Nicholas James: Yeah, thank you very much. So this is a randomized comparison of magnetic resonance imaging introduced upstream of TURBT in the staging pathway. So obviously, I'm presenting on behalf of my various cohorts as listed. And I've got no personal financial interest in this, but I do have to acknowledge the funding source, which is our National Institute for Health Research, as listed on the screen here.
So the general background will be well known to the audience. Bladder cancer diagnostic pathway is essentially unchanged for a century and it comprises an outpatient flexible cystoscopy for visual diagnosis of a bladder mass, and then patients have passed to a rigid cystoscopy and resection for pathological diagnosis and staging. The problem with this is that, as a one size fits all, it's not very good because everybody's debulked and often without muscle, so they have to be sent back for re-resection. Often the resections are happening in one place and the cystectomies, or radiotherapy, or chemotherapy might be happening in multiple other hospitals.
So you've got a complex pathway with multiple handovers and typically that adds up to delay. And we know these delays are not inherent just to the British system. They occur in all other health systems as well just because of the nature of the pathway. And despite recent advances in chemotherapy and stuff, bladder cancer outcomes are pretty much unchanged for the last three decades. And just finally to comment, mostly for cancer, we're happy to image to stage and biopsy for histology. We don't normally surgically debulk as part of our pathological or imaging staging.
So the design of the trial, it's an open label, multi-stage, randomized trial, testing whether, if we introduce a multiparametric MRI before TURBT, we reduce the time to radical treatment for suspected muscle invasive bladder cancer. And the thing to note here is that we're not replacing TURBT with MRI. We're adding MRI to the pathway. We're not saying that people should not have pathological diagnosis, but basically anybody who was found to have a bladder mass in the hematuria clinic was eligible. The only exceptions being not fit for an MRI scan because of metal implants and things.
One to one randomization, standard pathway, everybody got the standard of care. Pathway to the MRI pathway, we triaged patients into patients that were probably non-muscle invasive, and that was about 50% of the total, and then patients who were possibly muscle invasive, and that was about half as well. And a lot of those, about half of them were actually muscle invasive. So we're using the MRI to distinguish the non-muscle invasives from the muscle invasives where we've got a high suspicion that it might be muscle invasive, is kind of the subtlety of this.
When we set out, we'd envisaged that we would be doing an MRI and a biopsy, but actually we permitted TURBT according to clinician preference. And essentially, pretty much everybody did get a TURBT. So what it is not is a trial of MRI versus TURBT, or MRI and biopsy. Although that was what we intended, it wasn't what actually happened.
So three stages. Feasibility stage, intermediate stage, final stage, and all published apart from the last one. So we published the feasibility a while back. It's feasible. We published the intermediate stage earlier on this year. The references are listed at the bottom two papers, in fact. And what we showed was that the time to definitive treatment, which we defined as surgery, radiotherapy, chemotherapy, or palliative care for muscle invasive bladder cancer was the primary outcome measure.
You can see that the time to treatment comes down hugely with the introduction of the MRI. So what we're now showing is the final stage, which is progression-free and overall survival. And I must stress this is underpowered because the trial was cut short by COVID, but we nonetheless do have complete outcomes out to two years for all patients. This is the causes of death. So just to highlight, significant numbers of patients die from completely unconnected causes, but kind of related to the fact that these are older patients who are mostly lifelong smokers. That also means significant numbers of other cancers, in particular smoking related cancers of the upper aero-digestive tract, giving substantial proportion of all of the deaths being not down to bladder cancer. These are the bladder cancer deaths with a very obvious numerical difference between them.
If we look at progression-free survival, the numbers at risk are at the bottom. The hazard ratio for this is 0.75. Less than one is in favor of the MRI pathway. Confidence crossed one, however, and P-value is non-significant. If we split that up by non-muscle invasive and muscle invasive color coded, as per on the right, again, the numbers are all too small. They overrun each other.
However, if we go onto bladder cancer specific survival, we can see we've got a very striking hazard ratio in favor, with a P-value that's significant and confidence interval not crossing one. And if we split that by non-muscle invasive and muscle invasive, we can see that is entirely driven by an improvement in the muscle invasive bladder cancer outcomes.
And just to highlight, it takes around 100 days, a third of a year, median on the TURBT arm, so we can see that patients are having recurrences before they appear to have even got to their definitive treatment. So it does imply that delay is bad. You expect delay is bad. And the contrast with the the MRI arm is quite striking here. The separation is very early. No separation in these arms, really. And if we look at overall survival, we see the same. The hazard ratio is less than one because of the deaths from other causes being a substantial cause in proportion of this.it's not significant. But again, we've got separation on the muscle invasive part of it.
So our overall conclusions from this are that there's strong trends to improvement in all the key trial metrics. None of them trend the wrong way. All underpowered. There's no evidence, very importantly, of a detriment. We haven't calculated a P-value that MRI is worse, but it's very, very improbable that MRI is worse. And it might have been worse if we've been classifying patients wrong, and calling patients non-invasive and treating them as non-invasive cancers, and they were actually muscle invasive and therefore got the wrong treatment. There's no evidence of that at all. And there is a statistically significant improvement in bladder cancer specific survival driven entirely by muscle invasive bladder cancer outcomes being different on the MRI guided pathway. So we think this is a really very interesting result. I mean, you can speculate as to what's driving it. Might just be a fluke. It may be that delay is important, but I think you can't rule out that TURBT is itself potentially harmful here as well.
We think that the patients having the MRI-guided TURBTs are having less of a TURBT. I think they're having a sufficient sampling for histology, sampling for CIS, checking the ureters, checking the urethra, things like that, and not having an attempt at complete resection, because you've already decided from the MRI that you're going to be doing a cystectomy. Why would you debulk them?
So we think it's plausible that the more extensive TURBTs on the control arm may be actually harmful. I mean, I don't know that, obviously, but as a urologist, I'd be very interested in your thoughts on this, Ashish, clearly.
Ashish Kamat: So first off, Nick, I always am impressed with how well you condense a complicated topic into a succinct presentation, so thank you for doing that. And congratulations for initiating and leading such a complex study that, really, it is going to be paradigm changing. I mean, it's already changed how we think about the management of patients with bladder cancer when they come in to our clinics. And again, thank you for always being open and available for debate, and you and I have debated this many times before.
So first off, I just want to say, I'm so happy to hear you say this, and the evolution of the study was, like you said, initially meant to replace TURBT, but just the way the practice happens, it's a complement, a supplement, and a way to improve the care of our patients. Just bring it in before the initial resection.
Which brings me to a couple questions. First off, I agree. The management and the way we visualize the resection, the biopsy, the TURBT for patients has evolved over time, and I do recognize that if a TURBT is done inappropriately or incorrectly, can actually cause a lot of harm to our patients, as you know, so absolutely agree. And imaging in the form of CT where MR is not available, or MR where it's available, prior to TURBT really should be standard of care. Have you seen this become more widely used as people have been more aware of the relevance of MRI? Leave aside the VI-RADS, just multiparametric MRI. Have you seen this happen in your practice pattern or where you're seeing patients?
Nicholas James: Yeah, we're doing them very extensively now in our management. And actually, what we're seeing across our network is a difference in pattern of care between the centers directly within our orbit and the centers that send patients in.
So in the centers that are kind of more peripheral, what we're seeing is that they're all going down pathway one. I mean, again, you're a urologist, but the team are taught to debulk, to take all the tumor out, get as much out as you can, and all the rest of it, and with all the inherent inaccuracies and morbidity attached to that. Whereas the ones coming in through either through us directly or through St. George's, which is the big teaching hospital that the Marsden is linked to in terms of urology services, what they're telling me is that they're just not doing debulkings anymore.
They're doing a sufficient TURBT to get the information we need. We've got an MRI scan. We're very comfortable with it at staging. And we're very comfortable as well now that, if we're happy something is definitely a T3 or a T4 tumor, and somebody's done a TURBT and not got any muscle in the sample, we're not going back and re-resecting those. We're just going on to whatever the next stage of the treatment is.
And the thing that was striking about the time to treatment was that it came down by so much, even though everybody got a TURBT. So we think that what's happening is that you've done the MRI, you think, "Right, we know this is a bad tumor. We need to make sure they get on down the pathway quickly." And I suspect they're also being done by people like yourself, not by more junior people, because if you're going to operate, you're going to want to know what you're operating on.
So we think that it's a tool for improving decision making that allows you to say, "Right, well, this is a small non-invasive tumor. Straight to TURBT. That's fine." And I assume a less challenging operation. And the complex stuff needs to be planned by the complex skillset people, like you. Anecdotally, I can't prove that is what's happening. The trial very much supports that being as what's happening. And obviously we've got a range of different surgeons involved with this from multiple different sites, so it's not just a single site experience. It's 15 sites' recruitment.
Ashish Kamat: I mean, that's an insightful hypothesis. And I think, like you said, it's hard to prove that in this study, but it makes actual perfect sense from a clinical standpoint. I recall from the JCO paper, and again, full disclosure, we had a little editorial back and forth, but the composite endpoint time to definitive treatment included radiotherapy, palliative care, surgery, et cetera. In this, even with the caveats of small numbers, are you able to tease out which is actually contributing to the improved survival of the patients? Is it everything? Is it the radiotherapy, the surgery? What is driving it, in your opinion?
Nicholas James: So actually, I don't know the answer to that. We've had to rush this analysis through in order to get to the... We literally did the analysis the kind of the day before the late breaking abstract deadline. We're still doing the analysis, so I don't know the answer to that.
But one of the things that was very striking when we did the first analysis was that there was like four patients or something who got palliative care as their definitive outcome. Two of them, they were split two and two across the arms. For the two that were on the standard care pathway, the decision to treat with palliative care was made after they had died, whereas the decision to treat on the MRI arm was made the day after the MRI scan when people got the report.
So those sorts of things, they're not going to affect survival, but they're massively important to the patients, because those patients would have been referred for palliative care, whereas the other ones were still rattling around with no decision and actually died before anybody thought what they were going to do with them. So I think it's unlikely those are driving the survival differences, but I think that would have been a very important qualitative difference in terms of what happens to the patients. Those patients didn't get TURBTs on the MRI arm either, so they were spared a procedure, of course, they didn't need because they were metastatic and dying.
And so I think that there are some immediate, very obvious benefits, but I think it's more likely to have been driven by the ones getting surgery or radiotherapy, therefore, because they're the ones where we... You know, outcomes are not great. I mean, the striking thing... I know you'll say, "Oh, my cystectomy is all 80% reliable and so on." But it's so hinged on case selection. And the patients ending up in the big teaching hospitals, us included, are not going to be the same as the patients overweight, old, heavy smokers, managed in district centers.
Ashish Kamat: Right. No, absolutely. Nick, again, we could chat about this forever, but I do want to ask you one question in closing, which is a little bit outside the study, but I would love to get your insight. We, as a field, are moving towards bladder preservation, right? And we're using systemic therapy, getting clinical complete response, trying to then preserve the bladders. And as part of that clinical complete response, the TURBT complete resection of the tumor is an integral part of that pathway. How would you recommend using the data that you've generated, clearly imaging, but then the TURBT component, how would you recommend it, putting on your hat as a bladder preservationist, if we want to skip the whole TMP pathway and we just want to go on to bladder preservation with no local therapy?
Nicholas James: So in terms of bladder preservation, I think as we've seen with the EV pembro data at ESMO, the data's great. We're going to see more, and more, and more complete responses to systemic therapy.
I think the other thing I would say about MRI is that, if you don't do a big resection and just take enough tissue to get a diagnosis, one of the problems with bladder cancer with CT is that the tumor is isodense to to normal bladder, and it's hard to measure bulk with something that is not a lump, it's a sheet, really, of varying thickness. Whereas with an MRI, the diffusion changes if you respond. So the diffusion into the tumor improves as the tumor responds. So you can numerically quantify the change in diffusion on an MRI scan if you don't do too much surgery.
So it's another argument in favor of upping the imaging, reducing the debulking, increasing systemic therapy, and then you can, you've got an accurate way of assessing it, particularly, probably, if we bundle in liquid biomarkers, urinary DNA, circulating tumor DNA as well. I think there's a lot of scope to redesign the pathway here, very much for the benefit of the patients, in terms of less morbidity and almost certainly better outcomes.
Ashish Kamat: And as we saw at ESMO, urinary tumor DNA correlated with clinical complete response, path response, ctDNA did as well. So absolutely. I think everything's going to be a part of the machinery that helps us take better care of patients.
Nicholas James: Exactly, exactly. Yeah, that's very exciting.
Ashish Kamat: Nick, always a pleasure. Thank you so much for taking the time. Congratulations on everything you've done and being so busy at ESMO.
Nicholas James: Fantastic. Great.