Sanjay Popat:

Well, hello everyone and welcome to this episode of Video Journal of Oncology in conjunction with the British Oncology Group, BTOG, where we are diving into the intricacies of ctDNA NGS, specifically entitled How to Read a ctDNA NGS Report. I’m joined by a number of expert colleagues who are going to guide us through this complex maze so we can get the most out of our NGS reports if we are lucky enough to see them and we’re diving into where we are with commissioning as well. So I’m joined by a number of guests today. First of all, Rachel Butler MBE, who’s a consultant clinical scientist and also chief operating officer of the North Thames GLH. Also joined by Professor Alastair Greystoke from Newcastle University and consultant medical oncologist at the Northern Cancer Center. I’m also joined by Dr. Terri McVeigh, who’s a consultant clinical geneticist at the Royal Marsden Hospital and is doing a huge amount of work in genomics education and also my colleague Dr. Matthew Krebs, who’s consultant medical oncologist at the Christi and senior lecturer at the University of Manchester. So let’s just dive straight into the topic.

I’m going to ask, first of all, Rachel, let’s get to the basics. What is ctDNA? Is it the same thing as cfDNA, circulating free DNA? We use, we hear these terms. Talk me through it. What is it?

Rachel Butler:

Yeah, I understand it’s difficult to get your head around these. So you take a normal blood sample from any individual who doesn’t have cancer or any other disease, you’ll find that as well as white blood cells in the blood, you’ve also got circulating free DNA and that’s normal free DNA. It’s just shed from the white blood cells as they die, essentially. A patient who has cancer has a tumor, that tumor will also be shedding DNA into the circulation. So what you end up with is a mixture of circulating free DNA. Most of it’s normal and a small amount of it will come from the tumor, and we can obviously use that in terms of a non-invasive biopsy, but I would stress it’s a very small amount comes from the tumor.

Sanjay Popat:

And how do you actually test it? I mean, it’s just a simple, straightforward blood draw, isn’t it?

Rachel Butler:

Absolutely. It’s very easy. So we take a five mil blood draw. Sometimes we put it in a special tube which preserves it for a little bit longer. Then we extract that cell-free DNA, we get rid of the white blood cells effectively, and there’s a couple of different methods we can use. So you can use a targeted mutation detection if you know exactly what you’re looking for and you want to monitor a patient. But in the context of lung cancer we’re talking about today, it’s more likely that we’re going to use next generation sequencing, a gene panel with 50 or 100 genes and we’ll be able to screen all of those genes with that very small amount of circulating tumor DNA. The really important thing about these technologies is they have to be super sensitive, because there may be very, very little circulating tumor DNA in the blood sample that we take.

Sanjay Popat:

And when you take the blood sample, do you have to freeze it? Do you have to put it on liquid nitrogen? Can you keep it in the fridge? What do you do with it? How precious do you have to be with the sample?

Rachel Butler:

It’s not as easy as a straightforward blood sample for many other tests, but you can either take it in an EDTA tube or one of these preservative tubes. The important thing is that you don’t put in the fridge, which I know doesn’t feel quite second nature. And then it’s also important that you get that sample to the genomic testing lab as soon as possible. However, these preservative tubes give you a little bit longer. They give you sort of 24, 48 hours in terms of transportation time.

Sanjay Popat:

Great. And Ali, Rachel said it’s just a matter of taking a blood draw. Can we do this on our patients at any time or do we have to be careful about the impact of chemo or drug treatments that they’re on? And just tell me about when the best time is to sample a patient.

Alastair Greystoke:

Well, the important thing first is patient selection. So there’s no point doing this in patients that have early stage disease, stage I, stage II disease. There’s not going to be enough small amounts of tumor DNA for you to reliably detect it. I think most of us are using at present time in patients who have been newly diagnosed with either advanced stage III or stage IV cancer. When you can do the sort of next generation sequencing that Rachel was described to profile your patients. If you get them on chemotherapy immunotherapy, targeted therapy, the levels should drop if it’s working. And that’s been shown in a number of contexts. We are starting to look at it in the evaluation of resistant disease, particularly for targeted therapies. And we may look at some cases that situation later on.

Sanjay Popat:

But if you had a patient that had started a TKI, for example, and was two months into it doing well, that’s presumably not the best time to take a blood draw, is it?

Alastair Greystoke:

No, the results are going to be, you’d hope that they’ve had very little or no detectable circulating tumor DNA and it’s not going to tell you about the biology of the tumor when it does find the progression, be better off doing it at that point.

Sanjay Popat:

And Matt, in terms of patient selection, does tumor DNA from the brain seep out into the plasma for example, is that a reliable thing? So for example, patients who’ve got brain predominant disease or bulky brain disease, I personally would worry that they’re not really going to shed ctDNA. We don’t really understand the brain biology in ctDNA. Is that true or have I got it wrong?

Matthew Krebs:

No, I think that’s right, Sanjay. And if you sort of extrapolate from primary CNS tumors, those patients very rarely have much circulating tumor DNA, because obviously because of the blood brain barrier, it’s not getting into the bloodstream, it doesn’t sort of circulate in the same way. So you’re very unlikely to detect DNA in those patients circulating DNA. I find from my experience, patients with a lot of liver metastases, bone metastases, they generally tend to have a lot of circular tumor DNA from my experience. But CNS, I agree completely.

Sanjay Popat:

Ali, do you want to come in on that? Is that the sort of patients that we are looking for?

Alastair Greystoke:

Yeah, we’ve known for a long time, actually even from the early days of ctDNA, that those patients who only even with stage IV disease, if it’s just in the chest, they actually have relatively low levels. When you start to get outside the chest and certainly bone metastases, you can see much higher levels. There is a slight relationship with histology as well. So some adenocarcinomas may shed less than, for example, squamous cancers and certainly small cell, which tend to see very large amounts of circulating tumor DNA in those patients.

Sanjay Popat:

So listen, we’re all excited by ctDNA, because that’s why I’ve got these great experts around the table. I didn’t have to ask twice, but the key issue is it available? Where are we with NHS commissioning at the moment? Ali, can you just update us, all of us?

Alastair Greystoke:

Yeah, unfortunately this is a slight frustration I suspect to all of us. So on the National Test Directory, which is a list of tests that are funded through the NHS, the only circulating tumor DNA tests are EGFR. So you can look for both your primary and acquired resistance mechanisms in EGFR and ALK hotspot, and that probably doesn’t really drive patient care. It’s not looking for the fusion at present time. There is a proposal going forward and there is a pilot that we’re running at the moment trying to integrate circulating tumor DNA as a routine diagnostic in NHS England, but at present time it is not routinely available or funded outside of that pilot.

Sanjay Popat:

So this is being recorded in May 2023, and at the moment we don’t have routine commissioning at the moment, but we remain optimistic. So tell me what this pilot is all about. I’ve heard about this thing called GMSA pilot. Tell me more about this Ali.

Alastair Greystoke:

Yeah, so a number of centers have looked at taking circulating tumor DNA at the time when the patient sees oncologists, and you’ve published a lot of this work yourself and shown that you get the results back quicker than the result just doing your standard tumor analysis through your genomic laboratory hub and you get patients onto treatment quicker. What we’re trying to assess through the NHS England pilot is, well, what about doing it even earlier when the patient first presents to chest doctor with a CT scan showing stage III or stage IV malignancy way, you know they’re going to go through the diagnostic pathway and they’re end up seeing an oncologist. What about doing that, the blood sample at that point, getting your results back earlier, even potentially before your MDT discussion and trying to get the patients onto treatment as quickly as possible. The hope is that by doing this project, so it’s 700 patients across England, we can demonstrate which patients should have circulating tumor DNA when, and what’s the most cost-effective use of what is still at the moment, quite an expensive technology.

Sanjay Popat:

So pilot implementation to get more data to aim for commissioning is basically-

Alastair Greystoke:

Exactly, exactly.

Sanjay Popat:

Yeah. Okay. So let’s just talk about some of the aspects of what we find when we’re doing ctDNA analysis. Terri, I want to bring you in the conversation. So Rachel earlier said that there are these small fragments of DNA and some of them had derived from the tumor and some of them are derived from the blood. So when you look at a ctDNA report, how can you tell whether it’s from the patient themselves, it’s germline, something they’re born with or it’s from the tumor? Because a lot of docs are really scared about blood genomics, right? They’re really scared that they’re going to find something that’s germline or that they’re going to miss something that was germline. So do they need to worry? Is this a problem in lung cancer? I mean, obviously it’s a different issue in breast or endometrial, et cetera, but in lung, is this an issue and how can we tell?

Terri McVeigh:

Yeah. Lots of great questions there. So I guess the first question about whether you can figure out if something is of germline origin or from elsewhere is probably the way that I would phrase that question rather than saying, is it as straightforward as is it germline or tumor-derived? There’s a couple of things you need to consider. So I think the first kind of flag that sets the scene or perks people’s attention is the variant allele frequency. So without getting into the nitty-gritty of next generation sequencing, the basic principle is we have generally got two copies of most genes in our body. And if you’re a carrier of one copy of a variant, you should see that in about 50% of the reads from your next generation sequencing. Now because of technical issues and artifact, et cetera, if you were looking at tumor-derived DNA, we start to get a little bit interested if you see anything present at a variant allele frequency of 30% or higher when you’re talking about point mutations and for Indels 20% or higher.

Now to be honest, in the blood, it’s nearly always kind of closer to the 50% if something is of germline origin, because of the way the technology works. But that’s not the only thing that you need to consider. The other thing you need to consider is the gene in which the variant has been identified, because there are some genes in which germline variants are associated with syndromic disorders. So for example, something like a BRAF variant, those germline variants in BRAF are associated with RASopathies, and you wouldn’t typically miss a BRAF-associated RASopathy, because those individuals will have a syndromic phenotype that’s usually evident from childhood. So I don’t tend to get as excited about those kind of genetic variants. The ones that I do get a bit excited about are those where if somebody did carry a germline variant, you wouldn’t know just by looking at them.

So something like BRCA1, TP53, et cetera, they’re the ones you just can’t really tell based on the gene alone. The other thing to consider, basic principles, has the person got a family history of cancers that are associated with germline variants in the gene in question? Are they exceptionally young and have an associated cancer? The problem with lung cancer I guess is heritable risk factors for lung cancer, yes, a few have been identified. We know those case reports of constitutional EGF4 variants. We know constitutional TP53 is an association as well, but realistically, the vast majority of lung cancer risk is attributable to non-genetic factors. So the kind of common scenario that we find is incidental findings. So you’ve done ctDNA analysis in somebody with lung cancer and you find they’ve got a germline ATM variant or a CHEK2 variant or something like that. So in those kind of situations, whether it’s actually anything to do with their lung cancer is sometimes the question. So your earlier question about how frequently we see this, the studies are very difficult to pin down, because it depends what you characterize as a clinically actionable variant.

Is it anything pathogenic that’s of germline origin? Is it anything causal? Is it something that’s in a gene that’s associated with the significant level of risk? The rates of all kind of things that are pathogenic about 5%, but most of them will be incidental.

Sanjay Popat:

Okay. So we need to look at the sort of 30 to 50% variant allele frequency. We’re unlikely to find something that’s germline, but we might find something that’s incidental. And how do we, as clinicians that aren’t geneticists know which ones we don’t need to worry about and which ones to refer to? So should we be referring anything of we think that’s of significance to clinical genetics or should we be bringing to Genomics Tumor Advisory Board for discussion, for example?

Terri McVeigh:

Yeah, I think MDT discussion is really key here, because actually a lot of the time it’s not a straightforward answer. Now, ESMO have published some really useful guidance about this, about when should you do germline confirmatory testing if you have a tumor-derived variant of suspected origin. But even those guidelines, it’s not always that clear-cut, because they’re kind of set with this threshold of what we call what’s the likelihood threshold that something is of germline origin? You have to just set a line somewhere, you can’t test everything. So they’ve outlined very strict criteria. So if it’s on tumor and it’s likely to be causal and it’s greater than 30%, fair enough, go for it. But if it’s highly penetrant, so if it’s associated with the high risk of another type of cancer, then obviously for the greater good of that individual that may potentially be at risk of other cancers and for their family members, it makes sense to test.

So something like BRCA1, BRCA2, if I see something of likely germline origin, I’d certainly want to see them even if it has nothing to do with their phenotype. But you could make a case for a number of different genes that aren’t included within that guidance as well. It all comes down to what’s the family history, what’s the clinical benefit of knowing? If it’s something where you can action and potentially prevent cancers or at least facilitate early detection, then it’s worth doing. So I think multidisciplinary discussion, certainly reaching out to your local genetic service and seeing if they think it’s a worthwhile, nobody minds getting those kind of emails. They’re easy questions to answer. And I have to say, I’ve never got an email that has been unreasonable ever. It’s always a reasonable question.

Sanjay Popat:

That’s good to know. And presume that confirmatory testing happens at the genetics clinic, it’s not really an issue for the medical oncologist to sort out in their clinic.

Terri McVeigh:

Absolutely, and it depends on the gene in question as well, because sometimes I’ve actually, because it’s very difficult to concept to explain to a patient, to say, “We found this in your blood, we need to do another blood test to see if it’s really in your blood.” It’s a very difficult concept to get around. And depending on the gene in question, I have sometimes actually gone for skin, because there are certain scenarios where actually if you have a high burden of disease, you can actually pick up somatic variants at these very high variant allele frequencies even when you’re doing a germline test. So in those kind of situations, if I want to get something that is definitely not a cancer cell, I might actually recommend doing a skin biopsy to get another cell line that definitely represents the constitutional DNA. So yeah, certainly I would say that’s not really within your realm of responsibility. I’d be very happy to take over at that stage and proceed with the confirmatory testing.

Sanjay Popat:

Well, that’s good, that’s going to be a great relief. And also the fact that I don’t need to worry that much, because the probability of a germline finding in a blood test in lung cancer at least is relatively low, is what we’ve heard. And Matt, we’ve heard about this thing called a GTAB or Genomic Tumor Advisory Board or Genomic MDT. We have different terms. Tell me what is it? Do they happen in the UK? And what happens in them?

Matthew Krebs:

Yeah, sure. So they’re a bit like sort of modern MDTs, but for discussions specifically of genomic results, so each GLH or Genomic Laboratory Hub runs a GTAB either every week or every two weeks. They were originally set up for 100,000 Genomes Project for discussion of whole genome sequencing and largely do still discuss mainly whole genome sequencing results. But as we started to do more panel testing, both in tumor and in liquid and the pilot that Ali referred to before as a good example, we’ve started to use that in our GTAB to start to look at some of these ctDNA reports and you’ve got all the right people in the room to interpret them. So you’ve got the clinical scientists, you’ve got clinical geneticists, you’ve got clinicians, other people with vested interest, and with all the expertise in the room, you can really go into some detail to interpret the report. So sometimes it can be obvious. You might not need a GTAB to tell you’ve got an EGFR Exon 19 deletion, then fine if that’s all that’s there, it’s clearcut.

But quite often there are multiple other things present on the report and you may not be quite sure how to interpret them. And the GTAB is a great place to take those reports to and get the expertise from the clinical genetics and colleagues around how to interpret them properly. So a really valuable forum for discussing these results.

Sanjay Popat:

I think that’s right. I think especially in ctDNA, if you’ve not really used to seeing these results and interpreting them, always get every single one reviewed in your GTAB and then when you get a feel for it, you may be a bit more selective with what you send there. But yeah, Matthew, I’m going to challenge that. So do we need a GTAB? Because of course, if we’re doing ctDNA NGS and it’s run commercially, we’re get a beautiful report which tells us which drugs we need to be using. So what is the point?

Matthew Krebs:

We definitely need GTABs. So yeah, in terms of recommending treatments, I would take it with a slight pinch of salt would be my advice, or more than a slight pinch of salt. So it depends where some of these reports come from. So quite often they can be U.S.-centric or they could be European-based and not necessarily UK-specific advice. So often it’ll be clearcut, to go back to EGFR Exon 19 deletion, we all know which treatment we’re going to select for those patients. Outside of anything of your comfort zone of saying, yes, I recognize that mutation in the variant, I know the right drug to give fine. If you’re not so familiar with the gene or particularly the specific variant that’s being reported, don’t go by what’s on the report, because often those reports are just generated in an automated way where it might pick in terms of how it’s generated, it sees a certain gene and says, oh yes, you’re going to be eligible for drugs X and Y.

And quite often that is not the case or it’s just not reimbursed in our country, right? So the best thing to do is if it’s anything outside of your comfort zone and you don’t recognize it, even if it’s a gene that you’re familiar with like KRAS or EGFR, but you don’t recognize the variant, the right thing to do is take it to your GTAB really, or to your GLH or ask a colleague who’s more familiar with the reports around what does that mean and would that patient really benefit from that specific treatment? And I think the GTABs absolutely are the best forum to do that. Just a slight aside, Sanjay, there’s a national study going on in the UK looking at circulation tumor DNA in early phase trial patients to try and match patients to clinical trials.

And we have a national molecular tumor board run every couple of weeks where people can come along and bring those ctDNA reports, and it’s really, really valuable even for us who having much more experience with using them every single meeting we go to, we learn something new from our colleagues or from the clinical geneticist. So they’re really, really valuable meetings.

Sanjay Popat:

So I think that’s right. We learn from each other, don’t we, across specialties, right the way through. So Rachel, I’m going to ask your thoughts about the sort of predictions that we see on these reports. And I thought in your labs you’ve got all these clever people that sort of spend a lot of time looking at these things or is it all done by computer, you just put the report in? Tell me what happens behind the scenes.

Rachel Butler:

Yeah, so we’ve kind of got a mixed model at the moment of who’s delivering the ctDNA analysis. And with that mixed model, again, you get a mixture of different types of reports, which I accept is not particularly helpful right now. So at the moment, a lot of our ctDNA tests are going out to commercial companies and as you’ve just heard from Matt, these may not necessarily be in the UK and the reports from that analysis, they may sequence 50 or so genes and they’ll report back any variants that they see in those genes. And so, you haven’t necessarily stringently looked at each of those variants to say, do we think that this particular variant is actionable in this tumor type? So I think those reports we need to be a little bit more cautious with. Although I have to say those commercial companies are doing a great job in the delivery of the technology.

If we have UK-based, so GLH-based reporting, they’re likely to be far more circumspect in terms of saying, right, we’ve actually delved into this variant. We’ve worked out whether it’s actionable or not. And if we don’t know, we’d probably write that on the report as well if we are really not sure if a variant is actionable. But we are at a slight kind of a time of transition right now. We are using the commercial providers right now, even though it’s not funded within the NHS. But I think there will be a gradual move towards the GLHs delivering ctDNA technology. But I would want say one point of caution there is the reason we are reliant upon the commercial companies is because their technology detect gene fusions. And a lot of the local technologies really struggle to detect the gene fusions, and that’s obviously really important in lung cancer.

Sanjay Popat:

So Ali, we’ve heard about the reports and some of the issues around those, some of the ctDNA reports that we get recommend in the additional 25 pages that we get a whole list of clinical trials that the patient’s suitable for. Is that meaningful way forward? Is that helpful or what are your thoughts on that?

Alastair Greystoke:

So it can be helpful. It can point you towards a trial that a patient may be suitable for and will about encouraging trial entry and making options available to patients. However, very similar to what Matt said, unfortunately a lot of this just comes from a computer having the gene put in and doing an automatic readout. So we looked at this very closely and found that often the trials it recommends maybe in a different country, maybe for a distant disease type or is no longer interested in that particular biomarker. So again, just treat it with a degree of skepticism, talk to your local Experimental Cancer Medicine Center, look on the UK-based websites to see if this is a genuine option for your patient.

Sanjay Popat:

Thanks very much Ali. And so, let’s put this all together now and go through some real reports of patients that have undergone ctDNA NGS and let’s get our experts around the table’s thoughts on how we interpret these reports? Okay, let’s start with this first patient. This is a 75-year-old male smoker with a 40-pack year history who presents into the respiratory service, has a CT scan which shows a small suspected primary and a large pleural effusion. The effusion was tapped and confirmed a TTF1-positive adenocarcinoma. In parallel, he had a blood draw for ctDNA NGS and we had this result where it reads no tumor-related somatic alterations were detected in the patients. Rachel, so what do we think about that result? Is that a useful result? Is that not a useful result?

Rachel Butler:

It’s probably not a useful result. So what I would take from the report that we have here is that in fact the blood sampling has failed almost. So even if there was ctDNA present in that sample, there was so little of it, we haven’t been able to detect it. So in this situation, my recommendation would be go and do an invasive test, a tissue sampling, or even look at the pleural fluid.

Sanjay Popat:

Yeah, okay. So the conclusion is a non-informative result. We really do need tissue molecular to really help us forward. And Ali, what sort of percentage are we seeing this? I’ve seen, I mean, I guess it depends on the type of case that we’re doing the denominator. Do we have any idea of what sort of percentage of non-informative ctDNA and sort of routine lung practice?

Alastair Greystoke:

Yeah, I think if you’re looking at stage IV disease, it’s probably around about 15 to 20%. But as we talked about earlier, will partly depend on the pattern of this disease. This person only has intrathoracic disease, so it will be a lower pickup.

Sanjay Popat:

Yeah. And similarly, if the patient was at a T1 and only brain mets, I’d also not really expect them to be shedding much and expect-

Alastair Greystoke:

Exactly, exactly.

Sanjay Popat:

Okay, let’s move on to case number two. This is a 75-year-old female who presented with dyspnea back pain, cough, importantly they’ve never smoked. CT scan for staging demonstrated left upper lobe mass, there was a pleural effusion, adenopathy, bilateral lung nodules, they’d undergone bronchoscopy and the results of that and molecular were pending. And in parallel she had ctDNA NGS. And Matt, I’m going to get your thoughts on this particular one. There’s a few genotypes that you can see here. What’s your approach to interpreting this report?

Matthew Krebs:

Yeah, I think this is a very valuable result for this patient. So obviously, the first thing that jumps out to you is the EGFR L858R mutation, which is one of the common EGFR mutations of course. And then there’s a number of other findings. But just sticking on EGFR for a moment. If you draw your eyes all the way along to the right where it talks about the percentage of, they call it cell circulating free DNA here, it’s 0.72%. Now that is quite low. But nonetheless, this looks to be a very significant finding for this patient. Now it’s probably lowish. Looking at the clinical details you’ve provided, again, they’ve got a pleural effusion, it’s a left upper lobe mass, some adenopathy and bilateral nodules that presumably are limited to the lungs. So I’m presuming they’ve not got much metastatic disease outside of the chest.

So you’re already thinking they’re not likely to have very much circulating tumor DNA. So this is the kind of level you might expect for this patient. So I think that fits with a clinical picture. And then you’ve got a few other things on here. PIK3CA and three TP53 mutations, all again, relatively low variant allele fractions. This just tells me this is a valid result. We found genes that we would expect to find in a lung cancer patient that’s something actionable with the EGFR and I personally would be happy to treat that patient on the basis of that result with an EGFR inhibitor. And I don’t take much accounts of the other genes that have been identified there other than just acknowledging it. They’ve got a TP53 mutation. Would that impact how well they’re going to benefit? Maybe, but wouldn’t impact my decision overall to treat that patient.

Sanjay Popat:

Great, thank you. So about of prognostic value, the TP53 mutation. Ali, we don’t need to have tissue verification of the L858R. It’s straight onto EGFR kinase inhibitor, isn’t it?

Alastair Greystoke:

Yeah. So one of the benefits of the GMSA pilot that we talked about earlier is that NHS England have just updated all the blue tick criteria. So whilst three or four months ago, theoretically you would’ve required histological confirmation, now in the context of the MDT being happy, this metastatic lung cancer all fits with physical picture, you should go ahead and treat this patient.

Sanjay Popat:

Okay, great. And I think I know that this panel also has RB1 on it and there’s no RB1 reported here. So we’re fairly confident this is an RB1 wild-type TP53 mutant background for the somatic EGFR mutation. It gives us some sort of prognostic utility as well in terms of the NGS read as well. Let’s move on to case three. This is a 55-year-old female. Again, a never smoker who presents to the service. PS1 performance status 1. Cough dyspnea, thoracic discomfort CT scan demonstrates a moderate-sized tumor T3, mediastinal nodes M2 and M1c by virtue of lung pleura. Actually, I should say it was adrenal metastases there as well. EBUS was positive for TTF1-positive adenocarcinoma. And NGS has been requested and in parallel a ctDNA was performed. And we see this long, very scary list of mutations here. Ali, I’m going to ask you. Am I really just panicking now, because I can see this long list of words I’ve never seen before? How should I approach this one?

Alastair Greystoke:

So again, you take the approach that Matt looked at before and you look the top gene shows an Exon 19 deletion and we should all be very familiar with those from looking at tissue reports. And this fits with the never smoker, common sensitizing mutation. Here you’re seeing, if you look at the percentage of ctDNA, which is a slightly different place in this report, you can see just under 21%, which suggests that’s actually relatively high for the clinical scenario and maybe it’s telling you something a little bit about the patient, but again, this is something where you would be happy to treat the patient appropriately. But there is some extra information on there. So the CTNNB1 genes, you can see that all that very low percentages and this is where the biomarker nodes get involved, that’s actually Beta-catenin, probably doesn’t have any significance for the patient.

But going down to the bottom, you can see there is a reasonable level of mutated P53 8%, and also suggesting that lower levels of possible mutation in RB1. And we know that these patients who have RB1 and P53 mutations at present at baseline may be at slightly worse prognosis and may be at slightly more prone to transforming to small cell in the future. Doesn’t mean treat them any differently now, but that may just be information made, but want to just bank away for the future and you may just want to keep a closer eye on this patient.

Sanjay Popat:

So you’d be happy to start an EGFR kinase inhibitor straight on this. We don’t need to wait for tissue, we can just crack on. And we’ve discussed this type of case in some of our internal meetings before we see EGFR amplification. Do you think that’s contributing to the higher variant allele frequency that you’ve got there for the Exon 19 deletion?

Terri McVeigh:

Potentially. It gets very complicated when you see coexisting amplifications. It just means you have to kind of think maybe your VAF is furiously increased. So I don’t put too much weight on now, what low means versus high is a little bit nebulous given different types of technologies use different kind of metrics for how they quantify that. So yeah, regardless it’s definitely there. And as I completely agree with Ali, I treat on the basis of the finding, yeah.

Sanjay Popat:

Okay. So we just crack on with that and we keep calling, just work through methodologically through the list of genes there to figure them out. Okay, case four. This is a 52-year-old female, an ex-light smoker who’s already known and established to have stage IV non-small-cell lung cancer, which is EGFR mutant, EGFR Exon 19 patient. She actually was treated with first line afatinib and then relapsed and had proven EGFR T790M resistance mutation identified and it was transitioned onto second line osimertinib, had a very good response to osimertinib and is now progressing. The disease isn’t particularly biopsy, she’s undergone ctDNA NGS and we have this particular result. Now I see some red things on the screen that’s making me quite scared. I’m going to come to Ali. How should I be interpreting this particular result?

Alastair Greystoke:

It is just very much in the context of the clinical history. It’s just showing you how the patient has become resistant through lines of treatment that you’ve given them. So the cancer presumably had the Exon 19-del as the initial founder mutation and then when you gave the afatinib you progressed, you led to the formation of T790M, which presumably showing up a tumor. And that’s just showing you in the very top line. And then it’s also showing you some information about how it may have become resistant to osimertinib and that they’re finding small levels of the third mutation in EGFR, what’s called C797S that stops the binding of osimertinib. And that may be the mechanism of resistance of, although the levels are becoming quite low at that level and you may want to take it with a pinch of salt. There are trials starting to come through looking at this, but I’d probably want a biopsy to confirm it at that level. I don’t know what Matt thinks about that.

Sanjay Popat:

Yeah, Matt, we have fourth generation drugs in development looking at C797S in this setting. Are you getting excited that you might want to think about recruiting this participant to your trial?

Matthew Krebs:

Yeah, for sure. I think with this finding, yeah quite right that those studies are very early on at the moment. There aren’t many open if any yet just at the moment in the UK. But they are coming and this is exactly the sort of patient that would be eligible. Just another comment on this as well. So I agree entirely with Ali’s interpretation. I guess there’s some relevant negatives here as well, right? So we’re looking for potential resistance alterations to the osimertinib and I guess we’re looking for things like met amplification or possible fusions or some other mutation in the MAP kinase pathway. I guess it’s notable that they’re not there and that’s important, but I wouldn’t be 100% confident about that. And I’m on the same page as Ali, given that the cfDNA, the variant allele fraction is so low, you are not going to be absolutely confident that there definitely isn’t some amplification there in met or potentially a fusion. And I think really the biopsy is really critical.

I know you said this patients going to be hard to biopsy, but maybe you just need to wait for that patient until you can get some tissue and then be absolutely certain by tissue testing. But there’s nothing else there.

Sanjay Popat:

So it’s an important finding. So we have some potential positive results here, the C797S in the background of T790M. But because of the low variant allele frequency, we can’t exclude other resistance mechanisms being there. And so, you might want to consider a re-biopsy to look for those other things as well. Pertinent, important negative with ctDNA. So let’s move on to the next case. This is an interesting one. This is a 51-year-old female who again is a known EGFR mutant with Exon 19 deletion was treated with first line afatinib and who at that point on progression had a ctDNA looking for T790M, but it was non inform, so she transitioned to tissue biopsy which confirmed the T790M and as you’d expect then went onto second line osimertinib. Had a very good response to osimertinib, but is now progressing in the lung again, not biopsiable. And we get this long list of variants here. So Terri, how do I think about interpreting this list of variants here?

Terri McVeigh:

Yeah, so I guess your eye is immediately drawn to the high fraction of the first variant. And excuse me, from my perspective as a clinical geneticist, I also get a bit excited by the variant itself, because this is one that we very frequently see. So this R337H is a recurrent variant and one of the first questions I’d ask is the ethnicity of this individual, because this particular variant is actually a founder variant in South Brazil. So about one in 500 people in that part of the world have this particular germline variant in TP53. So those two things that the variant allele fraction and the fact that we see this quite a lot would make me very nervous that this is a potentially germline variant. This particular variant is really interesting actually, because there’s lots of debate about whether it’s low penetrance, because actually in Brazil one in 500 people, that’s not that uncommon.

There’s a very significant risk of adrenal cortical cancer in childhood, but actually in the people that get through childhood without developing that cancer, the risk of other TP53-associated cancers is reportedly lower than other TP53 variants. Saying that though, we have a couple of families where they have the bog-standard kind of more typically belief for many kind of presentation and we are increasingly seeing lung cancers happening as part of the TP53 kind of phenotype. So this is definitely a woman that I would want to see for confirmatory testing. The other variants sort of pale in significance, they may well be tumor-derived, they’re not likely to be of germline origin, but whether this is related to her lung cancer, who knows? I think that will be difficult to prove or disprove either way. And it probably won’t influence her immediate management, because of course TP53 isn’t target-able as yet.

Sanjay Popat:

So in summary, this is a list of variants and you’re looking at the near 50% variant allele frequency of the TP53 mutation making you think that this is most likely to be germline and that this particular mutation is one that has been reported as germline and so you would expect a clinical genetics referral for this type of case. And so, therefore this ctDNA NGS hasn’t really told us anything about the resistance mechanism to osimertinib, which is why it was done, but has told us something about perhaps the constitutional genomics of that individual which may or may not have implications for that family. Is that right Terri?

Terri McVeigh:

Absolutely. And I would just have a little word of caution about TP53, because TP53 variants obviously have huge implications when they’re pathogenic. It’s belief for many is a pretty horrendous cancer predisposition syndrome. But actually interpreting the significance of variants in TP53 is really challenging and we have a slightly lower threshold for calling something clinically actionable when we find it in the tumor context. Part of that is just the pipelines are much quicker. You have to go through the variant interpretation with not less stringency, but certainly in the germline context we look at multiple other lines of evidence compared to what’s looked at in the somatic context. The guidelines have changed very recently such that our threshold for calling something likely pathogenic in the germline is actually much higher than it used to be. So even if you do identify a variant that you think is likely to be of germline origin, it’s very important that you have a chat with your clinical genetics team, because they may not actually recommend it for confirmatory testing, because if it ends up being something that’s not considered pathogenic, it won’t be taken forward for confirmatory testing.

So I’d be a little bit cautious about over-egging a variant until you’ve kind of spoken to somebody with germline expertise and the scientists are amazing in the germline laboratories for TP53 interpretation.

Sanjay Popat:

Right. Okay. So discuss with clinical genetics from the TP53 viewpoint that we’ve identified there. But Matthew, from the lung oncology viewpoint, we haven’t found anything associated with the osimertinib resistance. So presumably this ctDNA’s of no value in terms of working up the next steps forward for therapeutics for this patient, is that correct?

Matthew Krebs:

Yeah, I think that’s right from clinical treatment perspective, nothing obvious to get your teeth into for a treatment option. I could see they’re not biopsiable. So I guess if they’re progressing in thinking about further treatment then perhaps something obviously standard of care options but in terms of chemotherapy, but I guess maybe some other trials that may not be, require a specific resistance mechanism that might be an option always worth exploring [inaudible 00:41:17].

Sanjay Popat:

So I’m going to move on to this case. This is a interesting one. This is an 81-year-old male, an ex-smoker with a 20-pack year history with a previous bypass graft surgery, hypertension, COPD, whose CT scan demonstrated T2N1 tumor with adrenal metastatic involvement alongside several other areas. EBUS confirmed TTF1-positive adenocarcinoma and whilst the NGS was cooking and had been a lot of blood samples taken for ctDNA, and we find this interesting variant. So Alastair, can I have your thoughts on this particular variant and what does this mean for what is meant to be non-small-cell lung cancer?

Alastair Greystoke:

Well I suspect this isn’t coming from the tumor or not from lung cancer anyway. You’re seeing a mutation in JAK2, which is an very common mutation in JAK2, which is commonly associated with hematological malignancies such as myelofibrosis. It’s at a decent level. I would want to be looking at this man’s full blood count and giving his age and that I would be discussing with my hematology colleagues and asking them if they would like to see him, because I’m concerned there’s something else going on here as well.

Sanjay Popat:

Thanks. So I’ll bring Terri into the conversation. So this isn’t a variant allele frequency of 50%. So this isn’t germline. Ali’s told me that he doesn’t think that it’s from the lung cancer. So what is it? I’m a bit confused.

Terri McVeigh:

Yeah. So this is really interesting and I think Rachel alluded to this earlier on. So when you do blood-based analysis, you’ll pick up variants from a number of different sources. One source is the germline, and this is what I meant earlier on about saying it’s not as easy just to say is a germline or tumor-derived, because it could be from anywhere. So there’s a very well-recognized phenomenon where it’s called mutational burden of aging. So hemolytic stem cells will acquire variants in different genes. If you have a variant allele frequency of greater than 1%, it’s present in about 1% of the population over 65. It’s about 10% if you’re over 85. And there are certain genes in which variants will be currently identified that are part of this clonal hematopoiesis of indeterminate potential is what we call it. So it’s a indeterminate potential if the variants are in genes that are associated with hematological malignancies so that the genes most commonly kind of implicated are DNMT3A, ASXL1.

But you also see it in genes that are typically associated with solid organ cancers as well, things like TP53 and KRAS. And that’s where it gets really messy, because if you find a variant that reasonably low variant allele frequency in something like TP53 in a 90-year-old, that could easily be tumor-derived, but it could well be part of clonal hematopoiesis as well. Something like JAK2 as Ali has said, like that’s very characteristic of myeloproliferative neoplasm. So again, the first thing I’d be looking at is a full blood count in this gentleman and potentially an onward referral to hematology as well. But it gets really tricky, because some of those genes, it could be from anywhere. And actually if you start talking about constitutional mosaicism as well, it actually gets even more complicated. So I won’t get into that we here all day, but yeah, could be from anywhere. But the gene here is pretty convincing that it’s hematological rather than lung anyway.

Sanjay Popat:

Okay, great. So we’ve seen something in JAK2 and this is something called a V617F, which I think is what you classify as a hotspot. And so, your summary is that when we see that we think, does the patient have a hematological malignancy or might they develop one in the future at some point? And we either need to check whether that’s the case or refer on to hematology to evaluate the patient further. But Ali, coming back to you in terms of the lung cancer question that we’re asking, which is what is the target in this patient’s cancer? This is non-informative, is that right?

Alastair Greystoke:

Yeah, it’s probably, well, looking at the pattern of disease, you suspect they should have circulating tumor DNA, but you must remember that these panels do some number of genes between 18 and 300. And it may just be that we haven’t picked up the gene that’s driving this cancer forward, either because it’s not on the panel or for example, it may be a fusion which can sometimes be more difficult to pick up. So it’s non-informative in terms of the cancer. You need more work done on his EBUS sample.

Sanjay Popat:

Okay, brilliant. Well look, we’ll close these cases now and we’ll wind up our discussion. I want to thank all our expert colleagues for inputting their time to this amazing discussion that we’ve had over the past hour. We can see that ctDNA brings a whole load of new data and can be clearly very beneficial to our patients. But the interpretation of it can really be quite challenging and I would suggest it’s not for the beginners. So I get your toes dipped into it and get your reports reviewed in Genomic Tumor Advisory Boards until we have some sense of confidence as to the pattern recognition, that starts to occur after all. I’d like to thank my expert colleagues who are giving up their time and helping with this excellent discussion, our colleagues at VJ Oncology for collaborating with BTOG for this series of discussions. And for you and the audience, do watch out for our next collaborative discussion, our panel-based discussion with Video Journal Oncology and BTOG, do tune in and look out for our announcements about this CG course. Thanks for your time.