A biomarker led approach to bladder cancer
Prof Andrea Necchi – The Vita-Salute San Raffaele University, Milan, Italy
Dr Guillermo De Velasco – Comunidad de Madrid, Madrid, Spain
Prof Arlene Siefker-Radtke – MD Anderson Cancer Center, Houston, USA
Dr Yohann Loriot – Gustave Roussy, Paris, France
AN: Hello, welcome to the ecancer.tv round table discussion. I’m Professor Andrea Necchi, I’m a medical oncologist from San Raffaele University in Milan, Italy, and director of the GU medical oncology department in San Raffaele Hospital in Milan, Italy. I am pleased to be joined today by an esteemed faculty of friends and colleagues to discuss the issues around the role of FGF receptor targeting in bladder cancer and a biomarker-led approach to patients with urothelial cancer. Arlene, please.
ASR: I’m Arlene Siefker-Radtke, a professor of genitourinary medical oncology at the University of Texas MD Anderson Cancer Center where I do all bladder cancer all the time. It is a pleasure to be here with my esteemed colleagues discussing biomarker work in bladder tumours.
YL: Hello Andrea, I’m Dr Loriot, I’m part of the Department of Cancer Medicine and Department of consult medicine at the Instutut Gustave Roussy in Paris, France. I’m leading the bladder cancer programme.
GDV: Hello everyone, I’m Guillermo De Velasco, I’m a medical oncologist based at University Hospital 12 de Octubre where I work with GU tumours as well.
AN: Thank you, thank you friends. So let’s start first from the data, from an important study, the THOR study, that investigated oral erdafitinib in biomarker-selected patients with advanced urothelial cancer. So, first of all, I would like to ask you, Yohann, to briefly summarise the design of the THOR study and to enter into the details of the main findings from cohort 1 of the study and the subgroup analysis that you presented at this meeting.
YL: Yes, so basically in THOR there are two different cohorts, cohort 1 and cohort 2. Cohort 1 are more patients with metastatic urothelial carcinoma that received one or two lines of prior treatment that included immune checkpoint inhibitors. So these patients were randomised into two groups, erdafitinib versus chemotherapy of choice – it could be vinflunine or docetaxel. The primary endpoint was overall survival and we reported a few months ago, at ASCO actually, that the overall survival was better for the patients treated with erdafitinib versus chemotherapy. The hazard ratio, just to remind, was 0.64 and the median overall survival was 12.1 months versus only 7.8 months for patients treated with chemotherapy. So the objective response rate and PFS were improved for patients treated with erdafitinib.
Basically, there’s another cohort, cohort 2, that Arlene will present this data. Here at ESMO we presented a subgroup analysis. Why? Because there was some question from the primary presentation regarding the difference that we can have on different specific subgroups. What we showed is that basically the OS benefit was observed in all the subgroups that we included in this analysis, so, for example, age, gender, primary location – upper tract versus bladder cancer. But also what is important is the type of alterations because of course there are a lot of different alterations, fusions, different types of mutation. What we showed is that basically all the patients, regardless of the mutation or the fusion they could have, benefit from erdafitinib versus chemotherapy. So it was quite important to look at. There was some analysis on the type of metastasis. As you know, some patients could have liver mets, lung mets and bone mets. What we showed again is that the benefit from erdafitinib was seen regardless of the location of the metastasis, including the patients with bone mets or liver mets, for example.
AN: And did you find any major difference with regards to the primary location of the tumour, the bladder tumour versus upper tract tumour?
YL: If you look at this Forest plot, it seems that patients with upper tract disease benefit even more from erdafitinib. But obviously it’s a subgroup analysis so the study was not powered to address this specific question. The numbers of patients in each subgroup is not the same. So it’s merely hypothesis generating; it’s important to look at, of course, but this kind of analysis, as you know, is difficult to integrate in general.
AN: And what about your experience also besides the clinical trial data? Do you have, for example, with regards to the safety profile of erdafitinib, do you have a sense of any difference of any clinical factors that may be impacting on the safety profile of the drug?
YL: So you mean if there is any predictive factor for toxicity?
AN: Yes, or certain characteristics that may predict poor tolerability of the treatment.
YL: In my experience not so many. It’s difficult to identify the patients who will develop toxicity. Of course, if you respect the inclusion criteria of course the patients with a long history of diabetes you have to check the eye, the macula, before giving erdafitinib to these patients. But otherwise there is no reported very strong factor that predicts a grade 3 or grade 4 toxicity in these patients.
AN: Thank you. So, Arlene, what about cohort 2 of the study? We were very much eager to see and to hear from you the results of the head-to-head comparison of erdafitinib versus pembrolizumab. What did you find?
ASR: Sure, well the goal of the cohort 2 portion of the trial was to gain an understanding of the biology of the FGF altered tumours. If you look at FGFR3 alterations, they appear enriched in an immunological cold subtype. We often call these the luminal 1 subtype. Early data with immune checkpoint inhibitors suggested that immunologically cold tumours had a lower degree of benefit with the use of immunotherapy. So we attempted to explore the impact of sequencing – would it make sense to give erdafitinib before a patient received treatment with an immune checkpoint inhibitor. So this trial randomised patients who had preselected FGF alterations, prior chemotherapy but no prior immune checkpoint inhibition. Patients were randomised in a 1:1 ratio and they were fairly well balanced between both groups. What we saw, based on overall survival which was the primary endpoint of this clinical trial, was that there was no significant difference in overall survival comparing erdafitinib versus pembrolizumab. The survival results appeared similar. If you look at the survival curve, the early portion of the survival curve did show potential benefit for erdafitinib which we have seen with other chemotherapy when compared to pembrolizumab. But when you look at the tail of the curve, immune checkpoint inhibitors have been associated with a durability of response that has not been achieved with cytoreductive therapies alone. It really was that tail of the curve that contributed most to the hazard ratio. We also see this in the objective response rate with a 40% objective response rate with erdafitinib and only a 21% objective response rate with pembrolizumab. So we see twice the responders but because immunotherapy has a durability, it really was the durability in those who responded where we saw the impact on survival.
AN: Thank you. So how do you envision the use of… or is there any role for PD-L1 expression as a biomarker actually for predicting a response to immune therapy? Because on the other side we have the activity that is still there for pembrolizumab in these immunologically cold patients. So would you envision any further steps in the future for applicability of PD-L1 expression testing in urothelial cancer?
ASR: I think it’s true, the outcomes that we observed were similar to what was observed in non-FGF altered cohorts. So it wasn’t a good means of detecting a group who would benefit from one versus the other. One of the challenges with PD-L1 expression levels is it’s a very dynamic marker. It can increase in the setting of chemotherapy, it can increase from infections, prior intravesical BCG. One might almost predict if you sneeze your PD-L1 levels will go up at least a little bit, that is a joke of course. So PD-L1 really is not a good biomarker and that’s why it has really not been useful in the treatment of urothelial cancer patients or in the ability to select patients who would or would not benefit from an immune checkpoint inhibitor.
AN: That’s a great point. Please remember that in Europe there is still an approval for adjuvant nivolumab in high risk patients after cystectomy in PD-L1 selected patients. So there is still an issue around the role of testing of PD-L1. What about the testing for FGF3 receptor genomic alterations? Based on the cohort 2 results from the THOR study, when do you think that it’s the right time to test the tumour or the patients for FGF receptor alterations?
ASR: In the US we’re able to test for mutations and typically have it covered by insurance programmes for any patient with a metastatic urothelial tumour, especially since we have accelerated approval of erdafitinib which targets FGF alterations. I personally test patients the moment they come to my clinic and they have metastases because with the current standards of care we’re still not curing the majority of patients. That way I have the mutation in hand so if their treatment is no longer working we know what their FGF status is and can plan the next line of treatment.
AN: That’s a great point. I think that it’s a very important point for daily practice so I would like to briefly hear also from Guillermo and Yohann what do you think about the right timing for FGF receptor testing. Guillermo?
GDV: Yes, I would agree that the best moment to get the information, probably not only of the FGFR but any alterations that the patient may have, is from the very beginning since they’ve got metastatic disease. I guess maybe in some cases already with high risk in the context of localised disease it maybe had an impact but definitely for all those patients with metastatic disease we should get all the molecular information that we can get.
YL: Yes, I agree of course and we have to remind that for most patients FGFR3 fusion and mutation are early events. So you can find these mutations on the cystectomy sample or nephrectomy sample. So when you see the patient with the metastasis first line it’s time to do the test actually. So thereby doing the test on the tissue or maybe it’s more convenient on ctDNA. It could be another way to do the test, the turnaround time is very short now – 12 days in my institution – so it’s very convenient for the patient and for the physician as well.
AN: Yes, the timing is very important, also because of the data that emerged from the THOR-2 study. The THOR-2 study was a study that was designed to test erdafitinib versus investigator choice of intravesical chemotherapy in patients with papillary high risk, not BCG unresponsive, disease. So, quite briefly Guillermo, do you think that oral erdafitinib or any systemic therapy, targeted therapy, in this patient population with very early disease will have a space in the near future?
GDV: So the sooner we go, the more careful we should be about how we are treating our patients. We know that all systemic therapy comes with a price. We know that erdafitinib is a drug with manageable effects but it will have some side effects. So we need to make sure that because we are starting to see that we can provide some specific intravesical treatments that may be much more effective with lower adverse events. So maybe the balance is quite important in those cases. So I guess it may have an impact but I think we need to start to select better the patients and maybe we will have a later conversation about which specific mutations may need a systemic therapy or which ones may need only an intravesical treatment.
AN: That’s a great point. Arlene, so the 77% relapse free survival rate reported with THOR-2 in this population. I suspect, or my feeling was, that it was not perceived and achieved in the urological community as an outstanding result based on the fact that as a counterpart we have side effects to manage that may be difficult to manage by the urological community basically. What do you think?
ASR: I think that is the challenge. When we’re giving erdafitinib to stage 4 disease these are people essentially with a death sentence. So they are willing to accept more toxicity compared to a patient with earlier stage disease where removal of the bladder can help them live comfortable lives without hand foot syndrome, dry mouth and the potential change in vision. So side effects become less acceptable in earlier stage disease. I would personally argue if you had a way of giving erdafitinib without having that systemic toxicity we’d see much more excitement about that high recurrence free survival rate. To be able to control tumour in the bladder helps patients retain their bladders for much longer and patients really want to live with the bladders they were born with.
AN: Yes, and we will cover this issue, this important topic, soon after. The next point that I would like to raise, the question that I have to you, Guillermo, is about extending and broadening the landscape of FGF receptor alterations, so looking at the different co-occurring mutations that we usually find together with FGF receptor alterations. Is there any data interesting that emerged from this meeting?
GDV: There is really interesting data coming especially from the RAGNAR trial. So that was a pan-tumour study where all basically tumours with FGFR alterations may be included. So now what has been published is that there could be a specific mutation associated with FGFR, especially p53, PI3K alterations. What is really important, in my opinion, is that this is the beginning. So we are looking at what is this pathway. If you think about other tumour types where we have a really good oncogene, we’ve been learning for many years. So now we are saying, ‘Okay, we have this group of patients being treated with erdafitinib, for example, but there are small cohorts where we have only this main driver, FGFR.’ We’ve learned that when you have FGFR mutation plus p53, usually those cases have a worse prognosis. So now probably it’s about to start to basically describe better in which cases with FGFR alterations we may have even longer benefit. Because when you look at the data and you go, for example, for the THOR trial, you see maybe some patients with really long durable responses. So you see the median duration of response is not as good but if you look at the tail there are some patients that really get a lot of benefit. I guess for those extreme cases we should start to think, ‘Okay, maybe it’s not about only the FGFR mutation, it could be one of those specific FGFR mutations plus which other genomic alterations may have those cases.’
AN: That’s great. And, very briefly, what do you think about the potential for testing urine for DNA alterations in these patients.
GDV: Liquid biopsy is the future. So we are talking all the time in any solid tumour about liquid biopsy. The good thing in bladder cancer is we have the urine. So there are some studies also presented, small studies with less than 100 patients, but quite relevant in terms of how good they are to test genomic alterations. So these studies are already able to capture differences between non-muscle invasive bladder cancer. They are also showing the difference with muscle invasive bladder cancer, even though there are a few patients with follow-up samples, they were able to try to predict what is going to be the recurrence based on the urine tumour DNA. I guess this is just the tip of the iceberg but we will see probably in the future will be in those liquid biopsies to basically follow patients with specific FGFR alterations.
AN: Thank you. And we also heard from this meeting important data emerging from the novel use of intravesical therapies in patients in particular with non-muscle invasive high risk disease. We had updated data from the SunRISe-1 study. The SunRISe-1 study is evaluating TAR-200 which is a newer way of delivering in a continuous way intravesical gemcitabine through an osmotic release with a silicone tube, a small silicone tube, that is put into the bladder. The SunRISe-1 study in particular is testing TAR-200 as monotherapy or in combination with immune checkpoint inhibitors in patients with BCG unresponsive disease. But the complete response rate with an updated cohort of all the patients confirmed initial data presented at the EAU this year with 7.7% complete responses. In particular, complete responses followed a very, very rigid definition of response that was including essential evaluation of urine and biopsy and at two different timepoints. So very, very robust definitions compared to other similar studies conducted in this field. We also have data from the TAR-210 which is the same way of delivering not standard chemotherapy like gemcitabine with TAR-200 but, for the first time, an intravesical targeted therapy like erdafitinib. So, Arlene, very briefly, what do you think about the initial data presented from the TAR-210 study?
ASR: First of all, what a wonderful way to develop erdafitinib and avoid that systemic toxicity. Essentially it’s a pretzel but when they insert it it’s kind of linear. It looks like a line but as they remove it from the catheter it curls into the shape of a pretzel which opens these pores and allows elution of drug into the urine. As a result of this elution you can get sustained drug levels over a long period of time. So you avoid the systemic toxicity, you maintain constant drug levels over a long period of time without the need for frequent cystoscopies and drug installations. So, first of all, what a benefit for the patients. What they saw was that we do see evidence of clinical activity. We see evidence of that response and benefit and patients who are not developing any of the systemic toxicities associated with oral erdafitinib. So I would argue these pretzels are just a wonderful way to go in getting drug directly to the bladder. And as we understand more about biomarkers or other pathways associated with resistance we can develop combination pretzel strategies targeting whatever a person needs in their bladder and help them maintain their bladders for much longer, maybe even put most of us out of business as medical oncologists, which would be a wonderful thing.
YL: What is intriguing is that in this specific setting you can also monitor the response with the urine ctDNA and they need to give the TAR-210 if there is, of course, mutation and then to stop if there is no mutation anymore and then to rechallenge the patient with mutation or cure. So it’s a new way to treat the patient, it will be very fantastic actually.
AN: Yes, I’m sure there are intricacies between the way we test for FGF receptor and then the material that we use for FGF receptor testing and the treatment that we use for these patients. So, an exciting time overall for bladder cancer patients and an exciting ESMO meeting, a bladder ESMO meeting for the first time, finally, for this disease. I would like to thank my esteemed colleagues and friends for joining me in this round table. I would like to thank ecancer.tv and I would like to thank you all for your attention. Thank you.
