Sam Chang: Hi. I'm Sam Chang, and I am lucky to be surrounded by greatness today. We have Dr. Boris Gershman and Dr. Sia Daneshmand, who have actually utilized real‑world data based upon the Cysview blue light endoscopy, PHOTO registry. I'm not sure of the exact name.

But looking at real‑world patient population and, Boris, work that you helped lead, helped develop a new kind of risk stratification system, tell us more about the abstract that you presented at the AUA 2025 in Las Vegas.

Boris Gershman: Of course. Well, first, thanks so much for having us here today. It's a pleasure to chat with you.

The genesis for this was really there's a need for better risk stratification, risk prediction in terms of recurrence and progression risk for patients with non‑muscle‑invasive bladder cancer. And that's to help tailor surveillance strategies, counsel patients in a risk‑adaptive therapeutic management.

Existing tools are old. They're over 20 years old and may not apply to patients treated with blue light cystoscopy, which has been shown to reduce recurrence rates.

And so we utilized the multi‑institutional blue light cystoscopy registry, which included about 1,100 patients in this cohort, to develop a new predictive model. First, we did some Cox regression, univariable and multivariable Cox regression, to evaluate the associations of a whole variety of clinical, patient, disease, pathologic features with—

Sam Chang: And those are all included in the registry.

Siamak Daneshmand: Yep.

Sam Chang: So the amount of data you all collect—obviously, when you set up something like this, you don't want to miss anything important, but you don't want to make it so onerous that people don't fill out the data to provide all the variables. So the fact that you had such a wealthy treasure of data, really kudos to you, Sia, and all that.

Siamak Daneshmand: It's been a collaborative effort, obviously, with many centers. And if I may just interject just a little bit about the registry, first of all, it's been revamped just quite recently actually. And it has to. As we develop more tools and new biomarkers and new treatment methods, we need to alter the data fields.

And it has become a very powerful database. We now have excellent follow‑up on the patients. There's additional sites, additional funding to get the follow‑up, which is absolutely critical in developing these kinds of risk models.

In the past, it was just about diagnosis and increased detection rates. Now we've moved past that. We detect more tumors with this technology. We need to know the recurrence and progression rates. And so these new variables that are being captured in the blue light registry—

Sam Chang: Well, it gives you information around that. And then the work, like Boris, in terms of, now we've got this data, let's see how we can help predict both recurrence and progression. I'm sorry, I interrupted.

Boris Gershman: Oh, no, please. Yeah, I completely agree. The data set is unique. I think it's real‑world data, prospectively enrolled patients, rigorous collection of what's really important.

And so we were able to evaluate a variety of clinical and pathologic features that reflect things like patient characteristics, smoking history, presence of hydronephrosis preoperatively, as well as a variety of tumor characteristics, some of which were included in prior risk stratification models, like tumor size, multifocality, stage, grade, et cetera, and then some newer ones.

And basically, we first did a Cox regression model to evaluate the associations of these features with the outcomes of recurrence and progression.

And then we performed a LASSO‑penalized regression to build a predictive model that would then be used to help inform patients, with the ultimate goal of being able to tell a patient there in clinic, if you put the values for the predictor variables in the model in a calculator online, to say, your risk of recurrence at 1, 3, and 5 years is x, y, and z, your risk of progression is x, y, and z.

And this is relevant to blue light patients, contemporary treatment paradigms. And therefore, we can potentially counsel you, because patients always want to know, what's my risk, Doc? Improve the way we personalize efficient surveillance strategies. Maybe I need to do it so frequently, or maybe we do. And then also risk‑adaptive management with all the new intravesical therapies that are coming about, both for high‑risk and intermediate‑risk.

Sam Chang: Anything that we can help offer to patients that is evidentially supported really gives patients some confidence regarding, this next step I'm taking makes sense. This next step I'm considering, it weighs in the fact that I've got this chance of the disease coming back or getting worse, et cetera. So I think incredibly important.

As you presented this data, and if people have discussed this with you, what are you going to do? Are you going to use it online? Are you going to have an app? Tell me the actual operational integration of this into your daily clinical workload?

Boris Gershman: It's a great question. So I think future directions—we haven't done this yet, but ideally it would be developed on a website that can be clinician‑facing and even patient‑facing. The value of having a granular model, where you just enter patient or disease characteristics and then it gives you the predictive estimates, is that you don't lose any of the precision and accuracy that you would if you truncated the parameter estimates for risk tables.

So that's one option. I think that's the easiest thing to do, and probably the next thing we're going to try to do. Certainly, I can see this being valuable as an app as well, but I think a website calculator would be probably the most easy to operationalize.

Sam Chang: And, Sia, as this registry gets larger, more robust, longer follow‑up, et cetera, you can only see how continued updating, different modeling, et cetera, will provide even more information to patients over time.

Siamak Daneshmand: Absolutely. We have 4,000 patients in the registry right now.

Sam Chang: 4,000.

Siamak Daneshmand: 4,000. This will become, I think, the world's largest contemporary registry for non‑muscle‑invasive bladder cancer going forward. And as I said, the concentration right now is on the follow‑up for the patients, which is absolutely critical.

But this is modern use. So we're going to see prospective use of—or I guess, retrospective use of gemcitabine, docetaxel, for instance, within the registry, within a blue light registry. Some of the novel therapies, some of the novel biomarkers that are real‑world data. So super exciting, honestly, to see it develop from the very beginnings.

Sam Chang: Oh, absolutely.

Siamak Daneshmand: Registry to this massive tool.

Sam Chang: I said in the past, huge kudos for your personal effort and perseverance because this is a huge lift. And the fact that now you can say 4,000‑plus, that number is only going to increase. And it really provides, like I said, incredibly important information in the real world.

Obviously clinical trials really are essential for our therapeutics, our diagnostics. But going forward, having the impact of, hey, this is what we've done in clinical practice in multiple centers. We've gathered all this data. This is what actually is helpful. This is what we've actually found. And then work like—[Inaudible]—so you've led in terms of, hey, this is how we integrate this data for helping our patients determine what's my risk of this, what's my risk of that, I think are incredibly important. So kudos to both of you. Thank you very much. We look forward to future presentations as well.

Siamak Daneshmand: Thank you, Sam.

Boris Gershman: Thank you so much.