Ashish Kamat: Hello everybody and welcome once again to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, urological oncologist in Houston, Texas, and it's a pleasure to welcome once again to our forum Dr. Roberto Contieri, who has been with us several times before. Welcome Roberto.

Roberto Contieri: Hi. Thanks for the invitation, and thanks also to UroToday for having me here.

Ashish Kamat: So you've talked about different things in the past. Today, we've invited you over to talk about your publication, this project that you put together on the enhanced prognostic value of a four-tier hybrid grading system in Ta non-muscle-invasive bladder cancer. It's a relatively controversial topic because many people are trying to make things simpler, and in some ways this does do that, but in some ways it makes it more complicated. So love to hear your presentation on your publication and then we'll have a quick discussion at the end. So take it away.

Roberto Contieri: Okay, thank you very much. My disclosure for this presentation is that I'm not a pathologist, I'm a urologist, so I will discuss the topic from a clinical point of view. However, when we talk about tumor grading in urothelial carcinoma, we have to know that it is based on the architectural and cytological features of the neoplastic urothelium. And of course, the grade reflects the aggressiveness of the tumor, and has prognostic implications. However, we have to understand that grading represents a biological continuum. Therefore, arbitrary cutoffs are necessary to categorize tumors in a way that is clinically interpretable. We know that the first grading system was introduced by WHO in 1973, which categorized patients in G1, G2, and G3 tumors. And then the ISUP/WHO Consortium proposed the two-tier system in 1999, which divided patients into low-grade and high-grade tumors.

And this system was adopted in 2004 by WHO. In recent years, interest has grown in hybrid systems that combine features of both grading classifications and can be structured in three-tier or four-tier systems. Of course, each system has some pros and cons. For example, the three-tier system, the G1, G2, G3, but also the new three-tier system which combines the two systems has a risk of middle category dumping. And also we have to consider that the more cutoffs we have in a system, the less reproducible the system becomes. When selecting the best grading system, we need to balance the clinical usability and practicality of the grading system, ensuring it provides simple and actionable risk stratification with the informative value of the data, which should adequately support oncologic prognostication and also most importantly, patient counseling. If we look at international guidelines, there is a huge difference between EAU and AUA guidelines. For EAU, of course, WHO 2004 provides a slightly better classification, but the 1973 system is a stronger prognosticator for progression compared to the WHO 2004.

However, they also state that three-tier and four-tier combinations prove to be superior to either classification system alone. So they recommend using both systems or a hybrid system. AUA guidelines are easier, don't even mention the old system, and just state to use the 2004. Just stratify the patients between low-grade and high-grade tumors. But my interest in this grading system arose when I worked on this publication. We recently published it in European Urology Oncology, which is completely on another topic, which is about follow-up of intermediate risk patients. But this issue of the EAU guidelines, which suggest using two systems, creates a problem when we define intermediate risk patients because, as you can see from the figure, the composition of the intermediate risk group is completely different if we use the WHO 1973 grading system or the WHO 2004 system. That's why the grading system we use has an important clinical impact when we stratify the patients. And it will also decide the follow-up, but also the treatment of the patients. So that's why we performed a collection of data from 18 Italian hospitals.

We included in this study, recently published in BJU International, more than 1,200 patients with Ta non-muscle-invasive bladder cancer tumors. But why Ta? Because in my opinion, different grading systems have a higher impact when deciding the sites for follow-up or treatment of Ta because if we also include T1, we know that the majority of T1 tumors are high-grade, and of course those are with the highest risk of progression. So I think that if we want to evaluate the grading systems, Ta patients are the best population. We stratified the patients according to the WHO 1973 grading system, the WHO 2016/2022 classification, but also according to two hybrid grading systems. The primary endpoints were the prognostic accuracy of the grading systems for high-grade recurrence-free survival and progression-free survival. We used Harrell's concordance index, the C-index for the accuracy. You can see the characteristics of the patients in the table on the right of the screen.

And we had a 26-month median follow-up, which I know is not a long follow-up when talking about non-muscle-invasive bladder cancer. However, we found that 418 patients developed recurrence during the follow-up, including 184 patients with high-grade recurrence and only 44 patients developed progression. If we look at the tables, it shows the C-index for the grading systems according to the oncological outcomes. And you can see that consistently the hybrid four-tier system has a higher concordance in predicting the oncological outcomes compared to the other systems. However, of course, we also have to look at the curves, the Kaplan-Meier curves, and I want you to notice for example that in the three-tier system on the bottom left of the screen, when you look at the difference in high-grade recurrence between G2 and G3 patients, there is no difference, the curves barely overlap. The curves overlap in the Kaplan-Meier. However, when we shift to progression, there is a significant difference between G2 and G3 patients. So we also have to understand what we want to know from the grading system.

If it is prognostic value, we have to decide, am I interested in high-grade recurrence, which is of course of interest in low-grade patients, but also in high-grade patients who undergo BCG because a high-grade recurrence will completely change their treatment. But of course, we have to understand that progression to muscle-invasive bladder cancer has a huge impact on the treatment and also on the risk of the patient. So I agree that this is very controversial, but also because it's not an easy topic of course, and any stratification leads to losing some information. So in conclusion, hybrid grading systems offer reproducibility and improved prognostic accuracy. I also want to raise some issues with the grading systems. As I showed before, the EAU 2021 risk stratification is validated separately on WHO 1973 and the WHO 2004 grading systems, but it's not validated on the hybrid grading systems. So this is a space for future research.

And also, current EAU guidelines for BCG-failure patients rely only on the high-grade patients, I mean only on the WHO 2004 system, and it includes only indications for patients who develop high-grade recurrence with no difference between G2 and G3 recurrence. I think that nowadays that we have some different conservative treatment options for patients who fail BCG and also many will come, I think that the stratification of the recurrence might be interesting to offer the best treatment to the right patients. Of course, integration of molecular biomarkers and advanced imaging can further refine the predictive accuracy and lead to personalized management. So I want to thank you for listening and I'm open to discussion.

Ashish Kamat: So Roberto, thank you for that presentation. And I like the way at the beginning you made your disclaimer that you're not a pathologist, you're a urologist. So that means I can't really ask you some of the hard questions I wanted to ask you. I'm joking, because I am going to still ask you those questions. Right? So one of the things that we, and everybody, has championed for many years is to make things as simple as necessary, but not too simple. Right? But the reason to make things as simple as possible is because we want to be able to replicate data across the world, and we don't necessarily want to handicap patients just because they might be in centers where they don't have expert pathologists. And that's where the low-grade high-grade dichotomy came up because it's easier for many pathologists in the community that may not be specializing in bladder cancer to just categorize a patient low-grade, high-grade. And it tends to give enough prognostic information for the patient that you can actually take care of the patient well. When you are proposing this four-tier system, which I understand is much more personalized, how would you then address the disparity in care that would occur for patients that don't have access to such specialized pathologists?

Roberto Contieri: So this is of course a great question. This is the main issue with the grading systems. But I think that this is also for staging, because grading is only a part of the pathological report, of course. And there was an interesting report from San Rafael Hospital, and they are a tertiary referral center, and they found that they do a lot of second opinions and in a very high rate of patients there was a restaging of the same bladder sample. So this is a big issue in bladder cancer, I think. We need expert pathologists, and of course, not all patients can refer to a referral center of course, but there's been also a recent publication that actually was a survey which proposed the same slides to different pathologists across the world, European, American pathologists, Australian pathologists, and the ability of pathologists to use the WHO 1973 system was comparable among the pathologists who daily use the system compared to the American ones. So I think it can be used. The problem with the WHO 1973 system is that the categories are not as well-defined as for the WHO 2004. So I think this is a pathological perspective, but of course an improvement would be the better definition of the categories.

Ashish Kamat: Roberto, you dodged my question a little bit. Right? Because you answered it well, but the question was how would you make these patients have access to this sort of tier system in places where they don't have expert pathologists. We can't have all the patients come to say MD Anderson or San Rafael. So I think that's one of the practical considerations. Have you considered providing this in a form of machine learning or AI? And again, I'm not saying that you didn't publish this, but I'm just saying this is something that might help mitigate those issues. Right? The second question I had for you is that there has been a preponderance of data. For example, the EAU prior to 2020, all the data, data from clinical trials. There's a lot of current data that looks at low-grade and high-grade and doesn't have the three-tier or the four-tier system. How would you take all that data that's been generated in prospective studies and prospective phase three studies and then retrospectively apply that to the four-tier system? Any thoughts there?

Roberto Contieri: Well, we are at the moment when decisions have been made for prospective years, of course, for other cancers, we are in a phase where we already have some interesting results that will change the disease in the future years. But still there are lots of issues. So I think that if we apply the hybrid system now, the effect will, we are in time to apply for the future treatments. I understand that a lot of work has been done with the two categories, which are okay, but I think that when you use a hybrid system, you still have the difference of low-grade and high-grade. So you can still use the two-tier system for clinical implications. Also the BCG unresponsive definition, but also all the data that have been generated on this definition can be applied even if we use the three-tier system. Also, because one of the proposed three-tier systems includes low-grade and then only stratifies high-grade into G2 and G3. So I think that it can be used also for that category.

Ashish Kamat: Yeah, and I agree with you. I was just putting you on the spot on purpose because I said I would. Right? I think work like this that you published emphasizes how bladder cancer is a complicated disease and it's not just low-grade, high-grade. It's not just grade 1, 2, 3. I think from a practical perspective, yes, we should have patients categorized as low-grade and high-grade, but for example, like we do at MD Anderson, we then also report is it low-grade grade 1, 2, is it high-grade, grade 2, 3? So having that information available in front of us really helps me counsel my patients.

And you've been here and that's what we do. And then, having data such as yours is also important because it helps us as clinicians talk to our pathologists and say, "It's important that you report to us everything you see. If you see low-grade, high-grade, tell us the grade. But within that, if you're seeing that this is a high-grade tumor, but there's more grade 2, less grade 3, tell us. If it's low-grade, there's less grade 1, more grade 2, or vice versa, tell us." And then, we, as clinicians, can use all this information and in some ways use the four-tier system. So again, Roberto, thank you very much. Very important topic. Really glad that you could join us again. Thanks.

Roberto Contieri: Thank you very much. It was a pleasure for me.