2.1 The revival of secondary cytoreductive surgery
This year, two prospective trials evaluating secondary cytoreductive surgery (SCS) were presented; Table 1
provides an overview of both trials. Du Bois and colleagues, after validation of their selection criteria and demonstrating a benefit in progression free survival (PFS), reported the final results of the phase 3, randomized DESKTOP III trial (#6000) evaluating the impact of SCS in recurrent ovarian cancer [
]. The investigators reported an overall survival (OS) benefit in the surgery arm with a median survival of 53.7 months compared to 46.2 months in those who did not undergo surgery. The true benefit was seen in women who underwent complete resection at the time of surgery (60.7 months versus 46.2 months). It is important to note that patients who were not able to undergo complete resection had a worse outcome (median survival 28.8 months) when compared to the no surgery arm. In a separate randomized phase 3 study from China, SOC 1/SGOG-OV2 (#6001), SCS was also evaluated. In this study, there were designated selection criteria based on the iModel score in conjunction with PET-CT scan imaging to determine complete resection [
]. Zang and colleagues reported that SCS was associated with a benefit in PFS of 17.4 versus 11.9 months, though OS outcomes are still maturing. Similarly to DESKTOP III, those with residual disease did worse than chemotherapy alone. Therefore, though these results differed from GOG 213 which reported no difference in OS outcomes with SCS, both groups concluded that SCS is of benefit when performed in carefully selected patients, and that SCS should be done at centers of excellence/experience in which the opportunity to resect disease completely is optimal.
Table 1Comparison of the phase 3 trials in secondary cytoreductive surgery.
Lastly, another study of SCS presented was a randomized phase 2 trial of SCS with or without carboplatin hyperthermic intraperitoneal chemotherapy (HIPEC) in women with recurrent platinum sensitive ovarian cancer (#6016) [
]. The study randomized 98 patients and reported 0% perioperative mortality and comparable outcomes to those undergoing SCS only. Though preliminary, PFS and OS rates were not significantly different and did not demonstrate that HIPEC was superior to SCS alone.
2.2 Further insight into the activity of PARP inhibitors (PARPi) and PARPi combinations
This year, SOLO2 (#6002), a randomized Phase 3 trial of olaparib maintenance therapy in women with relapsed platinum-sensitive BRCA-mutated ovarian cancer following response to platinum-based therapy, became the first report of mature OS data from a Phase 3 setting [
]. The study enrolled 295 patients and previously reported a 13.6 month improvement in the primary endpoint of PFS in 2017. In this final analysis of SOLO2, a median improvement in OS of 12.9 months was observed (38.8 versus 51.7 months), with a hazard ratio of 0.74 (95% CI 0.54–1.00; p = 0.0537). While this finding was just shy of the threshold for statistical significance, the magnitude of the observed OS benefit was striking. Additionally, it is important to note that this final analysis included patients who received subsequent PARPi therapy, including 38% of patients in the control arm and 10% of those in the olaparib arm. When adjusted for subsequent PARPi in the placebo group, the observed improvement in median OS was 16.3 months (35.4 versus 51.7 months), with a hazard ratio of 0.56 (95% CI 0.35–0.97). These results support the use of maintenance PARPi in patients with BRCA-mutated platinum sensitive recurrent ovarian cancer who are PARPi naïve.
The primary analysis of NRG-GY004, a Phase 3 trial comparing the combination of cediranib and olaparib or olaparib monotherapy to platinum-based chemotherapy in relapsed platinum-sensitive ovarian cancer was also presented (#6003) [
]. This study asked whether a non-platinum alternative could improve PFS over platinum-based chemotherapy in platinum-sensitive ovarian cancer. The combination cediranib and olaparib did not meet the primary endpoint of improving PFS compared to chemotherapy, with a median PFS of 10.4 months for the combination and 10.3 months for chemotherapy (hazard ratio 0.856, 95% CI 0.663–1.105, p = 0.077), although the observed activity was comparable for both PFS and objective response rates. Due to a hierarchical statistical design, the activity of olaparib monotherapy was not formally compared to chemotherapy, but a median PFS of 8.2 months was observed. Side effects with the combination of cediranib and olaparib were significant; while hematologic adverse events were higher with chemotherapy, rates of non-hematologic adverse events were higher with cediranib/olaparib, and the discontinuation rate due to adverse events was 21%. Pre-specified subgroup analyses were notable for the high response rates and significant activity seen with both olaparib monotherapy and cediranib/olaparib combination in patients with a germline BRCA mutation. Further investigation of the cediranib/olaparib combination is ongoing in NRG-GY005 and ICON9, and these studies will give us additional insight into the activity of this combination in these settings.
Two Phase 2 trials reported this ASCO also provided additional insight into PARPi activity as monotherapy or in combination. NSGO-AVANOVA2 (#6012) was a randomized Phase 2 study compared combination niraparib and bevacizumab to niraparib monotherapy in relapsed platinum-sensitive ovarian cancer [
]. In an updated analysis, there was continued improvement in PFS, with a hazard ratio of 0.34 (95% CI 0.21–0.54, p < 0.0001). With 52% event maturity, the hazard ratio for OS was non-significant at 0.75 (95% CI 0.44–1.28, p = 0.30). While these results support increased activity of anti-angiogenic/PARP inhibitor combinations over PARPi alone, the activity of this combination in comparison to standard of care platinum therapy has not been established. The LIGHT study (#6013) further characterized the activity of olaparib as primary therapy for relapsed PARPi naïve platinum-sensitive ovarian cancer [
]. As expected, activity in BRCA-mutated tumors (germline or somatic) was highest, with response rates of 64 to 69% and median PFS of 10.8 to 11.0 months, while activity in BRCA-wild type HR deficient tumors (as assessed by the Myriad MyChoice assay) was more modest (response rate of 29%, median PFS 7.2 months), and activity in BRCA-wild type HR proficient tumors was lowest (response rate of 10%, median PFS 5.4 months).
Two additional studies provided insight on the dosing of niraparib and raised questions about how best to follow patients on PARP inhibitor maintenance. Weight and platelet-based dosing of niraparib was incorporated into the PRIMA study after enrollment was approximately two-thirds complete. Mirza and colleagues (#6050) reported no impact on efficacy of this individualized starting dose, with a hazard ratio for PFS with maintenance niraparib of 0.59 (95% CI 0.46–0.76) in patients enrolled to PRIMA who received the fixed starting dose of 300 mg daily regardless of weight or platelets, and a hazard ratio of 0.69 (95% CI 0.48–0.98) for those patients who received the individualized starting dose [
]. The interaction test p-value for efficacy based upon starting dose was 0.30. Finally, a study by Tjokrowidjaja and colleagues (#6014) utilizing data from the SOLO2 trial intriguingly suggested that nearly half of patients experiencing RECIST progression on trial did not meet GCIG CA125 progression criteria [
]. Some of these patients still experienced a rising CA125; however, approximately one quarter of patients had a stable or falling CA125. This observation raises the question of whether patients on PARPi maintenance should have regular imaging as opposed to relying on CA-125 surveillance alone. However, the study did not report how many patients had symptoms, and it is possible that some patients without rising CA125 would have been identified due to symptomatic recurrence.
2.3 Antibody-drug conjugates and immunotherapy
Activity from the combination of the folate-receptor alpha-targeting antibody-drug conjugate mirvetuximab soravtansine in combination with bevacizumab in tumors demonstrating medium or high FRα membrane staining was reported (#6004) [
]. Overall, an objective response rate of 47% was observed, with a response rate of 64% in patients with high FRα expression. Response rates were high regardless of platinum status, with a response rate of 59% in platinum-resistant patients and 69% in platinum-sensitive cases. As reported at last year's European Society of Medical Oncology (ESMO) meeting, FORWARD1, the randomized Phase 3 trial of mirvetuximab soravtansine failed to meet its primary endpoint of improved activity compared to chemotherapy in platinum-resistant ovarian cancer; however, the selectivity of the assay used to determine FRα expression may also have contributed to this outcome. Results from the upcoming Phase 3 study MIRASOL of mirvetuximab soravtansine in true high FRα expressers will be of high interest in the further development of this interesting agent.
Final results from the KEYNOTE100 study of pembrolizumab monotherapy in relapsed ovarian cancer were also reported (#6005) [
]. Overall activity across the cohort remained low, with an overall response rate of 8.5%. However, there was a trend towards increased activity in patients with high PD-L1 expression (defined as a CPS score ≥ 10), with response rate ranging from 11.6 to 18.2% in these patients. These findings again highlight the limited activity of PD1/PD-L1-directed therapy in ovarian cancer and the need for biomarkers to identify the small percentage of patients who may derive benefit from these agents.