PRIMA: captivated by the wizard of OS?

다국어 초록 (Multilingual Abstract)

Poly(ADP-ribose) polymerase (PARP) inhibitors have transformed the first-line treatmentlandscape for advanced high-grade tubo-ovarian cancer. Endorsed by internationalguidelines, they are now the standard maintenance therapy for patients with BRCA1/2mutations or homologous recombination (HR) deficient tumors, either alone or incombination with bevacizumab. In addition, for patients with HR-proficient tumors,they offer an alternative to bevacizumab. PARP inhibitors have also driven the adoptionof biomarker testing (BRCA1/2 and homologous recombination deficiency [HRD]) as aprerequisite to guide first-line systemic therapy decisions [1].
Strong evidence from pivotal phase 3 trials—SOLO-1, PAOLA-1, PRIMA, and ATHENAMono—has led to the approval of olaparib, niraparib, and rucaparib as monotherapy for first-line maintenance and olaparib also in combination with bevacizumab [2-9].
These studies consistently demonstrate a substantial and persistent progression-free sur vival(PFS) benefit, ranging from 19% to 29% at 5 years for HR-deficient tumors (18% at 4 yearsin ATHENA), with hazard ratios between 0.33 in SOLO1 (BRCA mutated [BRCAmut] tumorsonly) and 0.51. The Kaplan-Meier cur ves run flat at this point and indicate a sustainedbenefit, showing no signs of converging, underscoring the enduring advantage of PARPinhibitors for these patients.
However, the benefit is less sustained in the HRD-negative population, and this discussioncenters on data for the HRD-positive patients.
As data matures, attention shifts to overall sur vival (OS) outcomes. Whilst mature OS datafrom ATHENA Mono and PRIME is still awaited, PRIMA, the latest study to release final OSdata, failed to demonstrate any OS benefit from niraparib in the overall as well as the HR-deficient subpopulation (hazard ratio=1.01; 95% confidence inter val [CI]=0.84–1.23; p=0.88and hazard ratio=0.95; 95% CI=0.70–1.29, respectively) [10]. These findings have sparkeddebate in the gynecologic oncology community, contrasting with more favorable outcomesreported in SOLO-1 and PAOLA-1 [3,6] and potential explanations are being sought.
SOLO-1, in a prespecified interim analysis at 7 years of median follow-up and a data maturityof 38.1%, has demonstrated a 21.5% difference in OS, a key secondar y endpoint, at 7 yearsin exclusively BRCA mutant tumors (hazard ratio=0.55; 95% CI=0.40–0.76; p=0.0004 [not significant]). The final analysis is pending, and patient selection excluded those with BRCAwildtype tumors [3].
In PAOLA-1 OS was a key secondar y endpoint too and as for the primar y endpoint, OS wasformally assessed in the intention-to-treat (ITT) population, with additional subgroupanalysis by BRCA- and HRD-status.
After a median follow-up duration of 61.7 months and a data maturity of 55% the addition ofolaparib to bevacizumab did not significantly improve OS (hazard ratio=0.92; 95% CI=0.79–1.12; p=0.41) in the ITT population at the final analysis. Within the HRD-positive subgroupthere was a meaningful trend for an improved OS in favor of olaparib with a difference in 5-yearOS rates of 17.1% (hazard ratio=0.62; 95% CI=0.45–0.85), however, the analyses by biomarkersubgroup are explorator y and no formal statistical testing was performed (Table 1) [6].
All of these trials used OS as a secondar y endpoint, and none provide unequivocal statistical proofof improved OS with first-line PARP inhibitor maintenance in advanced tubo-ovarian cancer.
While acknowledging the significant limitations of cross-trial comparisons and the lack ofstatistical evidence for OS benefits across the trials, it is difficult to disregard the clinicallymeaningful numerical differences in OS rates obser ved in SOLO-1 and PAOLA-1, particularlywithin biologically relevant subgroups, which were not evident in PRIMA. Therefore, exploringpotential explanations for the presumed discrepancies in OS among these 3 trials is worthwhile.
One possible explanation for PRIMA’s failure to show an OS bene...