Significantly Longer Duration of Overall Survival with Olaparib in Metastatic Castration-Resistant Prostate Cancer (mCRPC) - PROfound Study Journal Club - Zachary Klaassen and Christopher Wallis

October 16, 2020

In this Journal Club, Zachary Klaassen and Christopher Wallis present on the recently published New England Journal of Medicine article Survival with Olaparib in Metastatic Castration-Resistant Prostate Cancer, data from the Phase 3 PROfound study.  The primary outcome of the study was radiographic progression-free survival in cohort A, which was focused on BRCA1, BRCA2, and ATM mutant patients which were previously reported.  The results of the final analysis of overall survival had not yet been reported and it is these data being discussed by Drs. Wallis and Klaassen.  They highlight several discussion points from this paper. 


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. Today, we're talking about a recently published article from the PROfound trial, entitled, "Survival with Olaparib in Metastatic Castration-Resistant Prostate Cancer". I'm Chris Wallis, a Fellow in urological oncology at Vanderbilt and with me today is Zach Klaassen, an Assistant Professor at the Medical College of Georgia.

Here's a citation for the article we will be discussing. Again, this is the overall survival analysis of the PROfound trial led by Dr. Hussain.

By way of a bit of background, many UroToday readers will know that DNA repair mutations are increasingly important for patients with advanced prostate cancer. And these are important because they offer specific therapeutic targets. And so PARP inhibitors act particularly efficiently in patients who have DNA repair defects. And so this demonstrates the mechanism of these inhibitors. As you can see, stress and reactive oxygen species can cause DNA damage and the standard DNA repair pathway uses PARP as well as recruitment of other DNA repair enzymes to allow DNA repair and ongoing cellular viability. PARP inhibitors interact with the PARP complex and result in a so-called PARP trapping where there is an inability to release PARP, as well as an inability to recruit repair enzymes. And this leads to DNA instability and eventual cell death.

And so early data first in a Phase II setting looked at 50 patients with extensively pretreated mCRPC who received olaparib. And as you can tell in the rPFS, OS curves down at the bottom right. These metrics were significantly improved with the receipt of olaparib only among patients who had a biomarker positivity, whereas a biomarker negative patients had a significantly worse outcome.

And so this then led to the TOPARP-B trial, which was an open-label randomized Phase II trial. Again, in heavily pretreated patients with these DNA repair mutations. And this is a bit of a dose-finding study with randomization to 300 or 400 milligrams. And as you can see, overall, the composite response was about 45% with some variation between the dose cohorts of around 39% in the 300 milligrams and 54% in the 400-milligram group. And this led to 400 milligrams being the approach used going forward. And as you can also see by gene subgroups, there are significantly higher response rates in patients with BRCA1 and BRCA2 than other gene mutations.

And so this led to the rationale for the PROfound trial. And so this study recruited men with metastatic CRPC, who had had prior androgen-signaling agent, that is abiraterone or enzalutamide. Patients had to have one of 15 relevant homologous recombination repair mutations and then the cohort was subsequently stratified with a focus on BRCA1, BRCA2, and ATM mutations. Patients who were enrolled were randomized to receive olaparib as their next line of therapy or an androgen-signaling agent switch. That is a switch from abiraterone to enzalutamide or vice versa. The primary outcome of the study was radiographic progression-free survival in cohort A, which was focused on a BRCA1, BRCA2, and ATM mutant patients.

And so based on the primary outcome of radiographic progression-free survival here, we see a significant improvement in this metric for patients receiving olaparib compared to an androgen-signaling agent switch. And a secondary outcome on the initial publication for overall survival had fairly immature data. As you can see here, data maturity is only 38%. However, there is evidence of fairly early splitting of the curves, though the effect did not reach significance as the alpha significance threshold here is 0.01 due to alpha splitting.

We now get to the more mature data looking at overall survival. Again, to review inclusion criteria, these are adult men with mCRPC who have progressed on prior abiraterone or enzalutamide and may or may not have had prior taxane therapy. They had to have one of 15 pre-specified homologous recombination repair defects with a focus on BRCA1, 2, and ATM.

There was biomarker-driven stratification with this cohort A, among these focused set of genes and cohort B comprising an alteration to any one of the remaining 12 genes. Patients were randomized in a two to one fashion to olaparib versus androgen-axis targeting switch and this was further stratified according to taxane use and measurable disease. Treatment was continued until progression or toxicity and there was crossover allowed for patients receiving abiraterone or enzalutamide who had progression.

As we already talked about, the primary outcome was rPFS and key secondary outcomes included OS and again, most of these analyses were conducted in cohorts split, but there was also a pooled analysis and there was a pre-specified sensitivity analysis designed to explore the effect of crossover on the apparent benefit of olaparib on overall survival. And this is relevant because crossover was relatively common with upwards of 55% of patients overall and over 60% in cohort A switching to olaparib.

This just summarizes the study design as we've already discussed. And it's important to focus particularly as we highlighted before, on the alpha splitting on these statistical analyses. And so this utilized a hierarchical analysis strategy with dependence on outcomes in cohort A, that is patients with mutations in BRCA1, 2, and ATM. And so again, there was a hierarchy of outcomes and overall survival in cohort A here being the analysis we are focusing on today. And we've already talked about pre-specified sensitivity analysis for crossover.

And I will now hand it over to Dr. Klaassen to talk about the results and implications of these data.

Zach Klaassen: Thanks, Chris. The results in terms of the characteristics at baseline, these are certainly the same from the previous publication as the patients haven't changed. But just by way of review, 4,425 patients were screened, 4,047 had tumors available for testing, of which 2,972 had tumors tested with biomarkers reported, 778 had patients that qualified based on their genetic profile, and among those patients, 387 patients met eligibility and were randomized. You can see here in table one, the baseline characteristics taken from the original paper that the age was basically in the late sixties for the median. There was a preponderance of patients with BRCA2 mutations, as one would expect. You can see here that about one-third of patients had visceral metastatic disease. Most of these patients were of good performance status at ECOG zero or one. And you can see here, the split between the previous enzalutamide only, abiraterone only, or combination of enzalutamide and abiraterone at about 20% in each group.

Looking a little more into the genetic alterations, I'll draw your attention here to the BRCA2 in cohort A, it was more than half in the olaparib arm and 62% in the control arm. And you can also see that there was a good deal of patients in cohort A that comprised ATM mutation. Almost 40% in the olaparib arm in cohort A and 31.3% in the control arm of cohort A.

Moving to the results, this paper primarily was focusing on overall survival and they analyzed this in several ways. The first was overall survival in cohort A. This is patients with BRCA1, BRCA2, or ATM mutations. You can see here that in the olaparib arm on the top left, the median overall survival was 19.1 months with a median overall survival in the control arm of 14.7 months. The hazard ratio for death of 0.69 and a 95% confidence interval of 0.50 to 0.97 in the left.

On the right side, this was patients with regards to crossover adjusted analysis of the overall survival cohort. This is 67% of patients that crossed over. This is a sensitivity analysis on the bottom right and this showed once again, an early split of the curves, favoring olaparib with a hazard ratio for death of 0.42, 95% confidence interval of 0.19 to 0.91. Both of these analyses significantly benefited the olaparib group with regards to overall survival.

Moving to cohort B, which is the patients with the 12 other genes, not including BRCA1, 2, and ATM, you can see here that there was no difference in overall survival. Once again, on the left, this is the overall survival in cohort B. Median overall survival in the olaparib group was 14.1 months, in the control group, it was 11.5. A hazard ratio of death was 0.96 with a non-significant 95% confidence interval. And we see very similar results in the crossover adjusted analysis in the bottom right of the screen with a hazard ratio of 0.83.

And finally, what they did was they combined overall survival analysis in cohorts A and B. You see for the olaparib group in the top left the median overall survival was 17.3 months and the median overall survival in the control group was 14 months with a hazard ratio of death non-statistically significantly favoring olaparib. A hazard ratio of 0.79 but 95% confidence interval 0.61 to 1.03. Once again, very comparable results to the primary analysis in terms of the crossover adjusted analysis, the hazard ratio 0.55, 95% confidence interval of 0.29 to 1.06.

They did several subgroup analyses here as well. What you can see in this plot is a subgroup analysis in cohort A and I will draw your attention to the asterisks as these are the statistically significant findings for overall survival in this group. As we discussed previously, all patients did have a benefit of olaparib versus the control group. Patients that previously received taxane chemotherapy also benefited from olaparib with a hazard ratio of 0.56 and a confidence interval of 0.38 to 0.84. And to summarize the rest of this table, you can see here that ECOG 1 status significantly favored olaparib as did patients that were treated in Europe.

A similar plot for the overall population. This is a combined cohort A and B, showed that once again, a signal here for patients that had prior taxane chemotherapy with a significant benefit for olaparib with a hazard ratio of 0.66 and a confidence interval 0.49 to 0.91. And finally, the other significant result from this plot is patients benefited from olaparib that had PSA at baseline greater than the median.

And finally, what they did here was they looked at the results by alterations in single HHR gene deletions. And interestingly enough, this is a strong signal for BRCA2. As you can see here the hazard ratio of 0.59, 95% confidence interval of 0.37 to 0.95. And really this is probably what is driving the majority of these results. As you can see that other than a BRCA1, which had a hazard ratio of 0.42 with a non-significant 95% confidence interval, the rest of these genes either had too few patients to run analyses or were not statistically significant.

Just to summarize the adverse events in the overall population and compare that to patients at the crossover, to summarize, essentially there were no new signals from the primary analysis to this one. You can see the adverse event, all grades were 96% for olaparib, 88% in the control group, and 93% among those patients that crossed over. The most common adverse events in the olaparib group were anemia at 50% of the patients all grade, nausea 43%, and fatigue and asthenia at 42%. And you can see here that the patients that crossed over had a very comparable adverse event profile as to those that received olaparib upfront.

Several discussion points from this paper, we saw in PROfound, overall survival analysis that the risk of death was 31% lower with olaparib than with the control therapy, despite a substantial crossover from the control therapy to olaparib. And we found that patients with BRCA1 or BRCA2 alterations, specifically BRCA2, derived the greatest benefit from olaparib with respect to overall survival. And finally, as I mentioned on the previous slide, the trial has found no difference in the safety profile as what was originally reported. This was consistent with the primary analysis and there were no cumulative toxic effects observed.

In conclusion, PROfound found that among heavily pretreated patients with at least one alteration of BRCA1, BRCA2, or ATM, olaparib led to significantly improved overall survival. As we have shown previously, this was likely driven primarily by the BRCA2 responses. There is an OS benefit for olaparib versus next-generation hormonal agents, despite a substantial crossover from control therapy to olaparib. The previously observed adverse events were consistent despite longer follow-up in the subsequent analysis and ultimately the PROfound control group in this setting likely received substandard care given that there have been previous studies showing that the androgen receptor switch therapy has overall been unsuccessful.

We thank you for your attention and for attending this UroToday Journal Club.