Randomized phase II trial of weekly ixabepilone ± biweekly bevacizumab for platinum-resistant or refractory ovarian / fallopian tube / primary peritoneal cancer (NCT03093155): updated survival and subgroup analyses

Background: Ixabepilone may retain activity in paclitaxel-resistant disease. We previously reported improved response rates (ORR), progression-free (PFS) and overall survival (OS) conferred by ixabepilone+bevacizumab (IXA+BEV) compared to monotherapy (IXA) in heavily pre-treated ovarian cancers. We now describe a mature data set. Subset analyses were performed in patients with different taxane sensitivities and dose modi�cations. Methods Patients previously treated with paclitaxel were strati�ed by prior BEV and randomized to receive IXA 20 mg/m 2 days 1,8,15±BEV 10 mg/kg days 1,15 of a 28-day cycle in a multi-site prospective randomized phase 2 trial. Results: Thirty-seven patients were randomized to IXA and 39 patients to IXA+BEV. At the �nal data cutoff (05/27/2023), ORR was higher in the IXA+BEV arm (38.4% vs 8.1%,p=0.003). Dose reductions were necessary in most participants but did not diminish PFS/OS bene�t. Most patients were paclitaxel-refractory/-resistant (51%,n=19/37;67%,n=26/39); the remainder were taxane-responsive. The addition of BEV to IXA conferred bene�t in PFS (5.5 vs 2.2 mo; HR 0.31, 90%CI 0.20-0.49, p<0.001) and OS (10.3 vs 6.0mo; HR 0.56, 90%CI 0.38-0.84, p=0.02) that persisted after adjusting for prior taxane response. Conclusions: IXA+BEV has activity in heavily pre-treated ovarian cancers and offers signi�cant improvement in ORR and PFS/OS compared to IXA, despite prior taxane response and dose reductions. Clinical Trial Registration: NCT03093155


Introduction
Platinum and taxane combination chemotherapy represents the backbone of ovarian cancer treatment in the upfront setting. 1,2The combination of weekly paclitaxel with bevacizumab constitutes one of the most active regimens for platinum-resistant disease in patients treated with two or fewer prior lines of therapy. 3,4Unfortunately, heavily pre-treated patients continue to suffer from a lack of effective treatment options.After 5 lines, response rates are as low as 10-16% 5 and median overall survival after 3 lines may be as low as 5-9 months. 6oposed mechanisms of taxane resistance include enhanced drug e ux from the cell, rapid drug metabolism, dysregulation of microtubule-stabilizing proteins, altered binding a nity, upregulation of mediators of cell cycle progression, blunted spindle cell assembly checkpoint signals, 7 and alterations in signal transduction pathways that affect autophagy, senescence and in ammation. 8Though structurally distinct from taxanes, epothilones, such as ixabepilone (Ixempra® R-Pharm US, NJ), also hyperstabilize microtubules and may continue to exert effects in taxane-treated patients by overcoming these common resistance mechanisms. 9,10e conducted a phase II prospective multi-site comparison of ixabepilone with bevacizumab (IXA + BEV) versus ixabepilone monotherapy (IXA).With an initial data analysis date of 11/2020, the previous publication 11 showed IXA + BEV to be an effective combination for heavily pre-treated (median of 4 prior lines) platinum-resistant and taxane-treated ovarian cancer patients.We found that the bene ts of combination therapy on progression-free (PFS) and overall survival (OS) were not diminished by prior receipt of bevacizumab.IXA dose reductions were required by approximately 60% of participants in each arm most commonly due to peripheral neuropathy, neutropenia, and fatigue.The National Comprehensive Cancer Network (NCCN→) subsequently endorsed this combination as a category 2B treatment option for recurrent, platinum-resistant ovarian disease. 2 In the current manuscript, we present the nal survival analyses of the matured clinical trial data and further characterize the performance of the regimen in patients previously treated with weekly paclitaxel and among those requiring dose reduction.

Methods
Study design and conduct.This was an investigator-initiated phase II randomized open-label trial (NCT3093155) conducted at Smilow Cancer Hospital at Yale University and the Greenebaum Comprehensive Cancer Center at the University of Maryland.As previously described, 11 study participants were strati ed by study site and previous receipt of BEV with a 1:1 allocation using a dynamic randomization procedure to minimize strati cation-factor imbalance between arms.IXA monotherapy at 20 mg/m 2 intravenously days 1, 8, and 15 of a 28-day cycle was administered alone or with BEV 10 mg/kg intravenously days 1 and 15 administered until disease progression, death, or prohibitive toxicity (Fig. 1).[1] There were no signi cant amendments made during its conduct, though enrollment was brie y suspended during the COVID-19 pandemic.Power calculations assumed a median PFS for IXA monotherapy to be 5 months, given observations from GOG 126M. 12We required 80% power at 5% to detect a 2-fold increase in PFS via one-sided log-rank test while allowing for a single interim analysis for e cacy and futility.Calculations conducted in East v 6.4 (Cytel, Inc, Cambridge, MA) using the null variance estimator along with the O'Brien-Fleming spending functions for both alpha and beta required 28 PFS events in the interim and 56 PFS events in the nal analysis.The rst participant enrolled in March 2017.Because the COVID-19 pandemic slowed recruitment, we terminated enrollment after 78 participants and the occurrence of 61 PFS events.The present analyses represent a data cut-off of 05/27/2023.The full protocol is provided in the Supplemental Appendix.
Similar to previously published de nitions by others, 13 patients were considered taxane-resistant if they demonstrated disease progression within 6 months of paclitaxel/docetaxel administration.Patients were considered taxane-refractory if they progressed while receiving a taxane or demonstrated persistence of disease on end-of-treatment assessment that prompted initiation of a new line of therapy.
By default, the remaining patients were considered taxane-responsive.This investigation was conducted in accordance with the Declaration of Helsinki and approved by local Institutional Review Board.All patients provided written informed consent.
α Eligibility.Participants must have received prior treatment with 3 cycles paclitaxel, either 3-weekly or weekly.There was no limit on prior lines including BEV therapy.All participants were 18 years and had platinum-resistant (i.e., platinum-free interval <6 months) or refractory (i.e., disease progression during or 4 weeks after last dose of platinum) histologically con rmed epithelial (non-mucinous) ovarian, fallopian tube, or primary peritoneal carcinoma.All participants had measurable disease per RECIST (Response Evaluation Criteria in Solid Tumors) v1.1. 14Participants had to exhibit a performance status of 0-2. 15Any prior debulking status was permitted.Study Drugs.IXA was provided by R-Pharm US LLC, Princeton, NJ.BEV was supplied commercially.
Biosimilars were not permitted.Two 20% dose reductions were allowed for ixabepilone to 16 mg/m 2 then 12 mg/m. 2dpoints.Computed tomography was performed every 2 cycles.The primary endpoint was PFS, de ned as time from randomization to progression or death.Secondary endpoints were OS, de ned as time from randomization to death from any cause, and safety as de ned by Common Terminology Criteria for Adverse Events (CTCAE) v.4. 16Best response was based on RECIST v1.1.OR consisted of complete response (CR) or partial response (PR), and it did not have a durability requirement.Durable Disease Control (DDC) was de ned as CR, PR, or stable disease (SD) 6 months from date of best response.Statistical Analyses.PFS and OS were analyzed using the Kaplan-Meier method with one-sided log-rank tests and Cox regression.Multi-variate analyses were performed with paclitaxel response as a co-variate (sensitive, resistant, refractory).Fisher's exact tests were used to compare differences in best response between arms across subgroups.Individual patient responses were illustrated as a swimmer's plot.The datasets generated and analyzed during the current study are available from the corresponding author on request.

Results
Treatment and disease characteristics.Eighty-one participants were screened, and 78 were randomized (Fig. 1).One withdrew consent and one was found ineligible, leaving 76 evaluable for e cacy.Patient and disease characteristics showed no evidence of imbalance. 11Notably, 49% of participants had received > 3 prior lines of chemotherapy and 18% had platinum-refractory disease, with no statistical differences between the arms (p = 0.82 and p = 0.14, respectively).Within the IXA arm (N = 37), 9 (24%) of patients were taxane-resistant, 10 (27%) were taxane-refractory; within the IXA + BEV arm (N = 39), 13 (33%) were taxane-resistant, 13 (33%) were taxane-refractory (p = 0.44) (Table 1).Both arms contained patients previously treated with weekly paclitaxel, either in combination with carboplatin or in the platinum-resistant setting as part of an AURELIA 4 regimen: 35% (N = 13; IXA) and 26% (N = 10, IXA + BEV).[2] One additional patient in the IXA + BEV arm also received weekly nab-paclitaxel but was not included for purity of the analysis.Among all patients previously treated with weekly paclitaxel, two patients in each arm (5.% and 5.1%, respectively) had received weekly paclitaxel with bevacizumab and two patients in each arm had received weekly paclitaxel monotherapy.Of note, one patient received both ≥ ≥ ≤ ≥ weekly paclitaxel followed 16 months later by weekly paclitaxel with bevacizumab and is therefore counted towards both totals.
Secondary endpoints: OS, OR rate (ORR), safety.Since the original report, there was one additional response.ORR increased and remained higher in the IXA + BEV arm (38.4% vs 8.1%, p = 0.003).Median OS was 10.3 versus 6.0 months (HR 0.56, 90%CI 0.38-0.84,p = 0.02); this is virtually unchanged from the previous estimates.There were no complete responses, but in the combination arm, 14 patients achieved a durable response (stable disease or partial response > 6 months) (Fig. 2).

Discussion
We previously reported improved ORR, PFS, and OS conferred by weekly ixabepilone with biweekly bevacizumab (IXA + BEV) compared to monotherapy (IXA) in heavily pre-treated ovarian cancers.We suggested that prior BEV should not preclude re-treatment with the combination of IXA + BEV and highlighted pre-clinical observations that may explain the apparent synergy between these two agents. 11he present report also illustrates encouraging activity of IXA + BEV in patients despite prior exposure to weekly paclitaxel and in the setting of dose reductions to further inform clinicians following endorsement of the combination by the NCCN (category 2B). 2 The mature data reveal an impressive ORR of the combination arm (38.4% vs 8.1%, p = 0.003), with durable response in 14 patients and respectable median PFS of 5.5 months and OS of 10 months.Unsurprisingly, in a heavily pre-treated population, dose reductions were common, but ancillary analyses suggest this did not damper response and allowed patients to continue therapy.Our data suggest that a starting dose of 16 mg/m 2 can be considered and may warrant further study.
While favorable responses to weekly paclitaxel with 17 or without bevacizumab 18,19 after three-weekly paclitaxel are well-described, data quantifying the response to repeated treatment with weekly paclitaxel remain limited.Gunderson et al (2017) reported no attenuation of bene t for weekly paclitaxel despite prior exposure to weekly administration during primary therapy among 20 patients; compared to 79 patients previously exposed to three-weekly administration, the clinical bene t rate was actually improved (69% versus 36%, p = 0.05). 20In another small series of 26 patients 21 re-challenge with weekly paclitaxel after prior weekly paclitaxel produced a radiologic response rate of 34.6% with a median PFS of 3.7 months (95%CI 2.3-6.7) and OS of 18.1 (95%CI 9.6-28.6).Three patients received weekly paclitaxel for a 3rd time and 2/3 achieved stable disease.It is therefore not unreasonable to expect a subset of patients to bene t from dose-dense treatment with another microtubule-stabilizing agent such as ixabepilone following prior weekly paclitaxel.Ixabepilone may overcome paclitaxel resistance in several ways.Firstly, epothilones tend not to be substrates of drug exportation pumps that shuttle paclitaxel from the cell.Secondly, epothilones also appear to retain a nity for the microtubule despite upregulation of class III β-tubulin over the constitutively expressed class I β-tubulin, which reduces paclitaxel binding by altering the sterics of the pocket. 9,10Exploratory analyses employing whole exome sequencing among patients with a response to weekly paclitaxel for > 12 months suggested the importance of genes related to angiogenesis (VEGF, MMP9), tubulin superfamily (TSC2), apoptosis (BCL2L1, BAD) and interleukin pathways (CXCR1, CXCR2, IL1A, IL1B).Studies investigating predictors of response to weekly paclitaxel are surprisingly scant, and also lacking for epothilones, and should be the subject of future investigations.
In summary, the regimen of weekly ixabepilone and biweekly bevacizumab achieves high response rates and respectable PFS and OS in heavily pre-treated ovarian cancer patients.Response rates appear to be higher than those reported in this setting for chemotherapy alone, as well as for non-taxane chemotherapies combined with bevacizumab.Use of the combination does not appear to be precluded by exposure to prior bevacizumab or previous treatment with weekly paclitaxel.Biomarkers to predict sustained response to treatment with weekly ixabepilone and bevacizumab are needed.Declarations 2. A correction has been submitted to the original manuscript due to an error in Table 1.We previously reported 27% (N = 10) of patients in the IXA arm and 23% (of patients N = 9) in the IXA + BEV arm had been treated with weekly paclitaxel (either with carboplatin or as an AURELIA regimen with or without bevacizumab).Among these patients, we described 1 patient in the IXA arm and 4 patients in the IXA + BEV arm who had received weekly paclitaxel with bevacizumab and 1 patient in the IXA arm and 2 in the IXA + BEV arm who received weekly paclitaxel monotherapy.

Tables
Tables are available in the Supplementary Files section.Swimmer's plot of responses among 39 patients who received ixabepilone and bevacizumab.

Figures
Figures