Pazopanib and Fosbretabulin in recurrent ovarian cancer (PAZOFOS): A multi-centre, phase 1b and open-label, randomised phase 2 trial
Abstract
Objective: Vascular co-option is a known resistance mechanism to anti-angiogenic agents. This study aimed to determine the safety and efficacy of combining pazopanib, an anti-angiogenic agent, with fosbretabulin, an anti-vascular agent, in patients with recurrent epithelial ovarian cancer.
Methods: Eligible patients had recurrent epithelial ovarian cancer with a platinum-free interval (PFI) of 3 to 12 months. Patients were stratified based on PFI (greater than 6 months versus 6 months or less) and prior use of bevacizumab.
Results: In the Phase 1b trial, twelve patients were treated. The most common Grade ≥2 adverse events (AEs) were hypertension (100%), neutropenia (50%), fatigue (50%), and vomiting (50%). One dose-limiting toxicity (DLT) was observed, which was Grade 3 fatigue. The recommended Phase 2 dose level was fosbretabulin 54 mg/m² on Days 1, 8, and 15, combined with pazopanib 600 mg once daily (od) every 28 days. This regimen was then compared to pazopanib 800 mg od in a randomized Phase 2 trial. In the Phase 2 trial, twenty-one patients were randomized (1:1) to receive either the combination of fosbretabulin and pazopanib or pazopanib alone. Across both phases, four patients treated with the pazopanib and fosbretabulin combination developed reversible, treatment-related cardiac adverse events, leading to premature discontinuation of the study. In the Phase 2 trial, the median progression-free survival (PFS) was 7.6 months (95% CI 4.1–not estimated) in favor of the experimental arm (combination therapy) compared to 3.7 months (95% CI 1.0–8.1) for the control arm (pazopanib alone), with a hazard ratio (HR) of 0.30 (95% CI 0.09–1.03, P = .06).
Conclusions: It remains uncertain whether the combination of pazopanib and fosbretabulin is an efficacious regimen for treating epithelial ovarian cancer. Future trials will need to implement effective cardiac risk mitigation strategies to improve tolerability and ensure patient safety.
Introduction
Ovarian cancer is the leading cause of gynecological cancer-related mortality in North America and Europe, underscoring the urgent need for new treatment options. Vascular Endothelial Growth Factor (VEGF) has been established as a therapeutic target in ovarian cancer, but the benefits observed have generally been modest. One common mechanism of resistance to VEGF inhibitors is vascular co-option, where tumor cells exploit pre-existing blood vessels to obtain nutrients and oxygen. Combining VEGF inhibitors with vascular disrupting agents (VDAs), such as fosbretabulin, which target pre-existing tumor vasculature, is hypothesized to overcome this resistance mechanism, restrict tumor growth, and improve progression-free survival (PFS).
Fosbretabulin (formerly known as combretastatin A4 phosphate, or CA4P; Mateon Therapeutics, Inc., South San Francisco, CA, USA) is a synthetic phosphorylated prodrug of combretastatin A4, a naturally occurring derivative of the South African tree *Combretum caffrum*. It is a potent inhibitor of microtubule assembly. Initial evidence of its vascular disrupting activity was demonstrated in rodent cancer models and later confirmed in human solid tumors. Fosbretabulin targets pre-existing tumor neovasculature, causing an acute and reversible reduction in tumor blood flow, which leads to central necrosis within tumors. While VDAs like fosbretabulin have been shown to rapidly reduce blood flow to tumors, they exhibit limited activity as single agents due to rapid tumor regrowth. This is believed to be because circulating endothelial progenitor cells repopulate the tumor rim, which remains viable due to surrounding normal vessels. Inhibition of VEGF has been shown to prevent the recruitment of circulating endothelial progenitor cells, and combining VEGF inhibitors with VDAs has demonstrated additive anti-tumor activity in vivo. Clinical trials have confirmed the efficacy of the fosbretabulin-bevacizumab combination.
In this study, our objective was to determine the safety and efficacy of fosbretabulin in combination with the orally active anti-angiogenic agent pazopanib in patients with recurrent epithelial ovarian cancer.
Methods
Patient eligibility
The eligibility criteria for the Phase 1b and randomized Phase 2 trials were identical (clinicaltrials.gov reference NCT02055690). Eligible patients had recurrent epithelial ovarian, fallopian tube, or primary peritoneal carcinoma. All eligible patients had received at least one prior line of platinum chemotherapy and had progressive evaluable disease according to RECIST v1.1 and/or GCIG CA-125 criteria within 3 to 12 months of a platinum-based regimen.
Eligible patients were aged 18 years or older, had a WHO performance status of 0–1, adequate bone marrow function (hemoglobin ≥90 g/L, neutrophil count ≥1.5 × 10⁹/L, platelet count ≥100 × 10⁹/L), adequate renal function (either an uncorrected calculated creatinine clearance [Cockcroft-Gault equation] or corrected isotopic clearance measurement ≥40 mL/min), adequate liver function (bilirubin ≤1.5 × upper limit of normal [ULN], ALT or AST ≤2.5 × ULN), adequate blood coagulation parameters (PT or INR ≤1.3 × ULN and aPTT ≤1.2 × ULN), urine protein level ≤2+ or ≤2 g over 24 hours, and were clinically euthyroid. Eligible patients could have received bevacizumab prior to trial entry, provided the last dose was administered ≥6 months before the first dose of pazopanib/fosbretabulin.
Patients were excluded if they had undergone radiotherapy, surgery, or tumor embolization within 28 days of the first dose of pazopanib/fosbretabulin; had received endocrine therapy, immunotherapy, or chemotherapy during the previous 4 weeks before the first dose of pazopanib/fosbretabulin; had ongoing Grade ≥2 toxic manifestations from previous treatments; were known to be serologically positive for hepatitis B virus (HBV), hepatitis C virus (HCV), or human immunodeficiency virus (HIV); were pregnant or breastfeeding; had major thoracic/abdominal surgery from which they had not yet recovered prior to trial entry; had a history of cardiovascular disease within 6 months of trial entry, including coronary revascularization, acute coronary syndrome, symptomatic peripheral vascular disease, or congestive heart failure (NYHA Class III/IV); had sustained hypertension defined as three or more readings of systolic blood pressure (BP) ≥140 mmHg or diastolic BP ≥90 mmHg during screening; had clinically significant abnormalities on a screening electrocardiogram (EKG); had a history of cerebrovascular disease (including transient ischemic attack), pulmonary embolism, or untreated deep vein thrombosis within 6 months of trial entry; had clinical evidence of or symptomatic central nervous system (CNS) metastases/leptomeningeal carcinomatosis; had clinically significant abnormalities that could increase the risk of gastrointestinal bleeding or perforation; had evidence of active bleeding; were known to have endobronchial lesions or lesions infiltrating major pulmonary vessels; had clinically significant hemoptysis within 8 weeks of starting pazopanib/fosbretabulin; had been previously treated with pazopanib and/or fosbretabulin; had concurrent malignancies other than adequately treated conization-biopsied carcinoma in situ of the cervix or basal/squamous cell carcinoma of the skin. Written informed consent was obtained from each patient registered in PAZOFOS.
Study drugs
Pazopanib was administered orally in tablet form once daily, without food—either at least one hour before or two hours after a meal. The treatment was given in 28-day cycles. Fosbretabulin was administered as a weekly intravenous infusion via a peripheral vein, with each infusion lasting 10 minutes. It was given on days 1, 8, and 15 of each 28-day cycle, for up to six cycles. The dosage of fosbretabulin was determined based on the patient’s body surface area (BSA), with a maximum cap at a BSA of 2.2 m². When both drugs were administered together, pazopanib was given at least one hour before fosbretabulin.
In the experimental arm of the phase 2 trial, patients who completed six cycles of combined pazopanib and fosbretabulin treatment and demonstrated either a complete response, partial response, or stable disease per RECIST v1.1 criteria continued on pazopanib alone. This continuation lasted until either disease progression or the occurrence of unacceptable toxicity.
During the phase 1b trial, three dose levels were evaluated. Dose level 1 involved pazopanib 600 mg once daily combined with fosbretabulin 54 mg/m² on days 1, 8, and 15 every 28 days. Dose level 2 consisted of pazopanib 800 mg once daily with the same dose of fosbretabulin. Dose level 3 involved pazopanib 800 mg once daily and fosbretabulin increased to 60 mg/m² on the same schedule.
In the phase 2 trial, patients randomized to the experimental arm received the recommended phase 2 dose (RP2D), which was 600 mg pazopanib once daily and fosbretabulin 54 mg/m² on days 1, 8, and 15 of each 28-day cycle. Patients in the control arm received 800 mg pazopanib once daily, every 28 days, continuing until either disease progression or unacceptable toxicity occurred.
Study endpoints
The principal objectives of the initial phase 1b clinical investigation centered on the identification of dose limiting toxicities and the establishment of the maximum tolerated dose for the combined administration of pazopanib and fosbretabulin. The maximum tolerated dose was specifically defined based on the observation that no more than a single participant, within a cohort of at most six individuals receiving the same dosage, would experience a dose limiting toxicity. It was stipulated that any adverse event reaching a grade of two or higher in severity could be classified as a dose limiting toxicity. Furthermore, the criteria for a dose limiting toxicity were restricted to those grade two or higher adverse events that manifested during the first cycle of treatment. A dose limiting toxicity was further characterized as an adverse event for which a causal relationship with either fosbretabulin or pazopanib, or both, could not be definitively ruled out. The reporting of all adverse events was conducted in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.02.
The primary objective of the subsequent randomized phase 2 clinical investigation was progression-free survival. Progression-free survival was precisely defined as the duration extending from the date of a participant’s randomization into the study until the earliest occurrence of either progressive disease, as determined by RECIST version 1.1 criteria, or death. Tumor assessments, involving computed tomography scans and/or magnetic resonance imaging, were conducted at regular intervals of every eight weeks. The secondary objectives of this phase 2 trial encompassed the objective response rate, evaluated according to both RECIST criteria and RECIST/GCIG CA-125 criteria, overall survival, and safety. Overall survival was defined as the time elapsed from the date of randomization to the date of a participant’s death. The duration of therapy was defined as the period extending from the first day of the first treatment cycle involving either the combination of pazopanib and fosbretabulin or single-agent pazopanib until the date of the final administered dose of either pazopanib or fosbretabulin.
Statistical analysis
The principal aim of the initial phase 1b clinical investigation was to ascertain a recommended phase 2 dose for the combination of pazopanib and fosbretabulin through the determination of the maximum tolerated dose. The primary aim of the subsequent phase 2 clinical investigation was to establish whether the combined administration of pazopanib and fosbretabulin resulted in a statistically significant improvement in progression-free survival when compared to treatment with pazopanib alone in patients with recurrent ovarian cancer. The phase 2 clinical trial was specifically designed with sufficient statistical power to detect a progression-free survival hazard ratio of 0.65 when comparing the group receiving pazopanib and fosbretabulin to the group receiving pazopanib alone. This hazard ratio corresponds to an anticipated improvement in the progression-free survival rate at six months from 20% in the pazopanib-only group to 35% in the group receiving the combination therapy.
Angiogenesis biomarker study
The investigation into angiogenesis biomarkers was designed with the objective of determining whether a more pronounced anti-vascular effect could be observed in the group receiving both pazopanib and fosbretabulin when compared to the group receiving pazopanib alone. This comparison was to be made by analyzing changes in the levels of five circulating angiogenesis biomarkers: VEGF-A, VEGF-R2, Ang1, Ang2, and Tie2. To achieve this, longitudinal changes in the concentrations of these biomarkers in plasma were assessed at specific, pre-determined time points throughout the study. These time points included before the commencement of treatment, on day 1 of the first treatment cycle, on day 8 of the first treatment cycle, on day 15 of the first treatment cycle, and at the time of disease progression. Enzyme-linked immunosorbent assays were employed to quantify the concentrations of each of these five angiogenesis biomarkers.
The concentrations measured during the course of treatment were then compared, using a log ratio, with the pre-treatment plasma concentrations of the same biomarkers. This comparison was intended to identify any changes in biomarker levels that were associated with the administered treatment. Furthermore, the relationship between the observed changes in biomarker levels and the dose levels received by the patients was estimated through the application of linear regression analysis. The patterns of change in biomarker concentrations over time were assessed by fitting a general linear model to the biomarker concentrations, with time as a variable, after normalizing all concentrations against the baseline pre-treatment levels for each individual patient.
Results
Phase 1b trial
In the initial phase 1b clinical investigation, a total of twelve patients were enrolled. Initially, three patients received the first dose level, and another three patients received the second dose level, and none of these participants experienced a dose limiting toxicity during the first cycle. However, due to the occurrence of multiple delayed toxicities of grade two or higher in severity at the second dose level, which were experienced by patients after the completion of the first treatment cycle and notably included hypertension and neutropenia, the second dose level cohort was expanded to include an additional three patients. This expansion was undertaken to gain a more comprehensive understanding of the toxicity profile associated with this dose level.
Among these three additional patients, one individual developed a dose limiting toxicity, which was reported as grade three fatigue. Consequently, the second dose level was defined as the maximum tolerated dose. Following this determination, the first dose level cohort was also expanded to include another three patients, and no dose limiting toxicities were reported within this expanded group. Therefore, the first dose level was defined as the recommended phase 2 dose. The most frequently occurring adverse events, as classified by the Common Terminology Criteria for Adverse Events, in the phase 1b clinical trial were documented.
Notably, two patients participating in the phase 1b trial developed cardiac adverse events that were considered to be related to the treatment. Both of these patients experienced these cardiac adverse events after the completion of the first treatment cycle, and therefore, these events were not classified as dose limiting toxicities. A more detailed description of these treatment-related cardiac adverse events was provided in supplementary information.
Safety
The commonest treatment-related grade ≥ 2 CTCAEs are reported in Table 3. Importantly, in the randomised phase 2 trial, 2 more patients treated with pazopanib and fosbretabulin developed treatment- related cardiac AEs. A detailed description of the treatment-related cardiac AEs is provided in the supplementary data. The randomised phase 2 trial was therefore prematurely terminated because of the two grade 3 cardiac events in the phase 2 trial and the lower-grade treatment- related cardiac AEs reported in the phase 1b trial.
Efficacy
Prior to the termination of the clinical trial, twenty-one eligible patients had been randomly assigned to either the experimental or the control treatment group. The median progression-free survival in the experimental treatment group was 7.6 months, with a 95% confidence interval ranging from 4.1 months to a value that could not be estimated within the study period. In the control treatment group, the median progression-free survival was 3.7 months, with a 95% confidence interval ranging from 1.0 month to 8.1 months. The hazard ratio for progression-free survival, comparing the experimental arm to the control arm, was 0.30, with a 95% confidence interval ranging from 0.08 to 1.03, and a corresponding p-value of 0.06.
The median overall survival in the experimental treatment group, with only one event of death observed, had not yet been reached, with a 95% confidence interval ranging from 7.6 months to a value that could not be estimated. In contrast, the median overall survival in the control treatment group, where five events of death were observed, was 8.4 months, with a 95% confidence interval ranging from 1.0 month to a value that could not be estimated. The hazard ratio for overall survival, comparing the experimental arm to the control arm, was 0.10, with a 95% confidence interval ranging from 0.01 to 0.91, and a corresponding p-value of 0.04. Patient follow-up in the study ceased when the PAZOFOS trial was closed on December 5th, 2017.
Among the patients enrolled, twenty had disease that could be measured according to RECIST criteria. This resulted in an objective response rate of 18% in the experimental treatment group, with two out of eleven patients experiencing partial responses. In the control treatment group, the objective response rate was 22%, with two out of nine patients experiencing partial responses. The response rate as assessed by RECIST/CA125 criteria was 18% in the experimental treatment group, with two out of eleven patients responding, and 22% in the control treatment group, with two out of nine patients responding.
A subsequent analysis, conducted after the initial analysis, revealed that the median duration of therapy for all patients who received pazopanib and fosbretabulin was 156 days longer than for those patients who received pazopanib only. Specifically, the median duration of therapy was 268 days in the experimental arm compared to 112 days in the control arm, resulting in a hazard ratio of 0.43, with a 95% confidence interval ranging from 0.17 to 1.10, and a corresponding p-value of 0.08. The clinical trial was prematurely terminated due to safety concerns, without a full consideration of the outcome data. It was noted that if the safety concerns could have been effectively managed or mitigated, the study might have been considered to provide sufficient preliminary evidence to warrant proceeding to a phase 3 clinical trial, given the pre-specified design characteristics of the study, which included a beneficial effect estimate favoring the experimental treatment arm and a p-value less than 0.1.
Angiogenesis biomarkers
Longitudinal plasma samples were available for analysis from fourteen patients who received the combination of pazopanib and fosbretabulin in both the phase 1b and phase 2 clinical investigations. In contrast, only three patients who received single-agent pazopanib had both a pre-treatment, or baseline, plasma sample and at least one post-treatment plasma sample available for analysis. Due to this significant disparity in the number of available samples between the two treatment groups, a comparative analysis of biomarker changes between the study groups was not deemed statistically appropriate.
Within the group of patients who received pazopanib and fosbretabulin, statistically significant reductions in the concentrations of Ang1 and Tie2 were observed following treatment, specifically at day 15 of the first treatment cycle, with p-values of 0.003 and 0.030 respectively, as determined by paired t-tests. Conversely, the concentration of VEGF-A showed a statistically significant increase following treatment at both day 8 and day 15 of the first treatment cycle, with p-values of 0.012 and 0.002 respectively, as determined by paired t-tests. In the same group, the concentration of VEGF-R2 showed a statistically significant reduction by day 8 and day 15 of the first treatment cycle, with p-values of 0.025 and 0.008 respectively, as determined by paired t-tests. It was observed that at the time of disease progression in these patients, the concentrations of the measured biomarkers tended to return towards their initial pre-treatment levels.
Discussion
PAZOFOS represents the first clinical trial to document the safety and efficacy of combining fosbretabulin, a vascular disrupting agent, with pazopanib, an anti-angiogenic agent. The efficacy data derived from the PAZOFOS trial suggest that the combination of pazopanib and fosbretabulin has the potential to improve survival outcomes in patients compared to treatment with pazopanib alone. However, it is important to interpret these findings cautiously, considering the limited statistical power of the study due to its premature termination. The primary observation from the PAZOFOS trial was that while the combination of pazopanib and fosbretabulin demonstrated promising efficacy, it was also associated with serious, though reversible, cardiac toxicity, evidenced by elevated troponin levels and left ventricular dysfunction in two patients. Further investigation is needed to determine whether this observed cardiotoxicity arises as a secondary effect of the acute changes in blood pressure associated with fosbretabulin and/or pazopanib, or if it represents a direct effect of these agents on the coronary endothelium and/or the myocardium.
A recent preclinical study conducted in rodents provided evidence that CA4P, a related compound, induces treatment-emergent cardiac toxicity, exhibiting similarities to the toxicities reported in human clinical trials. In this rodent study, the administration of a single dose of CA4P at 120 mg/kg resulted in acute elevations in cardiac markers, such as CK and CK-MB, electrographic changes, and histological evidence of myocardial injury resulting from ischemia. Furthermore, a dose-escalation study of CA4P in healthy dogs demonstrated that six out of eight animals developed acute elevations in serum cardiac troponin I levels following a single intravenous infusion of either 75 mg/m2 or 100 mg/m2. These recent preclinical observations, when considered together with the documented cardiac toxicity reported in the PAZOFOS trial, underscore the necessity of implementing effective risk mitigation strategies to enhance tolerability and maximize patient safety before proceeding with phase 3 clinical trials investigating this combination.
The findings from PAZOFOS contribute to the growing body of evidence suggesting that combining an anti-angiogenic agent with a vascular disrupting agent can be effective in the treatment of recurrent ovarian cancer. A recent phase 2 clinical trial investigating the combination of bevacizumab and fosbretabulin (GOG 0186I) demonstrated encouraging improvements in progression-free survival in patients with relapsed ovarian cancer, with a reported hazard ratio of 0.69, a 90% confidence interval ranging from 0.47 to 1.00, and a p-value of 0.05. In this GOG 0186I trial, fosbretabulin was administered at a higher dose but with a less frequent schedule (60 mg/m2 on day 1 of every 21-day treatment cycle) compared to the PAZOFOS trial.
Moreover, the GOG 0186I trial incorporated prophylactic administration of anti-hypertensive agents two hours prior to the fosbretabulin infusion if a patient’s systolic blood pressure was 130 mm Hg or higher, or if their diastolic blood pressure was 90 mm Hg or higher, or if a patient had previously required blood pressure control for transient hypertension during a prior fosbretabulin infusion. While the GOG 0186I trial did not report specific cardiac toxicity, the incidence of grade 3 or higher hypertension was lower than that observed in the PAZOFOS trial. This suggests that implementing a different dosing schedule for fosbretabulin and pazopanib, coupled with a proactive and intensive algorithm for managing hypertension, might potentially minimize the cardiovascular risks associated with this therapeutic combination.
The angiogenesis biomarker study conducted within the PAZOFOS trial provided valuable insights into the molecular changes that occur following treatment with a combination of an anti-angiogenic agent and a vascular disrupting agent. Specifically, plasma concentrations of VEGF-R2 were significantly reduced following treatment, which is consistent with the expected inhibition of the VEGF signaling pathway by pazopanib. Similar pharmacodynamic changes have been reported in studies investigating other VEGF-receptor tyrosine kinase inhibitors. Furthermore, the concentrations of both Ang1 and Tie2 decreased following the initiation of treatment and subsequently increased at the time of disease progression.
This pattern aligns with previous observations made in our own research involving patients diagnosed with ovarian or colorectal cancer who were treated with bevacizumab. Indeed, Tie2 has been proposed as a potential biomarker for vascular response to VEGF inhibitors, where a reduction in its plasma concentration may indicate a benefit from treatment, while an increase may predict progressive disease. Due to the limited number of patients enrolled in the PAZOFOS trial, it was not possible to definitively verify this hypothesis in this study, although a similar pattern in biomarker trajectories was observed. An exploratory analysis indicated that different dose levels of fosbretabulin were not associated with the magnitude of the observed biomarker changes, suggesting that the pharmacodynamic changes observed may be primarily attributable to the effects of pazopanib.
Currently, validated therapeutic options for recurrent ovarian cancer are largely limited to traditional cytotoxic chemotherapy, which is often associated with reducing disease-free intervals and the inevitable development of chemotherapy resistance. Angiogenesis remains a critical aspect of tumor biology, essential for tumor growth, proliferation, and metastasis, and VEGF inhibitors have demonstrated activity in ovarian cancer regardless of platinum sensitivity. The results obtained from the PAZOFOS trial provide evidence that combining anti-angiogenic agents and vascular disrupting agents may potentially improve efficacy in this setting, albeit at the cost of cardiac toxicity.
This finding should stimulate further research and development of more potent combretastatin analogues, such as combrestatin A1 diphosphate (CA-1P). Future clinical trials investigating such combinations should carefully consider the complications reported in the PAZOFOS trial and explore the use of lower doses of the combination therapy, PFI-6 along with the implementation of stricter blood pressure control measures, to minimize the risk of cardiac toxicity. A focus on identifying the minimum biologically active doses of these agents may be crucial to fully exploit the potential of this active combination while ensuring patient safety.