Genomic Tumor Correlates of Response to Trimodal Therapy for Muscle-Invasive Bladder Cancer - Sophia Kamran

April 22, 2024

Sam Chang hosts Sophia Kamran to discuss her recent research on bladder cancer. The study, titled "Genomic Tumor Correlates of Clinical Outcomes Following Organ-sparing Chemoradiation Therapy for Bladder Cancer," focuses on genomic characterizations that could predict outcomes for bladder preservation strategies. Dr. Kamran details the use of genomic and transcriptomic profiling to guide treatment options for muscle-invasive bladder cancer, emphasizing the potential of biomarkers in determining the most effective treatment pathways—either surgical or a trimodal therapy approach. The research aims to refine treatment decisions and enhance patient outcomes through precision oncology, offering a promising step toward personalized medicine in bladder cancer treatment.


Sophia Kamran, MD, Radiation Oncologist, Massachusetts General Hospital, Assistant Professor of Radiation Oncology, Harvard Medical School, Boston, MA

Sam S. Chang, MD, MBA, Urologist, Patricia and Rodes Hart Professor of Urologic Surgery, Vanderbilt University Medical Center, Chief Surgical Officer, Vanderbilt-Ingram Cancer Center Nashville, TN

Read the Full Video Transcript

Sam Chang: Hi, I'm Sam Chang. I'm a urologist in Nashville, Tennessee, and work at Vanderbilt University Medical Center. We are quite fortunate to have actually one of the leading radiation oncologists when it comes to the treatment of bladder cancer. We have Dr. Sophia Kamran, who is a radiation oncologist and assistant professor at the Mass General Cancer Center in Boston, and she's actually going to talk about a recent publication that they've put together looking at genomic characterizations or classifications that may help predict outcomes associated with bladder preservation type strategies.

So Sophia, I've known you for some period of time. I want to thank you so much for spending some time with us, and look forward to your presentation.

Sophia Kamran: Thank you. Thank you so much. Thank you for the invitation to be here. I'm excited to share some of our exciting results with you.

So again, thank you for having me today. The title of this presentation, and of the paper that we published recently in Clinical Cancer Research, is "Genomic Tumor Correlates of Clinical Outcomes Following Organ-sparing Chemoradiation Therapy for Bladder Cancer," also known as trimodal therapy, which I'll talk a little bit about.

Muscle-invasive bladder cancer is a challenging entity to treat, and ideally, what we would like to get to in all of oncology, but precision oncology really is what we're looking for where we could utilize biomarkers to guide management or appropriate management for our patients. This is a little schematic from my colleagues that was published in 2018, but basically gives us an overview of what we're thinking.

A patient presents with muscle-invasive bladder cancer; they're seen in the multidisciplinary clinic. We look at clinical pathologic factors right now, but hopefully, one day we can look at molecular biomarkers either that are predictive or prognostic, and that can help steer a patient towards one of the two gold standard treatments, which are either radical cystectomy plus/minus neoadjuvant chemotherapy or trimodality therapy, which is TURBT followed by chemoradiation.

So right now, again, we use clinical pathologic factors really to help guide that management decision, but perhaps if we had additional biomarkers, we could really help hone in on what is the appropriate treatment for our patient. So that's really the holy grail that we would like to get to. And so this is one step in that process.

At Mass General Cancer Center, we've been treating patients with muscle-invasive bladder cancer with trimodality therapy for many, many years, and we have access to their tissues and other biological specimens. We hypothesize that there are underlying molecular features that contribute to varying outcomes to TMT.

So we aimed in this particular study to look at the genomic and transcriptomic landscape associated with a response to TMT. Namely, we used whole exome sequencing and expression profiling.

In terms of our outcome, our clinical outcomes, we define that as modified bladder-intact event-free survival. I just wanted to take a moment to define what this is because it is a bit different than the traditional outcome of overall survival, among other things. So it's mBI-EFS. This is actually the secondary endpoint of the very large energy SWOG 1806 trial that is ongoing, looking at TMT plus/minus atezolizumab.

The endpoint includes muscle-invasive recurrence, loco-regional nodal progression, distant mets, death from bladder cancer, and radical cystectomy for either cancer recurrence or toxicity. So essentially, we just looked at whether or not patients had one of these events. If they had a cumulative incidence of events, they were defined as having an unfavorable outcome, whereas if they did not have the incidence of any events, they were defined as having a favorable long-term outcome.

These are the methods. This is how we identified patients. We looked and made sure that patients had adequate TURBT tissue for whole exome sequencing, which we had 92 patients identified. They underwent whole exome sequencing. We looked at quality control metrics. Sixteen did not pass, and we ended up with 76 patients for final whole exome sequencing analysis. And then just the breakdown in terms of favorable and unfavorable. It was almost about half and half.

Of these patients, we also looked to see if we had enough tissue available for RNA extraction and transcriptome analysis, of which about 67 were available. And you could see again the breakdown of favorable versus unfavorable outcomes. The median follow-up was 74.6 months in the live patients.

One thing we first wanted to look at with this data set is whether or not it was a representative data set for muscle-invasive bladder cancer. And you can see here that it actually is. The overall tumor mutational burden was about 12.4 mutations per megabase, which is in line with what we see with other MIBC cohorts.

Again, the frequency of gene alterations was also in line with other MIBC cohorts, and prior signatures, mutational signatures were also consistent with other MIBC cohorts. So that just, again, suggests that our cohort is representative and can be utilized for these types of analyses.

One of the first things we wanted to look at was the difference in tumor mutational burden between the favorable and unfavorable responders. As you can see, there was no significant difference between the two. And you can see the breakdown of mutational burden between favorable and unfavorable patients.

Then we wanted to perform an exploratory analysis, specifically looking at alterations in DNA damage repair genes. The reason why we wanted to look at these specific DDR genes, DNA damage repair genes, is because there's been a lot of interest in these specific genes, specifically in terms of how to select or identify patients for MIBC clinical trials.

So there are two clinical trials that are ongoing that are looking at bladder-sparing regimens that are utilized a little bit differently. A lot of them are chemo-only regimens, but just looking to see how we can spare the bladder in different ways. And specifically, if patients have a combination of specific gene alterations, they would be offered one thing versus another thing, or they may even be offered participation in the trial itself.

So the genes are listed below as you can see here, and there's a lot of interest in DDR genes in terms of radiosensitivity as well, so radiation sensitivity. So we were very interested in these specific genes within our cohort, and whether or not that correlated at all with favorable or unfavorable outcomes.

So we looked at these DDR mutations in our cohort, and I'll just orient you really quickly to this figure. So CMB is along the top. We have every single individual patient listed from left to right, and whether or not patients received concurrent cisplatin. This line refers to whether or not patients had loco-regional failure, and then this is the mBI-EFS outcome.

Again, blue is favorable, red is unfavorable, and these are the genes that are the DDR genes of interest. And what we found is that there is a clustering of DDR genes among gene alterations among individuals that had a favorable outcome. And we basically looked at that with regards to those outcomes. So you could see that having any DDR mutation was associated with an improved mBI-EFS and loco-regional control. And you can see that was statistically significant compared to patients that didn't have any alterations, any DDR alterations.

And then when we dug down deeper into the specific DDR alterations, you could see ERCC2 came out as being highly associated or correlated with favorable outcomes. And so you could see that there were eight patients in total that had ERCC2 mutations, and seven of them clustered among the favorable outcomes.

So that was very interesting to us and we wanted to dig a little further. What is ERCC2? This is a key gene involved in the DNA damage response. It's specifically involved in the nucleotide excision repair pathway, and the ERCC2 mutations have been previously associated with an improved response to neoadjuvant cisplatin-based chemotherapy.

So again, we found that more patients with favorable long-term outcomes had ERCC2 mutations compared to unfavorable outcomes. Of the eight total ERCC2 mutations, seven were located in the helicase domain. And then this is just a nice schematic for... as you can see here.

Interestingly, six of the eight patients that had ERCC2 mutations received cisplatin-based chemotherapy. They all had good outcomes. Of the two that did not receive cisplatin-based chemotherapy, one who had a good outcome received mitomycin C, which also requires the NER pathway. And then the one person who had an unfavorable outcome did not receive any cisplatin-based or NER-type based requirement for chemotherapy.

So that's very interesting that it seems that when you have an ERCC2 mutation and you're receiving trimodality therapy specifically with cisplatin-based chemotherapy or another chemotherapy that requires the NER pathway, patients do very, very well.

Again, this is looking more specifically at our association with the mBI-EFS outcome and also bladder-intact event-free survival. The difference between these two outcomes is that bladder-intact event-free survival includes all-cause mortality as an event as well, whereas mBI-EFS does not.

And then finally, we wanted to see whether or not in cell lines that ERCC2 mutations were... how they behaved in ERCC2 wild-type and ERCC2 mutant cell lines in response to radiation and chemoradiation with cisplatin.

So you could see that on the left-hand side for both of these, these are two different types of cell lines, bladder cell lines, and there's one that's ERCC2 wild-type, and then the other side is ERCC2 mutant. And then along the top horizontally, we're seeing increasing radiation amounts, and then vertically going down, we're seeing increasing chemotherapy amounts.

And as you can see, to get the same amount of cell kill in the wild-type, you need a lot more cisplatin and irradiation to really see adequate cell kill. But with the mutant cell lines, you don't actually need as much irradiation or cisplatin chemotherapy. The same with this other cell line, we're seeing the same thing, whereas if it's wild-type, we're seeing that we need increased amounts of either chemotherapy or irradiation, and not so much if it's mutant.

And then these lines or these graphs are actually showing the same thing: that in the mutant cell lines, you can see that there's more cell kill when there's the combination of iridium and cisplatin.

So our conclusions or our takeaways are that this is the largest MIBC TMT cohort to date with integrative molecular characterization. We know that this is representative of other large MIBC cohorts based on gene alterations. We see that the frequency of FGFR3 alterations is low, which is reassuring because FGFR3 is more associated with non-muscle-invasive bladder cancer. And this is obviously a muscle-invasive bladder cancer cohort. Also, mutational signature analysis was consistent with other large MIBC cohorts.

We found that alterations in ERCC2, which is a critical DDR pathway gene, may be associated with an improved response to TMT. And in addition, just really having any DDR pathway genetic alteration was associated with improved response to TMT as well. We used cell lines for functional characterization, and as you could see, it suggests that functional ERCC2 loss is sufficient to drive increased sensitivity to chemoradiation, specifically with cisplatin.

So in conclusion, we say that ERCC2 represents a very promising biomarker for the selection of patients with MIBC who may be appropriate for treatment with trimodality therapy. Of course, this is just a single institution study and it needs further validation in larger cohorts, but this is very exciting.

And then if you want to read more, you can check out the paper that was just published this past December, and I just want to say thank you to all of my co-authors, study PIs, and of course our funding sources.

Sam Chang: Sophia, that was fantastic. I think that clearly we're scratching the surface on our understanding of prediction for patients with neoadjuvant therapy of any kind prior to surgery and prior to radiation treatment. So to have any signal, it definitely makes us excited about, okay, now we can better define a cohort that may best be treated in this way.

So where do we go next? We see this signal with ERCC2. What do you see next in terms of research? Is it enough to say, "Boy, we need a big study looking at this predictive mutational signal"? Where do you want to go next?

Sophia Kamran: Yeah, I think that's a great question. So I don’t think this is quite ready for prime time just yet. I do think that we need to validate it in an external large cohort. Luckily, we do have access to NRG cohorts. So NRG has a long history of performing trimodality clinical trials, and so they also have specimens available. And now that we have both the ERCC2 gene mutation to look at as well as just general DDR mutations and alterations to look at, I think that we can use NRG biospecimens to validate our findings.

And I think that would be a next big, big step to really define, yes, this is truly something that we should be incorporating into our everyday clinical practice. That is actually ongoing. I’m actually performing that with my colleagues. It’s being led by Kent Mouw, Dr. Kent Mouw, and Dr. Jason Efstathiou, as well as myself and some of the other colleagues on this paper where we’re actually looking at these specimens and trying to validate what we found.

Sam Chang: I think that's incredibly exciting. I know it's small numbers, but it was fascinating that the single patient who became unfavorable was the single patient that didn’t get cisplatin-based chemotherapy, which you showed in terms of your graphic and your cell line data. It's this combination of chemotherapy and radiation dose, and if you didn’t get that platinum-based chemotherapy, you’ve lost out, and you were in that single patient, that unfavorable group. So it's really... I mean, it is definitely hypothesis generating, but it is very exciting.

Let’s say in a large cohort like the NRG, you’re able to look at multiple variables that predict outcomes, an unfavorable outcome, that type of thing. Are you also then also going to try to take those into account, worse stage, worse volume disease, other things that you all know tend to be worse actors, now you limit them because they're enrolled in trials, but I look forward to seeing in a multivariate analysis how it all washes out. What are your thoughts?

Sophia Kamran: Yeah, yeah, no, absolutely. I think once we get bigger numbers, larger numbers, we absolutely want to include this in a multivariable analysis with all the other clinical factors that we know. I mean, maybe even if you have... Because like you said, these are clinical trial patients, but I think patients that have mild hydronephrosis, sometimes we kind of say, "Oh, that could be allowable." But maybe having an ERCC2 mutation or maybe a DDR mutation can overcome some of these typically, traditionally negative factors that we think of where we're saying, "Oh, no, we should steer them away from TMT because they have a large tumor or they have hydronephrosis."

But perhaps if their tumor is extremely radiosensitive, it'll just melt away and it doesn’t even matter. It just like you said, it could just completely washes out and we should be steering them towards TMT, or we might say the opposite. "Oh, actually it doesn't matter, and we should absolutely be steering them towards radical cystectomy because they're going to have better outcomes that way." So yeah, that's absolutely the right way.

Sam Chang: Yeah, I think that the whole molecular characterization, classification, whatever phraseology you want to use that helps make that next decision tree will be really important. And I think even more exciting as we integrate other, you talked about atezolizumab, but other things that we integrate in the neoadjuvant setting or peri-treatment setting, both for surgery and for radiation therapy, I think is incredibly exciting.

So I've had the pleasure of being able to work with you on different panels at different meetings, and as always, you've impressed me not only with your data, but also with the next steps and where you're going to go with the research. And so thanks for spending some time with us, and we look forward to your next work in terms of looking at larger cohorts and helping to better define patients who will get optimal therapy with trimodal bladder-preserving therapy. So thanks again, Sophia.

Sophia Kamran: Absolutely. Thank you for having me.