Neal Shore: Yeah, I agree with you completely, Alicia. Yeah, we've had this wonderful development of life-prolonging therapies, which has gone from, at least for prostate cancer, where we had, in 2004, docetaxel. Fast-forward 20-plus years later now, and there's arguably seven to eight novel MOAs that are life-prolonging. And then, you add into that there's maybe 12 to 14 agency regulatory-approved therapies that have survival benefits. So I think for our audience, it becomes pretty clear that it creates a complexity in trial development, making it hard to really get to that classic binary endpoint of overall survival, which we like. Certainly, in medical oncology, it's very clear-cut, it's binary, you're either alive or you're not.
But to your point, one of the things that we looked at in the publication that came out in the European Journal of Cancer this past year, that you and I were very fortunate to be co-authors on, with several other really great investigators, is what about these intermediate endpoints that you're talking about. And I think it speaks to not only in resistant disease, but also in sensitive disease, and it could be high-risk localized, it could be biochemical relapse, it can be low-volume/high-volume metastatic disease, de novo recurrent. So one of the things that we looked at was the intermediate endpoints, there's a lot of ways of parsing it, but radiographic progression-free survival is really important, PSA response, PSA dynamics, time to resistant disease, that's really important too. How we think about that vis-a-vis certain prostate cancer working group definitions and others, I think, is really important, because at the end of the day, there's an influence on real world decision-making.
Alicia Morgans: Absolutely. And in this work, we performed a systematic literature review of studies of metastatic hormone-sensitive prostate cancer. These were randomized controlled trials up through December of 2023, and they had reported hazard ratios for both radiographic progression-free survival and overall survival, and the goal was to understand how rPFS might serve as a surrogate for overall survival. So ultimately, we included 31 trials, 35 treatment comparisons, and nearly 19,000 patients in this particular meta-analysis, and we used an approach called meta-analytical methodology, and this was used to assess the correlation between rPFS and overall survival. And really, we sought to understand the accuracy of that survival prediction based on the rPFS that was observed, and really metric-ing it against the observed overall survival to understand whether this might be a reasonable surrogate. And we tested the robustness of our results by several sensitivity analyses to just ensure that we were confident in those findings.
Neal Shore: Yeah, that's a really nice succinct way of summarizing it. A lot of it was predicated on wonderful work done by a good friend and colleague, Susan Halabi at Duke, who's really one of the international pioneers on epidemiological understanding statistically about these surrogates. And to your point, we found a really strong correlation of rPFS and OS, the correlation coefficient was statistically significant and powerful. And so, for our busy community clinicians, it's a little statistically wonky, but it's a good read frankly. So we look at these predictive models, do we meet certain thresholds, to your point earlier, and we did, we met these validity thresholds, where most of the overall survival hazard ratios, if they fell within a certain prediction interval, created an understanding, or at least an understanding for validation.
One of the things that got us going on this was the ARANOTE trial, which really, because of the challenges that I was alluding to earlier regarding so many other crossovers to other approved life-prolonging agents, which limits some of the comparator arm, and even arguably sometimes patients will end up getting the same thing twice, not often, but sometimes. But it helped us understand really that when you take a therapeutic that's an another ARPI, for example, how can we see, if we get clear rPFS benefit, which we saw on ARANOTE, can we predict also the OS benefit? And so, I think that was a big reason for why we were thinking about this. But I like to think about it as you were saying earlier, Alicia, how we can think about this sort of strategy for other populations within the prostate cancer journey.
Alicia Morgans: I agree. And ultimately, I think that this study also helps us understand how we might be able to use that rPFS in place of OS, and also to recognize that certain studies are not really powered to accurately predict overall survival, but we can use rPFS as a reliable endpoint. As an example, the ARANOTE trial, which you just mentioned, was really designed only to find differences in rPFS, and we can use the surrogacy to predict a long-term OS hazard ratio of 0.64. But of course, in the actual study, this was underpowered, and so is not necessarily the same thing that we saw in an underpowered analysis.
So limitations exist here in terms of our real life ability to complete trials in metastatic hormone-sensitive prostate cancer, as well as limitations that exist for this particular study. As you mentioned, bias from crossover, use of the intermediate OS values in some cases may be off because of the size of the trial, heterogeneity across trials can always affect the results, and as we go to lower levels of aggregation, there are really more exceptions. And this analysis tries to overcome what it can, and just to be an additive contribution to the literature to help us, in clinic, feel confident in our rPFS endpoints.
Neal Shore: Yeah, I agree. And like in any really good paper, we carefully instructed the audience that we go through those limitations. But I think at the end of the day, we conclude, and rightfully so, and thus we were happy that we were published in the European Journal of Cancer, that rPFS is a reliable surrogate for OS in mHSPC trials. There's a rather interesting and statistical-derived surrogacy equation that gives them an ability to make that estimation for OS outcomes, and I think the fact that we took in the rigor of these 35 trials really made me feel good about the conclusions that we reached.
Alicia Morgans: So Neal, let's talk a little bit more. I wonder, if overall survival is so important, does rPFS really still play a role in ensuring that clinical trials are going to be feasible and efficient? Is there a reason that we still need rPFS and a purpose for the analysis that we just talked about?
Neal Shore: Yeah, it's a really important question, I think, from a clinician standpoint, and also from a regulatory standpoint. From a regulatory standpoint, we're starting to see a lot of studies. We just saw things at ESMO, looking, for example, the CAPItello study, which was solely based on rPFS. We have had in bladder cancer a series of high-risk NMIBC studies that are really essentially just looking at recurrence and progression. Even at five years, the maturity events for survival are very, very low. Interestingly, the EMBARK trial was based upon an MFS, which was largely driven by rPFS. So I think that from a regulatory standpoint, we're starting to see some movement to recognize that if we are only going to wait for OS, we're going to be waiting a really, really long time, and there's going to be all these confounders of crossover, and then there's always new things that are coming and new targets, and then, of course, our real world consideration as well, what is a meaningful rPFS? So I think the answer is yes, but with qualification.
Alicia Morgans: Yeah, I would agree. I do think that these kinds surrogacy findings that suggest that rPFS can be a reasonable surrogate for overall survival helps in my clinical conversations with patients too, because it is incredibly meaningful for people to have that first-line of therapy last for a long period of time. We can give that information with an rPFS endpoint, and then we can say, "You know what? This long period of time that you have really does, regardless of our ability to capture it cleanly and perfectly in a clinical trial, does reflect what we expect in terms of prolongation of overall survival and is meaningful for you." So patients appreciate that and appreciate the fact that it is really difficult to understand survival, given all of the other things that may intervene between the initial treatment and end of life, and so it is very helpful for me in those clinical discussions. But I wonder, do you think about, as we look ahead, how adopting surrogate endpoints like rPFS can help accelerate innovation and drug development and opportunities for patients and really the potential for their future care?
Neal Shore: Yeah, I do. Just to take it back to just that patient interaction, I don't typically have patients who say to me, "What's going to be my survival benefit from XYZ drug?" They're mostly really, in some way or fashion, asking, "Well, how long am I going to be on this? How long do I stay on this therapy?" And of course, we talk about most importantly subjectively, "How well are you tolerating therapy?" And then, we look at our metrics for signs of progression, and there's lab findings, of course, we're seemingly very wedded to PSA, but we recognize that it's much more than just PSA and PSA is not perfect, but other lab values are important too and other additional biomarkers. But then, the really big one is imaging progression, or rPFS, and as simplistically as it sounds, but new lesions are not good. And especially when we can follow the right criteria, whether it's RECIST or some new developments in PCWG4.
But I think for patients, it's a very important consideration, as opposed to, "Well, how much longer am I going to live?" Which as researchers and clinicians, we want to understand that, and so there's a little bit of a disconnect there. Likewise, I think the regulatory authorities really are moving more and more to recognize that if we only stay loyal to the notion that we must see OS, we're going to really stagnate a lot of drug development, as well as drug approval. So I think there's a healthy balance within the question that you ask.
Alicia Morgans: I agree. It's so interesting that you bring that up, because I think that patients do ask, that's the number one question, "How long am I going to be on this treatment?" And it is so important for their planning. And when they do sometimes ask about overall survival, my answer in clinic is always that it's very difficult for me to say, because I expect it to be years, based on the data I have now, but it may be many more years, based on the therapies that are being developed today, and even being developed yesterday that haven't come to the fore yet.
And so, it is nice to know that we can say that rPFS, that we can start the clock today, is something that I can estimate. And overall survival is changing so much day-by-day with all of the new treatments that we have, and certainly with the opportunity for you to get this, that or the other treatment in the future, and so that estimate in itself is already outdated at the time of publication. But at least we know we can rely on rPFS, and that the longer you're able to have that rPFS now, the better you'll live, and we have some surrogacy evidence to suggest that. So I think that's so true. And as more and more innovation happens, if we don't have surrogate endpoints, we simply will not be able to develop them, because not every drug that we give in an early line is ultimately going to meaningfully affect overall survival of patients as they should cross over and get those effective treatments over time.
Neal Shore: Yeah, very well said. I think, basically, we're experiencing all the recent couple of decades of success, particularly in the last decade, and it's true for prostate cancer and it's certainly true in other areas of GU oncology. And I believe that the regulatory authorities, whether in Europe or Asia or the US, are recognizing that as well. Advocacy groups are really demanding it too, appropriately so. But I think that this is starting now, and that's why I'm proud to have been part of this publication with you and this research. Others are working on this too, we mentioned our dear colleague and friend, Susan Halabi. But I think it's important for our colleagues and community to recognize this is what's around the corner. So I appreciate your comments very much, and again, this is a really good discussion, and hope that the listeners will take a look at the article in the European Journal of Cancer.
Alicia Morgans: Agree. Well, thank you for your time, Neal, and great to work with you, as always.
Neal Shore: Thank you.