2a. Study Design
We designed and conducted a multicentre, dose-escalation, open-label, phase IIa clinical trial to evaluate the safety and efficacy of olinvacimab in patients with recurrent glioblastoma. This study was an open-label, non-randomized trial conducted at two sites in Australia: Austin Hospital (Melbourne, Victoria) and Sir Charles Gairdner Hospital (Perth, Western Australia).
There were three dose levels for this study: dose level 1 (IV olinvacimab 8mg/kg weekly for 3 out of 4 weeks), dose level 2 (12mg/kg weekly for 3 out of 4 weeks) and dose level 3 (12mg/kg weekly). Patients were initially enrolled onto the lowest dose, after which sequential enrolment onto dose levels 2 and 3 occurred. The criterion for enrolment to the next dose level was based on the absence of ≥ grade 3 DLTs during the 1st cycle in the previous dose level (figure S2). A safety review committee (SRC) convened to determine the safety and decided on enrolment into the next level or changes in dosing frequency of study drug in case of occurrence of ≥ grade 3 haemangiomas or other DLTs during the 1st cycle.
Any subject who had to be withdrawn from the study before the completion of the 1st cycle, could be replaced by another patient. Subjects were treated for up to 1 year unless a cause for termination occurred, such as progression of disease (PD) or the withdrawal of consent.
2b. Patient selection
Eligible patients were ≥ 19 years of age with histologically confirmed primary Glioblastoma and MRI evidence of rGBM after initial treatment with chemoradiotherapy (CRT) using temozolomide. One previous recurrence/progression of glioblastoma with reintroduction/altered schedule of temozolomide was allowed. At least one confirmed measurable lesion or nonmeasurable lesion by Response Assessment in Neuro-Oncology (RANO) criteria was required . Expected survival ≥ 12 weeks was required.
Patients were required to have a Karnofsky performance status (KPS) score ≥ 80, acceptable haematologic parameters (ANC ≥ 1.5 x 109/L, Platelets ≥ 75 x 109/L, Haemoglobin ≥ 9.0 g/dL), coagulation tests (Prothrombin time ≤ 1.5 x ULN, Activated partial thromboplastin Time ≤ 1.5 x UNL) and biochemical studies (Total bilirubin ≤ 1.5 x UNL, Aspartate aminotransferase or alanine aminotransferase ≤ 3 x ULN and Creatinine clearance (CrCl) ≥ 30 mL/min).
Key exclusion criteria were: uncontrolled hypertension, uncontrolled seizures, severe heart failure by New York Heart Association classification, oxygen-dependent chronic disease or active major psychiatric disorder. No prior therapy with VEGF targeted agents was permitted, nor was treatment with systemic chemotherapy, hormonal therapy, immunotherapy or biologic therapy except CRT or temozolomide alone within 2 weeks prior to the baseline study visit. Concomitant therapies (i.e corticosteroids, anticonvulsants, analgesics etc.) were allowed at the investigator's discretion.
Prior single-agent phase I data for olinvacimab in non-CNS tumours established a Maximum Tolerated Dose (MTD) of 24 mg/kg  on days 1,8,15 every four weeks. However, this study did not include patients with GBM. PK data from this trial also revealed adequate trough concentrations (Cmin >20 µg/mL)- previously determined in vitro (HUVEC) and in vivo (COLO205)- using doses of 10mg/kg and 12mg/kg. The 8mg/kg dose level also remained above this trough concentration at all points but one. This was used as the starting dose for the current phase IIa study for patients with rGBM. Enrolment to the next dose level was allowed if no patients experienced grade ≥ 3 haemangiomas or other dose-limiting toxicities (DLTs) during the 1st cycle of the prior dose level. Grade ≥ 3 haemangiomas were defined as covering > 30% body surface area and/or requiring urgent invasive intervention.
Olinvacimab was initially administered as an intravenous infusion lasting approximately 90 minutes, followed by an observation period of 90 minutes. If the first administration was welltolerated, subsequent administrations consisted of a 60-minute infusion followed by a 30-minute observation period
2d. Endpoints and assessments
This was an exploratory study to evaluate the safety and dose-response of olinvacimab in patients with recurrent glioblastoma. The primary endpoint was to examine the safety of olinvacimab at pre-specified dose levels in patients with rGBM.
The study was planned to recruit a total of 12 patients: three for dose levels 1 and 2 respectively and six patients for dose level 3. Unless specified, the significance level of all analyses was set as 0.05 using two-sided testing.
RANO criteria were used for response assessment . Progression Free Survival (PFS) and Overall Survival (OS) data were examined for each treatment group using the Kaplan-Meier method with 95% confidence intervals. PFS was defined as from the date of the drug administration to disease progression time point. OS was defined as from the start date of administration of the IP to the subject’s death.
Objective response rates (ORR) for each treatment group was also noted. ORR was defined as the frequency of patients whose integrated assessment of response at terminational visit was either a complete response (CR) or partial response (PR). DCR (disease control rate) was defined by the frequency of patients who had either a CR, PR or stable disease (SD).
2e. Pharmacokinetic and pharmacodynamic assessments
Pharmacokinetic parameters were assessed using the same parameters and methods as the prior phase I study. The same serum angiogenic markers were also assessed given their relationship to VEGFR-2- the target of olinvacimab. The 1st and 3rd infusion samples were analysed for dose levels 1 and 2, while 1st and 4th infusion samples were analysed for dose level 3.
Dynamic Contrast-Enhanced MRIs (DCE-MRIs) were analysed for changes in perfusion parameters during treatment with olinvacimab. This was an exploratory endpoint for the study. Analysis and data management was undertaken by a central imaging core laboratory in Seoul, South Korea. DCE-MRIs were performed on up to three occasions, six weeks apart. Perfusion parameters assessed included K-trans and initial area under the gadolinium curve (IAUGC) values. For patients with multi-focal lesions, a separate analysis of repeated measures with a linear mixed models was performed. Careful collaboration between the two radiology sites ensured high quality data that was acquired with similar parameters by coordination between the two sites.