Hello, and welcome to ESMO 2017 here in Madrid. I’m Will Davies with ecancer and joining us is our blogger superstar Bishal Gyawali. Nice to have you with us again.
It’s my pleasure. Thank you very much for having me again.
Now I suppose we should start with what’s been happening at this year’s conference, and you’ve made an appearance with your own poster.
Yes, what I did was with the PD-1 inhibitors, nivolumab and pembrolizumab, we see atypical patterns of response in terms of longer duration of response for a selected few patients. The Kaplan-Meier curves of some of the nivolumab trials are crossing over which is a very unique thing to observe. So based on these findings I thought what could be the relation between progression free survival and overall survival for PD-1 inhibitors. Until now there have been a lot of studies like that that have studied the correlation between PFS and OS but they usually do it for a particular cancer type because they think the correlation between PFS and OS is a function of the tumour type. But now with PD-1 inhibitors which have different patterns of response, and we talk about the traditional RECIST criteria not being a good tool to measure progression with this type of antibodies, so it was a very logical question to ask in my opinion whether PFS and OS correlate for this type of drug and what I found was that they don’t. There was no relation between overall survival and progression free survival for nivolumab and pembrolizumab trials. In contrast to the targeted therapies such as bevacizumab in which we see a benefit in PFS which does not translate to OS, in this case we observe that there could be a benefit in OS without seeing any benefit in PFS.
So you’re going to have to keep your eyes open well after your trial has possibly failed to see if it’s…
Yes, exactly. We concluded that OS should be the primary endpoint but our reason was a bit different. It’s not because the PFS benefit may not translate to OS but because you could have a benefit in OS without having a benefit in PFS.
That leads on to a discussion I had earlier with one of the doctors about a possible new endpoint, PFS2 – progression free survival after progression free survival, that after your disease relapses and progresses you give them the follow-up treatment, maintenance or second line, and then you see if there is that long-term effect from whatever first line therapy they’ve had. Like you say, there have been some very long-term overall survival benefits not seen with initial PFS and that finding, that second relapse, that second failure as the new endpoint to aim for. Your thoughts?
The first thing we need to understand about progression free survival in itself if it’s a surrogate. We care about progression free survival because we believe it will lead to improved OS or it will lead to improved quality of life for the patient. If it does not do both these functions, if the improved PFS does not lead to improved OS or it does not improve the quality of life then what exactly does it mean? Because we define progression very arbitrarily; in the RECIST criteria we say if the sum of diameters of the tumour size grows by more than 20% it’s progression, 19% it’s not progression, 21% it’s progression. So it’s very arbitrary for the sake of clinical trials; we need to have a cut-off at some point. But for the patient, let’s imagine you are a patient with a tumour. You do not know whether your tumour has grown by 22% or by 18%, you have no idea. We take CT scans, MRI scans, we measure and then we tell you that information. But you, by yourself, you have no idea unless you have symptoms.
So I’ve got this tumour, it’s probably grown, I’m not sure, I can’t tell.
Yes, that’s why for the patient it doesn’t matter, progression free survival in itself doesn’t matter. For the patient what matters is whether they have some symptoms and whether they could get relief from those symptoms or whether they could live longer as a result of taking the drug. So progression free survival, we need to understand that it is just a surrogate, it’s not an endpoint in itself. So PFS2 is a further complicated surrogate. I’m debating whether PFS in itself is a relevant endpoint so I would be very reluctant to accept PFS2 as a surrogate endpoint for OS.
Certainly if there was that translation of PFS to OS then you could smooth things over and maybe get by but were there any disease indications at all where that translated from your investigation or was it across the board no correlation?
Vinay Prasad and colleagues have done a systematic review which was published in JAMA a couple of years ago where they studied all the correlation studies across all the tumour types, most of the tumour types. They have provided those data for correlation coefficients between PFS or disease free survival and OS for many tumour types and I don’t remember off the top of my head but, for example, for colorectal cancer there is a very good relationship between disease free survival and overall survival. So if you have such type of good correlation data then in such cases having a disease free survival or progression free survival as an endpoint is very reasonable because it saves you time and you have good data that it correlates with overall survival.
So why not?
Yes. But when you don’t have that correlation data then it doesn’t make much sense to use these surrogates.
Speaking of data, we should probably get on to some of the data that has been presented at this year’s conference, some of the headlines. Let’s start off with the STAMPEDE trial which we talked about at ASCO earlier this year.
Yes we did.
Let’s see what’s new in STAMPEDE, I guess.
Yes, it was a very interesting study presented in the prostate cancer session. This was not a direct comparison and this was not a proper randomised comparison of the two arms of STAMPEDE but this was what they call an opportunistic comparison. STAMPEDE they have multiple arms that can test against the control simultaneously. They had an arm for zoledronic acid, they had an arm for docetaxel, they had an arm for abiraterone acetate. So what they found was separately they found that the addition of docetaxel also improved OS and, as we found in the last ASCO, they showed that the addition of abiraterone also improved OS. But now we don’t know which of them is the better one because docetaxel also improved OS, abiraterone also improved OS. So they did what they call an opportunistic comparison and they tried to compare the data between the abiraterone acetate arm and the docetaxel arm. The progression free survival and other surrogate endpoints were better with abiraterone acetate but when it comes to overall survival there was no difference between the abiraterone acetate arm and the docetaxel arm. So the conclusion would be that… of course there is a caveat that this is not properly powered to see a survival difference and this is an opportunistic comparison, but this is the best evidence that we have got to make a comparison between either abiraterone or docetaxel. So overall survival is the same and it will all be a matter of patient preference and cost and toxicities, because the toxicity profiles are completely different between abiraterone and docetaxel. So with androgen deprivation therapy what do we add? It would be very appropriate to add docetaxel up front because it gives you the same survival and it’s much cheaper than abiraterone acetate. When you progress then you always have abiraterone to add later. But depending on your setting and depending on the patient’s other insurance system or the patient’s financial status, some patients can prefer abiraterone acetate because of the difference in toxicity profile or you can use docetaxel. It’s completely up to the patient and the physician and the financial status and the availability. So this is a very important study because now you can feel confident in using either of those options and docetaxel is the cheaper option.
And it’s good to at least have the choice rather than be bottlenecked into the one shot you’ve got of maybe having a chance of getting any better.
Yes, it’s virtually a question of using immediate docetaxel or using docetaxel later. Or you could say the same thing for abiraterone.
Then also we heard from the RANGE trial with Dr Daniel Petrylak.
This was a very interesting study but people have covered it as a very positive trial but I’m not very optimistic about it because we don’t have overall survival data, there is a benefit in progression free survival which is very minimal. I don’t remember it but it’s less than 2 months, it’s one point a few months. It is statistically significant, you can’t deny that, but it’s less than 2 months of benefit in PFS – does it actually matter so much to the patient? We also have now other options, we don’t necessarily need to use docetaxel plus ramucirumab in second line urothelial cancer, you can use immunotherapy agents. So I’m not very optimistic, I don’t think this is a practice changing trial.
We heard results earlier today looking at combined immunotherapies compared to sunitinib which is already being eclipsed by cabozantinib in the first place. So it’s a busy space to have a drug available in.
Yes. This is in the setting of renal cell cancer first line. Now we have a lot of options in renal cell cancer compared to what we had ten years ago. So we can’t now even count all the drugs with the fingers of our hand. So this is a very interesting time to be in renal cell cancer, of course, and we have so many drugs. Now, regarding what should be the first line treatment, if you compare the evidence between nivolumab plus ipilimumab, the trial that was presented at ESMO, and the cabozantinib trial then cabozantinib is a phase II trial and this is a phase III trial. So evidence-wise this is stronger and this has shown benefit in overall survival so there is no doubt about it. But one caveat, one particular point of interest for me, is nivolumab has already been approved for second line so we need to know how many patients in the sunitinib arm actually got immunotherapy in the second line. If these patients did not receive immunotherapy in the second line then it is not a real comparison of the real world now because if you give nivo plus ipi up front and the patients in the sunitinib arm if they don’t receive immunotherapy later then that’s a biased comparison which obviously will show very favourable outcomes for the nivo plus ipi arm. Another point I was concerned about was why didn’t we test nivo alone up front because we had the data for the benefit of OS with nivolumab in second line so it would be prudent to take nivo alone in first line where they could do better. But now we have data for nivo plus ipi so we do not know whether nivo alone would do good or nivo followed by ipi would do good. So it would have been much more reasonable to take nivolumab versus sunitinib in the first line because in the melanoma trial we have seen a study published in JAMA Oncology a few weeks ago in which they showed that most of the patients can’t receive the whole course of ipilimumab plus nivolumab because of toxicities. They also showed that it’s actually not essential to give the whole four courses of ipilimumab efficacy-wise. So it would be very interesting to see how nivo alone would perform against sunitinib in first line.
I am certainly growing cautious around CTLA4 molecules generally, certainly after the MYSTIC information that came out a few weeks back in which dabrafenib, which was doing fine by itself, combined with an anti-CTLA4 in tremelimumab did much worse than the single agent, I believe. We’ve heard lots of information from dabrafenib out of PACIFIC this year, we’ll come back to that in just a second. Because the combination of nivolumab and ipilimumab was one of the headlines for melanoma as well, the CheckMate-238 I believe it was. That was presented by Dr Jeffrey Weber, in fact I had quite a conversation with him about the early reports of death for CTLA4. He said that they were greatly exaggerated, that there would still be a space for CTLA4, certainly in second or third line relapsing after resection, after resection, after resection, after resection. But there are lots of data swirling around and it would be nice to pluck a single cohesive answer out of all of it but that’s just not going to happen, I suppose.
That’s really unfortunate, I agree. You talked about CheckMate-238 and there are some very intriguing points about that trial, as I saw Tuesday morning published in The New England Journal of Medicine. They are testing adjuvant nivolumab versus ipilimumab in this trial. Stage 3, stage 4 patients with melanoma where the tumour has been resected and we are testing adjuvant therapy. Last year at ESMO we had another practice changing trial published in The New England which was of adjuvant ipilimumab alone. So today we compared nivolumab versus ipilimumab. But the difference is last year when they published the trial of adjuvant ipilimumab, and it has already received FDA approval, it is used in the dose of 10mg/kg for three years. I was very vocal about it - this is a higher dose than necessary and this is a longer duration than essential. But 10mg/kg for three years was the trial and it got approved in the same way. But in this CheckMate-238 they are comparing now nivolumab against ipilimumab but the dose they are using is 3mg/kg and the duration they are using is for one year.
It’s not a direct comparison trial.
So you can’t have your cake and eat it too. So if you say adjuvant ipilimumab you need 10mg for three years then you stick to it.
Is it worth finding the data from their ipilimumab from this year’s CheckMate and comparing it to the approved dosage of 10mg for three years and seeing if even just the reduced dosage is giving some kind of advantage over what has been approved at this long, heavy duration?
Yes, it would be prudent but we would be making cross-trial comparisons and it doesn’t always give a definitive answer. Even in this case we have data for recurrence free survival but we don’t still have data for overall survival. So now we are talking about giving a drug as expensive and toxic as nivolumab for one year because it increases your recurrence free survival. But we need to know whether it improves the survival or not because when the patient finally relapses you are again going to use… the standard of care right now is first line nivo, pembro or nivo plus ipi depending on where you practice. So how does the scenario, the landscape, change if you give nivo straight away as adjuvant, then what do you do when it relapses? So we need to have overall survival data; if it improves overall survival then this can be the practice changing trial, the standard of care. But if it improves only recurrence free survival but does not impact on overall survival then this might not be practice changing.
You mentioned there the duration of one year of expensive drug in nivolumab and one year versus continuous nivolumab was subject to investigation as well in which disease was it?
In lung cancer.
In lung cancer. Shall we skip ahead to that or shall we talk about the Combi-AD data of BRAF and MEK inhibitors in melanoma first?
Let’s finish melanoma first and then we can move to the duration of nivolumab for lung cancer. Because this trial, which was again published in New England Tuesday morning, about using a BRAF inhibitor plus MEK inhibitor as adjuvant. It has already been approved for metastatic melanoma but now they are testing it in stage 3 melanoma after resection. This was a very interesting study but again I want to talk about before because it has exactly similar issues as that with CheckMate-238. This is again a standard treatment when you relapse so if you use it after relapse and you get the same survival then why use it up front because it will harm your quality of life? Because these are adjuvant patients, patients who don’t have tumour, whose tumour has been resected, and you are asking these people to take a BRAF inhibitor plus MEK inhibitor combination which is toxic in every sense of the word – physically toxic, financially toxic – and you are asking them to take it for one year with the hope of preventing relapse in future. But when it relapses and you take this maybe the patient has more motivation to take the drug or the threshold for toxicities. If he receives the same survival anyway then why do you want him to take it up front. There are a few patients who don’t relapse so you are over-treating them.
I’m sure this would contribute to… there was quite a lot of discussion on the panel between representing doctors, Dr Axel Hauschild and Dr Jeffrey Weber, about CheckMate or Combi-AD having this combination BRAF/MEK or nivolumab, which one when and for who. So I guess we’ll have to see how that OS bears out for either of them, whenever that data comes to fruition.
Yes, and it’s very surprising because you’d think as a patient, as a public, that it’s pretty obvious that all of cancer researchers are trying to improve overall survival, are trying to improve the quality of life for the patients.
Are you well? Are you still alive? Two key issues to every patient.
Yes, but it’s unfortunate that we have allowed the evidence base for our treatment to erode gradually and gradually so that we are accepting very low evidence levels for us to change our practice. I’m not sure if the patients are actually happy about it, maybe this is a very interesting area to actually study with the public and patients, whether you are happy with this type of lower thresholds for accepting treatment or whether you actually want to… If you take a drug which is toxic and expensive whether you actually want it to be a very good drug or whether you’re happy to gamble with a lower quality evidence with this drug.
Find a patient and ask, ‘Now, we’re going to give you a drug which is almost statistically significant in improving your partial complete response rate. How do you feel about that?’
I’m sure the mind of every patient hearing that information would be too full of, ‘What does that mean for me and my life?’ to talk confidently.
And you’ll also get nausea, vomiting, diarrhoea.
Oh good, oh that too, yes. It’s a very delicate conversation to have.
Yes, I think we would definitely need a study in that.