Read the Full Video Transcript

Alicia Morgans: Hi, my name is Alicia Morgans and I'm a GU Medical Oncologist and Associate Professor of Medicine at Northwestern University in Chicago in the United States. I'm so excited to have here with me today, Dr. Ewan Gibb, who is a Senior Scientist and the Bladder Cancer Lead at Decipher Biosciences in Vancouver, in Canada and Dr. Andrea Necchi, who is an Associate Professor and Medical Oncologist, as well as being the Director of GU Medical Oncology at San Raffaele in Milan, Italy. Thank you both for being here with me today.

Ewan Gibb: Thank you very much.

Andrea Necchi: Thank you, Alicia, and thank you, UroToday colleagues, for inviting us.

Alicia Morgans: Wonderful. So I wanted to speak with the two of you about a really exciting publication that has recently come out in European Urology, looking at actually molecular characterization of patient samples in patients who have been treated with neoadjuvant pembrolizumab, as well as untreated patients, to understand potential molecular signatures that we can use to direct therapy and treatment choice in the future. Andrea, can you tell us a little bit about how you collected these samples? Who were the patients included in your assessment?

Andrea Necchi: Yeah, thank you, Alicia, and welcome everyone. So this study actually pursued a collaboration between my center and Decipher colleagues, and Ewan Gibb in particular, regarding the biomarker analysis from the PURE-01 study. The PURE-01 study was a phase II study of pembrolizumab given neoadjuvant before radical cystectomy for patients with muscle-invasive bladder cancer. We have already published with Ewan Gibb and Decipher colleagues, prior to biomarker analysis, that was aimed to evaluate the association between molecular subtypes assessed and immune gene signature assessed on the TUR pre-therapy specimen. We have already reported in prior European Urology papers that, for example, patients having a tumor with a high level of immune gene signature scores, or named people with a tumor classified as a claudin-low tumor or basal-type tumor, were more likely to respond completely and to maintain the response to neoadjuvant pembro after cystectomy.

Now we look at a different thing. Now we look at the changes that occurred while on treatment and we started comparing the tissues from radical cystectomy post-therapy and the TUR specimen before therapy, in order to see what were the changes that occurred, induced by the treatment. And in doing so, and in order to frame the data into the context, we relied upon prior work that Decipher colleagues, and Ewan Gibb in particular, have done in previous papers by comparing the outcomes with those from a larger neoadjuvant chemotherapy cohort, from a multicenter study and another cohort with the samples from radical cystectomies from untreated patients, so patients who did not receive any neoadjuvant therapy.

So before entering the biomarker part, it is important to say that first, we provided the evidence that there may be changes between cystectomy and TUR, so between post and pre-therapy samples regarding biomarkers. This is important in order to see whether we can envision postoperative strategies, of course, long-term and for one of the first times, we evaluated the molecular subtyping on the post-therapy samples on radical cystectomy samples after prior treatment. And it's important to realize that we see no 100% overlapping between the subtypes between the radical cystectomy and TUR because there were some changes, some shifting subtypes. A proportion of patients with a shifting molecular subtype, according to treatment, and the shifting subtype should be considered by physicians, by investigators when dealing with personalized perioperative therapy.

And then we observe, going more in-depth on the biomarker data, I will leave that word to Ewan Gibb, to explain what were the major findings that were striking for me, and may orientate future research in the field.

So please Ewan, I leave the word to you regarding biomarker data.

Ewan Gibb: Thank you very much, Andrea. Yeah, so to reiterate, we had a cohort of 26 patients, we had pre and post-treated samples for pembrolizumab, so that was the PURE-01, and we compared these cases to a cohort of 133 cases that had cisplatin-based neoadjuvant chemotherapy. So we had pre and post for that as well and then we had a cohort of 94 patients who had radical cystectomy samples only. So we profiled all these cases on an array-based platform, which measures about 46,000 different gene points and generated molecular subtypes to begin the study.

So we saw initially, in pre and post in both the NAC setting and the pembrolizumab setting, that there was a shift in the subtype, they really were not consistent between the TURBT and between the RC cases. And so initially, we started to look at what is called a clustering algorithm. So what this does is basically puts the patients together in groups that are related to each other. So we noticed initially in the neoadjuvant chemotherapy setting after treatment, and so post-cisplatin-based RC samples, and this was published in Clinical Cancer Research a few years ago with Dr. Seiler, is that we found four different unique molecular subtypes.

So these are somewhat consistent with the known molecular subtypes that we talk about with the consensus model and such, but we thought it was, [inaudible 00:06:35] there is like a basal-like group, so this is very consistent with basal tumors, high keratin 5, keratin 14, immune filtration. There is a luminal-like group, so very differentiated, high uroplakin 2, PPAR gamma, things like this. And then we saw two unique groups, one was an immune-like subtype we called it, this is highly infiltrated, so a lot of immune activity, high PD-L1 staining, high CTLA staining. And then another unique group, which we called scar-like, we named it scar-like in part because when you look at it, it is relative to true scar tissue from the former tumor bed of pT0 patients, they are very consistent, the pattern gene expression, patterns are very consistent.

And so that sort of sets the background for what we saw in PURE-01. So in the PURE-01 study, actually, it was quite unique in that we identified a luminal-like group, a scar-like group, which has unique characters, which we can get into in a few minutes, and a very, very small basal-like group, which is somewhat basal I'd say, not necessarily as basal as one might look at in say the neoadjuvant chemo setting. So that is basically the start of the findings. We did not see an immune-like group in the post-pembrolizumab setting, which was quite intriguing.

Alicia Morgans: That is intriguing, especially since obviously you're trying to increase immune activity with that approach to therapy. So Andrea, what do you make of that? What do you make of these subtypes? Why we might see what we see and how those subtypes relate to the outcomes that I know you collected on these patients as well?

Andrea Necchi: Yeah, thank you, Alicia. I think that we put the basis, along with other researchers worldwide, to the development of a standard that is highly needed, that has to do with a personalized perioperative approach.

With our work, actually, we highlighted the fact that we should consider changing the treatment and considering sequential treatment, personalized sequential treatment, based on the changes that may occur in the tumor, and maybe partly induced by the tumor, maybe partly induced by the treatment, and then may occur as a resistance mechanism within the tumor. It's striking to note, for example, that an extraordinary opportunity that has come from ... It's clear from our analysis, and we provided a clear picture at the end of the manuscript, showing that, for example, in luminal tumors of which the residual disease post-pembro is enriched, there is a clear high expression of targets like NECTIN4 and Trop2, and the same data has been already provided by other authors in the prior meetings at ASCO GU, there were many posters from other colleagues, from other groups showing similar findings. So luminal-type tumors having pretty high, or higher than the other subtypes, expression of these genes and potential targets that represent the targets of ADC, meaning that we may envision a future of sequential, rather than combination, therapy of IO and ADC before and after surgery, for example.

And in the same subtype, luminal subtype, we have the well-known target of FGF receptor 3, and so also FGF receptor inhibitors given sequentially in a resistant tumor post-immunotherapy may be a good option. And there is still a huge option to provide a maintenance period of immunotherapy in patients with inflamed tumors, mainly after chemotherapy. So chemo, IO sequence may be a good sequence to test in future trials in selected patients.

So in my view, the main message coming out from our study is to try to invest in study designs that actually provide the patients with sequential therapy that is personalized based on the changes in the subtype.

Alicia Morgans: That is such a good message and so important. And I guess I would ask Ewan, as you are the person who would add the biologic data or the molecular analysis to a study like this.  Would you envision having treatment for a period of time, perhaps a biopsy post-immunotherapy or checkpoint therapy, for example, get the biopsy, do the molecular analysis, and then have a subsequent therapy, either pre-cystectomy or post-cystectomy that is based on that analysis, the TURBT sample that could happen prior to further treatment.  Is that something that might even be feasible? And if it was, how long would it take to get the reading or the report back from something like a Decipher test in the interim?

Ewan Gibb: Yeah, absolutely, that's exactly what I was thinking. So I think, as Andrea pointed out, with the sequential treatment, we need two different profiles because we are really not clear on how a tumor may change with treatment. There is consistency in what happens, but just not how it gets there. So if we profiled a patient from a radical cystectomy sample, for example, we need, of course, another single sample classifier to correctly put the patient to one of these subtypes, these new subtypes. But yeah, that is definitely an opportunity to inform adjuvant treatment decisions.

Alicia Morgans: That's fantastic. So I guess, Andrea, you have kind of already given us your bottom line and your statement, but could you just, one more time, reiterate, what would your message be? What would your summary be? The take-home message from this work, from your perspective as the clinician?

Andrea Necchi: Well, my perspective is that aside from the developments that may occur, that could be envisioned based on this biomarker data that we have, for example, the biomarker data that we have provided, we should be able to raise the bar and move towards studies that aim, not only to provide a cure for the patients but to provide a bladder in tact cure [inaudible 00:13:27]. This is the next and almost final generation studies that we should be aiming to and I think that moving on towards these studies, and we will see in the next ASCO meeting, the early data from this type of study, we should invest strongly in the biomarker developments in this perioperative space. And collaborators like Decipher, like Ewan and other colleagues, and other companies are the key to try to reach this goal, that is the major goal, the practice-changing goal for these patients, the true practice-changing endpoint for the patients, providing them with a strategy that may avoid, ideally, any type of radical local therapy, including cystectomy or radiotherapy and curing the disease with systemic therapy alone. It is in my view, the major goal in this disease.

Alicia Morgans: I think that's a wonderful goal and I hope we get there. And, Ewan, as we wrap up from your end, how would you say that investigators and clinicians can best partner? It seems like the future really is a partnership between molecular teams, like yours, or teams that assess molecular characteristics to really ensure personalization and hopefully eradication of the disease. How can we best partner?

Ewan Gibb: These collaborations with individuals like Andrea are, I think, instrumental to moving this field forward. I mean, we can't do this work on our own from the clinical perspective and we provide a lot of support for the bioinformatics genomics side. And I think really what we are doing now is we are intersecting biology and clinical characteristics of the tumors, and really best positioned to really inform what would be the best treatment for the patient. I mean, these subtypes do have consistent biology even pre and post-treatment and I think we can definitely leverage that to best inform what therapies would help the patient the most.

Alicia Morgans: That's a great message and a great message, I think, to all of us as we care for patients with bladder cancer and try to, as Andrea said, really move the bar and ensure that we get the best outcomes for our patients with the right therapies, which we are one step closer perhaps to being able to achieve.

So thank you both so much for your efforts, for your work, and for your time today in talking this through, I appreciate it.

Andrea Necchi: It was a pleasure. Thank you, Alicia and UroToday colleagues. Thank you.

Ewan Gibb: Thank you kindly.