This site is optimised for modern browsers. For the best experience, please use Google Chrome, Mozilla Firefox, or Microsoft Edge.

Case study: STAMPEDE - a turning point in trial design

The NIHR research infrastructure, embedded in the NHS, has been supporting the delivery of novel, complex and innovative clinical trials for over a decade.

Find out more

Clinical need

STAMPEDE (Systemic Therapy in Advancing or Metastatic Prostate Cancer: Evaluation of Drug Efficacy) is a landmark clinical trial and arguably the most cited trial in the history of “complex innovative design” trials in the UK. It is viewed by many as a milestone in the development of trial design, and with good reason. It opened for recruitment in 2005 and 16 years later recruitment is ongoing having recruited over 11,500 participants (albeit paused for periods during the pandemic). Most importantly, it has produced three practice-changing results and improved life expectancy for patients with advanced prostate cancer.

All of this was achieved using a Multi-Arm Multi-Stage (MAMS) adaptive platform design which was ground-breaking at the time the study opened. But why was this pioneering design developed?

Firstly, there was a clinical need. According to the Cancer Research UK website, there are around 48,500 new prostate cancer cases in the UK every year, that's more than 130 every day. Back in 2004, the standard-of-care for prostate cancer, hormone therapy, hadn’t been updated for several decades and median survival was estimated to be around three or four years for metastatic prostate cancer. Secondly, a number of new treatment approaches for cancer were then emerging which various pharmaceutical companies were interested in developing. Although having several promising treatments in the pipeline is a good position to be in, it can also pose problems when it comes to running clinical trials. Max Parmar is Professor of Medical Statistics and Epidemiology and Director of the Medical Research Council (MRC) Clinical Trials Unit (CTU) at University College London (UCL). He summarised the issue:

“If you have a number of drugs that look promising in early phase trials in a particular disease, when you come to set up larger, phase three trials for each of those drugs, these trials will be competing with each other. Each trial will need its own control arm, requiring its own patients, as well as the treatment arm. So the trials will slow each other down. What we needed was a practical approach - a way to speed up the evaluation of multiple new potential therapies, and improve success rates in identifying the most effective ones. We also needed a platform for adding new therapies to be tested as they emerge rather than setting up completely new trials”

Efficient design

To tackle this issue Professor Max Parmar came up with the idea of the MAMS design. At the time Professor Parmar was also an Associate Director of the National Cancer Research Network (NCRN) when it was established in 2001. He held this role for over 10 years, at which point the NCRN had already begun to evolve into what eventually became the NIHR Clinical Research Network. He explained how the MAMS design works:

“It is an evolution of the multi-arm Randomised Controlled Trial (RCT). Traditional RCTs have two arms which compare a new treatment (arm one) to either a placebo or the current standard of care (arm two). Multi-arm RCTs can compare several new treatments simultaneously against the control arm.”

Professor Matt Sydes is a Professor of Clinical Trials and Methodology at MRC Clinical Trials Unit at UCL and was involved in implementing STAMPEDE. He added:

“MAMS trials go one step further by building flexibility into the study protocol through the introduction points of pre-planned review, which we call interim analysis. These happen throughout the trial and the findings from the interim analysis are used to modify the protocol and therefore course of the trial. What was different with these interim analyses was the focus on “lack-of-benefit” - asking research arms to show some evidence that they were worthy of continued study. This allows a trial to adapt by moving away quickly from treatments that are unlikely to improve outcomes. Furthermore, the design permits the inclusion for study of new treatments that become worthy of testing during the trial. Each of these features provokes the phrase adaptive design.”

In conjunction with Nick James, David Dearnaley and Noel Clarke, the original design of STAMPEDE was drawn up. There was strong clinical support for the MAMS design. The trial went through a number of iterations before the final agents that went into the trial in 2005 were chosen, but the basic principle was never in doubt. Professor Parmar explained how this worked in STAMPEDE:

“For example, in the case of STAMPEDE, the trial opened with six arms: five research arms and one control arm. This used three different drugs: two recently licenced for prostate cancer (a chemotherapy drug and a bone protecting agent) and one re-purposing question (an anti-inflammatory drug with potential cancer modifying properties). The drugs were tested both alone and in two combinations, making five experimental arms in all.

“Recruitment was closed early in 2011 to two arms that did not yield sufficiently promising results, at their second analysis point. The other three research arms continued accruing patients.

In 2015, the first practice changing results from the trial were published in The Lancet showing that adding the chemotherapy drug, docetaxel, improves survival for this patient group. In 2017, the second practice changing results were published in the New England Journal of Medicine from the first added research arm showed that abiraterone improves survival for this group. The third set of results in 2018, from the trial’s second added comparison, published in The Lancet, showed that treating the prostate with radiotherapy alongside standard treatment substantially improves survival for a subgroup of men with advanced prostate cancer. These three sets of results led to a change in practice across the NHS and worldwide and also to the control arm of the STAMPEDE trial, as the control arm will always be the best standard of care available.”

“In this way adaptive trials are seamless: we do not need to stop and start a new trial for each new question being answered. To date the design has allowed STAMPEDE to deliver 11 randomised comparisons. No other trial has delivered that kind of data in trials with such prognoses. In addition, if these comparisons had been conducted as two-armed trials, it would have taken in excess of 40 years to achieve the same outcomes.”

As a way of quickly and effectively discovering which treatments work best, it’s a no brainer. For the pharmaceutical companies developing the new treatments there are also many advantages to be gained by bringing new drugs to the platform. To explain further Professor Nicholas James, Chief Investigator, zoomed in on the first added comparison of the STAMPEDE trial:

“One of the key innovations we made with STAMPEDE was to realise that ‘multi-arm’ did not have to mean all the arms started at the same time. As a clinician, it became clear that we were going to want to add new comparisons to the original platform. After a couple of false starts, abiraterone emerged as an excellent candidate to test this proposal. To our knowledge no other trial group had done anything on this scale. We started talking to Janssen in 2010 about having access to the drug, abiraterone, for research. We got an agreement in principle that year and the whole set up process was completed in a year so we started recruitment the same time as the drug was licenced for relapsed disease in late 2011.

“A French-led European consortium got the go ahead at the same time in 2010. We took two years, until early 2014, to recruit just under 2000 men for our 'abiraterone comparison'. It took the French just under two years to get their first patient in, because they had to set up the trial from scratch, the French trial has just reported early results from their trial at ASCO 2021, four years after STAMPEDE.

“Or you could look at it another way. It took us five years to have our first 80 centres recruit at least one patient. By the time we had the abiraterone trial ready to go we had 120 centres open. It took five months for 80 centres to recruit at least one patient.

“As a way of setting up a trial it is much quicker... Investigators and NHS trusts seem to like it because they don’t have to stop and start new trials, and the trial funders and sponsors like it because we can provide an answer to the research question much, much quicker than we would by setting up the trial in the conventional way.”

But it is not just the sponsors and research teams who benefit. The wider patient population benefits because, if a trial can generate results earlier, better treatments can be made available via the NHS quicker. Additionally, the design of the trial means that patients taking part are statistically more likely to be randomised to receive a new treatment. As patients generally enter trials with the hope of getting a new treatment, a multi-arm trial, with a higher chance of this happening, is attractive compared to a two-arm study.

Hanna McEvoy is Head of Clinical Research at Cancer Research UK, which is a key funder of STAMPEDE. She agrees that the trial marks a turning point in trial design:

“STAMPEDE is a real jewel in the crown of Cancer Research UK’s clinical research portfolio. The trial has delivered monumental findings that have changed clinical practice for men worldwide. It was the first MAMS trial Cancer Research UK funded and has really paved the way in efficient trial design, saving both time and money in the development of new treatments for patients. We are now seeing the efficiencies learned from the STAMPEDE design being implemented right across our clinical trials portfolio, and beyond.”

 

NIHR support

The trial is ongoing, despite being paused during the pandemic, and to date 127 sites have contributed to STAMPEDE. Royal Preston Hospital is one of two hospitals in Lancashire Teaching Hospitals NHS Foundation Trust. It remains one of the highest recruiting sites in the UK for STAMPEDE having enrolled more than 300 participants. Paul Brown, Head of Research and Innovation at the Trust, reflects on how the study was delivered locally and the support it received from the NIHR Clinical Research Network:

“I believe Royal Preston Hospital is the fourth highest recruiting single site in the UK. We began recruiting back in 2008. We were extremely keen to offer the trial to our patients to be able to give them access to cutting edge treatments. The design of STAMPEDE meant that participants had a really strong possibility of being randomised to a research treatment arm, as opposed to a traditional two arm trial where the odds tend to be 50/50. This is really important for patients who are facing a life limiting disease.

“We collaborated with three other trusts locally so that collectively we could offer STAMPEDE to patients across the entire North West region. Clinicians in the different trusts referred their patients to Royal Preston Hospital, which is how we became one of the top recruiting sites. It was great to be able to give so many patients that option of taking part.

“That said, it wasn’t always an easy study to deliver. Our Trust provides tertiary services to a large area in the North West, including some cancer services, but we didn’t have a dedicated or specialist centre like an ECMC - Experimental Cancer Medicine Centre. It was a resource intensive study to run. However, the study delivery was underpinned by support from the Local Clinical Research Network with two full time research delivery staff. It was a real success. Not only was it changing practice - it was keeping people alive.”

In the 16 years since STAMPEDE opened the UK has gained a wealth of experience in delivering complex and innovative clinical trials. The number of clinical trials using a MAMS design globally has escalated in recent years and the COVID-19 pandemic has brought into sharp focus the benefits of the adaptive platform approach. The principles of this design, seen in the world leading RECOVERY, REMAP-CAP and PRINCIPLE trials, has enabled the UK to answer questions of global significance quickly and efficiently.

Professor Parmar reflects on how STAMPEDE has shaped the way we do clinical research:

“The concept of wrapping multiple important questions into one protocol, moving as quickly as possible away from treatments that are not going to improve outcomes, and adding in assessment of new research treatments worth of testing seems to have hit the mainstream. The research community has bought into the idea. The challenge now is to ensure that we can perform these studies in areas where there has been little progress for decades and a really great need such as in neurodegenerative diseases such as dementia, progressive multiple sclerosis, Parkinson's disease and motor neurone disease.”

Professor James added that, as well as accelerating the assessment of new agents in prostate cancer, the ability to link over-arching translational studies to a single trial vehicle, rather than having to repeatedly link new studies to new trials has also been transformational.

“Within STAMPEDE we now have a linked tissue collection with over 5000 samples collected to date, a germline DNA collection with more than 8000 samples, an image bank with more than 5000 patient scans, a plasma DNA collection plus the ability to link all this to NHS Digital records. This will ensure that we are generating novel biological insights into the underlying reasons for the outcomes that we see, both good and bad. The multi-arm randomisations mean that we can correlate the effects, for example of particular biological changes or imaging patterns, not just with outcomes but on the differing effects of different interventions.”

Further information

Related information