Article Text
Abstract
Most women with ovarian cancer experience disease relapse, presenting numerous treatment challenges for clinicians. Maintenance therapy in the relapsed setting aims to extend the time taken for a cancer to progress, thus delaying the need for additional treatments. Four therapies are currently approved in the USA for secondline maintenance treatment of platinum sensitive, recurrent ovarian cancer: one antivascular endothelial growth factor agent (bevacizumab) and three poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitors (olaparib, niraparib, and rucaparib). In addition to efficacy, maintenance therapies must have a good tolerability profile and no significant detrimental impact on quality of life, as patients who receive maintenance are generally free from cancer related symptoms. Data from key bevacizumab trials (OCEANS, NCT00434642; GOG-0213, NCT00565851; MITO16B, NCT01802749) and PARP inhibitor trials (Study 19, NCT00753545; SOLO2, NCT01874353; NOVA, NCT01847274; ARIEL3, NCT01968213) indicate that bevacizumab and the PARP inhibitors are effective in patients with platinum sensitive, recurrent ovarian cancer but differ in their tolerability profiles. In addition, the efficacy of PARP inhibitors is dependent on the presence of homologous recombination repair deficiency, with patients with the deficiency experiencing greater responses from treatment compared with those who are homologous recombination repair proficient. Allowing for caveats of cross trial comparisons, we advise that clinicians account for the following points when choosing whether and when to administer a secondline maintenance treatment for a specific patient: presence of a homologous recombination repair deficient tumor; the patient’s baseline characteristics, such as platelet count and blood pressure; mode of administration of therapy; and consideration of future treatment options for thirdline and later therapy.
- ovarian cancer
- poly(adenosine diphosphate-ribose) polymerase inhibitor
- treatment outcome
- anti-vascular endothelial growth factor agent
Statistics from Altmetric.com
- ovarian cancer
- poly(adenosine diphosphate-ribose) polymerase inhibitor
- treatment outcome
- anti-vascular endothelial growth factor agent
Introduction
Most women diagnosed with advanced ovarian cancer experience recurrent disease, and about 70% have disease relapse within 3 years of frontline chemotherapy.1–3 Several treatments are available for recurrent ovarian cancer, but it can be challenging for clinicians to choose the best option. Historically, ovarian cancers have been termed ‘platinum sensitive’ if recurrence occurs ≥6 months after the last dose of platinum based chemotherapy, and ‘platinum resistant’ if recurrence occurs sooner.4 The fifth Ovarian Cancer Consensus Conference of the Gynecologic Cancer InterGroup advised that tumors should be defined by multiple factors, including platinum sensitivity, surgical outcomes, histology, mutation status, and response to non-platinum treatments. However, platinum sensitive terminology remains in use.5 Standard of care for platinum sensitive recurrent ovarian cancer includes rechallenge with a platinum based chemotherapy doublet, although most tumors ultimately become platinum resistant.4 6
Maintenance treatments aim to delay progression after response to chemotherapy, with the goal of prolonging overall survival. In the setting of recurrent ovarian cancer, maintenance is classified as continuing to receive a drug that was given concomitantly with cytotoxic chemotherapy or switching from cytotoxic chemotherapy to a different medication; however, maintenance is essentially an additional ‘treatment’. The concept for use of poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitors as maintenance in this setting came from the efficacy of these agents in the recurrent, measurable disease setting where patients with small volume residual disease (visible as a partial response to induction therapy, or not visible as a complete response to induction therapy) receive benefit from subsequent PARP inhibitor treatment.7 8 Olaparib and rucaparib were evaluated as monotherapy or ‘treatment’ in two open label phase II studies. A subanalysis of Study 42 (NCT01078662)9 10 investigated olaparib capsule monotherapy in patients with advanced BRCA1 or BRCA2 mutation ovarian cancer who had received three or more lines of prior therapy, and ARIEL2 (NCT01891344)11 assessed rucaparib monotherapy in patients with recurrent platinum sensitive ovarian cancer. Both studies demonstrated the efficacy of PARP inhibitor monotherapy in BRCA1 or BRCA2 mutation advanced ovarian cancers.10 11 Niraparib monotherapy has been evaluated as late-line treatment of ovarian cancer in the phase II QUADRA study (NCT02354586).12 Recently, data have also been reported for olaparib as late-line monotherapy compared with non-platinum chemotherapy in patients with BRCA1 or BRCA2 mutation relapsed ovarian cancer in the phase III SOLO3 study (NCT02282020).13 These studies further demonstrated the efficacy of PARP inhibitor monotherapy for the late-line treatment of patients with advanced ovarian cancer.12 13
Patients for whom maintenance therapy is an option are usually in response to chemotherapy, or in some cases may have stable disease following chemotherapy. They should therefore be free from cancer related symptoms. A maintenance treatment that improves progression free survival and/or overall survival should only be considered if toxicity is manageable and health related quality of life is not significantly affected. For example, although progression free survival significantly improved in GOG-178 (NCT00003120) for women who received 12 versus 3 cycles of maintenance paclitaxel after frontline chemotherapy,14 toxicity was increased and there was no improvement in the secondary endpoint of overall survival (this may be because the study was closed early following the positive interim progression free survival analysis).14 However, the GOG-212 study (NCT00108745) found that maintenance chemotherapy did not improve overall survival over surveillance among women with advanced ovarian cancer who had a complete response to firstline therapy and maintenance chemotherapy resulted in increased toxicity.1 Recently, attention has turned to targeted treatments, including antivascular endothelial growth factor agents and PARP inhibitors, in frontline and recurrent settings. Choosing a treatment for a particular patient can be complex.15 This review analyses decision making factors for choosing a secondline maintenance treatment for platinum sensitive recurrent ovarian cancer, limited to agents that are approved by the US Food and Drug Administration (FDA). Ethics approval was not required for this review article.
Bevacizumab as Secondline Maintenance
Bevacizumab (Avastin) is the only FDA approved antivascular endothelial growth factor treatment for platinum sensitive recurrent ovarian cancer. Its FDA approved indication is in combination with carboplatin/paclitaxel or carboplatin/gemcitabine, followed by single agent maintenance.16 Bevacizumab is also available in the USA as secondline and thirdline treatment for platinum resistant ovarian cancer (with paclitaxel, topotecan, or pegylated liposomal doxorubicin) and frontline therapy for stage III/IV disease (with carboplatin/paclitaxel, followed by single agent use).16 We review the trials that led to FDA approval of bevacizumab for platinum sensitive recurrent ovarian cancer, OCEANS (NCT00434642), and GOG-0213 (NCT00565851), and a trial in patients with prior frontline bevacizumab, MITO16B–MANGO-OV2b–ENGOT-ov17 (NCT01802749). These trials enrolled patients with platinum sensitive recurrent ovarian cancer who had not received chemotherapy for recurrent disease.17–19 Patients received bevacizumab plus chemotherapy, followed by maintenance bevacizumab, or chemotherapy alone.17–19 Trial design details are provided in the online supplementary material.
Supplemental material
Efficacy
In OCEANS and GOG-0213, a statistically significant improvement in progression free survival was seen with bevacizumab compared with chemotherapy (Table 1).17 18 MITO16B assessed bevacizumab rechallenge and reported a significant progression free survival increase with bevacizumab compared with chemotherapy alone (Table 1), indicating that rechallenge provides significant benefit for patients who have already received frontline bevacizumab.19
In ovarian cancer trials, it is difficult to assess whether investigational agents improve overall survival, primarily because of increasing expected postprogression survivorship and multiple lines of subsequent therapy.20 21 Of the trials discussed, only GOG-0213 was powered to assess overall survival.18 GOG-0213 underestimated median postprogression survivorship by nearly twofold (expected, 22 months; observed, 42 months).18 Nevertheless, the difference in median overall survival in the intention to treat population (based on pretreatment stratification factors) was approximately 5 months (hazard ratio (HR), 0.829; 95% confidence interval (CI), 0.683 to 1.005; p=0.056) (Table 1). Following the identification of incorrect treatment free interval stratification data for 7% of patients (equally balanced between treatment arms) a sensitivity analysis of overall survival based on audited treatment free interval stratification data was performed. The overall survival difference was improved by adjusting for corrected treatment free interval and participation in the trial’s surgical objective (HR 0.823; 95% CI 0.680 to 0.996; p=0.045).18 No significant difference was reported for overall survival between the bevacizumab plus chemotherapy arm compared with the placebo plus chemotherapy arm in the OCEANS study but the study was not powered to detect an overall survival difference (Table 1).22 Differences in the chemotherapy regimens (and therefore the comparator arms) between the GOG-0213 and OCEANS studies should also be noted, with paclitaxel plus carboplatin used in GOG-0213 and gemcitabine plus carboplatin used in OCEANS.18 22 Data are immature for MITO16B.19
Tolerability
OCEANS and GOG-0213 reported more grade ≥3 adverse events with bevacizumab (OCEANS 90%; GOG-0213 96%) than chemotherapy alone (OCEANS 82%; GOG-0213 86%).18 22 Adverse events associated with bevacizumab included hypertension, proteinuria, thrombosis, and non-central nervous system bleeding (Table 2).18 19 22 In OCEANS, grade ≥3 hypertension and proteinuria occurred more frequently with bevacizumab (17.0% (bevacizumab) vs 0.9% (chemotherapy) and 9.7% vs 0.4%, respectively).22 Similarly, 12% of patients receiving bevacizumab in GOG-0213 reported grade ≥3 hypertension (vs 1% with chemotherapy) and 8% reported grade ≥3 proteinuria (vs none with chemotherapy).18 Toxicity was similar in MITO16B (grade ≥3 hypertension: 27.5% (bevacizumab) vs 9.7% (chemotherapy); grade ≥3 proteinuria: 4% vs 0%).19 Gastrointestinal perforations are a known risk with bevacizumab, but these were rare in OCEANS and GOG-0213 (0.8% (bevacizumab) vs 0.4% (chemotherapy) and 2% vs <1%, respectively).17 18 22
Treatment discontinuation was more commonly caused by adverse events with bevacizumab than with chemotherapy in OCEANS (22.3% vs 4.7%) (Table 2), most frequently hypertension and proteinuria.22 Similarly, in GOG-0213, more patients in the bevacizumab than in the chemotherapy arm discontinued treatment because of an adverse events (25.5% vs 10.7%).18 FDA prescribing information lists adverse events for which bevacizumab should be discontinued, including hypertension and proteinuria, but does not recommend dose modifications.16
Health Related Quality of Life
GOG-0213 assessed the trial outcome index of the functional assessment of cancer therapy–ovarian questionnaire, which evaluates physical, social, emotional, and functional well being.18 Higher scores represent better health related related quality of life. The functional assessment of cancer therapy–ovarian trial outcome index decreased during therapy in both arms but returned to baseline about 6 months after the first treatment cycle.18 Mean functional assessment of cancer therapy–ovarian trial outcome index decreased more with bevacizumab than with chemotherapy alone, but this was not statistically significant (Table 3).18 OCEANS did not assess health related related quality of life; data were unavailable for MITO16B.
PARP Inhibitors as Secondline Maintenance
Three PARP inhibitors are FDA approved as maintenance therapies for platinum sensitive recurrent ovarian cancer: olaparib (Lynparza), niraparib (Zejula), and rucaparib (Rubraca). These are indicated as maintenance for women with recurrent epithelial ovarian cancer who had a complete/partial response to secondline or greater platinum based chemotherapy, regardless of BRCA1 or BRCA2 mutation status.23–25 All three PARP inhibitors are also approved for these patients in Europe.26 27 Additionally, rucaparib is FDA approved as thirdline or later monotherapy for germline or somatic BRCA1 or BRCA2 mutated ovarian cancer, olaparib as fourthline or later monotherapy for germline BRCA1 or BRCA2 mutated ovarian cancer, and niraparib as fourthline or later monotherapy for homologous recombination deficiency positive ovarian cancer.23 25 28 We review the trials that led to FDA approval of these PARP inhibitors as maintenance: Study 19 (NCT00753545), SOLO2 (NCT01874353), NOVA (NCT01847274), and ARIEL3 (NCT01968213). These trials enrolled patients with platinum sensitive recurrent ovarian cancer who had received ≥2 lines of platinum based chemotherapy and had a complete/partial response to their most recent regimen.12 29–32 In Study 19, SOLO2, NOVA, and ARIEL3, all patients received a PARP inhibitor or placebo as maintenance.12 29–32 Trial design details are in the online supplementary material.
Efficacy
In Study 19, progression free survival significantly improved with olaparib compared with placebo (Table 1).29 Patients with a BRCA1 or BRCA2 mutation derived the greatest progression free survival benefit, regardless of whether they had a germline or somatic BRCA1 or BRCA2 mutation.33 34 Maintenance olaparib also provided significant progression free survival benefit for BRCA1 or BRCA2 wild-type patients (Table 1).33 Exploratory analyses suggested that patients with a BRCA1 or BRCA2 wild-type with homologous recombination repair tumor mutations (see online supplementary table 1, T5 panel, Foundation Medicine) received greater progression free survival benefit from olaparib than those without homologous recombination repair tumor mutations, although the latter still benefited (HR 0.21, 95% CI 0.04 to 0.86 vs HR 0.71, 95% CI 0.37 to 1.35).35 Patients in SOLO2 also experienced significantly improved progression free survival with olaparib compared with placebo (Table 1). Figure 1 shows progression free survival across these trials; SOLO2 showed the greatest improvement in median progression free survival for BRCA1 or BRCA2 mutated patients, exceeding that seen in the BRCA1 or BRCA2 mutated subgroups of Study 19, NOVA, and ARIEL3.30–34 However, there were similar HR values reported for all studies, ranging from 0.18 for Study 19 to 0.30 for SOLO2.
NOVA reported significant progression free survival improvement for patients who received maintenance niraparib compared with placebo, regardless of BRCA1 or BRCA2 mutation status.31 Patients with a germline BRCA1 or BRCA2 mutation experienced the greatest progression free survival benefit, followed by germline BRCA wild-type patients who were homologous recombination repair deficient, and then by the overall germline BRCA1 or BRCA2 wild-type cohort (Table 1).31 Of the germline BRCA1 or BRCA2 wild-type patients, those with a somatic BRCA1 or BRCA2 mutation experienced similar progression free survival benefit to those with a germline BRCA1 or BRCA2 mutation, supporting the similarity between germline and somatic BRCA1 or BRCA2 mutated tumors.31 Germline BRCA1 or BRCA2 wild-type patients who were homologous recombination repair deficient but did not have a somatic BRCA1 or BRCA2 mutation also received progression free survival benefit with niraparib, demonstrating that a somatic BRCA1 or BRCA2 mutation did not completely drive benefit for germline BRCA1 or BRCA2 wild-type, homologous recombination repair deficient patients. Similar benefit was also observed for germline BRCA1 or BRCA2 wild-type patients without homologous recombination repair deficiency.31
In ARIEL3, the BRCA1 or BRCA2 mutated subgroup, the homologous recombination repair deficient subgroup, and the overall population had significantly prolonged progression free survival with rucaparib compared with placebo (Table 1).32 Prespecified analyses also showed improved progression free survival in a BRCA1 or BRCA2 wild-type cohort, indicating that overall benefit was not entirely driven by the BRCA1 or BRCA2 mutated or homologous recombination repair deficient cohorts.32 Rucaparib provided progression free survival benefit for BRCA1 or BRCA2 wild-type patients with both high and low loss of heterozygosity (HR 0.44 and 0.58, respectively).32
Overall response rates are shown in Table 1 for patients with baseline measurable disease.29 32 To estimate the long term effects of maintenance treatments, intermediate efficacy endpoints such as time to second disease progression and time to first and second subsequent therapy or death can be used, bridging progression free survival and overall survival.20 In Study 19, time to second disease progression and time to first and second subsequent treatment or death were both significantly improved with olaparib compared with placebo (online supplementary table 2). Kaplan–Meier curves for time to second disease progression remained separated for the treatment arms between 3 and 6 years of follow-up.36 In SOLO2, time to second disease progression, and time to first and second subsequent treatment or death significantly improved with olaparib compared with placebo (online supplementary table 2).30 Similarly, in NOVA, time to second disease progression significantly improved with niraparib compared with placebo (online supplementary table 2).31
Overall survival is immature for SOLO2, NOVA, and ARIEL3, but overall survival data for Study 19 represent the longest available follow-up (median 6.5 years) for a maintenance PARP inhibitor.36 Despite Study 19 not being powered to assess overall survival, data suggest an overall survival advantage for patients who received olaparib, although the predefined threshold for statistical significance (p<0.0095) was not met (HR 0.73; 95% CI 0.55 to 0.95; nominal p=0.021) (Table 1).36 The overall survival advantage appeared greatest for BRCA1 or BRCA2 mutated patients, but some separation was seen in the BRCA1 or BRCA2 wild-type Kaplan–Meier curves.36 For the overall population, separation in Kaplan–Meier curves became most apparent after about 3 years, suggesting that long term responders drove the overall survival advantage.36 Fifteen patients received maintenance olaparib for 6 years or more.35 Five of these were BRCA1 or BRCA2 wild-type (33%); one had an homologousrecombination repair tumor mutation, two were homologous recombination repair wild-type, and two had uncertain homologous recombination repair status.35
Overall, olaparib, niraparib, and rucaparib have demonstrated efficacy for BRCA1 or BRCA2 mutated and wild-type patients.
Tolerability
Perhaps the greatest difference between maintenance olaparib, niraparib, and rucaparib is their adverse events profiles. Of the adverse events they have in common, the most prevalent include nausea, vomiting, fatigue, and anemia (Table 2). Nausea, vomiting, and fatigue are usually mild or moderate in severity.30–32 Myelodysplastic syndrome and acute myeloid leukemia have been reported rarely with the three PARP inhibitors (1–2%), with comparable incidences to placebo (0–4%). In Study 19, SOLO2, NOVA, and ARIEL3, most patients received treatment despite toxicity, with PARP inhibitor discontinuation caused by adverse events in 6%, 11%, 15%, 21%, and 13% of patients, respectively (vs 2% for each placebo arm).30–32 36 Toxicities were generally managed using dose modifications and supportive treatment. NOVA most commonly reported dose modifications due to adverse events, followed by ARIEL3 and Study 19/SOLO2 (Table 2). FDA prescribing information for olaparib, niraparib, and rucaparib includes recommended dose modifications.23–25
Niraparib is associated with the highest incidence of grade ≥3 adverse events (Table 2), particularly thrombocytopenia and neutropenia, followed by rucaparib and then olaparib.31 In NOVA, the incidence of grade ≥3 thrombocytopenia was 33.8% with niraparib compared with 0.6% with placebo.31 A retrospective analysis found that baseline platelet count <1 50 000/μL and body weight <77 kg were risk factors for grade ≥3 thrombocytopenia.37 Of patients with these risk factors, 35% had thrombocytopenia in the first treatment month compared with 12% for patients without risk factors.37 Although not FDA specified, it is advisable to consider platelet count and body weight when considering niraparib. Patients at risk should receive a reduced dose in anticipation of thrombocytopenia.
Preclinical data for the PARP inhibitors show the most off-target hits for niraparib, including hits at a 50% inhibitory concentration of <0.25–2.5 µM for dopamine, serotonin, and norepinephrine transporters.38 FDA prescribing information notes that niraparib may cause central nervous system and cardiovascular effects related to inhibition of these transporters.24 Consistent with this, 24.3% of patients receiving niraparib in NOVA reported insomnia (vs 7.3% with placebo); data for ARIEL3 were 14% compared with 8%, and SOLO2 reported no difference in the incidence of insomnia between arms.30–32 Additionally, hypertension was reported with niraparib, but not with olaparib or rucaparib (Table 2).31
Rucaparib also differs in safety profile from niraparib and olaparib. Notably, 10% of patients receiving rucaparib in ARIEL3 reported grade ≥3 increased alanine aminotransferase or aspartate aminotransferase compared with no patients on placebo32; 34% of patients experienced increased alanine aminotransferase/aspartate aminotransferase of any grade with rucaparib (Table 2) compared with 4% on placebo.32 However, these elevated alanine aminotransferase/aspartate aminotransferase concentrations were generally transient and unaccompanied by other signs of liver toxicity.32
Indirect comparisons of SOLO2, NOVA, and ARIEL3 suggest that olaparib is associated with a reduced odds of grade ≥3 adverse events and dose interruption compared with niraparib and rucaparib (online supplementary table 3).39 In rat models, PARP inhibitors exhibited different plasma to bone marrow ratios, which may explain their clinical hematologic profiles.38
Health Related Quality of Life
Study 19 reported no significant difference in time to worsening of functional assessment of cancer therapy–ovarian trial outcome index (Table 3), overall functional assessment of cancer therapy–ovarian, or functional assessment of cancer therapy/NCCN ovarian symptom index with olaparib compared with placebo.40 SOLO2 also reported no significant difference between arms in the mean change in functional assessment of cancer therapy–ovarian trial outcome index (Table 3).41 Furthermore, SOLO2 examined quality adjusted progression free survival (progression free survival combined with a European quality of life 5 dimensions, 5 level questionnaire score and time without symptoms of treatment toxicity (time between randomization and progression over which patients did not experience significant grade ≥2 nausea, vomiting, or fatigue).41 Olaparib was associated with prolonged quality adjusted progression free survival (mean 13.96 vs 7.28 months; 95% CI, 4.98 to 8.54; p<0.0001), and patients receiving olaparib were free from significant toxicity related symptoms for longer than those receiving placebo (mean time without symptoms of treatment toxicity 15.03 vs 7.70 months; 95% CI 4.70 to 8.96; p<0.0001).41 This is noteworthy given that patients who receive maintenance treatment are generally well. Niraparib did not adversely affect health related related quality of life in NOVA; similar mean scores for functional assessment of cancer therapy/NCCN ovariansymptom index, European quality of life 5 dimensions, 5 level questionnaire score, and the European quality of life visual analog scale were reported in both arms.42 Detailed data are unavailable for ARIEL3, but for BRCA1 or BRCA2 mutated patients, no significant health related related quality of life difference was reported with rucaparib compared with placebo (Table 3).32
Considerations When Choosing Secondline Maintenance Treatment for Platinum Sensitive Recurrent Ovarian Cancer
It is crucial that maintenance options are discussed with patients with platinum sensitive recurrent ovarian cancer before initiation of secondline chemotherapy and after, to confirm their choice. We defined key considerations for choosing between the secondline maintenance options for platinum sensitive recurrent ovarian cancer.
Tumor Mutations
BRCA1 or BRCA2 mutation prevalence in platinum sensitive recurrent ovarian cancer can be up to 40%.43 44 Furthermore, about 50% of high grade serous ovarian cancers harbor homologous recombination repair defects.45 46 Clinical data show that all patients with platinum sensitive recurrent ovarian cancer derive benefit from maintenance PARP inhibitors, regardless of whether their tumors harbor BRCA1 or BRCA2 mutation or other homologous recombination repair deficiency. Patients with a BRCA1 or BRCA2 mutation have been observed to benefit most from PARP inhibition, and patients with homologous recombinant repair mutated tumors have been seen to receive greater benefit than those without.31–33 35 Thus when considering secondline maintenance, clinicians should counsel patients with platinum sensitive recurrent ovarian cancer on the specific amount of benefit to expect from a maintenance PARP inhibitor, based on BRCA1 or BRCA2 mutation and homologous recombination repair deficiency status.31 32 35 The efficacy of bevacizumab is not significantly affected by homologous recombination repair status, and this should be considered when choosing a maintenance therapy.47
Tolerability and Mode of Administration
The PARP inhibitors and bevacizumab have distinct safety profiles. Given the adverse events associated with bevacizumab, it may not be appropriate for patients with uncontrolled high blood pressure or cardiovascular comorbidities; similarly, such patients may not be suited to niraparib given its association with hypertension. For patients receiving bevacizumab or niraparib, blood pressure should be monitored regularly. Patients with baseline anemia may not be suited to PARP inhibitors, and patients with baseline neutropenia or thrombocytopenia may not be suited to niraparib. It is also worth noting that PARP inhibitors are administered orally; bevacizumab is given intravenously.
Other Considerations
Generally, platinum sensitive recurrent ovarian cancer eventually becomes platinum resistant, so clinicians must also consider future therapeutic options. Bevacizumab is approved as secondline and thirdline treatment for platinum resistant ovarian cancer; rucaparib and olaparib/niraparib are only indicated as thirdline (or later) and fourthline (or later) treatment, respectively, for platinum resistant ovarian cancer that is BRCA1 or BRCA2 mutated. Consequently, clinicians may favor a PARP inhibitor as secondline maintenance for platinum sensitive recurrent ovarian cancer and retain bevacizumab for later lines. Recently, olaparib became the first FDA approved PARP inhibitor for frontline maintenance of BRCA1 or BRCA2 mutated advanced ovarian cancer. This approval, based on SOLO1 (NCT01844986),48 will impact the choice of secondline therapy. The ongoing OReO trial (NCT03106987) is investigating olaparib retreatment for patients whose disease progressed after prior PARP inhibitor maintenance. MITO16B showed that bevacizumab has significant efficacy for patients who relapsed after prior bevacizumab, confirming that this agent can be used more than once.19
Cross Trial Comparisons
Cross trial comparisons have many caveats, even for trials with similar eligibility and treatment paradigms. Comparisons of progression free survival between the trials reviewed here are complicated by the time when progression free survival was recorded: the bevacizumab associated trials measured progression free survival from initiation of secondline chemotherapy; the PARP inhibitor trials measured progression free survival from initiation of maintenance (Figure 1). Progression free survival curves from the bevacizumab trials encompassed all randomized patients, including those with progressing or stable disease on chemotherapy, those who withdrew, and those who suffered a non-disease related death. Furthermore, median progression free survival and overall survival are difficult to assess compared with the PARP inhibitor studies because BRCA1 or BRCA2 mutation status was unknown in the bevacizumab trials. Additionally, given the proportion of BRCA1 or BRCA2 mutated patients in the PARP inhibitor trials (Study 19, 51%33; SOLO2, 100%30; NOVA, 37%31; ARIEL3, 35%32), enrichment for this favorable prognostic factor clouds direct comparison. Comparison of tolerability in the bevacizumab and PARP inhibitor trials is also complicated because data for bevacizumab include the chemotherapy combination, which would generally increase toxicity.
Of the trials discussed, we noted differences in progression free survival assessment and patient populations. OCEANS, Study 19, SOLO2, and ARIEL3 used investigator assessed progression free survival as the primary endpoint; NOVA used blinded independent central review.17 29–32 Generally, when a tumor progresses, scans are no longer sent for blinded independent central review, which may create bias and increase median progression free survival.49 However, these trials reported consistency between investigator assessment and independent central review (Figure 1).17 29–32 In terms of patients, the bevacizumab trials enrolled those who had received frontline chemotherapy only. The PARP inhibitor trials included patients who had received multiple lines of treatment17–19 29–32: 54% of patients in Study 19, 40% in NOVA, and 37% in ARIEL3 had received three or more chemotherapy regimens. Patients in NOVA had to have observable residual disease of <2 cm, which was not stipulated by the other studies.29–32 Finally, the surgical objective of GOG-0213 meant that some patients receiving bevacizumab had no measurable disease because of secondary cytoreduction (see online supplementary material)18; many patients in the PARP inhibitor trials were in partial response to chemotherapy.29–32
Future Directions
In the near future, bevacizumab may be combined with PARP inhibitors. The PAOLA-1 (NCT02477644) study of firstline bevacizumab in combination with platinum based chemotherapy followed by bevacizumab with or without olaparib maintenance treatment was recently reported and demonstrated a significant improvement in progression free survival when olaparib was added to the maintenance regimen (median progression free survival 22.1 months with olaparib vs 16.6 months with placebo; HR 0.59; 95% CI 0.49 to 0.72; p<0.001).50 Other studies are also assessing frontline chemotherapy plus bevacizumab, followed by maintenance niraparib (OVARIO; NCT03326193) or rucaparib (MITO-25; NCT03462212) in combination with bevacizumab. The recently reported results from the VELIA study (NCT02470585) of veliparib in combination with firstline platinum based chemotherapy and as maintenance treatment have also raised the question of whether PARP inhibitors should be combined with firstline chemotherapy for patients with advanced ovarian cancer.51 Another exciting option is the potential to use immuno-oncology agents, combined with chemotherapy and then as maintenance. Data from a phase II study (NCT02766582) suggest that the antiprogrammed cell death protein 1 agent, pembrolizumab, has activity in ovarian cancer when administered with carboplatin/paclitaxel and then as maintenance.52
The most important future directions for patients with ovarian cancer will be to define the treatment regimens that are most effective in: (1) patients treated with a PARP inhibitor who have not progressed; (2) patients exposed to PARP inhibitor treatment who have progressed; and (3) patients who are treatment naïve—can we determine a regimen that demonstrates improved efficacy to what has already been demonstrated? Treatment decisions will likely become more complex, but we hope that beneficial therapies can be found for all patients.
Conclusions
In conclusion, the treatment approach for patients with recurrent ovarian cancer has significantly evolved in recent years. The addition of maintenance treatment in the relapsed setting has extended the time taken for a cancer to progress, thus delaying the need for patients to receive additional lines of chemotherapy and its associated toxicities. Four therapies are currently approved in the USA for secondline maintenance treatment of platinum sensitive, recurrent ovarian cancer: olaparib, niraparib, rucaparib, and bevacizumab. Data from key trials of these treatments indicate that they are all effective in patients with platinum sensitive, recurrent ovarian cancer but differ in their tolerability profiles. In addition, the efficacy of PARP inhibitors is dependent on the presence of homologous recombination repair deficiency, with patients exhibiting homologous recombination repair deficiency experiencing greater responses from treatment compared with those who are homologous recombination repair proficient. Multiple factors should be considered when deciding if a patient is a candidate for secondline maintenance therapy and which option is appropriate. As such, clinicians should account for the following points when choosing whether and when to administer a secondline maintenance treatment for a specific patient: presence of a homologous recombination repair deficient tumor; the patient’s baseline characteristics, such as platelet count and blood pressure; mode of administration of therapy; and consideration of future treatment options for thirdline and later treatment.
Acknowledgments
Medical writing support was provided by Rachel Patel, MBiochem, and Claire Routley, PhD, from Mudskipper Business Ltd, funded by AstraZeneca and Merck & Co Inc.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @Shannon.Westin, @rcoledude
Contributors All authors contributed to the conception of this manuscript, and the collection and interpretation of published data. All authors drafted, reviewed, and provided their comments on this manuscript, and approved the final version.
Funding Medical writing support was funded by AstraZeneca and Merck & Co Inc.
Competing interests RA: consultancy/advisory (AstraZeneca, Clovis, Pfizer, Puma, Tesaro, VBL therapeutics). RC: integration panel (DOD-CDMRP); grant review committee (NCCN); consultancy/advisory (Clovis Oncology, Esperance Pharmaceuticals, Genentech/Roche); travel/accommodation/expenses (Amgen, Array BioPharma, AstraZeneca/MedImmune, Bayer, Clovis Oncology, GOG, Research to Practice, Merck, Millennium, Roche/Genentech, New Mexico Cancer Center, University of California, Irvine, University of Cincinnati Cancer Center, University of Miami); research funding (Abbott/AbbVie, Array BioPharma, AstraZeneca/MedImmune, Clovis Oncology, Esperance Pharmaceuticals, Johnson & Johnson, Merck, OncoMed, Roche/Genentech). SNW: consulting (AstraZeneca, Clovis, Merck, MediVation, Ovation, Pfizer, Roche/Genentech, Takeda, Tesaro); research support (AstraZeneca, ArQule, Bayer, Clovis, Cotinga Pharmaceuticals, Novartis, Roche/Genentech, Tesaro).
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.