Genomic Heterogeneity as a Barrier to Precision Oncology in Urothelial Cancer - Timothy Clinton

June 9, 2023

Sam Chang introduces Timothy Clinton to discuss his groundbreaking work on the genomic characteristics of urothelial carcinoma. Dr. Clinton's research explores the genomic heterogeneity within bladder cancer, specifically focusing on the differences between primary and metastatic patients. His work reveals significant discordance between the primary and metastatic tumors in the same patient, suggesting that relying on a single biopsy could overlook over half of the alterations within a patient. This study also unveils the significance of plasma cell-free DNA in detecting additional alterations not identified in the tissue, underlining the importance of dual tissue and plasma testing. Dr. Clinton’s work promises to inform a precision oncology approach and enhance the treatment of bladder cancer patients. The conversation ends with Dr. Chang probing into the influence of time and disease progression on the concordance rate between primary and metastatic biopsy.


Timothy Clinton, MD, Associate Surgeon, Assistant Professor, Department of Urology, Brigham and Women’s Hospital, Boston, MA

Sam S. Chang, MD, M.B.A. Patricia and Rodes Hart Endowed Chair of Urologic Surgery Professor Department of Urology at Vanderbilt University Medical Center

Read the Full Video Transcript

Sam Chang: Hello everyone, I'm Sam Chang. I'm a urologist at Vanderbilt University in Nashville, Tennessee and we're quite fortunate to have a rising superstar in urologic oncology, Dr. Tim Clinton. Tim was a resident at UT Southwestern and finished his fellowship at Memorial Sloan Kettering Cancer Center and he's recently started as assistant professor at Brigham and Women's Department of Urology in Boston. He will be talking about some of the exciting work, looking at genomic characteristics and actually the heterogeneity of the genomic characteristics of urothelial carcinoma. And I think he's got some slides to share to kind of find out some of the work that he and the lab under Dr. David Solit have kind of expanded on and have been working on for actually quite some time. So Tim, I'll turn the program over to you and thank you for spending some time with us.

Timothy Clinton: Well, thanks so much Sam. This was a real pleasure to come and talk with you and the group and I look forward to sharing our recent publication here. So, I'll go ahead and share my screen. All right, great. So, this is a recent publication that we put out from the lab and just as Dr. Chang mentioned, this is looking at the genomic heterogeneity within bladder cancer and ultimately really seeking to determine what the degree of heterogeneity and how that's going to affect our ultimate goal of precision oncology in the treatment of bladder cancer patients. And what this led to was really looking at the concordance of genomic alterations within primary and metastatic patients, looking at the differences in alterations and frequency of those alterations, but also just differences as we progress from a localized to a metastatic state. And this looked at... Not only tissue sequencing but down the line as we get into this presentation a little bit more, our first look at plasma cell-free DNA within bladder cancer as well.

Now, this project is not just a labor of love of myself, it's really a team effort. So, I do need to give some credit to my co-author Jessica Chen, a number of current and past MSK fellows and then of course the clinicians who have really helped amass a large sequencing cohort of patients at MSK. And then of course our PI, Dr. David Solit, who really has spearheaded a lot of the genomic work within bladder cancer and certainly this project. So, thanks again for having me.

So, I'd like to start off with a look at our Memorial Sloan Kettering Bladder cancer cohort. And by the time of this publication, we had amassed already 1,300 patients who have had some amount of genomic sequencing of bladder cancer. And this allowed us to really look at what are the differences in genomic alterations as you go from a low-grade disease to a high grade non-invasive, invasive and then ultimately metastatic cancer. And you can see that there's some striking differences across all of these genomic frequencies. And this is very different than a lot of the other large scale genomic studies that have been previously done such as TCGA where it's really been sort of siloed and looking at localized bladder cancer, muscle invasive bladder cancer. It's quite rare to have... In this setting, we had over 200 metastatic biopsies and this is just a very large cohort, and really gives us a good insight into what are the alterations that develop through treatment along the way through progression of disease.

There are some things that stood out that you would expect. So, genes such as P53 and RB1, really much more prevalent there as you can see in the middle within high risk and metastatic disease. Interestingly, a new target that's coming online, HER2 or ERBB2 much more prevalent in high risk or high stage in metastatic disease as well. And then, some things on the low-grade disease such as FGFR3, much more prevalent. But what you can see is that there's still about 20% of patients who still harbor some amount of an FGFR3 alteration within the metastatic setting. And so, this really kind of keeps going back to that precision oncology mindset where even though it's much more prevalent in low-grade disease, there's clearly a subset of folks as we know, who can be treated with FGFR inhibitors such as erdafitinib.

And then, most interestingly for us, where these chromatin modifying genes all the way over here on the right. So, this is the KDM6A, ARID1A, KMT2D, and this is gaining a lot of interest in the sort of epigenetic world of bladder cancer. And there's been a couple of recent studies in the last year that have actually shown that these alterations within chromatin modifiers are highly prevalent even in normal urothelium. And so, do these contribute to the development of bladder cancer or the progression of bladder cancer as really trying to be teased out in a couple of different labs. And actually, just this month, there was a paper by Dr. Byron Lee at the Cleveland Clinic, I think now at MD Anderson, and he very eloquently showed that KDM6A with this first sort of column here really leads in an epigenetic fashion to the development of normal urothelium to bladder cancer.

And so, this is the first time that we're actually seeing these genes and delineating exactly what they do. So, what I'll point out here in this, is as you go from a low grade to metastatic setting, that KDM6A, just as you expect it seems to be higher in the low grade or lower stages, really leading us to believe that this helps in the early development of bladder cancer. But a unique one is ARID1A right next to it, which is clearly rising as you get into the higher grade and higher stage diseases. And so, maybe this is a gene that's leading to the progression or metastatic progression of tumors as well. And so, this was a broad look at again, over 1,300 patients, and we really wanted to dive into what was some of the drivers of progression within an individual patient. So, this looks like sort of broad across different patients.

Now, let's look at individual patients. And so, we took primary and metastatic tumors from the same patient, and we had 23 of these, what I'll call matched pairs of primary metastatic tumors. And this is a little bit of a busy heat map here at the top in the figure A here, but each column represents an individual patient with a left side being the primary tumor and the right side being the metastatic tumor. And what we found in this setting was that there was still quite a bit of discordance between the primary metastatic tumors even within the same patient. In fact, the concordance between these alterations was only about 42%. That means over half of the mutations seen either in the primary of the metastatic tumor were discordant between the two. And this leads a lot. If you're using only one biopsy to guide your treatment, you may be missing over half of the alterations found within that patient.

But the other interesting thing is actually this top bar here where you can look at the TMB or tumor mutational burden, and you can see largely the tumor mutational burden between primary metastatic tumors is almost stable, almost identical in most of these. And that brings me to the figure down here, C and D, which is a phylogenetic analysis. So, we see this similar tumor mutational burden, which is already very high in bladder cancer as we know, but within the same patient in the primary in the metastatic setting, it stays very stable. And I think that this contributes to what we call early branched evolution of these tumors. It's clear that from that primary tumor when they... Whatever the early clone is, when it starts to develop a metastatic clone, it differentiates very quickly and develops its own set of mutations, hence the high discordance rate and hence the similar sort of tumor mutational burden.
But this has been sort of discussed previously and sort of well shown here for one of the first times. What we sought to sort of validate these findings from the whole exome sequencing that we did in those 23 patients. And we expanded our cohort to a much larger heat map here now in 119 patients. So, I don't expect you to be able to see each individual box, but take my word for it. This is just a very large cohort and we performed not whole exome sequencing, but next generation sequencing with MSK impact sort of an in-house sequencing initiative there. And again, we took individual patients with a primary and a metastatic tumor and compared the concordance rates of these very common alterations within patients, and we found that there was only 23% concordance amongst actionable alterations for these patients. So now, it's even lower.

And again, this just has high implications of the treatment options for patients if you're reliant on a single biopsy, a single primary tumor for these patients. And again, we sort of show down here this is the tumor mutational burden largely similar between the primary and metastatic tumors, thereby giving us sort of more sense that early branched evolution is clearly happening within these bladder cancer patients. And I bring you to this final figure down here in figure C. And the one unique thing is we'll go back to the beginning where we talked about ARID1A that was only developing, we saw it sort of rising as it got into the metastatic and high grade setting, that chromatin modifier, if I go back to the whole exome sequencing for a moment, you can see in this sort of middle row ARID1A when it's discordant is only prevalent within the metastatic tumor.

You can see that here on the right with each one of these boxes. We actually went back and looked at the primary tumor and resampled the primary tumor extensively just to make sure we didn't miss any reads of that mutation and we still didn't find any. And so again, ARID1A is only really being seen in the metastatic or studying for these matched patients. And again, we saw the same here within the much broader cohort as well. And this just gives us more of a sense that these epigenetic modifiers are really contributing to the progression of disease and specifically maybe ARID1A. And so, this is a big focus of ongoing research within some of these more epigenetic type modifiers. And then finally, we wanted to really expand this even further. So, there's a lot of discordance between primary metastatic tumors.

What are we seeing within the plasma cell-free DNA as well? These top two figures, figure A and B here are really looking at a very unique cohort that probably hasn't been seen before where we have a patient who has a primary metastatic biopsy and plasma all banked, and we sequenced all three of those to look at the concordance across alterations. And what I'll point out here in this A figure is that the screen bar shows the alterations that were isolated to just the cell-free DNA, so alterations that were not even picked up in the tissue at all. And in fact, we found about 20% of alterations were exclusive to just the plasma. So now, we're finding even more alterations that may guide treatment that were not even seen in the tumor biopsies. We expanded this of course a little bit more to just anybody who had a tissue-based biopsy and cell-free DNA, and we found that 17% of alterations were exclusive to the cell-free DNA, but I'll make a point in that there was also 23% of alterations that were only seen within the tissue.

So, one of my conclusions here is that I think you almost need both if you're really dealing with a refractory metastatic patient looking for options for treatment to really help guide you as far as determining any potential targets for future treatment. This is an example of ongoing sort of clinical efforts, but you can see here's a single patient who has a single metastatic biopsy and a primary biopsy, but they only found the FGFR3 alteration within the metastatic biopsy. And then, we did serial cell-free DNA measurements throughout their treatment course with erdafitinib. And you can see that the different alterations within FGFR3, including this FGFR3-TACC3 fusion, were really picked up only in the plasma. And so, I think this is for anybody who's ongoing treatment in a metastatic setting, serial cell-free DNA may pose as a real biomarker as to how that patient is responding to treatment.

And additionally, are they developing any resistance mechanisms. And we get into that a little bit in this paper as well, but ultimately that that's pretty much it. So, I think the results suggest that an analysis of tumor tissue collected from a primary bladder tumor will fail to detect really over 3% of actionable alterations. And I do think that patients who are ongoing into metastatic or refractory type treatments really warrant a metastatic biopsy and likely also cell-free DNA or some amount of a plasma biomarker in the future to help guide our treatments.

Sam Chang: Tim, that's great. Tim, that's fascinating work. It clearly shows us that we're just scratching the surface in terms of our precision oncology efforts and that precision oncology has opened up a lot of lack of precision. If I could say that in terms of... Let's start with that last point regarding looking at this trifecta of the primary site, the metastatic site, and the cell-free DNA in the serum. Did you notice, and it's so amazing to have all that as you pointed out, were there differences that you saw temporarily in terms of when you collected the cell-free DNA? You gave the specific example there with the FGFR3 treatment but were there differences in expression that you saw if we collected the serum, just when the primary side before metastasis or just after metastasis, or tell me about the differences in terms of timing of collection of that cell-free DNA.

Timothy Clinton: Yeah, this is a really great point. As we know, these tumors throughout progression are constantly... Either one clone is getting selected out or maybe there's some sort of amount of treatment selection, and especially metastatic patients, right? They're going from chemo to immunotherapy, maybe an FGFR inhibitor, whatever it may be. And so, these tumors are constantly progressing. So, the timing of when you get your either metastatic biopsy or the cell-free DNA makes a huge difference. For the purposes of this study, we took the cell-free DNA that was closest in timing to the metastatic biopsy to limit any of that temporal heterogeneity that you may see because of course, it probably would be more if you got it further away from the timing of your biopsy. But we know already in these patients, and we do point out a few of these individuals where if you get that cell-free DNA much later in the treatment course, well it almost doesn't even look like the same tumor anymore just because of the alterations that have changed.

When we looked at it in whole exome sequencing between the timing of the primary to the metastatic biopsy, that could be a big difference too. If you have a patient who became metastatic within a few months versus a couple of years, we know that there's actually some amount, we didn't see a statistically significant difference, but there was some amount of differences in the actual concordance rate between those given the amount of time between the primary metastatic biopsy.

Sam Chang: As you looked at this, were there any patterns that you noticed that these types of patients tend to had have a much higher concordance, be it who knows low-grade disease, that the few that became more aggressive there, were there any patterns that you noticed that affected this concordance rate between primary metastatic?

Timothy Clinton: Yeah, that's interesting. Yeah, I don't know that we found any specific patterns. What I'll say is that there are some select alterations where the concordance was a little bit higher. So, things like TSC1, even PIK3CA happen to have a little bit higher concordance, but ones that were actually more discordant were things like FGFR3 and HER2, which is just very unique because a lot of attention's being put on that from a precision oncology standpoint.

Sam Chang: Sure.

Timothy Clinton: And if those are the ones that are more discordant within these patients, well that can really sort of change your management for a lot of them. And so, that's about the only sort of interest correlation that I saw. Yeah.

Sam Chang: I mean, clearly that interest is, we have and now medications, they become actionable targets now we get more interest in, and you would love for obviously those targets to be ones where there could be applicable early for localized disease, for maybe even non-invasive disease, et cetera. So, that's why I think FGFR becoming such an interesting target, non-invasive disease, invasive metastatic, et cetera.

Timothy Clinton: There was an interesting paper too, nectin-4 is another one of course that's been looked at. And of course, it's not really captured in a DNA sequencing setting, but there was a recent paper that came out that looked at the concordance between primary metastatic sites, even for that sort of target like nectin-4, and there's still was some discordance there as well. And their conclusions were very similar to ours in that, again, I think what we're seeing across the board for precision oncology is really, you need to use your primary tumor to help guide you see if there's an inkling of an FGFR or a nectin-4 target, but make sure you do your metastatic biopsy as well. So, you can see maybe it is there and you just didn't capture it on the primary biopsy. It's quite interesting that others are seeing this too.

Sam Chang: I mean, if we're learning as much grief as we surgeons would give our oncologists regarding biopsying everything prior to any systemic therapy, I agree with you totally more and more the importance of that tissue biopsy. I think is becoming obviously more paramount. And I applaud all the studies even though they are more difficult, that require the tissue prior to any intervention. And then, careful monitoring afterwards in terms of response, I think is really amazing. And then, I think what's also important is that these findings regarding discordance or lack of concordance, not different from the breast cancer literature or melanoma literature or for other solid tumors, looking at the fact that the heterogeneity exists, we're going to have to obviously recognize that and then do everything we can to try to identify actionable targets. So Tim, last question. Where do we go next? Are you going to continue this work in Boston or are you going to continue your collaboration with Dr. Solit and the folks at MSK? Where are the next steps?

Timothy Clinton: Well, hopefully a little of both. I mean, obviously it's a great group. They've done a really nice job. But here at Brigham and Women and Dana-Farber, we have really fantastic folks, and in fact, there’s a lot of overlap between the two institutions. So right now, still in this plasma-based location, so we're looking at mostly epigenetic modifiers within cell-free DNA and a couple of different assays. But as a urologist, obviously a lot of this work was looking at really metastatic patients and their progression, and I'm trying to see can we identify some of these early identifiers of progression in more earlier stage disease as well. So, I think hopefully this will be more beneficial for the urologist in the future.

Sam Chang: Yeah. No, I think for sure, not only looking at invasive disease, but clearly, we need better risk predictors and biomarker evaluations of non-muscle invasive disease. I mean that that's clearly, we struggle to really evaluate what's going on and what are the best treatments. And so, I look forward to your future findings and your work, and congratulations on starting your career. And I know it will be one full of accolades, I can already tell. And so, thanks for spending some time with us and look forward to talking to you again in the future.

Timothy Clinton: Thanks so much, Sam. It was a real pleasure.