Innovative Design and Real-World Validation: The STAMPEDE Trial for Prostate Cancer - Noel Clarke

January 20, 2023

Noel Clarke delivers a comprehensive discussion about the STAMPEDE trial model, focusing on its innovative approach to prostate cancer treatment in the UK. He elaborates on the multi-arm, multi-stage (MAMS) trial design, emphasizing its flexibility and efficiency, and the collaboration with large data sets like the National Prostate Cancer Audit. Dr. Clarke highlights significant results, such as the combination of abiraterone, ADT, and radiotherapy reducing prostate cancer death, and a previously unrecognized survival benefit in using docetaxel. He also explores translational research's role, including genomic exploration, and addresses the potential for fractures in patients receiving ADT, suggesting early bone protective doses. Dr. Clarke presents the future of STAMPEDE, including the launch of STAMPEDE 2, focusing on metastatic disease, and STAMPEDE International to extend the protocol globally. The presentation concludes with fresh insights into docetaxel's efficacy and an invitation to participate in the ongoing effort.


Noel Clarke, MBBS, FRCS, ChM, Professor of Urological Oncology, Salford Royal Hospital & The Christie NHS Foundation Trust, Manchester, UK

Read the Full Video Transcript

Noel Clarke: And it's a great pleasure to be here. I'm going to talk a little bit about the STAMPEDE trial model, and I've been involved with STAMPEDE for many years now. And what we have tried to do with STAMPEDE is engage the UK's healthcare system to effectively work with the grain of what happens in the country, building with a novel trial design, STAMPEDE, and also more recently, engaging with large scale data sets such as from the National Prostate Cancer Audit. I'm the joint lead for this for England. We collect information on between 40,000 and 50,000 new cases every year and have done since 2013, so we can match that with hospital episode statistics. And I'll talk a little bit about that in a minute or two.

But first, let's go back. The rationale for the model of STAMPEDE was to try to improve outcomes for prostate cancer. We had limited options both for the M0 and for the M1 patients. And we had an idea for a flexible trial design, which largely came out of the MRC. Here's the gang of six that designed it. Max Parmar, who's head of the MRC Clinical Trials Unit, and his colleague, had this idea for a multi-arm, multi-stage trial. You can see the group, that's Nick James, Malcolm Mason, David Dearnaley a clinical oncologist, myself, Matt Sydes, and Max, and now we have a team of about 40 working at the MRC, so it's a big outfit.

Now the principle is this, that with a multi-arm, multi-stage trial, it was very different from a conventional phase two, phase three type thing. Now this model, which you see on your right, usually takes about 15 years to yield a result and there are some drugs which fail at this point. The MAMS trial was to set everything off at the same time and have a common control and to have different phases of the trial, phase 1, 2, 3, and 4, and ultimately, when a drug dropped out because it was either too toxic or ineffective, then we could put another drug in. And you'll be familiar with the agents, which we had. Not all the drug companies were able to give us their drugs, so there were some drugs which we simply weren't able to access. And we went with what had evidence for efficacy in the clinic and/or the lab. And these were our five, with a standard of care.

Now the first thing I'm going to show you is a graph of how long it took us to get up to 140 centers, just about. And you can see that took about five years. The advantage of the MAMS trial is when one agent drops out and we rebadge another agent and put it straight in, is that the trial infrastructure machinery is all there. For different arms, you can see when we started to add arms, this was the effect when looking at recruitment. You can see here is how long it took us to get the first tranche in, but then we were up and running with the next tranche in three months and those are the next two arms of the trial after that. We're very quickly able to reengage with 140 centers and keep the trial running, something which we've done successfully.

Now the other thing that you have to consider when running a trial over a long period of time is that things change and the standard of care in the control arm changed and has continued to change as the trial has evolved. In 2015, we were able to consolidate the data on docetaxel, that became a standard of care, so the control arm had to change. And again, when we published on abiraterone here in 2017, again the standard of care had to change. And that was helped by collaboration with large international groups. Many of you will be familiar with this graph, which is from the ICECaP Collaboration, which STAMPEDE and the MRC were integral with the Boston team and others around the world.

What that shows very clearly here is the linear correlation between metastasis free survival and absolute survival. And again, that was hugely helpful in rebadging the M0 component of the STAMPEDE Trial because we could then look at these two arms, which was the abiraterone and enzalutamide arm. And rather than wait for survival, which is what our original plan was, we went for a meta-analysis of the combined trial with metastasis free survival.

And of course that resulted in the publication in December 2021 online, subsequently in print in '22 of the M0 study of the combination of abiraterone, ADT and radiotherapy, which roughly halves the rate of prostate cancer death and has changed the standard of care yet again.

Now, it was always our principle to have a major translational element to the STAMPEDE study and there are two phases of this. As always, these things get planned in a bar somewhere. This was at ASCO in Chicago when Malcolm, myself, Nick and Gert Attard worked out how to run STRATOSPHERE. And we've added to that, the biomedical imaging group or big team. We have two major translational research projects amongst others. We have a number of other collaborations. One is the biomedical research group, which is coordinated through UCL in London and headed by Gert Attard. We're all contributing to this, but Gert leads the team, and there's a biomedical imaging group, which I lead at the Christie in Manchester, where we've managed to collect about 24,000 scans from the STAMPEDE patients. And we put that together to define a variety of elements and I'll show you a bit about that in a minute.

This is how The STRATOSPHERE tissue acquisition works. All the tissue from the various centers is brought into a central tissue bank, which is actually in Wales and then transferred to London, where it goes through a process of evaluation, quality control. And you can see, the combination of people here who are working with this, Dan Berney is the senior pathologist that we've got in addition to Gert Attard and team, a variety of research fellows of considerable talent, I have to say.

And to give you an example of the sort of stuff that comes out, this is a Genome Medicine paper in 2022. We have papers which will follow from this and papers which preceded it as well, which looking at copy number alteration, which is factored for tumor volume. If you concentrate on this graph up here, it's a somewhat complicated thing, but you can see we've got M0s, M1s for high and low volume. And what we can see is the differential response to treatment and the fact that the copy number alteration, even with pelvic lymph nodes is exactly the same as metastatic disease. And so that is helping us to unpick some of the genomic aspects of the disease, its sub-stratification and how it responds to different types of treatment.

The BIG Group challenged the existing dogma that metastatic disease couldn't be treated by local therapy. It started really with this notion, we published this paper in 2014, which looked at the way in which high volume and low volume behaved in castrate resistance. And so we decided to work on this with STAMPEDE and two very bright research fellows in collaboration with the radiology department at The Christie. And this is how it works. The scans come in through a central repository, they're assessed, they're allocated to treatment arms. The data is cross correlated with all the treatment therapies that the patients have had through the MRC clinical trials unit, and we work them out into low and high volume, low and high volume for the treatment and control arms. And of course that has worked out, as I mentioned this morning, that the bone scan actually is a direct predictive surrogate marker for response. This is the follow-on paper from this piece of work, which is largely from Adnan Ali, looking at the bone metastasis number 1, 2, 3, 4, 5.

And you can see here that this is a continuous variable. It doesn't just go naught to four, and then you've had it after that. You do get responses all the way up. But the biggest response, there's the equivalence line, is about here, three to four metastases. If you look at progression free, it's a little longer. The standard of care in M1 disease for low burden as defined simply by counting the number of metastases on a bone scan, is radiotherapy to the primary site. We published this in 2018, the follow-on paper with a longer term follow up at eight and a half years, we published this year in PLOS ONE Medicine.

Now when looking at Burden, no one takes account of the lymph nodes and this was a glaring omission in our view. We've pulled in all the CT scans, we've quantified the nodal burden, we've quantified the location, we've looked at differential response to abiraterone and docetaxel and rather to our surprise, this is what we found, that actually if you have bone and nodal disease, then docetaxel doesn't seem to work. But if you've got bone only disease, then it really is quite effective. And so the notion that docetaxel is dead, is really not true.

Now another thing that's important in a trial like STAMPEDE is the harnessing of the power of the meta-analysis team. And these people you see here, are all members of the MRC Clinical Trials Unit meta-analysis team, which undertook and Ashish and co. will probably remember the MRC bladder meta-analysis team, that was these people, did the same work. And what they do is pull together the data from trials and they pull together individual patient data and that's an important distinction.

This is work which is about to come out. You've not seen this yet, but the MRC meta-analysis team put together all the data from the different trials in M1 looking at metachronous and synchronous prostate cancer. And you can see here this is progression-free survival by disease volume at randomization. Clearly there's an effect on disease volume, but actually when you look at whether it's metachronous or synchronous, that metachronous is very clearly a different disease to synchronous disease.

Now taking that a stage further, we can also look at other things which present something of a surprise. And I draw your attention to this, and again, this is going to be in the same paper, but this looks at the clinical stage and this has not yet being done. This is the clinical stage of patients receiving docetaxel. You've got low volume cT1-3, T4, high volume, and look at this, these are patients who've had docetaxel. Now this is previously unrecognized, and we all know that prostate cancer produces in some circumstances, really bulky tumors in the pelvis and nobody pays any attention to that, because it's a systemic disease. But actually, the use of docetaxel in this setting, produces a really huge survival benefit. And I think we may need to revisit exactly how we manage these patients.

Now in these final few slides, I'm just going to show how you can link the national data sets which are available in your countries, to the trial data sets for real world validation. And we've been doing this with STAMPEDE, with data sets like the National Prostate Cancer Audit in England and Wales. And you can see the kind of stuff that we get information from. We have very accurate hospital episode statistics. Every time a patient comes into hospital for a procedure, it's documented by a code and we can access that code. We get data from the national radiotherapy. Every machine in the UK is linked to a central database and we know what patients have had, what their radiation has been, and we can put all this together and link it to treatments within trials like STAMPEDE.

Now I'll show you this series of slides which gives you a feel for the power that you can get from this. We published this in Cancer Epidemiology. This is on fractures in patients receiving ADT. Now, this is not just radiological fractures, these are fractures requiring hospital admission. These are genuine fractures, joint replacements and so on. And you can see, don't try and read this table, but these are all codes which equate to a procedure. And you can see the numbers that we work with, 152,000 boils down to about 80,000, so we've got low intermediate risk here, 40,000, 25,000 locally advanced, and just under 14,000 M1s. And you can look then at the fracture rate out at five years. And it really is quite striking, what happens to a patient with M1 disease presenting in England, which is, that the fracture rate at five years is approaching 30% and even in M0s, it's not insignificant when you scale that up for the numbers involved.

Of course, then what relevance is this to STAMPEDE? Well, we can back titrate this, and this is work which is going to be published early next year, where we pulled in the fracture rate in M0s and M1s in the bisphosphonate arms of the STAMPEDE study. And once again, you can see what we've done, which is we've taken the STAMPEDE data, we've linked it to the hospital episode statistics data, and we've got really quite substantial numbers which will show what the fracture rate in M0 and in M1 disease is. And just off the record, I can tell you that if you don't give a bisphosphonate early, a bone protective dose, the fracture rate is roughly double what it would otherwise have been, so it comes down to about 15% from 30% at five years.

Now finally, this is where STAMPEDE is going. We've closed off STAMPEDE 1 now. We have to complete and analyze the data from metformin and estrogen patches with a lot of translational work to do. But STAMPEDE 2 is now going to focus on this approach. We will take men with metastatic disease only, no M0s this time, and they'll be split into patients who are low volume according to conventional imaging and high volume according to conventional imaging, PSMA will be allowed, but it will not interfere with the randomization. And those with low volume will be randomized to standard of care or standard of care plus SABR, up to five metastases. In the high volume, these will be randomized into arms A and P, which is basically PSMA-based lutetium and PARP inhibition in BRCA mutated patients.

We're fully funded, everything's ready to go. We'll start this element of the trial probably in January 2023, and it will probably run for another 15 years and I hope will yield the kind of information that STAMPEDE 1 has yielded. We're looking also at the development of the translation aspects, centralization of images, centralization of tissue, etc. And we're also looking at STAMPEDE International. This is going to be a new departure, and if you consider it like a McDonald's franchise, so there's a brand which is STAMPEDE, there's a central data repository, which is the MRC Clinical Trials Unit, and there's a protocol.

And that protocol can then be applied in individual countries, whether it's in North America, Europe, Australia, wherever it is in the world, the individual country will finance their own trial. The data will be processed to the MRCCTU. And I hope then, that we can look at other aspects like surgery, the true role of PSMA, many other things, that we can't do in the UK that certain other countries can do. We would invite you to consider joining in this enterprise as and when we get it up and running. Thank you.