Significantly Longer Metastasis-Free Survival and Overall Survival in the ARAMIS Trial, Nonmetastatic Castration-Resistant Prostate Cancer - Christopher Wallis & Zachary Klaassen

September 30, 2020

In this Journal Club, Christopher Wallis and Zachary Klaassen discuss the updated analysis in the most recent publication of the ARAMIS trial. In this publication, overall survival (OS) and all other secondary endpoints were evaluated. The ARAMIS trial looked at the use of darolutamide for delaying metastasis and death in men with nonmetastatic, castration-resistant prostate cancer. Drs. Wallis and Klaassen highlight several important discussion points gathered from this updated analysis including the OS survival data for ARAMIS showed that metastasis-free survival (MFS) and OS were significantly longer with darolutamide than placebo among men with nonmetastatic, castration-resistant prostate cancer.


Christopher J.D. Wallis, MD, Ph.D., Instructor in Urology, Vanderbilt University Medical Center, Nashville, Tennessee

Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center

Read the Full Video Transcript

Chris Wallis:  Hello, and thank you for joining us for this UroToday Journal Club.  We are discussing the updated results of the ARAMIS trial, recently published in The New England Journal of Medicine. The article entitled, "Nonmetastatic, Castration-Resistant Prostate Cancer and Survival with Darolutamide". I'm Chris Wallis, a Fellow in Urologic Oncology at Vanderbilt and along with me today is Zachary Klaassen, an assistant professor in Urology at The Medical College of Georgia. 

Here is the citation for the paper we're discussing today. Dr. Fizazi leading the team of ARAMIS investigators, looking at darolutamide in patients with nonmetastatic castration-resistant prostate cancer. So by way of background, darolutamide differs somewhat from the other androgen-receptor inhibitors. It is an oral high-affinity agent, but notably, it has activity against the androgen receptor mutants for which there is resistance to other agents. Additionally, as a result of minimal penetration of the blood-brain barrier, there are some lower risks and concerns regarding seizure activity with the use of this agent compared to, for example, enzalutamide.

The key registration trial for darolutamide in this disease space was the ARAMIS trial in the initial publication, focused on metastatic progression-free survival. And so this was a relatively new FDA approved endpoint at that time, but it was used for registration on the basis of demonstrated improvements in metastasis-free survival with a median metastasis-free survival of 40 months among treated patients and of 18 months among those receiving placebo with a hazard ratio of 0.41, indicating a 60% improvement in metastasis-free survival. And this was obviously strongly statistically significant.

And so, as of July 30th, 2019, darolutamide was granted fast-track approval for the treatment of nonmetastatic CRPC on the basis of these improvements in metastasis-free survival. However, at the time of that initial publication, there was an interim overall survival analysis. At that point, 136 deaths in the overall cohort of approximately 1500 men had been observed. And so, as you can see here, median overall survival had not been reached in either the darolutamide or placebo arm. However, there was an indication of improved overall survival among men receiving active agents with darolutamide with a hazard ratio at that point of 0.71.

However, the updated analysis presented here is a more mature version of those data. And so to briefly review the methodology of this trial, men were included if they had non-metastatic CRPC, baseline absolute PSA level of at least two nanograms per milliliter, a PSA doubling time of fewer than 10 months, and a performance status of zero or one. Notably, unlike trials assessing enzalutamide or apalutamide, patients with seizures or seizure predisposing conditions were not excluded from the ARAMIS trial. So the generalized ability of this trial is potentially a little bit larger as the inclusion criteria are somewhat broader. As we discussed, the primary outcome of the ARAMIS trial was metastasis-free survival with a number of key secondary and exploratory endpoints.  Key secondary outcomes included overall survival, time to pain progression, time to first cytotoxic chemotherapy, and time to the first symptomatic skeletal-related event. Exploratory endpoints included the time to first prostate cancer-related procedure and the time to the next prostate cancer-directed therapy.

So, this was a 2-1 randomization design with twice as many patients allocated to darolutamide as to placebo as a double-blind design. And all patients continued on androgen deprivation therapy, typically in the form of LHRH agonists. Randomization was stratified by PSA doubling time, so while all men included had to have a PSA doubling time of fewer than 10 months, those with notably shorter doubling times, that is less than six months, were stratified against those with the PSA doubling time between 6 and 10 months. Additionally, randomization was stratified by the use of osteoclast-targeting agents at the time of randomization. And at the time of primary analysis for metastasis-free survival endpoint, randomization was unblinded and men in the placebo arm who had not progressed were allowed crossover to the darolutamide arm.

So this briefly summarizes the study schema, again, men with non-metastatic CRPC and short PSA doubling times, were randomized, darolutamide or placebo along with continued ADT and following unblinding at the time of analysis for metastasis-free survival patients continued to follow up to assess overall survival, which is what we're reporting here.

This looks at the patient flow, and as we can see, our 1,509 men who were randomized, just over 950 were allocated to darolutamide and the vast majority have received their allocated therapy and 550 or so were allocated to placebo and all received their allocated therapy.

There were relatively few protocol deviations, 13 in the intervention arm, seven in the control arm. The statistical analysis plan is slightly complicated here, but just to summarize, the authors based, on their power calculations, plan for final overall survival analysis at the time of 240 deaths or 240 events and they undertook the hierarchical testing of all these secondary endpoints using alpha splitting. And so, overall survival was the first of these hierarchically tested endpoints within overall survival. The authors conducted a Kaplan-Meier analysis using a stratified log-rank test to test for statistically significant differences in overall survival and then used Cox proportional hazards models to estimate the hazard ratios.  The efficacy analysis here of overall survival was taken into the intention to treat population and safety analysis was performed among all individuals who received at least one dose of their allocated therapy. I'll now transition to Dr. Klaassen to discuss the results and implications of this trial.

 Zachary Klaassen:  Thanks, Chris. So, at the time of the data cutoff, 49% of patients originally assigned to darolutamide were still receiving darolutamide. And as Chris mentioned, at the time of unblinding, the people in the placebo group were given darolutamide.  At the time of the data cut off, 86% of these patients in the placebo group were still receiving darolutamide. As you can see here in Table 1, this is "Life-Prolonging Therapy in the Intention-to-Treat Population." You can see that discontinued assigned treatment, because these patients all crossed over to darolutamide, they were 100% discontinued from the placebo group and 51% in the darolutamide group. And you can see the sort of life-prolonging therapies subsequent to the trial. You can see docetaxel, 14% in the placebo group, 9% in the darolutamide group, as well as abiraterone, enzalutamide and a few patients receiving sipuleucel-T.

This is the main figure for this presentation, which is the overall survival result. You can see here that the number of patients or the percentage of patients alive at 3 years was 83% for darolutamide, 77% for placebo, corresponding to a hazard ratio for death of 0.69, and a confidence interval of 0.53 to 0.88. And you can see here that this is quite early splitting of these curves, favoring darolutamide.

The results of the Subgroup Analysis show that the majority of subgroups did favor darolutamide. You can see here that the PSA doubling time of greater than, and less than six months was significant for darolutamide. You can also see here that a PSA at baseline less than 10, PSA at baseline below the median, Gleason score greater than seven, less than seven. Interestingly, in the younger patients, these will be important to follow in subsequent analyses, as there was not a benefit among men less than 65 years of age, but there was among older men in the subgroup analysis as well. Other topics of interest you can see, ECOG 0 was significant for darolutamide. ECOG 1 was statistically insignificant but favoring darolutamide and some other interesting signals that we may have interest in follow-up analysis as well, is the racial component, certainly favoring white men was darolutamide, including a strong signal in Black or African-American men. Additionally, there was no benefit for darolutamide for Asian men. So, the majority of these subgroup analyses favoring darolutamide with some interesting points which I just mentioned, which will be interesting to see and follow up on how these play out.

So the results of the secondary and exploratory endpoints you can see here. This is basically a clean sweep for darolutamide. As we mentioned at the top here, overall survival favoring darolutamide. The other endpoints, time to pain progression favoring darolutamide with a hazard ratio of 0.65 and a statistically significant confidence interval. Similarly, time to first use of cytotoxic chemotherapy with a hazard ratio of 0.58, time to the first symptomatic skeletal event, a hazard ratio of 0.48, and then to summarize at the bottom here, the exploratory endpoints, time to a first prostate cancer-related invasive procedure, strongly favoring darolutamide, as well as time to initiation of subsequent antineoplastic therapy, a hazard ratio of 0.36.

This is the Kaplan-Meier curve for time to the first chemotherapy. As we mentioned on the last slide, a hazard ratio of 0.58, 95% confidence that rose 0.44 to 0.76, and similar to the overall survival curve, darolutamide with an early split in the curve, significantly favoring time to first chemotherapy. Similarly, with time to the symptomatic skeletal event, a hazard ratio of 0.48 favoring darolutamide with a statistically significant confidence interval, not quite as early split of the curve as, and not quite as dramatic, but still statistically favoring darolutamide.

In terms of the adverse events, you can see darolutamide is on the left and placebo is on the right. There was only one adverse event that was more than 10% of the patients, any grade, and that was fatigue for darolutamide at 13.2%. But you can see when you look at the  grade 3 or 4, adverse events for fatigue, comparing the two, they're quite similar at 0.4% for darolutamide and 0.9 for placebo. Just to summarize this table, you can see, looking at this briefly that the adverse event profile for darolutamide was very comparable to the placebo arm.

In terms of an overview of the summary of adverse events, any great adverse event for darolutamide was 85.7 compared to 79.2 in the placebo group. So very well tolerated.  Slightly more grade 3 adverse events.  Grade 3 or 4 adverse events and darolutamide of 26.3 compared to 21.7, comparable grade 5 adverse events as well as serious adverse events. So in summary, these last two slides, darolutamide, in addition to ADT versus ADT, continuing alone, in the placebo group, was very well tolerated.

In terms of the results of the Summary of Death, you can see here that any death was 15.6% in the darolutamide group and 19.1 in the placebo group, fewer deaths from prostate cancer at 8.4% for darolutamide tempo, and 1% for placebo. And you can see here very comparable, specific deaths, cardiovascular disease, respiratory disease, cerebrovascular disease between these two groups.

So I think this trial brings up several important discussion points. And first and foremost, this was an OS benefit for darolutamide with a 31% lower risk of death. This is important because it was observed in this trial despite more than 50% of the patients in the placebo group receiving darolutamide or other life-prolonging therapy. So, even with these placebo patients subsequently becoming unblinded, receiving darolutamide and other therapy, there was still a 31% lower risk of death in the darolutamide group. As I mentioned in the previous slides, after 29 months of follow-up, there was a favorable safety profile, which was comparable to their ARAMIS group's initial metastasis-free survival publication. And finally, the percent of patients who discontinued therapy due to adverse events was also unchanged from the primary analysis. And it also showed that patients with non-metastatic CRPC can receive therapy for prolonged periods of time.

So in conclusion, the OS survival data for ARAMIS showed that MFS and OS were significantly longer with darolutamide than placebo among men with nonmetastatic, castration-resistant prostate cancer. The overall survival benefit was achieved even though more than half of the patients in the placebo group had received subsequent therapy. As I showed in the tables and figures, secondary endpoints all favored darolutamide, and the incidence of adverse events was similar in the treatment and the placebo groups. We hope you enjoyed this UroToday Journal Club on the ARAMIS overall survival data and we thank you for your attention.