We know the narrative: liquid biopsy promises a safe, painless, non-invasive, less-costly method of interrogating tumor DNA to identify targeted therapies for patients who are not amenable to biopsy. Patients will have a real-time view into the resistance patterns of their disease (companies are promoting the idea that patients should have a biopsy after every line of therapy, which was never feasible with tissue biopsy) and be equipped to select treatments that are more likely to work.
All this from a couple vials of blood? Well. Not so fast.
This post reflects only my experience, because I actually haven’t met anyone else who’s had a liquid biopsy. In February 2015, I flew to Philadelphia for a Guardant test; there was one breast onc who seemed to be using the the test frequently, and he happens to be an IBC expert, so he understands how critical these tests are for this patient population (IBC is by definition inflammatory, and you’re supposed to avoid cutting into it). I traveled for this test, as I was essentially out of therapy options and needed to turn up something new. They drew two vials of blood, and that was that.
Too easy, right? Right. Guardant found zero mutations.
This result is attributable to a lack of circulating tumor DNA (which is, of course, not a bad thing) and not an actual failure of Guardant’s sequencing process. I have an old Carys report that I was using as a control. Carys has its own shortcomings, which was especially true when I had my Carys analysis in 2012, but the test did find a PI3K mutation. I’m HER2+, and I suspected the HER2 positivity had not reversed, since I’d consistently been stable on HER2-targeted therapies and remain that way today. Guardant found neither mutation, nor any other mutation, and I subsequently got confirmation of my HER2 positivity: in desperation, I followed the Guardant disappointment by going back on Herceptin + Tykerb (both HER2-targeted therapies), and I responded.
Fast forward a year and a half to a competitor’s test. Brand new, broader panel, purportedly more sensitive. I was excited. I sent in my two vials of blood.
The same thing happened. I got my report yesterday. The test didn’t find anything.
The data on these things is decently compelling (this ASCO abstract found that 100% of MBC patients, albeit a very small sample, had mutations turn up via Guardant), so when you combine the reported performance with the many benefits of avoiding a biopsy, this should be a game-changer. But I can’t be the only person who’s having this experience – I’m early, but I’m not that early – so I’m suspicious that adoption is aligning with the hype. If the goal is not that patients get the test once, which is basically the traditional model for tissue biopsy, but that genomic monitoring becomes SOC, we can’t even be close to realizing that vision if X percentage of patients are getting no result at all (I would love to see that stat). I’m also not buying that there’s a significant market for early diagnosis, since it wouldn’t change treatment recommendations. From a tumor burden perspective, I have six and a half years of advanced disease kicking around in my bloodstream, and two tests have failed to characterize it, so I doubt patients with early-stage disease would get a reliable result.
So what’s the problem? Too soon? Patient selection? Can we tell who’s shedding sufficient volumes of tumor DNA? Even if the DNA is present, do we know the source? Genomic features may not be concordant lesion to lesion (which is a shortcoming of tissue biopsy as well). Risks and limitations aside, tissue biopsy isn’t going anywhere, and we may actually be seeing increased utility of obtaining tissue: patients are open to it for reasons like sequencing, which aligns with my own preferences. I’m not going to cut into my skin just to confirm that my HER2 status is unchanged, but I’m considering it to identify the presence of other markers. There are benefits to taking the risk that we just didn’t have even five years ago. The number of clinical trials enrolling based on genomic markers is increasing dramatically, and I can’t think of a study I’ve seen that will substitute a liquid biopsy result for a tissue result. A lot of studies require not just archived tissue, but a fresh biopsy – and anytime you’re taking tissue, you have the opportunity to get a comprehensive sequencing test. As for a role for ongoing monitoring, even if it is just two vials of blood and the burden on the patient is low, the test is expensive, so I don’t know if tracking the status of common markers is sufficient justification for the cost. There’s an elegant pitch and the seduction of one thing in oncology being easy, but if my experience is at all common, liquid biopsy may not have the most obvious place in everyday practice.