Neeraj Agarwal: Welcome to UroToday. We have such an honor and a pleasure having Dr. Neal Shore join us today, who is the director of the GU Oncology Program at Carolina Urologic Research Center. Welcome, Neal.

Neal Shore: Thank you so much, Neeraj. It's great to be with you today. So much fun to collaborate with you, not only on educational moments like this, but in all the different trials that we're doing, early- and late-phase, so thanks for having me.

Neeraj Agarwal: So Neal, you presented very interesting data recently, which was a real-world head-to-head analysis of overall survival in patients with metastatic hormone-sensitive prostate cancer who were initiated on apalutamide versus enzalutamide in the United States. Can you please tell us more about this study?

Neal Shore: Yeah, thanks, Neeraj. Look, we've had this proverbial embarrassment of riches: four different androgen receptor pathway inhibitors. Great trials that led to approvals now for enzalutamide and apalutamide, abiraterone and now darolutamide. And so what we wanted to do in this study was, even though we don't have direct comparator trials, how could we look at some retrospective analyses? And what we chose in a series of different analyses—we did one with apalutamide and abiraterone—we're going to talk about today looking retrospectively at any differences in overall survival with statistical analysis, and I'll explain that in a second, of the TITAN and the ARCHES trials, apalutamide and enzalutamide respectively. And the null hypothesis on this was there's no difference, no overall survival difference at 24 months post-treatment with enzalutamide versus apalutamide, or an alternative hypothesis is could we actually find a difference?

We looked at a couple of different data sources. One was known as the PPS. It's a US data source of large uro-oncology community practices. PPS stands for Precision Point Specialty, and we collected a lot of demographic and clinical data, very urology-specific. And the other big data source we used was the KRD, which is also known as the Komodo Research Database, which covers over 300 million US patients, and so we looked at the mortality data from that to really inform the overall survival analysis. And so really, none of this was dependent upon healthcare providers making the assessment.

There's a nice—in our paper, you can really get into the statistical modeling, but for purposes of time, basically we had over 1,800 patients, the demographics very nicely balanced with age, almost identical in these two groups and basically consistent with the demographics of this retrospective analysis of patients who were diagnosed with mHSPC, those who got apalutamide versus those who got enzalutamide. So age, race, their index date at the time of diagnosis, their Charlson Comorbidity Index, all remarkably well-balanced. The percentage of de novo disease, about 56% in both arms. Their tumor distribution, very well-balanced.

And we used this modeling that's called the IPTW or the inverse probability of treatment weighting, and again, I would encourage my colleagues to read the paper to really get into it, but the bottom line, and you see it in this Kaplan-Meier curve, is at 24 months post-indexing of apalutamide patients versus the enzalutamide patients, again, for mHSPC, low- and high-volume, de novo and recurrent, there was a statistically significant 23% reduction, hazard ratio of 0.77, favoring a reduction in the risk of death for the apalutamide versus the enzalutamide patients. And so this was a really interesting finding. When you look at the Kaplan-Meier curve, you start to see even a curve separation even before one year.

So again, retrospective, there are limitations of course. There could be miscoding or misclassifications, because this was all administrative claims. These are from a community urology perspective, these 1,800 or so patients in each arm, different from the phase III TITAN and ARCHES trials, which had a lot of academic sites and were also global. But when we look at it, we were able to basically do this comparison. And the thing too, in all fairness, is when you do a regression analysis, you can only adjust for certain covariates and there could be some unknown confounders, but who's going to go out and do one of these head-to-head prospective studies?

So I think that even though the hazard ratios were essentially identical in TITAN and ARCHES in those studies, when we looked at this real-world evidence, we did see a hazard ratio of 0.77 favoring apalutamide over enzalutamide. So I think it's somewhat compelling, interesting to look at, and I think we are not concluding that one drug is necessarily better or inferior to the other. We just saw this as a healthy analysis of real-world data.

Neeraj Agarwal: Thank you, Neal, for taking the time to go through those slides and explaining what you did and how these two drugs were compared in the real-world setting, and of course, you mentioned all those limitations which are associated with this kind of analysis. Please tell us more, especially for those who want to do a similar kind of analysis. Are there any recommendations by the FDA regarding these analyses, and are these prespecified?

Neal Shore: Yeah. So not entirely prespecified in the formal way that we think about it, and that's a really good question, but basically, it's clearly retrospective in that regard with the binary endpoint of survival or death. And as you said, we did have almost 1,900 patients, real-world experience looking at this Komodo database, which really captures literally several hundred million patients. The PPS was there to capture the indexing, the time that they started their ARPI, and in this analysis, this retrospective head-to-head, real-world, not direct, not prospective, was enzalutamide and apalutamide. But having said that, I think that more and more, the FDA is getting more comfortable with the fact that prospective trials are very time-consuming, they have crossover conundrums, and then just the overall expense of it.

So I think that the answer to your question is that these are FDA-approved therapies, and then more and more, I think these analyses are becoming informative, not so much because they're already approved, but I think as it relates to payers potentially and how they think about prioritizing how drugs are preferred. I think it's a bit of a challenge though at the end of the day, and as I said at the very beginning, we have this embarrassment of riches: four different androgen receptor pathway inhibitors, if you look at the direct inhibitors—apalutamide, enzalutamide, darolutamide—and then abiraterone. So I think that it's evolving, to answer your question more directly. The real-world evidence is really starting to inform how we prioritize potentially. Maybe it's based upon some subpopulations of patients. In this particular study, we didn't have any particular biomarkers that we analyzed, but I think there's an opportunity for future studies as well.

Neeraj Agarwal: Well, that was a very nice explanation of the rationale for the study, and thank you Neal, again, for sharing this massive, really large real-world dataset, which really gives new insights into the treatment of metastatic hormone-sensitive prostate cancer.

Neal Shore: Thank you, Neeraj.