Read the Full Video Transcript
Zachary Klaassen: Welcome to this UroToday Journal Club. We'll be discussing enzalutamide and survival in non-metastatic castration-resistant prostate cancer. My name is Zach Klaassen. I'm an Assistant Professor in the Division of Urology at the Medical College of Georgia, and joining me is Chris Wallis, he's a Fellow in Urological Oncology at Vanderbilt University in Nashville.
As I mentioned, we'll be discussing this paper that was recently published in the New England Journal of Medicine following presentation at the 2020 ASCO Virtual Meeting. The first author is Cora Sternberg for the PROSPER investigators.
By way of background, enzalutamide is an oral androgen receptor inhibitor, and it's been approved in men with metastatic castration-resistant prostate cancer in both the pre-chemo and post-chemo state based on the AFFIRM and PREVAIL trials. It was also recently approved in men with metastatic castration-sensitive prostate cancer based on ENZAMET and the ARCHES trial.
Enzalutamide was first approved by the FDA in 2018 in this disease space of non-metastatic CRPC on the basis of improved metastasis-free survival compared to ADT alone. You can see here that the hazard ratio for metastasis or death was 0.29 with a 95% confidence interval of .24 to .35, and this was published by Hussain and colleagues in the New England Journal of Medicine in 2018.
Subsequent to this publication was a quality of life study from this data, which was published in Lancet Oncology in 2019, looking at several factors that showed that enzalutamide compared to ADT alone improved health-related quality of life, including the BPI-sf item 3 tool, the EORTC QLQ-PR25 urinary symptoms metric, as well as that same metric for bowel symptoms, the FACT-P total score, and finally the European quality of life visual analog scale. So, you can see an early separation of the curves for all of these quality of life metrics favoring enzalutamide.
In the initial publication in 2018, which was the primary analysis of PROSPER, there were 23 months of follow-up, but the OS data was immature and the median OS was not reached. At this point in time, they had 165 deaths, which was 28% of the pre-specified survival events for the final analysis. The publication that was published in June of 2020 is the survival analysis with longer follow-up in the PROSPER trial, which we'll be discussing today.
This was a multinational, double-blind, randomized, placebo-controlled, Phase III trial of 300 sites and 32 countries. The key inclusion criteria were confirmed prostate adenocarcinoma with an increasing PSA despite a castrate level of testosterone, a baseline PSA level of greater than two nanograms per milliliter, a PSA doubling time of less than 10 months, and no evidence of metastatic disease.
In terms of their methods, stratification was based on PSA doubling time of less than or greater than six months. They also stratified based on previous or current use of bone-targeting agents, and randomization was two to one for ADT plus enzalutamide versus ADT plus placebo. Their primary endpoint was MFS or death, which, as I mentioned before, was reported in their 2018 New England Journal of Medicine article, and there were several secondary endpoints as you can see here, one of which was overall survival, which was the key endpoint discussed in their 2020 update.
Their efficacy endpoint was in the ITT population. They evaluated safety, which was described as time from the first dose to 30 days after the last dose or to the day before initiation of a new antineoplastic therapy. For their final analysis of overall survival, they needed 590 deaths to provide 85% power to detect a hazard ratio of 0.77 with a two-sided significance level of 0.05. This was the third preplan analysis for overall survival. After adjusting for multiplicity, they detected if there was a P value of less than or equal to 0.021, this indicated statistical significance.
Christopher Wallis: Thanks for the setup, Zach. I'll talk now about the results of the updated overall survival analysis and help put these in a little bit of context.
The original publication of the PROSPER trial described some of these initial background features, but we'll just review them quickly. Patients were screened and accrued between 2013 and 2017, with enrollment stopped after about 450 MFS events, giving adequate power for the primary outcome. There was two to one randomization schema to enzalutamide versus placebo and, as Zach described before, this was stratified on the basis of receipt of prior bone-targeting agents and PSA doubling time. As we might expect from the fact that enzalutamide prolonged metastasis-free survival, the median duration of treatment was longer for patients who were on therapy than on placebo.
Now, these baseline characteristics come from the initial publication, but as we would expect, they ought to be the same in the updated analysis, and so we see a fairly typical prostate cancer population, the median age in the mid-70s, good performance status, median PSAs in the range of 10, and a doubling time in the range of four months. Bone-targeted therapy was used infrequently in only about 10% of the population at baseline.
Now, following primary analysis of the metastasis-free survival endpoint, trial data were unblinded. And at that time, given that it demonstrated benefits of enzalutamide in prolonging metastasis-free survival, patients who were receiving placebo who hadn't yet progressed, that is who were still nmCRPC, were allowed to cross over. So, among 114 eligible patients randomized to placebo who had not progressed, 87 of those crossed over to enzalutamide. And this, as you might imagine, affects our ability to detect overall survival differences in an ITT analysis.
Despite the crossover, as we see here, overall survival was significantly improved for patients receiving enzalutamide compared to placebo. With a median follow-up of 48 months, we see a median survival of 67 months in those patients on therapy and 56 months in those initially randomized to placebo. As a result, the hazard ratio is 0.73 and a P-value of 0.001. As Zach alluded to before, due to multiplicity, the threshold for statistical significance here was 0.021, so this reaches statistical significance.
In subgroup analyses, we see a relatively consistent effect across subgroups defined based on geographic region, patient age, performance status, PSA doubling time, as well as absolute PSA values, the use of bone-targeting agents, Gleason Score, LDH values, and hemoglobin values, indicating that there are no specific subgroups for whom a greater or lesser benefit would be expected.
As of the data cutoff of October 15, 2019, about 60% of patients who were randomized to enzalutamide had subsequently stopped therapy, and a reasonable portion of those who had crossed over later had also stopped. Treatment was stopped for both disease progression and adverse events, and we'll talk about that a little bit here.
So, the next lines of therapy, following progression or discontinuation for toxicity, subsequent therapy was used in about one-third of the patients randomized to enzalutamide and about two-thirds of those randomized to placebo. Abiraterone was the most commonly used agent in patients who were initially randomized to placebo, where docetaxel was the most commonly used agent in those initially randomized to enzalutamide. And this makes sense given what we know from other work from Kim Chi's group as well as the CARD study, suggesting that failure on one androgen receptor-targeting agent should lead us to consider cytotoxic chemotherapy rather than an androgen receptor-targeting switch.
In terms of adverse events, this is a summary of overall any adverse events and serious adverse events. We see that these are quite common in both groups, although somewhat higher in the enzalutamide-treated population with a 94% rate of any adverse events versus 82% in the placebo group. Serious events, again, were more common in the enzalutamide group at 48% as compared to 27% in the placebo group, and adverse events leading to treatment discontinuation occurred in 17% of those in the enzalutamide group and only nine in the placebo group. Deaths were uncommon in both, but slightly more prevalent in the enzalutamide group.
Based on data from the use of enzalutamide in other disease states, including metastatic castration-resistant prostate cancer, there are specific adverse events of interest, including fatigue or asthenia and seizures, and these are summarized here. Fatigue is somewhat more common in the enzalutamide group with a proportion of 46% as compared to 22 in the placebo group. However, very few seizures were noted in the study population.
Now, despite the fact that we've been able to demonstrate a statistically significant difference in overall survival, these data are still relatively immature. And as we see from data on the use of enzalutamide in early metastatic castration-resistant prostate cancer, as data matures, we may expect changing effect estimates. And so, these show that from the initial analysis in 2014, 22 months through extended analysis published in 2017 of 31 months, and now onto 70-month median follow-up, the hazard ratio of the benefit of enzalutamide in early metastatic disease has moved towards one, although remains statistically significant, so Dr. Beer highlighted that this may occur in the non-metastatic space as well when he was discussing this paper at the ASCO 2020 virtual meeting.
In conclusion, an updated analysis of the PROSPER trial shows improved overall survival for patients with nmCRPC who receive enzalutamide as compared to placebo. There are now three agents approved for non-metastatic CRPC on the basis of metastasis-free survival improvements, and each of these has now demonstrated subsequent proven benefits in overall survival. The specific magnitude of benefit is likely to change as the data matures, but these data provide encouraging information for our patients, suggesting that early use of androgen-targeting agents in patients with castration-resistant disease is likely to provide a survival benefit.
I'd like to thank you for your time and hope that this session has proved both interesting and informative for you.