Geoffrey R. Oxnard, MD: What if you send a liquid biopsy and it’s negative—you don’t find anything? For me, I worry. There was no DNA there. That patient still doesn’t have a genotype. They’re still not a patient where it has been solved. And so, at that point, the FDA approval of liquid biopsy says a negative test should warrant the tumor testing as the fallback plan. So, if you get a negative test, think to yourself, “A liquid biopsy did not answer the question today. I need to get a piece of tissue, I need to do the biopsy, and I need tumor genotyping to get an answer for that patient.” Maybe at some point, a nonshedding cancer, when it’s more aggressive, will become a shedding cancer. I don’t think we have a good sense of that. And so, the rule of thumb should be that a negative liquid biopsy falls back to a tumor test as the reference standard.
Sandip Patel, MD: Plasma-based approaches to assaying the tumor, as opposed to tissue-based approaches, offer certain unique advantages. With tissue-based next-generation sequencing approaches, patients have to have an accessible biopsy. Many patients have bone-only disease, or metastases to the brain that are not accessible. And for patients with driver mutations, such as EGFR or ALK, when they develop resistance mutations, it may be difficult to biopsy the site of progression. It can often be in the bone, the brain, or other areas that are not amenable to biopsy.
And so, liquid biopsy-based approaches are, at least in my practice, at least initially used for these resistance mutations, particularly for T790M; C797S for EGFR mutations; and with ALK, the various resistance mutations that confer sensitivity or resistance to all the next-generation ALK inhibitors. The goal is trying to profile the patient based on a blood-based test, which has a quicker turnaround time and is less invasive. For patients in whom we do not get a substantial result, which often can be the case in patients with smaller tumors or slow disease burden, a biopsy of the growing lesion or the most PET-avid lesion to send for tumor-based next-generation sequencing then becomes my preferred approach.
There are 2 approaches to liquid biopsy that are currently utilized in cancer. The first involves the sequencing of circulating tumor cells. This involves the capture, through various selection methods, of circulating tumor cells in the blood, isolation of these cells, and the sequencing of these cancer cells directly. The second is cell-free DNA, or the actual DNA, that’s shed into the plasma and sequenced therein. Each approach offers unique advantages and disadvantages in terms of sensitivity, specificity, and breadth of sequencing, as well as the ability to capture different cell populations in terms of germline versus tumor-cell DNA.
Liquid biopsy can detect DNA that’s shed from any cell type. So, these are often cancer cells that are dying and will shed DNA as well as germline DNA. And so, oftentimes patients with BRCA mutations, we can detect with cell-free DNA. The hint there is the percent of DNA that comes from that specific germline allele often approaches 50%. For patients with a BRCA1, BRCA2, or P53 mutation, the percentage in cell-free DNA of 40% to 50% may in fact have a genetic predisposition to cancer, either a BRCA syndrome or Li-Fraumeni syndrome. And so, cell-free DNA approaches are able to assess this by the percentage of DNA that’s present in the blood.
I think the discrepancies between tumor and cell-free DNA may be explained by a couple scenarios. One is tumor heterogeneity. Any time we’re doing a specific biopsy of an area of the tumor, that may not represent the totality of the mutations within that cancer. We’ve seen several studies show clonal evolution over the time and space of lesions, and with the biopsy, we’re only capturing a specific lesion. The other aspect is that the ability to detect different types of mutations may be preferential to certain types of assays.
Different assays have different strengths in terms of the ability to detect amplifications or fusion events compared to point mutations. And so, the ability to use RNA, for example, from a tumor biopsy for sequencing cells—really a tumor NGS—likely enhances the detection of fusion events in those patients. However, improved technology in primer set cell-free DNA, in my experience, has been very robust at detecting ALK and ROS1 aberrations, which are fusion events.
Transcript Edited for Clarity