Alicia Morgans: Hi. I'm so excited to be here with Associate Professor, Alex Wyatt, where we're going to speak about ctDNA, and really the strategies that we can use in our clinical practices to order and then understand the results related to the genomic differences that we might see and really target in terms of our therapies. And you are a genomic scientist, so really so well poised to help us think through these questions. Can you talk to us a little bit about ctDNA in the first place, just as a starting point so we all understand what are the pros? Why would we use this? Why wouldn't we just use tissue all the time for our patients with prostate cancer?

Alex Wyatt: Yeah, so absolutely. So in the metastatic setting, obviously getting tissue is not always straightforward. It can be difficult to re-biopsy a patient, lesions can be inaccessible, and further to that, archival tissue is not always easy to track down and the quality of the DNA in the tissue may be poorer than is ideal for genomic tests. So obviously, conversely a blood sample is relatively easy to collect, is minimally invasive for the patient. Now, the key thing for the success of a liquid biopsy test is that you have material in that blood sample that comes from the tumor. So circulating tumor DNA is shed by tumor cells in different metastatic locations into the bloodstream, and so it's present within a few milliliters of blood. The key challenge is to make sure that you have enough ctDNA in your sample to actually inform on the tumor itself.

So we often think about the ctDNA fraction of a blood sample, how much tumor DNA is in there? So some patients with, let's say, localized disease or subclinical cancers, they'll have hardly any tumor DNA in their blood, whereas those with clinically progressing metastatic CRPC, sometimes close to 100% of the cell-free DNA in their bloodstream is tumor derived.

Alicia Morgans: So let's really just emphasize that point, because I think when we're thinking about ordering ctDNA for our patients, it's really important to think about the timing, and also the patient, as you said. So this is not something we would routinely recommend or is not recommended by guidelines, at least in a localized setting. It's more commonly used when we have therapeutics available, because this is, again, somatic genetic alterations, and so in many cases these are going to be metastatic patients. But it's important for us to recognize that we should probably order this from blood that's drawn before we start our next great therapy, because at that time, if the patient has a wonderful response to that treatment, the fraction of ctDNA in that sample is going to go down. Is that correct?

Alex Wyatt: That's absolutely correct. So ctDNA is really closely linked to the proliferative capacity of a tumor, and it's somewhat counterintuitive because we know ctDNA comes from apoptotic cells. So it's easy to fall into the trap and think, "Well, if I have an effective treatment, surely I'll be causing ctDNA to go up," but actually that's not what's been observed in any cancer. And so ctDNA levels in the blood track very closely with the proliferation of a tumor. So at times of clinical progression, we'll see very high levels of ctDNA, and really within just a few days of initiating effective treatment, you can have log reductions in the amount of ctDNA in the blood. So what this means is for a patient with progressing CRPC, that's your best time to get a blood sample. If you've already begun treatment, perhaps now you need to be thinking of a tissue test instead. Where this becomes complicated is in the castrate sensitive or the hormone sensitive setting, because oftentimes when you're making decisions about treatment, you've already begun the backbone of ADT, and so that will be impacting ctDNA levels. So you need to think about when testing upfront.

Alicia Morgans: Absolutely. So important. So thank you for talking that through. One of the other challenges that we have with ctDNA, and there are several that I appreciate you highlighting today, is that there can be... We just talked about false negatives, which are going to happen because we don't have sufficient ctDNA to actually capture these, but there's also the possibility for false positives. So can you talk to us a little bit about that?

Alex Wyatt: Mm-hmm. Absolutely. So ideally, we like to think that if we have a tube of blood, the DNA can come from two sources. One normal cell-free DNA from our normal cells, our germline essentially, and the other would be your cancer, the circulating tumor DNA. But the reality is, especially in an elderly patient, there's many other somatic expansions in the body that could be contributing to your cell-free DNA. The main one is actually expansions within the hematopoietic stem cell population, and we call this clonal hematopoiesis, sometimes clonal hematopoiesis of indeterminate potential. And those somatic expansions obviously occurring within the stem cell niche lead to expansions in your white blood cell population, and so those cells are contributing to cell-free DNA. So what this means is that somatic mutations can be present in the cell-free DNA, but they didn't come from the cancer itself.

Alicia Morgans: Absolutely. So how do we figure out when we're dealing with that CHIP, as we call it, that acronym, how do we know if CHIP is what's really showing up as an alteration, or if it's actually something in the tumor? And also, if you could just mention if there are specific genes that are more likely to have this interference from CHIP.

Alex Wyatt: Yeah, so typically when we see mutations in CHIP, they fall in genes that would be associated with leukemias, for example, so DNMT3A. Typically when we look at genes that are commonly mutated in CHIP, they'll be those that are related to leukemias, so DNMT3A, TET2, ASXL1, and others. So these are not typically the profile that you'd associate with some solid cancers, so that can help you understand if a mutation is potentially related to CHIP versus other cancers. The easiest way to know whether the variants in cell-free DNA are related to CHIP or your cancer is to, at the same time that you are profiling the cell-free DNA, also look at the white blood cells in your blood tube. And this is what many research groups would be doing, because at the same time, of course, you get the germ line, but any mutations that are represented also in the white blood cell fraction as well as the plasma, you can be pretty sure are related to expansions in the hematopoietic lineage and not necessarily coming from the cancer.

Alicia Morgans: So important. So thank you for that. And really, as we're ordering these tests, that might be important for us to think about ordering them in a setting where we can get both of these fractions really sequenced and understood so that we are automatically subtracting out that contribution from CHIP. Now, there is one gene that is associated with prostate cancer pathogenesis that we have seen some higher levels of CHIP, and that's ATM. Do you have a sense or can you explain why that gene might have higher rates?

Alex Wyatt: Mm-hmm. Yeah, you're absolutely right. There are some leukemia genes that overlap a little bit with the prostate cancer profile. ATM is one such gene. TP53 mutations do occur in CHIP. CHEK2 mutations occur. ATM and CHEK2 seem to have a prevalence of about 10% or so in CHIP, so we know they are particularly common. I'm not sure we know exactly why they're key drivers of clonal hematopoiesis, but of course, the key worry is that if you don't know for sure that this is a CHIP variant in origin because you've profiled the matched white blood cells, then you may think that this ATM mutation or this TP53 mutation is representative of your cancer.

Alicia Morgans: And that's so important because certainly with ATM and CHEK2, we have olaparib, which could be used in this patient population and would not be recommended if this is really just something from CHIP. So as we wrap up, what would your message be to clinicians who are trying to use ctDNA in their clinics, either because they don't have access to tissue or really just don't have the ability to maybe order that tissue from another hospital where it may be available, or maybe there's no tissue at all? So what would your recommendation be to them? What's your bottom line?

Alex Wyatt: Well, I think the first bottom line would be to try to pick a test that you think, or that appears to appropriately filter out CHIP related mutations. Those that profile white blood cells in parallel are obviously ideal. Others have bioinformatic approaches to try to filter out CHIP related mutations. But a key thing you can do when looking at a report is to ask yourself the question of whether the mutations that are there look like those that would come from prostate cancer. So we know about things like SPOP, TMPRSS2:ERG fusion events, but if all you are seeing is mutations in leukemia genes, then maybe ask yourself the question, "Is this really prostate cancer that's been profiled?"

Alicia Morgans: Well, that is a wonderful thing to recommend. It really makes sense, I think, from a clinical perspective. And we are lucky to have a genomic scientist like you to really point out these nuances. I so appreciate your time and your expertise.

Alex Wyatt: Thanks, Dr. Morgans.