PARP inhibitors are poised to acquire a stronger foothold in the frontline setting for advanced ovarian cancer, for which chemotherapy remains the backbone of treatment, according to a panel of gynecologic oncology experts who discussed the issue during an OncLive Peer Exchange® held during the European Society for Medical Oncology Congress 2019.
Bradley J. Monk, MD
PARP inhibitors are poised to acquire a stronger foothold in the frontline setting for advanced ovarian cancer, for which chemotherapy remains the backbone of treatment, according to a panel of gynecologic oncology experts who discussed the issue during an OncLive Peer Exchange® held during the European Society for Medical Oncology Congress 2019 (ESMO 2019). Chemotherapy ultimately fails most patients, but PARP inhibitors, which already have multiple indications in the recurrent setting, have a greater chance at providing cure when used earlier in the lineup, the panelists said.
In ovarian cancer, PARP inhibitors have joined an increasing number of therapies in the treatment landscape, and indications for approved agents are continuing to expand as results of clinical trials show benefit in different patient subsets. Treatment decision making is therefore set to become more nuanced, particularly as investigators continue to evaluate the utility of certain biomarkers and their diagnostic assays in clinical trials.
The panelists also examined how certain biomarkers predict the efficacy of PARP inhibitors and how these agents compare with bevacizumab (Avastin), another targeted agent approved for advanced, newly diagnosed, and recurrent high-grade serous or endometrioid ovarian, primary peritoneal, or fallopian tube cancer. They also discussed a novel antibody-drug conjugate being assessed in the recurrent setting that likely would have ceased being developed had a biomarker assay error not been identified.
Although biomarkers in ovarian cancer are still being investigated, they are starting to shape care, with most indications of currently approved targeted agents requiring use of a companion diagnostic to guide patient selection. “I think we’re turning ovarian cancer into a rarer and rarer disease with these subsets, which will help us do better, smaller, smarter trials to move things forward. It’s an exciting time,” panelist Kathleen Moore, MD, said.
Moving PARP Inhibitors to the Front Line
In the frontline ovarian cancer setting, the current standard of care is a combination of surgery and chemotherapy. Clinicians may add bevacizumab to chemotherapy and give it as a maintenance treatment; they can also consider the PARP inhibitor olaparib (Lynparza) as a maintenance treatment. It is approved for patients with BRCA-mutated advanced epithelial ovarian, fallopian tube, or primary peritoneal cancer in complete response (CR) or partial response (PR) after first-line platinum-based chemotherapy, and it is currently the only PARP inhibitor with a frontline indication in ovarian cancer.1-3 However, the panelists discussed 3 potentially practice-changing phase III trials presented at the ESMO 2019 meeting that are likely to lead to several PARP inhibitor approvals: PRIMA, assessing niraparib as a maintenance therapy; VELIA, assessing the still-investigational PARP inhibitor veliparib in combination with chemotherapy and as a maintenance therapy; and PAOLA- 1, assessing olaparib in combination with bevacizumab as a maintenance therapy.4-6
PRIMA randomly assigned 733 patients with newly diagnosed advanced high-grade serous or endometrioid ovarian, primary peritoneal, or fallopian tube cancer with a CR or a PR to first-line platinum-based chemotherapy 2:1 to niraparib (n = 487) or placebo (n = 246) maintenance therapy.4 Of these patients, 373 (51%) were homologous recombination deficient (HRD; niraparib, 247; placebo, 126) and 35% had stage IV disease. “[This study] enrolled women who we’d consider the highest risk,” Moore explained.
PRIMA’s primary end point was progression-free survival (PFS) in the HRD group, and a secondary end point was PFS in the entire study population. In the study, HRD status was measured using the Myriad myChoice HRD test. “It is important to know that this assay is still evolving,” panelist Michael J. Birrer, MD, PhD, said. He explained that the cutoff score for positivity was changed from previous studies, increasing from ≥33 to ≥42 (scale, 0-100). “That means, theoretically, it should predict for patients who are going to benefit from PARP [inhibition] better than the old assay,” he said.
Niraparib-treated patients in the HRD subgroup as well as the overall study population showed significant reductions in the risk of disease recurrence or death, demonstrating substantial improvements in PFS. The HR for PFS in the HRD group was 0.43 (0.31-0.59; P <.0001), with median PFS more than doubling, from 10.4 months in the placebo group to 21.9 months in the niraparib group. In the overall population, the HR was 0.62 (0.50-0.75; P <.0001), with a median PFS of 8.2 months in the placebo group compared with 13.8 months in the niraparib group.
“Just to put that in perspective, if you have an HRD assay and high-risk patients, who need [treatment] the most, you have basically a doubling of the PFS. It’s better than 6 months with bevacizumab,” moderator Bradley J. Monk, MD, FACS, FACOG, said.
Earlier in the discussion, the panelists explained that although all BRCA-positive patients are HRD positive, not all patients who are HRD positive are also BRCA positive. Subsequently, investigators have questioned whether HRD-positive, BRCA wild-type patients may also benefit from PARP inhibitors. PRIMA data indicate that they do, Birrer said. Subsequently, Monk suggested HRD should be considered “as an intermediate between no biomarker and BRCA.” In contrast, the niraparib benefit in the overall population, which included patients with no biomarkers (BRCA, HRD) was equivalent to what has been observed with bevacizumab (PFS, ~6 months). Nevertheless, because benefit was observed across all subgroups, Monk said he felt confident the FDA will approve niraparib for all-comers.
VELIA enrolled 1140 patients with stage III or IV high-grade serous carcinoma.5 Of these patients, 26% had BRCA mutations and 55% were HRD positive. Patients were randomly assigned 1:1:1 to platinum -based chemotherapy plus placebo followed by placebo maintenance, platinum-based chemotherapy plus veliparib followed by placebo maintenance, or platinum- based chemotherapy plus veliparib followed by veliparib maintenance.
“VELIA is a little different from PRIMA because we are making a decision right when we start chemotherapy regarding what we’re going to do,” panelist Shannon N. Westin, MD, MPH, FACOG, said, indicating that this study is particularly exciting because it introduces a PARP inhibitor that does not yet have an indication in ovarian cancer. “I think what’s really interesting about this study is, compared with the PRIMA study, a larger proportion of these patients were stage III,” she said. “It was about 70% or so, and a large proportion of them had surgery to no gross residual disease.” Subsequently, the study population was more representative of the patients seen in clinical practice.
One of VELIA’s primary end points was PFS, but the study was statistically designed as a hierarchical analysis starting with the patients with BRCA mutations, then moving to the HRD patients, and then moving to the overall population, with data having to be positive in the preceding group before analysis for the next group could start. “In [the BRCA-mutation cohort], we had a PFS of 34.7 months on the median compared with 22.0 months [for controls]. That was a hazard ratio of 0.44, so it was quite significant,” panelist Robert L. Coleman, MD, FACOG, FACS, said. “The next step down was to look at the total HRD patient population, which had a hazard ratio of 0.57. Again, there was about a 12-month improvement in the medians between the groups. Then in the all-comers population, we had a hazard ratio of 0.68, which is obviously statistically significant,” he said, noting that the next step will be to assess overall survival (OS). He added that the trial allowed neoadjuvant chemotherapy and use of dose-dense versus every-3-week paclitaxel because of toxicity concerns over combining a PARP inhibitor with chemotherapy. “The dose intensity that we were able to maintain throughout the entire chemotherapy phase was essentially standard-of-care dosing,” he said.
Regarding whether giving the PARP inhibitor with chemotherapy adds benefit, the current VELIA trial data indicate the benefit is mostly in the maintenance setting. However, Coleman said that PFS is likely “not a sensitive enough variable to determine whether [the PARP inhibitor is] actually adding anything.” Therefore, the OS analysis will be crucial in providing a more definitive answer.
Based on these new data, the panelists concurred that a PARP inhibitor is preferable over bevacizumab in patients with an HRD molecular signature. Although HRD testing is generally not done up front, PRIMA and VELIA findings support up-front testing of women with newly diagnosed advanced ovarian cancer to better inform treatment decision making. The panelists felt confident that if veliparib is approved, it will be approved for all-comers, which happened with niraparib, and will make choosing among PARP inhibitors challenging. Although data from both PRIMA and VELIA may be used to help guide decision making, Birrer warned the data are difficult to compare because the assays used in PRIMA and VELIA were not comparable. The assay used in VELIA had an HRD positivity cutoff of ≥33 compared with ≥42 in PRIMA, resulting in VELIA’s HRD group having more wild-type or non-HRD—reclassified patients. Until more data become available to guide selection among PARPs, treatment decision making is likely to boil down to factors such as toxicity, Monk said.
Table. (Click to Enlarge)
Another challenge with PAOLA-1, and what further sets it apart from PRIMA and VELIA, is that it has an active comparator as the maintenance therapy. “That already changes the bar for your hazard ratio because this group is not getting placebo or anything where you know how they’re going to perform. They’re getting a drug that we know works well, perhaps not as well as PARP inhibitors but pretty well. That changes things,” Moore said.
Moore proceeded to discuss several exploratory end points that were examined in PAOLA-1. “If you look at the BRCA wildtype, HRD-positive cohort, the hazard ratio is 0.43. You get 12 months. Is that just the PARP inhibitor effect?” she asked. “Is it just the PARP or is the bevacizumab additive? We can’t answer the question. I don’t know that we can really interpret this without a singlearm olaparib trial. But 0.43 is pretty nice.”
Panelists noted the data are comparable to those observed in PRIMA, in which the addition of niraparib to maintenance increased the median PFS by approximately 12 months. Based on these data, several panelists said they would be tempted to use this combination in their patients.
“Once I start bevacizumab, I don’t like to stop it if I don’t have to because I feel as though you keep getting that benefit,” Westin said. “It’s working, and she’s tolerating it. Now I feel good about adding olaparib to that patient’s regimen. I feel like that patient is really going to benefit. It’s exploratory, but exploratorily, I feel good about it,” she said.
Moving to Novel Agents in Recurrent Ovarian Cancer
The armamentarium for recurrent ovarian cancer has benefited from several advancements, including the approval of bevacizumab and several PARP inhibitors, the panelists noted. In the recurrent setting, bevacizumab is approved for use in combination with paclitaxel, pegylated liposomal doxorubicin, or topotecan in patients with platinum-resistant disease that has been treated with ≤2 prior chemotherapy regimens and in combination with carboplatin and paclitaxel or carboplatin and gemcitabine, followed by bevacizumab as a single agent, in patients with platinum- sensitive recurrent disease.7 The PARP inhibitors niraparib (Zejula), olaparib, and rucaparib (Rubraca) have various indications in the recurrent setting (Table).1-3 Although these are important treatment options for women with recurrent disease, they are not curative, Monk explained. This has led to the search for novel agents to improve outcomes. One such agent in clinical development that the panel discussed is the antibody-drug conjugate mirvetuximab soravtansine.
“Antibody-drug conjugates target something that their programs can find. In this case, it is the folate receptor α [FRα], which is a transmembrane protein that is almost ubiquitously expressed on ovarian cancer. It is medium to highly expressed about 60% of the time and negligible on normal tissue, so it’s an ideal target to get to the tumor,” Moore said. Mirvetuximab soravtansine was assessed for recurrent ovarian cancer in the FORWARD I trial.8
FORWARD I Trial
FORWARD I (NCT02631876) randomly assigned 366 women with recurrent platinum-resistant ovarian cancer whose tumors were characterized as having medium or high FRα expression 2:1 to mirvetuximab soravtansine (n = 248) or investigator’s choice of chemotherapy, which could include paclitaxel, pegylated liposomal doxorubicin, or topotecan (n = 118).8 The primary end point was PFS in the ITT population (medium and high FRα expression) and, separately, in patients with high FRα. Secondary end points included objective response rate (ORR) and OS. The median follow-up time was 12.5 months.
The study did not meet its primary end point. “The assumptions we made were that the control group would have a PFS of 3.5 months, which they did, and that for the mirvetuximab group, it would be 6.0 months. We were looking for a hazard ratio of 0.58,” Moore said. Instead, no difference in PFS was observed between treatments in the ITT group, and although there was a difference numerically in the PFS medians in the cohort with high FRα expression, yielding an HR of 0.68, the P value was not statistically significant, she said.
Although these results were surprising and disappointing, a reassessment of patients’ FRα expression found that one-third had low FRα expression, making them ineligible for the trial, and half of those characterized as having high FRα expression had medium FRα expression. The assay that had been used to determine FORWARD I eligibility was different from what was used in the phase I program, Moore explained. The study investigators had changed the assay to make it more of a companion diagnostic that would be more convenient for clinicians. This caused dilution of the primary end point, she said. When the investigators reran the data using the correctly characterized FRα-high subgroup, the PFS was statistically significantly improved, the OS increased from 11 to 16 months, and the ORR was much higher, at 23% versus 6%. “The study, on exploratory analysis, looks to be positive,” she said.
Based on these findings, the FORWARD I investigators will conduct a second, confirmatory trial called MIRASOL, which will characterize patients with the correct assay and enroll only patients determined to have high FRα expression. “It will be a 1:1 randomization, and there’s a more conservative statistical plan and a more robust power analysis,” Moore said. The trial is expected to enroll its first patient before the end of 2019.9
Unlike the PRIMA and VELIA studies, PAOLA-1 assessed whether adding a PARP inhibitor to bevacizumab would have a synergistic effect. The study enrolled patients with newly diagnosed stage III or IV high-grade serous or endometrioid ovarian, primary peritoneal, or fallopian tube cancer with CR or PR after standard platinum-based chemotherapy plus bevacizumab. These patients were randomized 2:1 to maintenance therapy with olaparib plus bevacizumab (n = 537) or placebo plus bevacizumab (n = 269).6 The primary end point was PFS in the intentionto- treat (ITT) population, which Monk said was an unusual end point for a PARP inhibitor trial. “I like the ITT concept, but we’re past that,” he said, indicating that the difference in effect with PARP inhibitors in patients with BRCA-mutant disease is so dramatic that the efficacy of these agents is obscured. Nevertheless, results of the PAOLA-1 trial showed an HR of 0.59 (CI, 0.49-0.72; P <.0001), with a median PFS of 22.1 months in the olaparib arm compared with 16.6 months in the control arm.