Read the Full Video Transcript

Jeffrey S. Ross: So, I chose to speak about liquid biopsies today, 50 years ago, I think it was the year 1972 that Professor Spiess was born. I was a pathology resident in training on my hematology rotation and looked at a peripheral blood smear in a patient with advanced metastatic breast cancer and saw some cells on the right slide that I certainly didn't think were blasts and I couldn't figure out what they were. And the attending hematologist went over them with me and said, "Those are circulating breast cancer cells in that unfortunate patient with advanced disease." And that sparked an interest in the ability of blood to guide therapy for cancer patients that has recently I think, reached fruition with the development of the actual liquid biopsy. My disclosures, I think we've had a very, very exciting evolution in cancer treatment of precision medicine and personalized medicine.

And the number of agents often linked to specific biomarkers or genomic alterations has grown almost logarithmically of late. But in blood, we have a mixture of a variety of different potential sources of circulating DNA. It can come from white blood cells, of course, in that patient. It could come even from bacteria in that patient's colon. It could come from apoptotic benign cells in various parts of the body. But in cancer patients, it can also come from leaking DNA into the bloodstream from the tumor itself, which we of course call ctDNA. And here you see this schematic drawing describing other potential sources of DNA in the blood.

So, we have tissue biopsies and we have liquid biopsies. And DNA can be extracted from both for the purpose of sequencing it usually now by next-generation methods in order to help guide therapy for patients, especially with clinically advanced disease. But now of course we know also for patients with early stage disease who are candidates for neoadjuvant regimens. The tissue biopsies have of course the advantage of a much larger amount of DNA that can be extracted and the ability to control for focusing on tumor versus benign. But they involve getting the tissue. Now, the tissue may be available archived because the patient had a resection or a transurethral biopsy of the bladder or needle biopsies of the prostate, and it already exists. On the other hand, the use of DNA to guide therapy often requires that therapy's already been given and the early original pre-treatment tumors may not be informative in many clinical settings.

That's when the thought of liquid biopsy comes to mind. Do we do a metastasis biopsy or do we do a phlebotomy? One is very easy, very safe and very fast, is very quick to take DNA from blood with paraffin embedded tissue. There is a much more elaborate series of steps to get the DNA out. When do we decide which to use, what we like? What's more convenient or do we decide based on the patient, their disease, the nature of their disease, the clinical stage of their disease, and maybe even other tests that guide us whether to use liquid or whether to use tissue? So, let's talk quickly about the history of the liquid biopsy. I've put circulating tumor cell assays just because that's where we began 20 years ago. We had the cell search looking for immunomagnetic bead, a captured circulating tumor cells and in some disease types, especially prostate cancer.

This was very informative of response to treatment, but unfortunately it was difficult to get DNA out of the captured cells in order to be able to design or plan therapy. And the CTC assay remains only rarely used. Today what has taken its place is circulating tumor DNA, ctDNA, and that's what my focus will be for the next oh 12 or 13 or so minutes designed for mutation, detection, and characterization. And one should always keep in mind that a mutation and alteration in the genomic sequence of the DNA from the tumor is not a uniform accessible change, that short variant mutations are easy to detect. Short insertions and deletions are much harder. Fusions depending upon the nature of the fusion, can readily be detected in liquid biopsies. But homozygous deletions and copy number changes like amplifications are heavily dependent upon how much DNA has been extracted since they are comparative assays to built-in white blood cell controls if you have it. And they challenge the system for its sensitivity and accuracy.

So, not all mutational changes are equal when it comes to the ease of detecting them in a liquid biopsy. I'm not going to talk about bone marrow, and I'm not going to talk about urine as a liquid biopsy for patients with bladder renal and even prostate cancers. But there is a literature for urine-based testing. My talk is blood based only. So, here we have the liquid biopsy listed here. There's two major commercial vendors of liquid biopsies in the United States, and that's Foundation Medicines Test, the F1 Liquid CDX and its competitive assay from the Guardant 360. These can serve as a disease burden, as we just heard from Signatera using circulating informed circulating tumor DNA as far superior than next-generation sequencing as a method of determining residual disease because of the extraordinary cost of repeated NGS tests on blood versus the ability to do it with an informed PCR or similar based test.

But the primary role for the liquid biopsy, at least today, is to try to guide therapy. It can be to decide primary therapy or it can be how do we overcome resistance to the primary therapy and that will time when the sample should best be collected. The biomarkers like blood TMB and MSI status as well as others are also detectable in blood and their clinical utility constantly is being refined as more studies employ them. So, here's more on the liquid biopsy, I'm going to talk a little bit about clonal hematopoiesis in a few minutes. That's an important confounding problem that we have sometimes, although we now are quite knowledgeable about how to avoid calling clonal hematopoiesis as tumor DNA derived as well. And finally, as you see at the bottom of the test, "It's not tissue first, it's not liquid first, it's the patient first," meaning the decision is made based on the clinical assessment of the patient.

And I'll give even some examples in genitourinary cancer of using even other ancillary tests to try to help guide this. So, there are a lot of problems when we do this. There's not very much circulating tumor DNA in blood, even in patients with very advanced disease. And some tumors, even when advanced don't elaborate very much circulating tumor DNA. So, we have to be very careful which patients we decide to do the test on because nothing can harm the patient more than the lost time of trying a liquid biopsy that will fail and the significant cost of doing that. And in some insurance settings, for example, in the United States, there's only one NGS test that the insurer will pay for. And if it's a failed blood-based NGS test, that patient is no longer covered perhaps for a future tissue biopsy that would be done now that the blood didn't inform. So, a great deal of thought has to be put into the decision by the treating physician, whether to do liquid and when to do liquid, so that it doesn't cost in time and money.

One of the advantages of liquid has been always that could overcome the issue of tumor heterogeneity, which we're going to hear about later in today's symposium. In general, it does help quite a bit. There are some problems, but here's an example of a melanoma patient in which the liquid biopsy pretty much got the driver mutations and the passenger mutations that were partially shown the driver of course in all of the sites, but different passenger mutations in different sites. But the liquid got them all, even the story going on in the brain, which is unusual since we think that most mutations in tumors in the brain are not going to be elaborated into the ctDNA that we will assess when we do a liquid biopsy, but it can overcome a tumor heterogeneity. Here are the two, just quick comparison of the two commercially available liquid biopsy tests in the United States. I don't believe, I don't know if the Guardant test is heavily marketed in Europe. The Foundation One liquid of course is.

So here's an example of a large tumor array of how much circulating tumor DNA you might find. I've put arrows for the three major GU cancers, bladder, renal, and prostate showing that all three of them with bladder and prostate a little bit more significant than RCC elaborating significant amounts of DNA on most liquid biopsies. So, for this particular tumor type GU cancers, liquid biopsies are quite readily used when selected patients are best assayed.

A lot of variabilities in ctDNA studies, a lot of factors can influence not only the clonal hematopoiesis issues, but issues involving a variety of methodology. How the DNA is extracted. The goal is to maximize the amount of ctDNA obtained from a blood sample using the best chemistries you can design. Because it's such a fleeting amount and below certain levels, one cannot make really significantly accurate assessments. So, selecting patients which have an unfortunately high ctDNA in blood is the best way to get very informative liquid biopsy results. And that's going to lead to this concept of the tumor fraction.

So, here are the kinds of alterations of course we look for in liquid biopsies and they're not a surprise of the same that we look for in many tissue biopsies. We see prostate and bladder here and for prostate cancer, liquid biopsies of late have been very focused on how they can guide the use of PARP inhibitors, meaning beyond BRCA1 and 2, the ATM mutation, which still remains an approved indication for PARP inhibitors in prostate cancer in the US. But we are now starting to find other alterations that are useful in planning therapy, some impacting the endocrine approach, the androgen ablation approach, and some even predicting responsiveness to docetaxel. Our most recent liquid biopsy study has been on the SPOP, the SPOP gene and prostate cancer and how it predicts that chemotherapy is likely to be less effective and maintenance of anti-androgen therapy, more effective in that subset of patients. For bladder cancer, of course, it's heavily focused on the target we've already learned about the FGFR.

So, as we go forward here and look at liquid biopsies, I should again mention what I spoke about a few minutes ago, which was germline testing. Now, a liquid biopsy for cancer therapy decision-making is not a germline test. It is using a germline sample. There's no question that the blood is a major germline sample for sequencing, mostly the white blood cell DNA for inherited mutations that could not only drive the patient's current cancer treatment decision, but also inform the family on their familial risk for that disease or other associated diseases. We don't have patient permission when a liquid biopsy is ordered to report germline results. And although, it's much easier to detect germline status on a liquid biopsy than a tissue biopsy because you have the white blood cell DNA in there that you can separately assess, you have to do the reporting very carefully, so that it doesn't violate legal standards about germline status.

But when we do, we certainly can cover all of the ESMO guidelines for what genes should be assessed for genetic susceptibility to cancer, many of which of course also drive treatment selection, including the PARP inhibitors with the BRCA germline mutations and many others.

So, here we see this tumor fraction study. This is emerging, this ability to calculate using an aneuploidy denominator, the actual portion of ctDNA that's coming from the tumor versus coming from non-tumor sources like WBCs. And it's a very significant prognostic factor. As the tumor fraction index increases the patient's risk of having disseminated disease, of having even therapy refractory disease goes up significantly. And as you can see from the chart to the right on the survival curves, it doesn't matter what tumor type it is, high tumor fraction leads to adverse prognosis. Here's an example of a patient with prostate cancer who had a liquid biopsy.

A couple of quick backgrounds, and again, this was a search for BRCA, this was a germline patient who went on a PARP inhibitor and did well. But I want to go to the next stage of this. I'll skip through the report, which is a little too detailed to understand that not all BRCA alterations in prostate cancer are the same. In the BRCA mutation spectrum and diseases like ovarian and breast cancer, where BRCA1 dominates over BRCA2, a homozygous deletion is extremely uncommon. But in prostate cancer, which is dominated by BRCA2, not BRCA1, homozygous deletion is about one-fifth around 20% of the alterations, which makes it 2%, 3% of clinically advanced prostate cancer has a BRCA2 homozygous deletion. But why does that matter? And I'll get to it in a second here. Here we just see the different sites with it. We found in this particular study that when PSA is below five nanograms per deciliter, the chances of an uninformative liquid biopsy are so high as to formally recommend that in prostate cancer a liquid biopsy should not be ordered for patients whose PSA is below five.

Now of course, a treating physician will then say, "Yes," but I have Gleason 10 or I have Gleason 9. Knowing that PSA is not as accurate a predictor of disease burden for those patients since each tumor cell is making so, so much less of it than a classic Gleason 7. And so, you may override that and go ahead with the liquid if you know it's a very high grade. But in a classic seven, if it's below five, we recommend you don't try a liquid biopsy. So, here is this patient who had a liquid biopsy that showed a BRCA2 homozygous deletion after years of therapy with hormonal therapies, taxotere and other drugs, PSA was I think 600 at its peak and disseminated disease, lungs, liver, I mean a high disease burden. Finally, got sequenced from blood and had a homozygous deletion and the patient went on a PARP inhibitor and had a durable dramatic sustained response.

Why is this difference? Well, when you have a BRCA2 base substitution mutation in activating this tumor suppressor, the most common, your resistance emerges due to the production of BRCA2 reversion mutations in which under the stress of the PARP inhibitor, the BRCA gene mutates and produces a new protein. It's not the perfect wild type protein, but it's good enough to restore the loss of overnight homologous recombination defect and the patient becomes resistant to the PARP inhibitor. But if you don't have the gene because you have a homozygous deletion, you can't make the reversion mutations. These men... If you identify them early, can go on a PARP inhibitor maybe instead of certainly cytotoxic chemotherapy and maybe even without androgen ablation therapy. They are so sensitive to PARP inhibitor and stay that way. And this was an example of this particular patient who went into complete remission with such a wide burden of disease and then sustained that remission for years.

In urothelial carcinoma, liquid biopsies are also very informative, especially in patients with late stage. And here is a plasmacytoid urothelial carcinoma with an E-cadherin in a CDH1 mutation, which we could pick up on liquid biopsy. We recently, we have a new study we just sent to the ASCO GU meeting with Professor Necchi as first author, in which we were able to do an outcome analysis using Foundation Medicine's Flatiron Health partnership for what are called real world clinical outcome studies. And what we showed was something that had first been published in upper GI cancer that when CDH1 is mutated, the patients are resistant to immunotherapy but retain sensitivity to chemotherapy. And we've shown pretty much the same thing with this outcome study in urothelial plasmacytoid CDH1 mutated bladder cancer that the outcome when IO drugs were used is significantly less efficacious than when classic platinum-based chemotherapy is the cornerstone of the treatment.

Great. A couple of quick words about clonal hematopoiesis. When we first discovered it was a bit of a surprise, but I don't think we should have not realized that as patient ages they start to produce these clonal alterations in genes associated with hematopoiesis and genes not associated with hematopoiesis. And it turns out that several of them, especially ATM, are very much a part of prostate cancer. So, we have to be very careful in a liquid biopsy not to call an ATM mutation that's actually coming from the patient's white blood cells coming from their prostate epithelial cells and have the patient call a PARP inhibitor who really is wild type for homologous recombination defect.

I have a lot more information about this, but I want to just focus on one last thing about CHIP is that it does occur in older patients and it has been linked in multiple studies to the presence of cardiovascular disease, particularly ischemic cardiovascular disease and that clinicians caring for cancer patients who order liquid biopsies who get back a report that includes significant CHIP genes identified should stop and think for a moment, is this patient under the care of a cardiologist?

Is this patient under care on cardiovascular drugs? And just to make sure that this isn't something as a surprise because we don't want to have a patient suffer a cardiovascular event that we could have prevented, and the CHIP would've told us that. Okay, I'm coming toward the end. So, is it liquid first? Is it tissue first? No, it's the patient first. We have to decide the clinical status, hopefully for bladder. We'll also get biomarkers that guide when to do it or not when it's in the clinically advanced stage, so we don't waste time and money. So, blood-based testing is emerging as a clinically useful approach for managing cancer patients and helping to drive their therapy and predict the responses to both targeted and immunotherapy drugs and whether to decide to do blood or tissue. First is based on the patient's clinical status, not what's easier to do.

I have to say as a pathologist, I know for medical oncologist ordering blood is good because you don't have to deal with the pathology department to get the block and submit it for sequencing, but you can't do that to the patient because doing a liquid that's not going to give you an answer is deleterious for care and for relapse and progression. Of course, we heard from the Signatera presentation and there are so many that ctDNA tested frequently using a non-expensive, not an NGS test, but a test that NGS informed how to design is a new breakthrough in how we're going to manage cancer patients in the future. And when appropriately used at the right time, liquid biopsies have a great potential to continue to improve precision and the ability to personalize the treatment for cancer patients.

Thank you very much.