Repeated blood-brain barrier opening with a nine-emitter implantable ultrasound device in combination with carboplatin in recurrent glioblastoma: a phase I/II clinical trial

Abstract

Here, the results of a phase 1/2 single-arm trial (NCT03744026) assessing the safety and efficacy of blood-brain barrier (BBB) disruption with an implantable ultrasound system in recurrent glioblastoma patients receiving carboplatin are reported. A nine-emitter ultrasound implant was placed at the end of tumor resection replacing the bone flap. After surgery, activation to disrupt the BBB was performed every four weeks either before or after carboplatin infusion. The primary objective of the Phase 1 was to evaluate the safety of escalating numbers of ultrasound emitters using a standard 3 + 3 dose escalation. The primary objective of the Phase 2 was to evaluate the efficacy of BBB opening using magnetic resonance imaging (MRI). The secondary objectives included safety and clinical efficacy. Thirty-three patients received a total of 90 monthly sonications with carboplatin administration and up to nine emitters activated without observed DLT. Grade 3 procedure-related adverse events consisted of pre syncope (n = 3), fatigue (n = 1), wound infection (n = 2), and pain at time of device connection (n = 7). BBB opening endpoint was met with 90% of emitters showing BBB disruption on MRI after sonication. In the 12 patients who received carboplatin just prior to sonication, the progression-free survival was 3.1 months, the 1-year overall survival rate was 58% and median overall survival was 14.0 months from surgery.

Fig. 1. Trial flow diagram.

A total of 38 patients were enrolled in the trial, with 33 patients being implanted and receiving at least one sonication with the SonoCloud-9 device to disrupt the BBB. One patient had early discontinuation after implantation due to a pulmonary embolism before any device activation occurred. A total of 27 patients were treated with all 9 emitters of the device, with 15 patients treated with ultrasound (US) before carboplatin administration and 12 patients treated US immediately after carboplatin administration. A total of 4/33 patients completed the study and received six cycles of BBB disruption with the SonoCloud-9 at the time of carboplatin administration.

Fig. 2. BBB disruption case studies.

Pre and post-sonication images from two patients with nine emitters active showing the region of BBB disruption induced by the SonoCloud-9 System (red arrows indicated region of BBB disruption; blue line indicates position of SonoCloud-9 device).

Fig. 3. BBB disruption by site and closure dynamics.

Post-sonication MRI was performed for all patients enrolled on trial after the first three sonication sessions to verify BBB opening. A A significant difference in sonication-induced T1 contrast enhancement was found between sites due to MRI acquisition parameters and gadolinium contrast agent used, with Site 4 using Gadavist® and all other sites using Dotarem®. N = 61 sonications (in 27 different patients from cohorts C and D). Site 1: N = 22; Site 2: N = 3; Site 3: N = 2; Site 4: N = 16; Site 5: N = 15; Site 6: N = 3. The violin plots indicate the median (with dot), first and third quartiles (gray line), and min and max (colored contour). ANOVA testing was performed with post hoc Tukey–kramer (site 1 and 4: p = 0.0004). B The time between sonication and gadolinium bolus for T1w image acquisition was 10–77 min at Site 1, which was due to the availability of MRI after sonication. There was a significant negative correlation between enhancement intensity and sonication to gadolinium injection time for treatments performed at this site (in 10 different patients from cohort C and D treated in site 1, p = 0.05), with an exponential decay fit indicating a half-closure time for the BBB of 1.3 h. The dashed line indicates the 95% confidence interval (CI) for the exponential decay fit. Source data are provided as a Source Data file.

Fig. 4. Patient case study of radiological response.

Overall, tumor growth was better controlled within the field of sonication compared to outside the field of sonication in patients treated in Cohort D that received carboplatin infusion prior to sonication to disrupt the BBB. An example of the T1w contrast-enhanced evolution of the tumor volume from six monthly pre-sonication images is shown. This patient had an increase in T1 enhancement up to Cycle 2 that then decreased over time with each monthly cycle of treatment. The region outlined in green corresponds to nine cylinders, each 20 mm × 80 mm in front of each of the emitters of the ultrasound implant, which corresponds to the sonicated volume with an additional diffusion margin of 5 mm. Source data are provided as a Source Data file.

Fig. 5. Tumor growth was slower in patients receiving carboplatin before sonication.

The evolution of the tumor-related hyperintense T1w volume in the region targeted by the implant (shown in green in Fig. 4) was evaluated and is shown in A for both cohorts treated with 9 emitters active. A significantly higher growth rate over the study duration was found in cohort C (median = 2.31 mL/month) than in cohort D (median = 0.54 mL/month), as shown in B (Two-sided Wilcoxon–Mann–Whitney test: p = 0.04). The violin plots indicate the median (withe dot), first and third quartiles (gray line), and min and max (colored contour). When the region targeted by the implant was excluded from the analysis, there was no significant difference between the evolution of the T1 enhancement (p = 0.55). The local probability of tumor control was evaluated using T1w images at progression. A visualization of this analysis is shown in C in which each of the circles depicts an emitter axis from a real SonoCloud-9 implant in a patient. D The percentage of ring-shaped ROIs surrounding emitter axes covered with hyperintense tumor at the end of the study were compared between Cohorts C and D. The probability of T1w enhancement was lower in Cohort D than in Cohort C (two-sided Mann–Whitney U test, N = 26, p < 0.05 for radiuses up to 7.5 mm). The sonicated zone with BBB disruption corresponds to the 0–5 mm bin (10-mm cylinders), and effect on local tumor progression is observed up to 10-mm from the emitter axes (statistically significant up to 7.5 mm). Source data are provided as a Source Data file.

Fig. 6. SonoCloud-9 system.

The SonoCloud-9 System consists of three components: A an implant containing nine 1-MHz, 10-mm diameter ultrasound emitters that are powered by B a transdermal needle used to connect the device at each activation to C an external generator that includes a touchscreen interface to guide the treatment and provide the energy to the implant. At each activation of the device, the nine emitters of the implant are activated sequentially using 25-ms long pulses every 2 s (duty cycle = 1.2%) at the same time as an intravenous infusion of ultrasound resonators (Definity®, 10 μL/kg) for a total duration of 270 s. D The simulated region of BBB disruption (corresponding to a region of pressure >0.2 Mpa) in a hypothetical patient. E The activation procedure to disrupt the BBB was performed monthly at time of carboplatin infusion, with carboplatin infusion performed either immediately before (cohort D) or after (cohorts A, B, C) sonication. A, C reproduced with permission from Sonabend et al..