Molecular Landscape of NMIBC and Risk Factors for Progression A Lecture & Discussion - Eugene Pietzak

April 12, 2021

Subtypes in non-muscle-invasive bladder cancer are not well-defined, whereas the molecular landscape of muscle-invasive bladder cancer is relatively well-established. During the optimizing personalized management of non-muscle-invasive bladder cancer session at the American Society of Clinical Oncology Genitourinary Cancers Symposium (ASCO GU), Eugene Pietzak, MD, discussed the molecular landscape of non-muscle-invasive bladder cancer. On UroToday's Bladder Cancer Center of Excellence, Dr. Pietzak answers questions from Ashish Kamat, MD, MBBS, giving further insight on the recent high-profile presentation.


Eugene J. Pietzak, MD, Assistant Attending Surgeon, Memorial Sloan Kettering Cancer Center, Department of Surgery, Urologic Oncology Service, Assistant Professor, Weill Cornell Medicine, New York, NY

Ashish Kamat, MD, MBBS, Professor, Department of Urology, Division of Surgery, University of Texas MD Anderson Cancer Center, President, International Bladder Cancer Group (IBCG), Houston, Texas

Read the Full Video Transcript

Ashish Kamat: Hello, and welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, professor of urology and cancer research at MD Anderson Cancer Center in Houston, Texas. And today, it's my pleasure to be welcoming Dr. Eugene Pietzak, who is an assistant and attending surgeon at Memorial Sloan Kettering in New York, and assistant professor at the Weill Cornell School of Medicine.

Dr. Pietzak has had a couple of really high-profile, interesting presentations at the recent virtual meetings. And I want to thank you, Eugene, for taking the time to present to our audience today on the Molecular Landscape of Non-Muscle Invasive Bladder Cancer and Risk Factors for Progression. If you are ready, the stage is yours.

Eugene Pietzak: All right, thank you for the opportunity to present on this important topic, looking at the molecular landscape of non-muscle invasive bladder cancer. These are my disclosures. None of them are relevant for this presentation.

So this was the outline for my GU ASCO talk. At that point, I had given a discussion of the importance of the molecular subtypes in non-muscle-invasive bladder cancer. However, in the interest of time, I'm going to focus this presentation primarily just on the genomic landscape of non-muscle invasive disease, and discuss some of the genomic predictors of BCG response.

So focused on moving towards targeted therapy for non-muscle invasive disease, we recently completed a project where we did target exome sequencing of 105 pre-treatment tumors from patients with newly diagnosed non-muscle-invasive bladder cancer. In this uncle print here, can be seen, it spans across the entire spectrum from low-grade TA, high-grade TA, high-grade T1, and then some representative muscle-invasive tumors as well.

We did do the sequencing of carcinoma in situ, but we found tissue-based approaches to be relatively unsuccessful, due to some of the stromal contamination issues with these smaller specimens. So certainly, alternative approaches to sequencing, whether it's urine-based or using microdissection, will certainly be needed.

But what we did find was that there was a high rate of chromatin-modifying gene alterations, as well as TERT promoter mutations, across the entire spectrum of both grades, as well as stage. This is very interesting because there are two recent science papers that looked at micro dissections of the histologic normal-appearing urothelium, and basically found that in chromatin-modifying genes in the normal-appearing urothelium, there were a lot of mutations that were found within these chromatin-modifying genes. But it wasn't until a second hit, in one of the more commonly known oncolytic drivers, oncogenic drivers that a bonafide tumor actually formed. So, sort of a two-hit type of approach.

And this is interesting, because urine-based approaches to both screening, as well as surveillance for bladder cancer, as highlighted by the group here at Stanford, you could see that in the urine of patients with bladder cancer, there are a lot of alterations within these chromatin-modifying genes, and TERT promoter mutation located up here. Where, in healthy adults, you don't really see that within the urine itself, which is interesting. These molecularly informed screening approaches, looking at the urine, proceed with the development of a clinically identifiable tumor itself, in some cases. And so urine-based approaches using chromatin-modifying genes could potentially be a way for both screening and surveillance of non-muscle invasive bladder cancer. And some of these mutations, some of these genes, may potentially be actionable therapeutic targets.

And so if you look across patients with high-grade non-muscle invasive bladder cancer, we see that more than 75% of them have at least one potentially actionable target. And these patients, including those with FGFR3, which is generally thought to be more favorable, still have recurrences after BCG, to the same rate as those that have wild-type tumors. And so looking at FGFR3 is sort of the most rational therapeutic target at this point, in non-muscle invasive disease. We see that FGFR3 alterations are very common within low-grade TA, about 70% to 80% of them, still common in patients with high-grade TA, and without carcinoma in situ, up to 60%, and even seen in patients with high-grade T1.

And this is particularly of interest in irrelevance because erdafitinib, an FGFR inhibitor, became the first FDA-approved targeted therapy for patients with metastatic bladder cancer who progressed after prior platinum chemotherapy. And so these are patients with an FGFR3 mutation or fusion. And we see that up to 40% of these patients who were treated with erdafitinib had an objective response to treatment, while only 3% had a complete response for patients with non-muscle invasive disease, because these tumors are generally less genomically complex, compared to metastatic tumors, it is very possible that the response rates would be even higher in patients with non-muscle invasive disease.

However, the caveat being is that toxicity was noted in these patients, especially with the eight-milligram dose. And so patients with non-muscle invasive bladder cancer may be less accepting of these treatments, and that risk-benefit balance really needs to be worked on for patients with non-muscle invasive disease.

And these are just two examples of clinical trial approaches looking at oral FGFR inhibitors. This is a study from Hopkins, led by Noah Hahn, that is looking at a window of opportunity between office cystoscopy and TURBT, with only four to six weeks of treatment. And on the other end of the spectrum, this is the Janssen industry-sponsored trial, of looking up to two years of treatment with erdafitinib as adjuvant therapy for patients with recurrent tumor after prior BCG.

Moving on to looking at DNA damage repair genes and their mutational burden and their significance in non-muscle-invasive bladder cancer. We see that ERCC2 is the most commonly altered DNA damage repair gene, and it is quite common in patients with high-grade non-muscle invasive bladder cancer, compared to those with low-grade non-muscle-invasive bladder cancer. This translates into a higher tumor mutational burden for patients with low-grade non-muscle invasive bladder cancer, which is quite similar to those with the muscle-invasive disease. And these are both statistically higher than those that have low-grade non-muscle-invasive bladder cancer. And for those with a DNA damage repair gene alteration, this tumor mutational burden is much higher than those that have unaltered tumors in high-grade non-muscle invasive bladder cancer.

Now, these findings were validated by the Dana Farber group fairly recently, and their cohort of high-grade non-muscle invasive bladder tumors. And again, they saw the same relationship with higher tumor mutational burden for high-grade non-muscle invasive bladder cancer, which is comparable to muscle-invasive disease, compared to low-grade non-muscle invasive tumors.

And this is, of course, of great interest, as we know that tumor mutational burden tends to correlate with response to immunotherapy and that so-called long tail on the survival curves for patients receiving immunotherapy. So, we know that BCG is nonspecific immunotherapy that provokes an influx of both innate as well as adaptive immune cells.

We've also known for a very long time, since Dr. Lamb and colleagues reported in 1991, of this relationship with the Long Tail for BCG immunotherapy. In fact, they comment on that in this comparison between BCG and intervascular chemotherapy. So certainly one of them, or the original immunotherapy for cancer treatment is BCG.

And so at this time, we had looked at whether there was a relationship with ERCC2, and an improved response rate to BCG, and there is certainly a relationship there, but it was not statistically significant. Our group, and others, continued to do profiling and sequencing of this, and will further investigate whether this ERCC2 gene or a tumor mutational burden, could potentially be a predictive biomarker for BCG response.

The group at Dana Farber looked at 62 high-grade T1 tumors treated with BCG. And what they found is that ERCC2 and BRCA2 were associated with a good outcome to BCG, which they defined as no recurrence with greater than four years of follow-up. They also found that a higher tumor mutational burden was indeed associated with this good outcome. What's interesting is they also found that ARID1A was associated with recurrence as well as a progressive disease for high-grade T1. So ARID1A being this chromatin-modifying gene that we also reported was associated with earlier recurrence after BCG. In fact, at least in our analysis looking at 341 genes, ARID1A was the only gene that was associated with recurrence risk after BCG itself. So this is definitely an area of interest and something worth further exploring.

Other groups have also looked at the role of tumor mutational burden, new antigen burden, and DNA damage repair gene mutations, in response to BCG. And just highlighting one such of these investigations, just the caveat being that this group did whole-exome sequencing without matched germline. When you are looking at mutational burden calling without matched germline, the algorithm that you use really does matter. It's somewhat hard to figure out whether, how to interpret this data without the ability to subtract the germline from the actual mutational burden itself. So something just to be a little bit mindful of as we continue to develop these studies.

But regardless, all this is quite interesting to us, as we have recently identified that patients with non-muscle invasive disease, who receive BCG and then, unfortunately, progressed to secondary muscle-invasive bladder cancer, have a paucity of ERCC2 missense mutations, compared to patients initially presenting with muscle-invasive bladder cancer, so-called primary muscle-invasive bladder cancer. We saw this in both a retrospective discovery cohort, as well as a prospective validation cohort, where the ERCC2 is, essentially these mutations are only occurring in patients with the primary muscle-invasive disease. And now this certainly could have clinical implications, as ERCC2 missense mutations are sensitizing to cis-platinum-based chemotherapy. And we saw that patients with secondary muscle-invasive disease certainly have worse outcomes with neoadjuvant chemotherapy, at least in our cohort. So this idea that maybe perhaps there is a clonal selection with BCG treatment, where it's selectively eliminating ERCC2 missense mutations, in these patients who potentially are progressing to muscle-invasive disease, versus whether patients who are, unfortunately, at higher risk for progressing to muscle-invasive disease, may not have ERCC2 missense mutations present in their pre-treatment tumors.

So all of this taken together is quite provocative, and while the leading hypothesis for BCG is, the mechanism of action is thought to be BCG-specific T cells, and that is the biological rationale for the large SWOG PRIME trial that was recently fully accrued, and we are anxiously awaiting the results of it. A lot of this emerging data for the role of DNA damage repair genes, and tumor mutational burden for BCG response, along with this very elegant recent mouse model of bladder cancer investigation published in PNAS by some of my collaborators at MSK, all are suggestive that there may potentially be a role for tumor-specific T-cells for BCG response. So despite over 40 years of clinical use, there is still a lot we need to learn about BCG, and that we are currently learning.

So in conclusion, patients with high-grade non-muscle-invasive bladder cancer do have a high rate of potentially actionable alterations, but this needs to take into consideration the risk to benefit ratio for these patients. As well, this may be a potential area for further development of urine-based screening and surveillance protocols, looking at the circulating tumor DNA present in the urine.

And there is certainly emerging data for a role for tumor mutational burden and DNA damage repair gene mutations that are suggesting that there may be a tumor-specific T cell mechanism for BCG as well. And then this all taken together highlights the role for multidisciplinary care in patients with non-muscle invasive bladder cancer, which will become increasingly more and more important as systemic therapies, like the FGFR inhibitors, as well as these immune checkpoint inhibitors, become more and more relevant for patients with non-muscle invasive bladder cancer.

Thank you very much for your time, and the opportunity to give this talk.

Ashish Kamat: That was great, Eugene. Thank you for giving a presentation in such a nice succinct manner. You covered a lot of data in your talk, and clearly, with the online format, folks will be able to go back and re-listen to sections and get the gist of what you said, which is really very insightful.

In the interest of time, I will keep my questions short. First off, if you could clarify for the audience, what is your take, based on all the work that you've done and the review of the literature, as to whether today, in 2021, are any of these markers actually ready for primetime use?

Eugene Pietzak: I don't think any of them are ready for prime time use at this point in time. But I think what is very nice is, assuming we will have the data for SWOG PRIME, and the other large prospective cohorts, that we will be able to do further biomarker validation. So I do think in the not so far off future that we will be able to bring more of a precision medicine type of approach to our patients.

Ashish Kamat: Yeah. I kind of agree with you. You also mentioned that BCG obviously has been used for decades, and it is in fact, if you compare with all the data, it's still the most effective immunotherapy for cancer, period. We happen to use it for bladder cancer. And the tools that we have nowadays have clearly evolved to where we actually understand it a lot more. And we are understanding that it is a very non-specific broad immune stimulant. In fact, that is why we are, and others, are using it as a vaccine to boost the efficacy against COVID-19, for example.

But keeping that in mind, and the work that you presented, just some of your thoughts on the actual plasticity of the tumor and the tumor microenvironment in relation to the BCG administration. Do you think that, for example, just hypothetically, induction alone would be sufficient? Do you think that any of these markers would be fluid and plastic enough that you could actually monitor them before, during, and after BCG? What are your thoughts there?

Eugene Pietzak: Yeah. I think that is obviously a very excellent question. Yeah. I think at different time points, there may be different markers that may be relevant, because obviously, it's not just the tumor, as you mentioned, it is the tumor immune microenvironment as well.  There is data that may be pre-established before BCG and predictive of response. There may be ways that we could potentially modulate that.

And then there's obviously the host response as well, and that may be independent or related to the tumor microenvironment. And so the different time points, before you initiate induction BCG, versus the response while receiving induction BCG, may be able to guide who certainly does need maintenance BCG, versus who may potentially be triaged away from maintenance BCG in order to conserve it during this BCG shortage, obviously, rationed up appropriately. But yeah, even the urinary biomarker that you and your group has been working on, looking at the response at the six weeks of BCG, maybe it will guide who needs subsequent treatment, who would need escalation of treatment, so to speak, is definitely something that needs further investigation, for sure.

Ashish Kamat: Again, this is great work, and I want to compliment you on this. It's funny, as recently as I believe was 2018, I had convened a multi-center international panel to look at markers of response to BCG. It was published in European Urology, and the conclusion was, we don't have anything yet, other than just monitoring those patients and following them. And then boom, after that, a lot of work that you and others have done, suggests that we might actually now have the tools where we could actually make that distinction earlier on in the course of a patient's course through their journey, essentially. So again, excellent work, and thank you once again for taking the time and sharing this with our audience.

Eugene Pietzak: Thank you for the opportunity. I greatly appreciate it.