PG: Hello, I’m Petros Grivas. I’m a medical oncologist at the Seattle Cancer Care Alliance, I’m an Associate Professor at the University of Washington and associate member of the Fred Hutchinson Cancer Research Center. I’m very excited today to sit down here with Michiel Van der Heijden, a very prominent medical oncologist in the Netherlands Cancer Institute. Michiel, very nice to see you.

MVH: Thank you, Petros. Nice to be here.

PG: It’s fantastic to talk at ESMO 2019, exciting data in the field of bladder cancer biomarkers. I was really intrigued yesterday to see your very impressive study in the neoadjuvant setting of bladder cancer regarding a combination of anti-CTLA4 and anti-PD-1. Could you tell us a little bit more about that and pathological complete response rates and biomarker work that you have been doing?

MVH: Yes, absolutely. This is a study we are highly excited about. What we are doing here is taking what we think is the most active combination of immune therapy – anti-CTLA4 with anti-PD-1, in this case ipilimumab and nivolumab, into the neoadjuvant space. Of course this combination is associated with also some toxicity so we decided to weigh the risk-benefit ratio a bit better to go for the most aggressive and risky stage which is stage 3 cancer. So we did not include T2 tumours but we did include node positive patients. What we did is we gave them a sequenced schedule of ipilimumab and nivolumab. This was based on work done in melanoma that this was better tolerated and still had the same types of response rates. What we found is that, first of all, it was possible to have surgery on all of these patients which is absolutely not a given, as you know, in bladder cancer.

PG: Of course.

MVH: We set the primary endpoint at surgery before twelve weeks and all but one patient made that primary endpoint and one patient was operated four weeks later and had a complete remission. So we got all of that done. If we look at our secondary endpoint which is complete remission we saw a complete pathological remission in 46% of patients and, moreover, in a few patients we saw only a very small foci of CIS, so a non-invasive cancer, or a Ta which is also a non-invasive cancer. When you combine that, so no invasive cancer at pathology, we had 58% that had that.

For biomarkers this is all very new data, the last patient had their surgery in May. We were able already to look at PD-L1 at baseline and there we find quite a good association with response. PD-L1 enriches for response in this group of patients. We’re going to look at a whole set of other biomarkers so we’ll look at tumour mutation burden, RNA sequencing, expression patterns, those kinds of things, and that will hopefully come out soon, in a few months I hope.

PG: That’s very, very exciting. So just this brought to my mind the ongoing clinical trials in metastatic disease that utilise the same combination, like CheckMate-901, they are using ipilimumab at 3mg/kg with nivolumab 1mg/kg as well, so a similar dose. Also previous dose from Dr Rosenberg looking at the same combination, ipi/nivo, in platinum refractory disease. Also in that study there was a 38% overall response rate in platinum refractory disease with ipi/nivo and the response rate was up to 58% in PD-L1 positive patients. What’s your opinion about the PD-L1 role in urothelial cancer? Do you use it? Do you think there is a practical clinical utility or it needs more data?

MVH: Of course the biomarker we should only use if it’s really prospectively validated and the studies that you’re referring to are still dose finding studies and exploratory studies. But there is, of course, a lot of data for PD-L1 and it is a bit conflicting in the metastatic setting. So, for example, for the atezolizumab trials, the IMvigor trials, PD-L1 was validated and tested in the phase I and phase II and they nicely optimised the biomarker and it seemed to associate with response. However, in a randomised trial there was basically no discriminatory power of this biomarker and it was prognostic. Strangely, in the pembrolizumab trial which uses another PD-L1 assay, of course, the biomarker was also mostly prognostic but in the other direction. To make the confusion even complete, in the first line setting, the IMvigor 210 trial, there was no association with response of the biomarker but still we now have a label for PD-L1 positive patients in the first line, cis ineligible population, based on data that is from ongoing studies and that we actually don’t know about, that is undisclosed so far. So at this point it’s pretty difficult to say what to make of all this data and we’ll need to see the randomised data in the first line setting to see what the outcomes exactly are in the different PD-L1 subgroups. But it seemed to have some association with response, how to use this in practice is still a bit difficult and we definitely should look at other biomarkers as well.

PG: It’s very exciting, I agree fully with you, and there is an urgent need for PD-L1 validation and also other biomarkers. You alluded about TMB, tumour mutational burden, gene expression profiling and you plan to do RNA-Seq in your own study. Of course mutations in DNA repair genes, there is some data from Memorial-Sloan Kettering, Dr Teo and colleagues, looking at mutations in DNA damage response genes, DNA repair. Correlating those mutations, especially the deleterious mutations, to response to immunotherapy, any comment on that? Are you going to look at those mutations in your study and do you see a future using those mutations?

MVH: There have been several studies looking especially at chemotherapy response. We had our own study with ERBB2 mutations; there has been a study which combined from whole exome sequencing three gene sets and looked at that. Then there’s ERCC2 and that later transformed into a DNA damage response pathway which I think is the right direction. We have to keep in mind that perhaps for the ERCC2 and the DDR genes these are the larger set of DDR genes. These have been tested also in a new series for the combination of RB1, FANCC and ATM as well as the ERBB2 data has not been tested in a large series although our own ERBB2 data also seemed to fit in other people’s data from the ERCC2 study. All in all, to make a long story short, I think this is not ready for prime time yet but it is very exciting as is, for example, all the work on molecular subtypes.

PG: I fully agree with you and I think there are at least three clinical trials in the neoadjuvant setting trying to prospectively validate, to your point, the clinical role of DNA repair gene mutations, ERCC2, ATM, RB1, FANCC. I don’t know if they evaluate HER2 based on your data. There are trials from Dr Plimack, Dr Iyer, Dr Galsky in the United States looking at the neoadjuvant setting, whether patients with those DNA repair gene mutations could have a pathological complete response and if there is any potential future of keeping the bladder in those patients. But, for now, this is all experimental, as you mentioned, in its validation. Do you have any comment about that?

MVH: Absolutely. I think they are taking the right path forward and testing it in prospective trials and that is definitely what needs to be done.

PG: I fully agree with you. Do you see a future with a kind of multi-omics approach, putting PD-L1, TMB, gene expression profiling all together kind of thing in an algorithm? I know you had a recent very nice paper about the immunogram that you suggested in urothelial cancer, do you see this being a future pathway?

MVH: Yes, absolutely. I think that’s where we as a field should be moving towards so we’re now really focussing on PD-L1 and it has to be tested in pretty much every trial that we do. But if you look at all the data then it seems that there are several factors that determine anti-cancer immune response. One of them is, of course, checkpoint inhibition or the hotness of a tumour, inflammatory pathways. Tumour mutation burden is another important one but there’s also molecular subtypes and it seems that there are also inhibitory signalling pathways in the microenvironment like TGF-β and EMT. All of these factors if we can find a way to assess all these parameters we should move towards a situation where we can put them into a quantitative model and hopefully in the end come up with the best therapy for that patient. We indeed tried to put that forward into a theoretical framework called urothelial cancer immunogram which aims to get a more holistic view of the anti-cancer immune response. So that’s something that we are very excited about but a lot of work still needs to be done to get there.

PG: I fully agree with you and I really enjoyed reading your paper about that. It’s very important to have this conceptual framework and try to have a quantitative metric going forward. But I agree fully, it’s very difficult to validate that in a prospective study. If I can just move a little bit our discussion about targeted therapies in urothelial cancer, in the United States we have now accelerated approval by the FDA of erdafitinib, an FGF receptor inhibitor for tumours, metastatic bladder cancer, that have FGF receptor 2 or 3 activating mutation or fusion. Any comment about that approval? Any potential future of this compound or other FGF receptor inhibitors, there are many, even outside the US? And any comment on the role of mutations and fusions and potentially I don’t know if other amplifications or something else?

MVH: I think is a highly exciting development. This is something that has been in the air for quite a long time as FGFR mutations are clearly one of the drivers in a subset of bladder cancer. Only for a long time the inhibitors were not active enough and also patient selection was also not always optimal. In this case erdafitinib has taken really the right approach where they focussed on the genetic abnormalities which are known to usually be the drivers of these types of signalling and push that forward in very good way. There’s no reason why other inhibitors wouldn’t be able to do that if the inhibitor is active enough and they take the right biomarker approach. But for now erdafitinib seems to be the most advanced of those.

PG: And you touched upon this question that I have always regarding the optimal biomarker for selection of patients for FGF receptor inhibition. Do you think that mutations or fusions is the way to go or overexpression? Any thoughts about that?

MVH: Generally with other tumour types and targeted therapy it’s the genetic abnormalities that can be targeted and one should get into the zone of signalling pathways or of overexpression. Then it becomes much more difficult and you can imagine that because it’s much easier for a cancer cell to adjust their transcriptional programme than to change a genetic driver. So I think genetic drivers are, in this case, the way to move forward. There are other drivers as well in bladder cancer that can be tested, other receptor tyrosine kinases. We have seen a little bit of data on that but there is a lot of work still to be done.

So I would like to also ask you a question now. Since you’re practising in the US and you have in this space more opportunities than we have in Europe at the moment. There are multiple drugs now in this setting, in the third line setting after chemotherapy and IO, so what is your feeling? Because you also have the option of enfortumab vedotin is probably coming soon and you don’t need to test any biomarker for that. We, as scientists, of course like much more the approach of testing a biomarker. What is your view on this?

PG: That’s a great question and actually we are all waiting to see what will happen with a regulatory review with the data with enfortumab vedotin, an antibody-drug conjugate against Nectin-4 which is expressed in urothelial cancer cells. The data looks very promising – overall response rate in the phase II study exceeds 40%. So many of us believe that this agent might be approved in the future, possibly as accelerated approval because there is an ongoing phase III trial comparing enfortumab vedotin to chemotherapy. There is definitely space for this compound, probably in the third line setting as you mentioned. So far there is no biomarker to select patients because Nectin-4 is expressed probably by most, if not all, urothelial cancers. So there is some variability in the expression of Nectin-4 but has not been used in the clinical trials for the selection of patients. They go with an all comer approach.

There is another antibody-drug conjugate called sacituzumab govitecan which is against TROP-2 and there was some updated data here at ESMO 2019 showing overall response rates approaching 30% in patients with multiple prior lines of therapy. So in the future I can envision a landscape where both of those antibody-drug conjugates could potentially have a role and, who knows, one maybe after the other because there’s a different mechanism of action, different antibody, different linker and different toxin payload being carried. Of course the big question is if patients have mutations, for example FGF receptor, what is the efficacy of those antibody-drug conjugates and we do not know yet. This is something that will be done hopefully in the future. But I think the antibody-drug conjugates will definitely have a future.

Also we have HER2, anti-HER2 antibody-drug conjugates. We showed data at ASCO and also we have studies right now of those anti-HER2 antibody-drug conjugates which reminds me of your work in HER2 mutations down the road. But it seems that for the antibody-drug conjugates against HER2 expression alone might be enough just to carry the toxin to the cells. It’s a little bit different than trying to inhibit HER2 that has mutations or amplification with more targeted therapies like afatinib or trastuzumab or pertuzumab. So it’s very exciting to see this evolving in the future. I think we’re going to experience an influx of hopefully many new agents in the next few years.

MVH: Yes, absolutely. The landscape has changed tremendously in the last few years and it’s very exciting times for bladder cancer.

PG: I agree with you. The drug development with all these agents, targeted therapies, immunotherapies, antibody-drug conjugates and many others, along with the development of biomarkers is super-exciting. I think the challenge we have is, to your point, to do the validation of biomarkers prospectively in clinical trials, how to use them in clinical practice in the optimal way. Through collaborations and clinical studies hopefully we will be able to do that. It’s exciting to be here so thank you so much for watching us and hopefully we’ll have more data in the future. Thank you for being here.

MVH: Thank you, Petros.