Long-term Follow-up of Non-metastatic (M0) Patients in the STAMPEDE Randomized Trial - Interview with Nicholas James

October 9, 2019

Nick James and Alicia Morgans talk about the newly presented STAMPEDE data just released and presented at the 2019 ESMO meeting including, long-term follow-up of the docetaxel patients as well as the much anticipated metastatic data. 

Professor Nicholas James BSc, MB, BS, FRCP, FRCR, Ph.D., Institute of Cancer and Genomic Sciences NIHR Senior Investigator, Consultant in Clinical Oncology at the Queen Elizabeth Hospital Birmingham and Professor of Clinical Oncology at the University of Birmingham.

Alicia Morgans, MD, MPH, Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.

Read the Full Video Transcript

Alicia Morgans: Hi. I'm so excited to talk today with Professor Nick James, who's a Professor of Clinical Oncology at the University of Birmingham and the Queen Elizabeth Hospital in Birmingham in the United Kingdom. And I'm very excited because I understand we're going to have the opportunity to talk about some of the new STAMPEDE data that has just been released. So let's hear it. What is new in STAMPEDE?

Nick James: We have two sets of data coming out. We have a long-term follow-up of our docetaxel patients, and we split it into M0 and M1. For the M0, in essence, it's a repeat of the previous data, so it's docetaxel still improves failure-free survival. That doesn't translate into an overall survival benefit, essentially because the survival has turned out to be too good. So it's an important update, and I think it probably in truth doesn't change anything very much, but nonetheless, we think is important to get that stuff out there.

The thing we are really excited about though is the metastatic data which we're publishing separately. So because there's been this ongoing controversy about is volume predictive of response to docetaxel, which frankly we never thought was true-

Alicia Morgans: Yes, but we Americans have definitely thought it's true.

Nick James: I know, but other people do. So the only way to resolve these sorts of things is with data. So we've gone back, retrospectively pulled out as many of the scans as we could find, which turns out to be about 800 or 900, and retrospectively classified the patients on their baseline scans into high and low volume. And it works-

Alicia Morgans: Great endeavor.

Nick James: It was an endeavor. So Noel Clarke and his team at the Christie, who pulled in all the scans and done the analysis, and then the other stats team at the MRC had done the usual wonderful job with the stats. And so essentially, about 40% of low volume by CHAARTED criteria into high volume. And in essence, what we see is pretty much identical hazard ratios for benefit for docetaxel in low and high volume, in line with the original analysis. So with a longer follow-up, the result holds up. Docetaxel still improves survival. It improves survival by the same amount, whether you are low or high volume.

We don't, in other words, see any volume effect whatsoever in terms of outcome of docetaxel. The only thing we do see with the volume is that the high volume has a worse prognosis than low volume, which of course everybody agrees with.

I think the other thing for us is that ... If you compare what's in STAMPEDE with what's in CHAARTED, almost all the STAMPEDE patients, both low and high volume, are newly diagnosed.

Alicia Morgans: De novo metastatic.

Nick James: Whereas in CHAARTED, most of the low volume are recurrent, and most of the high volume are de novo. So it is plausible that the differences that Chris Sweeney and others had been reporting in the low by high volume split are driven by the proportions of new versus relapsed being different in the two subgroups, rather than the biology of the disease per se being different in newly diagnosed metastatic. So I think the results are not necessarily incompatible with each other.

Alicia Morgans: That's good to hear. And it's actually ... The biology is still probably ... If you have a patient who you've caught in a localized setting, and then that patient has now relapsed, that biology may be very different than a patient who is then presenting with de novo metastatic disease, especially a high volume de novo metastatic disease.

Nick James: Our take on that is that if you are low volume de novo metastatic, you should get docetaxel, or some other treatment, abiraterone, apalutamide, whatever. But you should definitely not just get ADT. And I think given that none of the AR-targeting therapies are available in the UK for example, we would recommend that is docetaxel, therefore. And in fact, that is ... Our own surveys of practice in the UK show that is exactly what happens. Our investigators have consistently been giving docetaxel, and we think that remains the right thing to do.

We've also ... If you look at the radiotherapy as the primary data, and look to see whether that effect is there with or without docetaxel, because the later patients going into that randomization had docetaxel as part of their standard of care, it appears the benefit in low volume patients is the same whether or not you have docetaxel as part of your standard of care. So for our forthcoming oligometastatic arm, the recommended standard of care for UK patients is going to be ADT, docetaxel if your chemo fits, radiotherapy as the primary plus or minus radiotherapy to the mets.

Alicia Morgans: Yes. I think that when we talked maybe a year ago or so ... When we talked about the radiotherapy data, that is what we discussed, certainly that it looked like it held regardless of whether ... The radiation was a helpful thing, regardless of your systemic therapy, whether that contained docetaxel or not, but you will then, of course, you're going to present further survival data as that becomes more mature, or is this the final analysis that you've already presented?

Nick James: This is probably going to be the final analysis of the docetaxel data, I think because we don't ... That data is now relatively mature, and so we will probably completely ... We have closed the trial to perform a follow-up, but we will probably put the lid on that for now, because we have so much data that we can't keep tracking all of it.

Alicia Morgans: No, of course. And I think you even mentioned that when we were at the APCCC, following 10,000 patients, or however many thousands of patients, becomes something, tt's a good challenge to have. The other question I had regarding the high volume, low volume: the hazard ratios actually for both appeared across one. So the way that I read that is that in these groups, we just don't have the power to show that they are ... Clearly, the hazard ratio is clearly below one and that although when we put the data together, obviously ... and if there's no heterogeneity, it still seems that this would hold ... the effect holds across both of these groups. But I just wanted you to comment on that.

Nick James: Yeah, so I think people ... There's a lot of misunderstanding I think around what it means if the hazard ratio is across one, because if you start with a trial that is powered to show an effect, and then split the samples in half, each sample is intrinsically-

Alicia Morgans: Not powered.

Nick James: ... underpowered. So what you're looking for with the forest plots, is you're looking to see whether the effects are consistent or inconsistent. And so what the fact ... The fact that the hazard ratio is lined up, and the test for interaction is negative, that's to say it looks as if the effect is in the same direction in both groups, that's the important message. The fact that the hazard ratio is across one is inevitable, if you slice-

Alicia Morgans: Yes, because you're not powered.

Nick James: ... things down into enough subgroups, because you're not powered. And unless you powered the subgroups upfront individually, you're never going to be powered. And so people ... You often hear discussed and saying, "Well, this clearly works in T2 but not in N1," and they're misinterpreting the point of the forest plot.

Alicia Morgans: Yes. I just wanted to make sure that we got that out.

Nick James: Yeah, no, I think it's really important.

Alicia Morgans: Because I do think people will say that, and I think that it's important to hit that head-on. And actually I think the hazard ratio, if you look at the absolute numbers, I think for low volume it's actually maybe a little bit lower than it is for-

Nick James: It's softened slightly, but only slightly.

Alicia Morgans: And and so essentially it's just ... It's consistent.

Nick James: Yeah, I kind of ... I'm in danger of turning into a statistical bore about this, but we published a paper ... We re-analyzed the abiraterone data according to the day of the week in which the patient was born, and the day of the week in which the patient was diagnosed. So if you think that if the hazard ratio crosses one, you shouldn't treat patients, it means you shouldn't treat patients with abiraterone whom you'd biopsied on a Monday, for example.

Alicia Morgans: Well, we've got to get that out.

Nick James: So yeah, it's an important message.

Alicia Morgans: It's important information to get out.

Nick James: The paper is called Thursday's Child, and it's myself, Matt Sydes and Mel Spears, who are two of the STAMPEDE statisticians. But I think it's a very important read, because-

Alicia Morgans: It is, yeah.

Nick James: ... it highlights the ... People over-interpret where the confidence intervals go.

Alicia Morgans: Absolutely. And having a bit of fun with all of us too. So thank you so much for sharing these updates.

Nick James: Pleasure as always.

Alicia Morgans: Really exciting, and of course always changing our standards of care. Thank you so much.

Nick James: Pleasure.