The final analysis showed a significant difference in second progression-free survival (PFS2) with niraparib versus placebo, but only in patients with germline BRCA mutations.
The overall survival analysis was limited because the study was not powered to detect differences in OS. Still, the investigators found a trend toward better OS with niraparib in patients who had germline BRCA mutations.
These results were presented at the Society of Gynecologic Oncology’s Virtual Annual Meeting on Women’s Cancer (Abstract 11139).
The final analysis of this study was muddied by missing data for many of the 553 patients originally randomized, according to Ursula Matulonis, MD, of the Dana-Farber Cancer Institute in Boston, who presented the study’s results at the meeting.
“Since 2014, approvals of PARP [poly (ADP-ribose) polymerase] inhibitors worldwide in different lines of ovarian cancer therapy have revolutionized the treatment of women with ovarian cancer,” Dr. Matulonis said. “During the time from the initial results of the NOVA study in 2016 to this final analysis, PARP inhibitors have become available commercially globally.”
The wider availability of PARP inhibitors led to premature unblinding and withdrawals from the trial, resulting in missing data for a large proportion of patients.
Initial results
In the primary analysis of the trial, the median PFS was 21 months among patients with germline BRCA mutations who received niraparib and 5.5 months for patients with germline BRCA mutations who received placebo (P < .001).
Among patients with no germline BRCA mutations, the median PFS was 9.3 months for those who received niraparib and 3.9 months for placebo-treated controls (P < .001).
At the time of the primary PFS analysis in 2016, 155 of the 553 patients originally enrolled had left the study for reasons other than death.
In both treatment cohorts, a large number of patients were unblinded as to their treatment assignments by investigators, which required them to withdraw consent from the study under the protocol, resulting in the aforementioned loss of some data on subsequent therapy and survival status.
Final results: PFS2 and OS
At the time of the final data lock on Oct. 1, 2020, the mean follow-up was 67 months (5.6 years).
In all, 64 patients originally assigned to niraparib were still on study (28 in the germline BRCA mutation cohort and 36 in the nonmutated cohort), as were 28 patients originally assigned to placebo (9 and 19, respectively).
For the final analysis, the investigators were able to retrieve survival data on 51% of the patients originally enrolled, either through protocol amendments and obtaining consent again or through national databases.
Although crossover to the niraparib arm was not allowed on study, patients could receive a PARP inhibitor following disease progression or study withdrawal. Because of discontinuations, data on postprogression therapy were missing for 138 of the 553 patients (25%).
The hazard ratio for PFS2 with niraparib versus placebo was 0.67 for patients with germline BRCA mutations (95% confidence interval, 0.479-0.948) and 0.81 for patients without the mutations (95% CI, 0.632-1.050).
In an analysis adjusted for subsequent PARP inhibitor therapy, there was no significant difference in median OS between the niraparib and placebo arms, regardless of mutation status. In the cohort without BRCA mutations, the median OS was 31.3 months with niraparib and 35.9 months with placebo (HR, 0.97; 95% CI, 0.74-1.26).
However, there was a trend toward improved OS with niraparib among patients with germline BRCA mutations. The median OS was 43.8 months with niraparib and 34.1 months with placebo (HR, 0.66; 95% CI, 0.44-0.99).