Repeated blood-brain barrier opening with an implantable ultrasound device for delivery of albumin-bound paclitaxel in patients with recurrent glioblastoma: a phase 1 trial

Background Low-intensity pulsed ultrasound with concomitant administration of intravenous microbubbles (LIPU-MB) can be used to open the blood-brain barrier. We aimed to assess the safety and pharmacokinetics of LIPU-MB to enhance the delivery of albumin-bound paclitaxel to the peritumoural brain of patients with recurrent glioblastoma. Methods We conducted a dose-escalation phase 1 clinical trial in adults (aged >= 18 years) with recurrent glioblastoma, a tumour diameter of 70 mm or smaller, and a Karnofsky performance status of at least 70. A nine-emitter ultrasound device was implanted into a skull window after tumour resection. LIPU-MB with intravenous albumin-bound paclitaxel infusion was done every 3 weeks for up to six cycles. Six dose levels of albumin-bound paclitaxel (40 mg/m(2), 80 mg/m(2), 135 mg/m(2), 175 mg/m(2), 215 mg/m(2), and 260 mg/m(2)) were evaluated. The primary endpoint was dose-limiting toxicity occurring during the first cycle of sonication and albumin-bound paclitaxel chemotherapy. Safety was assessed in all treated patients. Analyses were done in the per-protocol population. Blood-brain barrier opening was investigated by MRI before and after sonication. We also did pharmacokinetic analyses of LIPU-MB in a subgroup of patients from the current study and a subgroup of patients who received carboplatin as part of a similar trial (NCT03744026). This study is registered with ClinicalTrials.gov, NCT04528680, and a phase 2 trial is currently open for accrual. Findings 17 patients (nine men and eight women) were enrolled between Oct 29, 2020, and Feb 21, 2022. As of data cutoff on Sept 6, 2022, median follow-up was 11.89 months (IQR 11.12-12.78). One patient was treated per dose level of albumin-bound paclitaxel for levels 1 to 5 (40-215 mg/m(2)), and 12 patients were treated at dose level 6 (260 mg/m (2)). A total of 68 cycles of LIPU-MB-based blood-brain barrier opening were done (median 3 cycles per patient [range 2-6]). At a dose of 260 mg/m(2), encephalopathy (grade 3) occurred in one (8%) of 12 patients during the first cycle (considered a dose-limiting toxicity), and in one other patient during the second cycle (grade 2). In both cases, the toxicity resolved and treatment continued at a lower dose of albumin-bound paclitaxel, with a dose of 175 mg/m(2) in the case of the grade 3 encephalopathy, and to 215 mg/m(2) in the case of the grade 2 encephalopathy. Grade 2 peripheral neuropathy was observed in one patient during the third cycle of 260 mg/m(2) albumin-bound paclitaxel. No progressive neurological deficits attributed to LIPU-MB were observed. LIPU-MB-based blood-brain barrier opening was most commonly associated with immediate yet transient grade 1-2 headache (12 [ 71%] of 17 patients). The most common grade 3-4 treatment-emergent adverse events were neutropenia (eight[47%]), leukopenia (five[29%]), and hypertension (five [ 29%]). No treatment-related deaths occurred during the study. Imaging analysis showed blood-brain barrier opening in the brain regions targeted by LIPU-MB, which diminished over the first 1 h after sonication. Pharmacokinetic analyses showed that LIPU-MB led to increases in the mean brain parenchymal concentrations of albumin-bound paclitaxel (from 0.037 mu M [95% CI 0.022-0.063] in non-sonicated brain to 0 .139 mu M [0 .083-0.232] in sonicated brain [3.7-times increase], p<0 .0001) and carboplatin (from 0.991 mu M [0.562-1.747] in non-sonicated brain to 5.878 mu M [3.462-9.980] mu M in sonicated brain [5.9-times increase], p= 0.0001). Interpretation LIPU-MB using a skull-implantable ultrasound device transiently opens the blood-brain barrier allowing for safe, repeated penetration of cytotoxic drugs into the brain. This study has prompted a subsequent phase 2 study combining LIPU-MB with albumin-bound paclitaxel plus carboplatin (NCT04528680), which is ongoing.