Read the Full Video Transcript

Sam Chang: Hi, my name is Sam Chang. I'm a urologist at Vanderbilt University in Nashville, Tennessee, and we are privileged to have a young investigator from the GU Malignancies Branch of the NCI. Dr. Elias Chandran is going to be actually focusing on kind of a genetic evaluation of urothelial carcinoma and looking at and explaining an article that has come out recently in upper tract as well as bladder cancer primaries. Elias, I know you and the senior author, Dr. Apolo, have spent time looking at this genetic mapping of these different tumors to give us an idea of what may or may not be related, but more importantly, what may or may not be predictive. So thanks so much for spending some time with us.

Elias Chandran: Thank you so much for having me on UroToday. It's such an honor. I'm happy to present our systematic review and meta-analysis on mismatch repair deficiency and microsatellite instability in urothelial carcinoma.

As a bit of background, mismatch repair deficiency, which is abbreviated as dMMR or sometimes MMRd, is the loss of at least one of the four mismatched repair proteins, which are MLH1, MSH2, MSH6, and PMS2. It can result from a somatic mutation in one of the genes or a germline mutation. If it's germline, it is also known as Lynch syndrome or more uncommonly Muir-Torre if it's in association with skin or sebaceous tumors as well.

dMMR leads to microsatellite instability, which is an immunogenic phenotype with high neoantigen expression. Pembrolizumab has tumor agnostic FDA approval in tumors with dMMR or MSI-H. These characteristics have been most well described in colorectal cancer. They occur in five to 15% of colorectal cancer and occur at a higher frequency in localized stage disease. They're also favorably prognostic in colorectal cancer, but there is a paucity of evidence or data in urothelial carcinoma.

I'd like to show two studies which have examined immune checkpoint inhibition in dMMR and MSI-H cancers. The first one is the NICHE-2 study, which was recently published by Myriam Chalabi in the New England Journal of Medicine. In this study, patients were given two cycles of immunotherapy, and all underwent surgery. 98% of patients had at least 50% reduction in the tumor, and 95% of patients had 10% or less residual tumor left in the surgical specimen. So remarkably, almost 70% of patients had a pathologic complete response after only having two cycles of immunotherapy.

The next study is a study of neoadjuvant dostarlimab in dMMR rectal cancer. This was first presented by Dr. Cercek at ASCO 2022, and recently she presented updated data. 42 patients have been treated with dostarlimab, and amazingly all of them have achieved a clinical complete response. You can see from the swimmers plot that these clinical responses have been durable, lasting two, three, and even four years, and none of these patients have required chemo, radiation or surgery, which are the standards of care for localized rectal cancer and are associated with significant morbidity. All patients have achieved organ preservation, which is just remarkable.

Our systematic review and meta-analysis aimed to estimate the prevalence of mismatch repair deficiency and microsatellite instability in bladder cancer and upper tract urothelial carcinoma. We wanted to estimate the prevalence of these characteristics in localized stage and metastatic disease. We also wanted to estimate the response rate to immune checkpoint inhibition and chemotherapy and to estimate the survival of dMMR/MSI-H urothelial carcinoma.

We included studies with patients with bladder cancer or upper tract urothelial carcinoma and reported on dMMR or MSI-H. We included clinical trials, other prospective studies as well as retrospective studies, and we searched the major databases which are shown on the screen as well as conference proceedings and pre-prints.

Our search yielded more than 1,700 studies, of which we included 110, which spanned a period of 30 years. Most of these studies were not eligible for inclusion as they were on other diseases, or they used methods for MSI-H which we deemed non-standard and also did not meet our strict criteria for mismatch repair deficiency, which required complete loss of the proteins rather than just reduced expression. There were 29 studies which reported on dMMR, and these included 4,700 patients and 38 studies on MSI-H including 17,000 patients.

Here we show the prevalence of dMMR in bladder cancer versus upper tract urothelial carcinoma. Just to orient you, these are two forest plots showing the meta-analyses we carried out. The forest plot on the left is for bladder cancer, and the one on the right is for upper tract urothelial carcinoma. Each row shows a study that was included, and the events column shows the number of patients who had dMMR in these studies, and the next column shows the total number of patients that were tested for the characteristics. Further along you see the events column, which shows the prevalence that was reported in these studies, and you can see a wide range.

We performed a meta-analysis, and we came with a point estimate of 2.3% for dMMR in bladder cancer and 8.95% prevalence of dMMR in upper tract urothelial carcinoma. So we found that dMMR occurred more frequently in the upper tract by at least three times.

Next we'll show the meta-analyses in MSI-H. The point estimate for MSI-H in bladder cancer was 2.11%. As there were more studies reporting MSI-H, we were able to take a deeper dive and break down the data by stage as well. This one on the top right of the screen shows MSI-H in localized bladder cancer, for which the point estimate for prevalence was 5.26%, and the one bottom right shows the point estimate of 0.86% in patients with metastatic bladder cancer. We found that MSI-H occurred more frequently in localized bladder cancer than in metastatic bladder cancer.

Next are our findings for upper tract urothelial carcinoma. For patients of all stages of disease, the point estimate for prevalence was 8.36%, which is about four times more than we saw previously in bladder cancer. Next we did meta-analyses in localized stage disease, and the prevalence was 18%. Lastly, for metastatic disease, the prevalence was about 5%, and again, the prevalence was higher in localized stage disease than in metastatic disease. This is similar to what has been reported previously in colorectal cancer as well.

Next, we wanted to estimate the response rates of dMMR/MSI-H urothelial carcinoma to systemic therapy. So these two pie charts show the response rates to, on the left immune checkpoint inhibitor and on the right platinum-based chemotherapy. Essentially what we did was we pooled all the individual patient data for responses to these different treatment modalities. As you can see, the majority of patients responded to immune checkpoint inhibitor, whereas only a small minority did so to platinum-based chemotherapy. If you look at the breakdown under platinum-based chemotherapy, the majority of those who did not respond actually had progressive disease to platinum-based chemotherapy. The only caveats I would mention are that pooled response rates from multiple studies are a crude way of estimating response, and the second of all is that the numbers are very small. But these do suggest that perhaps dMMR and MSI-H are predictive for response to immune checkpoint inhibitor and are predictive of resistance to platinum-based chemotherapy.

Next, I want to show you the prevalence of germline MMR mutations in bladder cancer and upper tract urothelial carcinoma. In bladder cancer, we had one study which reported on this, and two out of 463 patients had a germline MMR mutation, so that's 0.43% compared to a point estimate of 6.53% in upper tract urothelial carcinoma, again, showing that this is more common in upper tract urothelial carcinoma than bladder cancer. In terms of the pattern of MMR loss, the commonest protein loss was the dual loss of MSH2 and MSH6, and the commonest germline mutation was in MSH2. To summarize, our systematic review and meta-analysis found that the prevalence of dMMR and MSI-H were higher in upper tract urothelial carcinoma than bladder cancer. MSI-H was more frequent in localized disease compared to metastatic disease, and dMMR and MSI-H may be predictive for response to immune checkpoint inhibitors and may be predictive of resistance to platinum-based chemotherapy.

So given the two studies in colon and rectal cancer showing exquisite sensitivity of these patients to immune checkpoint inhibition, we think it is important to develop clinical trials in this subgroup of patients in urothelial carcinoma, as these patients may be able to retain their organs and may be able to avoid the morbidities of platinum-based chemotherapy. As such, we are developing a trial in the Alliance Cooperative Group hoping to achieve organ preservation. So with that, thank you very much.

Sam Chang: Elias, thanks so much for that presentation. It makes me wonder several things in looking at the meta-analysis and the next steps. In terms of the meta-analyses, why if you look at your forest plots for both MMR as well as MSI-H, you see a variety or the vast majority near the top of your forest plots ... It had a few small percentages in terms of changes, but there were always some significant outliers that really I think, I don't know if it skewed the data, but if you look, there were always two or three studies that have had much higher levels of these mutations. Why do you think that was the case?

Elias Chandran: So yes, we definitely saw these wide ranges in prevalence, and we did look into this and we did contact each study individually, these outlying studies, just to get a better idea of their research population or their patient population. Most of these studies had either only localized disease or a higher proportion of patients with localized disease. Some of them were also just much smaller in terms of patient numbers, and so maybe that's why they had such a high prevalence. The studies were weighted based on the number of patients, so the pooled estimate does take that into account as well.

Sam Chang: When you looked at the differences in the localized sites having higher expression versus the metastatic sites, so two questions regarding that. First, the tissue that was actually evaluated, were they the actual metastatic sites, either lymph node or liver or lung, or were they from the primary that developed metastatic disease? That's question number one. Question number two, any ideas regarding the pathophysiology of why metastatic sites would be less likely to have these mutations?

Elias Chandran: The source of the tissue is a great question. For the studies that mostly reported localized disease, most of them were from TURBTs, and some of them were also from radical cystectomy specimens, and so those were quite clearly reported. For the studies reporting metastatic disease, it wasn't very clear a lot of the time where the tissue was coming from. In practice, when we get patients with metastatic disease, sometimes the only tissue they have was from the primary surgery.

Sam Chang: Yes.

Elias Chandran: But often though, the gold standard is to biopsy when there is metastatic recurrence, especially if it's quite some time since the initial surgery that has occurred. We can only infer, I suppose, that it is going to be mixed sources from either the primary or the metastases.

Sam Chang: Right. Any idea why that would be? Metastatic sites express or seem to have a significantly lower level of these mutational changes.

Elias Chandran: Yeah. So this is very interesting. The dMMR and MSI-H produce a very immunogenic phenotype. Often if you look at the tumors under the microscope, you see tumor infiltrating lymphocytes in these specimens. So the thought is that they occur ... In colorectal cancer, the similar observation has been made in that because of this immune reaction, dMMR and MSI-H are potentially protective from lymph node spread or metastatic spread. So the cancers tend to stay local. If it does metastasize, it's just a less common phenomenon.

Sam Chang: It may be you're self-selecting those cells that in fact would be less likely to be successfully attacked by the immune system. It's fascinating. Also, this data confirms perhaps the efficacy of why for certain patients, systemic therapy like pembrolizumab for certain non-muscle invasive bladder cancer may in fact be effective. We've got to go back and see and look at the genetic material from some of these, especially in localized disease.

Historically, surgeons haven't thought of systemic therapy as being something that might be successful, but this data looking at not only the colon but the rectal data that you presented as well as the expression, that may help actually give a rationale behind the success of these systemic therapies. I'm very intrigued with what you all are going to do next prospectively in terms of the evaluation of this and the treatment for this. It also makes one consider the fact that the neoadjuvant studies going on with immune checkpoint inhibitors would be in fact maybe quite effective depending upon expression, depending upon those mutations. It would be really good as we tease out and get the data from this in terms of the specimens that are removed or from the initial TUR, if these were the patients that had MMR or MSI type mutations-

Elias Chandran: Absolutely.

Sam Chang: Upfront.

Elias Chandran: Absolutely.

Sam Chang: Yeah. It's really fascinating. Last question. Should we as or should I or we as urologic surgeons, should we start looking for these mutations? Should we start for every higher risk non-invasive cancer or invasive cancer? What's your thought on that, and what's your prediction regarding that?

Elias Chandran: I think that's definitely going to be the way of the future. For now, I guess we usually only do sequencing once the patient is metastatic, and that's usually to try and find mutations to think of their therapeutic options. But I think it will be, given that there is a difference in perhaps mutations in the localized setting versus metastatic, I think it will be important to be doing more of these, at least in a prospective research setting so we know so we can better understand it. But also testing for MSI-H and dMMR particularly are very important, especially in upper tract disease where it occurs more commonly. So actually the most recent AUA guideline does recommend universal testing of patients.

Sam Chang: Yeah, I was going to actually add that. As a member of that guidelines panel, that was one of the things that we really wanted to advocate, separated from other guidelines, which have not in fact made that step, but clearly the possibility of that and then influencing positively therapy down the line I think is very important. Now there's so many different ways that we're evaluating with different types of biomarkers, but to know upfront if truly patients that have the dMMR or MSI-H changes, if they really are in fact resistant to platinum, the whole kind of emphasis that we've had with neoadjuvant platinum-based chemotherapy prior to cystectomy, those patients really are not as likely to respond to platinum. It clearly helps us differentiate between those that may or may not benefit from that neoadjuvant therapy. Same thing with upper tract as well. So-

Elias Chandran: Yeah.

Sam Chang: We really look forward to your future studies, and we know that you have a bright, bright future. Thank you so much for spending some time with us. We really appreciate it.

Elias Chandran: Thank you so much. It was a pleasure.