Read the Full Video Transcript

Ashish Kamat: Hello everybody, and welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, professor of urologic oncology at M.D. Anderson Cancer Center. And it's a distinct pleasure to invite to the forum Professor Weranja Ranasinghe, who is currently at Monash University in Australia, but spent some time here with us. And we have really fun memories of you here, Weranja, at M.D. Anderson. So, welcome to the forum and looking forward to hearing what you have to say on the IHC profile to predict response to BCG therapy and patients.

Weranja Ranasinghe: Thanks so much, Dr. Kamat. Very nice to be here. As you know, I'm going to get straight to the talk if that's all right. So as you know, intravesical BCG is the gold standard for high-grade non-muscle invasive bladder cancer. But about 40% of patients recur or don't respond to BCG. And we also know that those who progress will do worse compared to having a cystectomy upfront.

There's also lots of novel emerging therapies in the horizon, and a very exciting time for bladder cancer. And in this context, it's really important to know who's going to respond to BCG. And it also allows us to sequence therapies upfront.

So our aim in this study was to see whether we can develop an IHC profile to predict which patients would respond to BCG. In view of this, what we did was we looked at our institutional cohort of all the patients undergoing BCG between 2018 to 2022.

We defined non-responders as per the EAU guidelines for BCG unresponsive disease, and identified 12 responders and non-responders. And we matched them for baseline patient characteristics and the EAU risk group, and looked at the pre-BCG tumor specimen as well as the post-BCG tumor, and performed IHC as well as RNA sequencing. These were the baseline patient characteristics. And as you can see, there's a mixture of papillary as well as CIS.

So first thing that we wanted to do was to validate our population. We know that with BCG, there's going to be increased in CD8-driven pathways as well as Th1-related pathways. And when we looked at the principal component analysis on the left, you could see that the BCG responsive group, which is in green, separated out very nicely from the BCG refractory group.

Now, we then looked at what pathways contributed to these differences. And you could see most of the pathways were related to CD8 expression or Th1-related pathways. So that essentially validated our study cohort to what's known in the literature.

But what we really wanted to do was to look at the pre-BCG samples to see whether we could identify responders from non-responders. But when we looked at the IHC samples, there were no absolute values in the CD8, CD4, Th1, or Th2 levels on IHC between the responders and the non-responders.

But when we looked at the ratio, you could see that the BCG non-responders had a higher CD4 to CD8 ratio. And when we interrogated that further using what differential expression of CD4 cells there were, using T-Bet as a surrogate marker for Th1 and GATA3 as a surrogate marker for Th2, we found that the non-responders had a higher rate of T-helper 2 to T-helper 1 ratio.

Also, we looked at the median PD-1 levels in the responders versus non-responders. And interestingly, the responders had a higher PD-1 level to start with. And when you expose them to BCG, the levels of PD-1 dropped. In contrast, in the non-responsive group, the levels of PD-1 increased, suggesting that the T-cells were getting exhausted. And there was a 78% increase.

So we looked to see whether we could utilize this integrated set of CD4:CD8 ratios, along with the T-helper ratios and the PD-1 ratios. And we came up with this integrated IC biomarker set using certain thresholds using the Youden's index. And the area under the curve in this was 0.86, which looked quite promising.

So in summary, the BCG responders had a distinct gene expression profile, which was enriched for the Th1 pathways after you exposed to BCG in keeping with the literature. The BCG non-responders had a tumor micro-environment which was lower in the CD8 to CD4 count, and, again, lower in Th1 to Th2 ratios at baseline. And the responders also had a higher level of PD-1, which gave us this integrated cohort that we could essentially use as a biomarker in this pilot study to identify BCG response.

And currently, this is being prospectively evaluated as this is only a pilot study. And mainly, the take-home messages from this is that we really need to identify patients who are unlikely to respond to BCG, given the prognosis, as well as, additional treatments coming up. And we think that understanding the biology can help develop these biomarkers, especially in this context, where this IHC based biomarker panel could be useful in predicting BCG response. And currently, it's being prospectively evaluated.

So I'd like to acknowledge all the uropathologists and other teams who've been involved in this study, as well as the funders and the patients and families who've been supporting this project. Thank you, Dr. Kamat.

Ashish Kamat: Thank you so much, Weranja. And just to remind the audience, you presented this recently at the ASCO GU in San Francisco. One of the things that I wanted to ask you, because we and others, and you as well, have been working on this, trying to be able to predict response to BCG. And it's a very complex interplay between the different pathways. You showed the Th1, Th2. Others have looked at cytokines. We have looked at cytokines. We've also looked at Th1, Th2.

I want to ask you a broader question. Looking at the pilot data that you have, tell us your plans for the prospective validation, because I think that's where the audience is really interested. What are your plans? And how do you plan to validate those prospectively?

Weranja Ranasinghe: So I think I'm just going to start off with why we are going to use IHC. In Australia and lots of other places, IHC is much more readily available. It's faster than genomic sequencing, kind of overcomes the tumor heterogeneity. And we wanted to focus on the tumor micro-environment because BCG is predominantly the mechanism of action is by immune activity. So based on those factors, the IHC profile we thought was the ideal option.

Now, the tricky part for us was that there were no absolute values, which made it a bit more trickier. And that's why we had to resort to ratios, which also makes sense, because it's always a fine balance between the immune suppression and the immune activity in your tumor micro-environment.

So to do this, we had to identify certain thresholds where we could use these ratios. And we did this using something called the Youden's index, where you look to see what the optimal ratio would be to identify the maximum number of patients who would have a response to BCG. So it's a little bit more work than just counting the number of cells. But we think that this pilot study and the primary aim was to identify what thresholds we could use for each of these markers would be the way forward.

We also used three markers, rather than two ratios and one rather than one, because there's a lot of variability in what's been published out there. And I think looking at things from the combined approach, we felt was a better approach than just looking at a single marker, per se.

Ashish Kamat: Great, yeah. No, absolutely right, because it's the ratio. It's the inducible levels or the inducible suppressive levels that really matters, because it's hard to normalize patients that have different baseline levels. And like you said, the ratio or the inducible levels, if you're looking at cytokines, et cetera.

But clear out a little bit on the patient population. And when you say prospective validation, are you planning on having two cohorts of patients that you're going to offer BCG, not BCG? Or are you studying, collecting tissue prospectively, and then looking at it in a retrospective manner?

Weranja Ranasinghe: So we recently got funded, Dr. Kamat, for both of those aspects. So firstly, we needed to prospectively validate all of the data on a bigger cohort, which is adequately powered using this pilot study. So that was the primary purpose of this. Then we've also started using gemcitabine docetaxel, and based on what the bridge trials, there's a lot of retrospective studies, and what bridge comes up as.

Then we do have a planned small pilot study, looking to see whether we could randomize patients to BCG versus gem/doce up front to see whether there is a response purely based on the marker. And that's also a secondary point to this, is that most of the patients who did not respond in this cohort actually went on to have gem/doce and had a really good response at two years. So we think there's some evidence to support a randomized trial as well.

Ashish Kamat: Great. Yeah, because as I'm sure you're aware, there's AI-based and machine learning-based slide data from a regular IHC. There's also other data that is showing that patients who may not do well with BCG might actually do better with gem/doce, and whether non-response to BCG is a biomarker in itself or it's these parameters, obviously needs to be studied.

Another question for you. You showed the area under the curve. How does the panel provide you incremental benefit? Or what is the incremental benefit over the existing, I guess, nomogram or stage-grade age factors? Did you assess for that?

Weranja Ranasinghe: Yes. So we matched the patients for the standard criteria. So that way, this was an added valuation, if that makes sense. We also looked at individual markers of the three markers that we use. And PD-1, interestingly, seems to be the highest prognostic indicator.

But we felt that combination approach would probably be a better way to look at it, because PD-1, and all of these markers do evolve over time. And it also depends on what time the TURBT is done, what specimens are used, and so on. But essentially, I think, it is an incremental additional value using these biomarkers compared to the existing nomograms. So that's where we think it's quite exciting.

Ashish Kamat: Excellent. Congratulations on the work and the presentation. And thanks for spending time with us.

Weranja Ranasinghe: Thanks so much, Dr Kamat.