382P The potential utility of end-binding protein 1 (EB1) as response-predictive biomarker for lisavanbulin: Final results from a phase I study of lisavanbulin (BAL101553) in adult patients with recurrent glioblastoma (GBM)

Lisavanbulin (BAL101553, prodrug of BAL27862) promotes tumour cell death by modulating the spindle assembly checkpoint. BAL27862 is a lipophilic small molecule shown in rodents to penetrate the brain. In preclinical models BAL27862 inhibits microtubule plus EB1-dependent migration and invasion of GBM stem-like cells, with a profound EB1-associated survival advantage. Patients with histologically-confirmed recurrent GBM or high-grade glioma received QD oral lisavanbulin in a 3+3 dose-escalation design to determine the maximum tolerated dose (MTD). Adverse events were assessed by CTCAE v4.03 grade (G) and tumour response by RANO every 8 weeks. EB1 immunohistochemistry was undertaken in available tissue samples and whole-genome sequencing, methylation and transcriptome analysis was performed in a subset of patients. EB1 expression was also explored in separate retrospective tissue microarray (TMA) studies using three independent GBM cohorts. 28 patients (16M/12F; median age 50 y, min-max 25-71 y) number of prior regimens = 2 (1-5), received lisavanbulin at 8, 15, 20, 25, 30 or 35 mg/d. At 35 mg/d, DLTs of G3 hallucinations, G3 confusion and G2 gait disturbance were observed. The MTD was 30 mg/d with no DLT in 6 evaluable patients. One patient with an IDH-mutated, C-MYC amplified GBM with high EB1 expression had a partial response at 25 mg/d (88% area reduction per RANO) and continues on study >24 months. At doses of 25-30mg/d another 2 GBM patients had stable disease; and one patient had a radiological mixed response after 2 cycles, discontinuing the trial but remaining alive >24 months, resulting in an overall clinical benefit rate at 6 months of 44% (4/9 GBM patients). TMA data showed a prevalence of high EB1 expression in 2-5% of GBM. Oral lisavanbulin is well tolerated up to 30 mg/d in patients with GBM, and shows indications of clinical activity. Tissue analysis of an exceptional responder supports preclinical data of EB1-positivity as a potential biomarker of response. A phase 2a biomarker-driven study in GBM is expected to commence recruitment in 2020.