DNA that could actually be from a germline source, as opposed to being from the tumor itself.
Variants of unknown significance are mutations in known cancer-causing genes. However, the specific mutations themselves are unknown to affect that gene’s function, such that it may or may not actually be clinically relevant for that given patient. We often see these mutations in non-exon 18 to 21 non–small cell lung cancer for EGFR, various other domains of BRAF, point mutations, and ALK. And, effectively, these are molecular aberrations that are in cancer causing genes, but may not actually cause cancer—they may not be driving the patient’s cancer. And so, it’s very important that any time you review a molecular report for a patient, you assess not only the specific gene that’s being affected, but also the specific mutation that’s occurring. Because some of those mutations may not actually be susceptible to targeted therapy. Similarly, many mutations are not actionable. They’re synonymous mutations, or they’re mutations in genes for which you don’t have therapies currently available.
Many times, our patients will undergo liquid biopsy and have clinically permanent mutations in cancer-causing genes, such as activating KRAS mutations or loss-of-function mutations in P53 and RB that we currently don’t have effective therapies for. For these patients, more traditional therapeutic approaches are likely to be effective, even though the assay is able to pick up a mutation in another cancer-causing gene.
One example of this are point mutations in ALK. For example, in non–small cell lung cancer, the driver mutation that’s susceptible to ALK inhibitors is actually a fusion event. However, oftentimes, these assays will pick up point mutations in ALK that actually are non-driver mutations. So, a careful review of the liquid biopsy will be important to determine which mutations are actually clinically actionable, versus those that are just passengers and not pertinent to intervene on to help the patient. Another example of this is EGFR mutations in codons outside of exon 19 and 21. Various mutations in exon 18, 17, and 20 are in the EGFR pathway and the EGFR gene. However, these are not susceptible to EGFR inhibitors that exist to date. And so, a careful review of the biopsy report will often have annotation that discusses the specific mutation and whether targeted therapy is appropriate for this patient or not.
Geoffrey R. Oxnard, MD: There have not been many studies that have taken a single patient and tested, for example, multiple NGS panels at the same time to see if they all align. There was one interesting report at ACR this year from AstraZeneca, actually, where they looked at 4 different NGS panels and found a striking amount of disagreement. What they said in that analysis was that for variants found above 1% AF, things seem to agree pretty nicely, until you actually get up to around 50%. Around 50%, you start seeing germline variants that some assays report and other assays hide, because they don’t want to talk about the germline variants. And so, I will say that there’s sometimes disagreement around 50%. And then, there’s the disagreement below 0.5%, as some assays are more confident in culling low-level things while other assays are a little more suspicious.
What I will say is that if you took a patient and sent 2 assays at the same time, you probably won’t get the same results. But, if it is an EGFR mutation or a BRAF mutation or an ALK rearrangement, hopefully in the right range—that sweet spot is around 0.5% to 20%—you will get similar findings. And so, I think there are areas where we can have confidence and areas where you need to be a little more suspicious. It is too bad that it’s this complicated. But the truth is we need to know these complexities in order to offer these safely for our patients. If you can’t figure out these complexities, go with tumor testing, because it’s better established. The fact is, tumor testing is not always an option. So, we need to adopt liquid biopsies, but also understand their limitations and complexities as we use them.