**Background:** Microtubule-targeting agents (MTAs) are effective cytotoxic chemotherapies, but resistance remains a major challenge. Lisavanbulin is the water-soluble lysine prodrug of avanbulin, which binds tubulin at the colchicine-binding site, inhibiting microtubule assembly and activating the spindle assembly checkpoint. Its mechanism also includes anti-vascular effects on the tumor microenvironment. A prior phase 1 study (CDI-CS-001) established the RP2D of 30 mg/m² as a 2-hour infusion, but dose-limiting vascular toxicities occurred at peak serum concentrations (Cmax), while anti-proliferative effects correlated with AUC. This led to the hypothesis that longer, lower-dose infusions could maximize efficacy while minimizing toxicity. Part 1 of the current study (CDI-CS-003) determined the MTD/RP2D for a 48-hour infusion to be 70 mg/m², achieving higher avanbulin exposure without significant vascular toxicity.
**Methods:** This was an open-label, multi-center phase 1/2a study (NCT02895360) conducted between August 2016 and August 2020 at six sites in Switzerland. The phase 2a portion enrolled two parallel cohorts: patients with platinum-resistant/refractory ovarian, fallopian tube, or primary peritoneal cancer (ovarian cancer cohort, n=11), and patients with glioblastoma at first recurrence (glioblastoma cohort, n=12). Key inclusion criteria included life expectancy ≥12 weeks, acceptable organ function, ECOG performance status ≤1 (ovarian) or ≤2 (glioblastoma), and measurable disease per RECIST v1.1 or RANO criteria. Patients received lisavanbulin 70 mg/m² as a 48-hour IV infusion on Days 1, 8, and 15 of 28-day cycles. Tumor assessments were performed at screening, end of study, and after every even-numbered cycle. The primary objectives were safety, tolerability, and anti-tumor activity.
**Key Results:** Twenty-three patients were enrolled (11 ovarian, 12 glioblastoma). Mean age was 61.9 years; all were Caucasian. In the ovarian cancer cohort, 81.8% had high-grade serous histology, and 63.6% had stage IV disease. All glioblastoma patients had prior surgery and radiotherapy. Median prior systemic treatment lines were 2 (range 1-4) for ovarian cancer and 1 (range 1) for glioblastoma. Mean treatment exposure was 84.1 days (81.6 vs 86.5 days). Treatment compliance was 98.3% overall. Most patients (22/23, 95.7%) reported at least one treatment-emergent AE. The most frequent AEs were fatigue (43.5%), constipation (34.8%), decreased appetite (30.4%), and abdominal pain (26.1%). Related AEs (any grade) occurred in 55% of ovarian cancer patients (grade 3-4: 27%) and 25% of glioblastoma patients (grade 3-4: 8%). Grade 3 hypertension occurred in two ovarian cancer patients (18%). No grade 5 toxicity occurred. Sixteen SAEs were reported by nine patients (39.1%), none considered treatment-related. No AEs led to treatment discontinuation. In the glioblastoma cohort efficacy-evaluable population (EEP, n=8), one patient had a confirmed partial response with target lesion reduction to <10% over 16 cycles, and one patient had stable disease for 10 cycles, yielding a disease control rate of 2/8 (25%) after eight cycles and an objective response rate of 1/8 (12.5%). In the ovarian cancer EEP (n=8), best response was stable disease in three patients (37.5% after two cycles, 12.5% after four cycles). All patients experienced disease progression. Six-month PFS rates were 9.1% (ovarian) and 16.7% (glioblastoma). Median time to progression was 54.0 days (ovarian) and 49.0 days (glioblastoma).
**Clinical Implications:** The 48-hour infusion of lisavanbulin at 70 mg/m² demonstrated a favorable safety profile compared to the prior 2-hour infusion, with lower rates of gastrointestinal, nervous system, and vascular toxicities. The study confirmed that extended infusion mitigates Cmax-dependent vascular toxicity while maintaining higher drug exposure. Anti-tumor signals were observed in both cohorts, notably a durable partial response in a glioblastoma patient. The authors note that EB1 expression may serve as a predictive biomarker for lisavanbulin response in glioblastoma, as suggested by prior studies, which could inform future patient selection. The favorable safety profile and signals of efficacy support further investigation of lisavanbulin, potentially in biomarker-selected populations.