PARPs: All for One and One for All? Enhancing Diversity in Clinical Trials

In this issue of Clinical Cancer Research, Gao and colleagues present the results of the L-MOCA study, a nonrandomized trial of maintenance olaparib given following chemotherapy to 225 Asian women with recurrent, platinum-sensitive, high-grade ovarian cancer (1). It cannot be assumed that, worldwide, all patients with cancer handle treatment similarly. This was exemplified in the JGOG 3016 study which showed that dose-dense paclitaxel-containing chemotherapy was superior to standard at prolonging survival in Japanese women with newly diagnosed ovarian cancer; a benefit that was not reproduced in the equivalent European study, ICON8. A likely pharmacogenomic explanation is that polymorphisms in the transporter gene ABCB1 (also known as permeability-glycoprotein), enriched in Asian populations, allow higher and more intense paclitaxel dosing.

Metabolism of PARP inhibitors may also be subject to pharmacogenomic variability. Originally developed to induce synthetic lethality in germline BRCA 1 or 2 mutation (gBRCA1/2)-driven tumors, olaparib, niraparib, and rucaparib are PARP inhibitors that are licensed for use in primary or recurrent ovarian cancer. Each has a different affinity for the PARP1–3 isoforms and are metabolized either by carboxylesterases (niraparib) or cytochrome P450 (CYP) enzymes such as CYP3A4/5 (olaparib) and CYP2D6/1A2/3A4 (rucaparib). When looking specifically at the four studies that led to their approval by FDA and European Medicines Agency as postchemotherapy maintenance for platinum-sensitive recurrent ovarian cancer: Study 19, SOLO2/ENGOT-OV21, ENGOT-OV16/NOVA, and ARIEL3 (2–5), the three PARP inhibitors appear to have similar levels of efficacy but distinct toxicity, pharmacodynamic, and pharmacokinetic profiles. However, a lack of information about the ethnicity of participants limits further exploration or meta-analysis within specific ethnic or racial cohorts. In addition, where these data have been presented, the racial uniformity of participants—80% and 95% were defined as “White” in ARIEL3 and Study-19, respectively—raises concerns that ethnic groups, particularly Asian and Black/Afro-Caribbean women are underrepresented and reflects the concentration of trial centers in Europe and North America (6). In SOLO2, of the 126 study locations involved, 29 were centers in China, Japan, or Korea, yet ethnicity information was not included in the patient characteristics, nor in the analyses (summarized in Table 1).

To address this knowledge gap, the NORA trial, conducted in 30 centers in China, was the first phase III randomized trial to demonstrate the safety and activity of maintenance niraparib in platinum-sensitive relapse in an exclusively Asian cohort. Responses were compared with the NOVA trial, conducted in a similar clinical context with a majority (87%) of White patients. Despite differences in the statistical design of the two trials, there were comparable outcomes in terms of reduction in risk of disease progression or death among the BRCA1/2 mutation subgroups (NORA: HR = 0.22; NOVA: HR = 0.27) and gBRCA1/2 wild-type (wt) group (NORA: HR = 0.40; NOVA: HR = 0.45).

Using a similar strategy to NORA, the L-MOCA investigators explored maintenance olaparib in an exclusively Asian population (91% from China and 9% from Malaysia) with recurrent, platinum-sensitive ovarian cancer. The existing approval for olaparib in this indication meant they could not institute a placebo-control arm. Instead, L-MOCA was run a single-arm trial, drawing on data from randomized PARP inhibitor studies for comparison. Patients with gBRCA1/2 who received maintenance olaparib in the L-MOCA trial had similar median progression-free survival (PFS) as those in SOLO2 (olaparib) and with rucaparib in ARIEL3 (21.2 months in L-MOCA compared with 19.1 months in SOLO2 and 16.6 months in ARIEL3). Median PFS among BRCA 1/2 wt patients in L-MOCA was 11 months (8.3–15.8), which compared favorably with the 9.3-month PFS in nongermline BRCA/non–homologous recombination deficient (HRD) patients treated with niraparib in NOVA and 7·4 months for the wt BRCA1/2 patients treated with olaparib in Study 19. While it would appear that olaparib has similar efficacy, toxicity was less equivalent; grade 3/4 anemia was higher at 25% in L-MOCA compared with 19% in SOLO2 and 5.1% in Study-19 (where olaparib capsules rather than tablets were administered), neutropenia was equivalent and grade 2/3 nausea lower (53% in L-MOCA vs. 73% SOLO2 and 66% in Study-19). Adverse event–driven dose adjustment or treatment discontinuation was also lower in L-MOCA at 44.6% and 9.4%, respectively, compared with 61.5% and 16.9% in SOLO2. Collectively, these findings indicate that, compared with the four pivotal PARP inhibitor trials, olaparib had a broadly similar PFS advantage among the exclusively Asian participants in L-MOCA and provides benefit without foreknowledge of the patients’ BRCA or HRD status. Although olaparib was tolerable for the majority of participants in L-MOCA, aspects of the olaparib toxicity profile seem to differ when compared with SOLO2 and Study-19.

PARP inhibitors will have an enduring place in the future management of ovarian cancer and other malignancies such as non–small cell lung and prostate cancer. They are also being evaluated in the treatment of neurogenerative diseases such as Parkinson's and Alzheimer's. It is therefore important that studies strive to represent the future patient population in which they will be administered. The FDA has recently reviewed the factors that may negatively impact recruitment of ethnic minority groups and released guidance to enhance diversity in clinical trial populations, including greater use of electronic communication and improving inclusion through public outreach, education and fostering community engagement. Ensuring that trial site distribution includes geographic locations with a higher concentration of racial and ethnic minority and indigenous populations would address the drive for wider inclusivity. In the absence of inclusive recruitment to these registration trials, results of independent studies such as L-MOCA are required to delineate responses and refine treatment guidance in specific ethnic groups. However, they are disadvantaged by their need to draw cross-study comparisons leaving essential questions around efficacy and/or tolerability unanswered.

S. Nicum reports grants and personal fees from AstraZeneca, GSK, and Clovis outside the submitted work. S.P. Blagden reports other support from Nucana PLC, Redx, Merck & Co, Astex Pharma, UCB, Sarah Cannon Institute, BerGenBio, and MiNA Therapeutics; personal fees and other support from Amphista Therapeutics and Ellipses; and personal fees from Theolytics and RAportss outside the submitted work.