Christopher J.D. Wallis: Hi, and thanks for joining today. I'm Chris Wallis, a urological oncologist from the University of Toronto. On behalf of my co-authors and investigators, I'm really happy to take a few minutes to go over our recent publication entitled The Association of Deep and Durable PSA Responses with Outcomes in Metastatic Hormone-Sensitive Prostate Cancer: Insights from the ARASENS and ARANOTE Trials. As we begin, I think most of us who treat prostate cancer patients understand that PSA or prostate-specific antigen has been long associated with prostate cancer diagnosis and treatment responses. And so in metastatic hormone-sensitive disease, it's clearly a critical biomarker. And we know that early PSA response usually assessed at time points like six or nine months provide important prognostic information. And as a result, may potentially allow us to tailor our therapies.

Unfortunately, as with all sort of statistical approaches that take continuous data and categorize it, or impose strict time cutoffs, we lose a little bit of the nuance. And so we need to acknowledge that PSA levels are inherently dynamic; they change over time. And understanding those changes can give us insights into disease behavior and the effectiveness or response to therapies. And as a result, a longitudinal analysis, looking at PSA trajectories tied to clinical outcomes like survival, I think allows us an important opportunity to gain further insights into disease biology, treatment response, and potential windows into treating our patients better. And so in order to do this, we used data from two large phase 3 randomized controlled trials. ARASENS, which evaluated the addition of darolutamide to ADT plus docetaxel. And ARANOTE, which evaluated the darolutamide doublet with ADT in the absence of chemotherapy. And across both of these trials, we modeled PSA trajectories using a linear mixed model. And in doing so, we wanted to capture how PSA levels change over time. We then modeled overall survival using a Weibull model, and we linked these two models using an approach called joint modeling.

In doing so, we could deconstruct the effect of darolutamide on overall survival. The most obvious or intuitive one is going to be the indirect effect, which is mediated through PSA response. But there's also a potential direct effect of treatment on OS that is not a portion of the PSA response. And in doing so, we're able to deconstruct this and look at the relationship between PSA responses and survival in a more nuanced way. And so perhaps not surprisingly, we found that baseline PSA was strongly associated with overall survival. And importantly, the rate of PSA change, that is the speed of decline, when initiating ARPI therapy was associated with overall survival as well. Steeper and more durable PSA declines were associated with improved outcomes, such that a 29% reduction in mortality was seen in the ARASENS trial and a 25% reduction of mortality was seen in the ARANOTE trial when the rate of PSA decline doubled.

So the faster patients have a PSA response, the better we anticipate they'll do in terms of survival outcomes. Now, perhaps not surprisingly, patients who received the ARPI, that is darolutamide in this case, demonstrated more favorable PSA trajectories. And by that we mean they had faster declines and slower subsequent increases in their PSA. As a result, this translated into sustained PSA nadirs, and perhaps not surprisingly, patients had the greatest benefit from the addition of the ARPI darolutamide in the first one to two years when this PSA decline is most rapid. And then there's some stability, but we note that there's a plateau in the time-varying hazard ratio between 0.6 and 0.7 in ARANOTE. So even many years out, we see that sustained survival benefit over time.

As we try and integrate these data with what we already know about advanced prostate cancer and the role of PSA as a prognostic factor, I think the addition of joint modeling helps us to understand in a more nuanced way that strong and time-dependent association between PSA trajectories and overall survival, here leveraged across two complementary phase 3 trials. As we move beyond static PSA time measurements, we get a more complete picture of treatment response. In particular, we can see that early, deep and sustained PSA declines, associated with the use of an ARPI like darolutamide, translate into long-term survival benefits. And I think this allows us to establish a really important and valuable framework that links these dynamic PSA changes to clinical outcomes and gives us the opportunity to add a little more nuance beyond assessing an absolute PSA level at a nadir time period like six or nine months. Work like this could never be done without the conduct of the pivotal phase 3 trials.

And so I want to thank the patients, their families, and all investigators and study personnel who participated in both ARASENS and ARANOTE, as well as my esteemed collaborators who worked on this analysis. So with this, I want to thank you for your time and hope that our discussion today and this paper has added just a little bit to your understanding of prostate cancer and hopefully translates into the care of your patients.