Clinical trial of blood-brain barrier disruption by pulsed ultrasound

See allHide authors and affiliations

Science Translational Medicine  15 Jun 2016:
Vol. 8, Issue 343, pp. 343re2
DOI: 10.1126/scitranslmed.aaf6086

A sound attack on brain tumors

Brain tumors are difficult to treat with chemotherapy because the blood-brain barrier greatly limits the delivery of drugs into the brain. Carpentier et al. have developed a pulsed ultrasound device, which they implanted into the skull of patients with glioblastoma, an aggressive and difficult to treat brain tumor, in a first-in-human trial. At regularly scheduled treatment sessions, the researchers activated the ultrasound device by connecting it to a power source, disrupting the blood-brain barrier long enough for subsequent chemotherapy to reach the brain. The authors confirmed that this approach was well tolerated and showed evidence of effectiveness to disrupt the blood-brain barrier, paving the way for further development of this therapeutic approach.


The blood-brain barrier (BBB) limits the delivery of systemically administered drugs to the brain. Methods to circumvent the BBB have been developed, but none are used in standard clinical practice. The lack of adoption of existing methods is due to procedural invasiveness, serious adverse effects, and the complications associated with performing such techniques coincident with repeated drug administration, which is customary in chemotherapeutic protocols. Pulsed ultrasound, a method for disrupting the BBB, was shown to effectively increase drug concentrations and to slow tumor growth in preclinical studies. We now report the interim results of an ultrasound dose-escalating phase 1/2a clinical trial using an implantable ultrasound device system, SonoCloud, before treatment with carboplatin in patients with recurrent glioblastoma (GBM). The BBB of each patient was disrupted monthly using pulsed ultrasound in combination with systemically injected microbubbles. Contrast-enhanced magnetic resonance imaging (MRI) indicated that the BBB was disrupted at acoustic pressure levels up to 1.1 megapascals without detectable adverse effects on radiologic (MRI) or clinical examination. Our preliminary findings indicate that repeated opening of the BBB using our pulsed ultrasound system, in combination with systemic microbubble injection, is safe and well tolerated in patients with recurrent GBM and has the potential to optimize chemotherapy delivery in the brain.


The blood-brain barrier (BBB) is a complex passive and active structure that surrounds the cerebral microvessels and protects the brain from exposure to potentially damaging substances, whether exogenous or endogenous in nature. However, because of the presence of the BBB, the development and application of potentially therapeutic substances in neurodegenerative and oncological disease remain difficult (14).

Preclinical studies have demonstrated that low-intensity pulsed ultrasound (US) directed to discrete brain areas is a minimally invasive technique that temporarily disrupts the BBB. Coupling US with systemic administration of micrometer-sized bubbles further facilitated and enhanced the delivery of a variety of therapeutics to specific brain areas in multiple animal models and prolonged survival in glioblastoma (GBM) models (515).

Despite more than 15 years of preclinical research, translation of focused US technology to clinical studies for repeated BBB disruption has been problematic. Small animal models have relatively thin skull bones in comparison to humans, which allows for focusing in the brain, even with simple, single-element US transducers. In humans, the thickness of the skull bone, which distorts and absorbs up to 90% of the US energy, requires the use of large hemispherical phased arrays with hundreds of individual US elements (16). This entails an extensive procedure that takes at least several hours, requires magnetic resonance imaging (MRI) monitoring during the procedure, and necessitates whole-head shaving and stereotaxy, which patients find unpleasant (16).

Translation of extracranial, focused US devices to clinical studies may be optimal for a single session disrupting the BBB in precise, deep-seated diseases of the basal ganglia, but it will be problematic for repeated monthly BBB disruption for diffuse and superficial pathologies. In anticipation of the limitations of extracranial focused US devices, we have developed a small implantable US transducer, SonoCloud, which allows for diffuse BBB opening with easily repeatable activation for repetitive chemotherapy/drug injections.

We previously reported the development of a small implantable US transducer, SonoCloud, and recent studies conducted in nonhuman primates (NHP) have confirmed the safe and effective transient disruption of the BBB with this unfocused low-intensity pulsed US system (7, 8, 12, 14). Anticipating the intractableness of most neuro-oncological and neurodegenerative diseases, we demonstrated that repeated exposure to the US system over a 3-month period had no behavioral, immunological, or neurological consequences in NHP (12, 14).

On the basis of extensive preclinical work, we were granted permission to transition our device system into clinical studies. Here, we report the interim results of our ongoing phase 1/2a trial aimed at determining whether repeated transient opening of the BBB is achieved by pulsed US with the implantable SonoCloud device and whether this procedure is safe and well tolerated in patients with recurrent GBM before receiving systemic chemotherapy with carboplatin.


Patient enrollment and treatment characteristics

Recruited patients had the 11.5-mm SonoCloud US device (Fig. 1) implanted within the skull bone in the extradural space and in the area of the tumor and its surrounding infiltrative region, even when the tumor was adjacent to or within eloquent regions (parts of the brain that control speech, motor functions, and sensory processes). The device was implanted either during a planned debulking surgical procedure or under local anesthesia in a 15-min procedure. Patient characteristics, dosing schedule, and radiological outcomes, including the depth and quality of BBB opening in the first five groups of patients enrolled and treated from July 2014 to January 2016, are shown in Table 1. A summary of the observed BBB opening is detailed in Fig. 2.

Fig. 1. The implantable US device.

(A) The US device is fixed in the skull bone and connected to an external power supply by a transdermal needle connection during activation. (B) The 11.5-mm-diameter US implant consisted of a 10-mm-diameter and 1-MHz US transducer encased in a biocompatible housing. The implant was passive and was connected by a transdermal needle to an external radiofrequency generator at each treatment to disrupt the BBB. Silicon spacers with thicknesses of 1, 2, and 3 mm were used between the implant and the skull bone to ensure that the front face of the transducer was flush with the inner surface of the skull bone.

Table 1. Patient treatment summary

. Five groups of patients were treated. Patient 3 (excluded from the table) did not receive US treatment because of a later diagnosis of radionecrosis. Patient 8 (excluded from the table) had spontaneous microhemorrhages on presonication MRI, and sonications were canceled. M, male; F, female; AEDs, anti-epileptic drugs; AUC, area under the curve; n/a, not available.

View this table:
Fig. 2. Contrast-enhanced T1-weighted MRI results.

(A) MR images obtained 2 days before and immediately after US treatments. The regions of enhancement after US sonications are indicated by white arrows in the postsonication images. The different classifications used for BBB disruption observed on contrast-enhanced T1-weighted MRI were as follows: grade 1, contrast enhancement in subarachnoid space (MRI of patient 6; session 3 at 0.8 MPa); grade 2, contrast enhancement in subarachnoid space and gray matter (MRI of patient 9; session 3 at 0.95 MPa); grade 3, contrast enhancement in subarachnoid space and gray and white matter (MRI of patient 10; session 4 at 1.1 MPa). (B) Summary of BBB opening observed relative to acoustic pressure for 41 treatments in 15 patients. The initial group of patients was first treated at an acoustic pressure of 0.5 MPa with the 1-MHz US implant. The subsequent groups were initially treated at 0.65 and 0.8 MPa and progressed up to a maximum level of 1.1 MPa. The relative grade of BBB disruption increased with acoustic pressure, with grade 3 openings only observed at 1.1 MPa.

The first cohort of three patients was started at an acoustic pressure of 0.5 MPa, which was conservatively determined on the basis of previous animal studies of BBB disruption (7, 8, 12, 14). Patients in the second cohort started at an initial acoustic pressure of 0.65 MPa with intrapatient dose escalation at subsequent sonication up to 0.95 MPa. Patients in the third, fourth, and fifth cohorts started at acoustic pressures of 0.8, 0.95, and 1.1 MPa, respectively, with intrapatient dose escalation at subsequent sonications up to 1.1 MPa. In the five cohorts, 15 patients had a total of 41 sonications performed, with 3 cycles at 0.5 MPa, 6 cycles at 0.65 MPa, 11 cycles at 0.8 MPa, 7 cycles at 0.95 MPa, and 14 cycles at 1.1 MPa. All sonication cycles were followed by systemic infusion of carboplatin. The pressure levels used and carboplatin cycles for each patient are shown in Table 1.

An adaptation of the microbubble dose to 0.1 ml/kg (instead of the fixed dose of 4.8 ml) was carried out starting with the second cycle for patients 9, 10, and 11 and for all subsequent cycles administered to all patients in the absence of dose-limiting toxicities (DLTs) to optimize the BBB opening (Table 1).

BBB disruption at acoustic pressures of 0.8 MPa and higher

In patients receiving 0.5 and 0.65 MPa, no BBB disruption was observed on contrast-enhanced T1-weighted MRI. Overall, BBB disruption was observed in 28 of 41 sonications, including 8 of 11 sonications at 0.8 MPa, 6 of 7 sonications at 0.95 MPa, and all 14 sonications at 1.1 MPa (Fig. 2 and Table 1).

As illustrated in Fig. 2A, which shows contrast-enhanced T1-weighted MR images taken 2 days before and 30 min after US treatment, different grades of BBB disruption were observed in patients 6, 9, and 10 (grades 1, 2, and 3, respectively). We also quantified the mean T1 enhancement using a 1-cm-diameter region of interest (ROI). Among the patients for whom BBB opening was detected, the mean enhancement was 15% (Table 1). Notably, the quality and depth of BBB opening increased with higher acoustic pressure, as seen with 6 of 14 sonications at 1.1 MPa, which were assessed by MRI as having grade 3 BBB opening (Table 1 and Fig. 2B).

Tumor progression

In previously reported clinical trials (1719) with carboplatin, recurrent GBM patients usually demonstrate tumor progression after 2 months of treatment. Here, one patient (#15), who had a grade 3 BBB disruption in four of four sessions, had no evidence of T1-weighted tumor progression and a slight reduction in the fluid-attenuated inversion recovery (FLAIR) signal after 4 months on the study (Fig. 3). Although this is an interesting observation, potential efficacy of the treatment cannot be determined with respect to constraints of phase 1 study design and unknown variables of each patient history.

Fig. 3. MR images from patient 15.

MR images of contrast-enhanced T1-weighted, FLAIR, and SWAN sequences for sonications 2 (top row; 2 months on study) and 4 (bottom row; 4 months on study) in patient 15. The T1-weighted images are from 2 days before sonication and 15 min after sonication. The FLAIR and SWAN sequences were obtained immediately after US disruption of the BBB by the SonoCloud implant. The rectangle shown is 20 mm × 60 mm and highlights the region where BBB disruption was observed. Each of the axial images was aligned and centered on the acoustic axis of the US transducer. The SonoCloud implant had a slightly larger artifact on the SWAN sequence because of the higher susceptibility of this sequence, but there was negligible artifact on all other MR sequences. No tumor progression or adverse effects such as bleeding (SWAN) or additional inflammation (FLAIR) were observed in this patient after four sonications, and the patient continues to be treated in the current study.

Safety of repeated BBB disruption by US in patients

Here, US treatments were well tolerated in all patients, and no DLTs were reported regardless of whether the BBB was observed to be disrupted on immediate postsonication contrast-enhanced MRI (28 of 41 sonications) or not (13 of 41 sonications). No evidence of acute hemorrhage, ischemia, or edema was observed in immediate postsonication susceptibility-weighted angiography (SWAN), diffusion, or FLAIR sequences. Despite the US emissions being directed to eloquent areas, patients did not present clinical symptoms or report any unusual sensations during the 2.5-min sonication period. When US treatments were performed in the language or motor areas, patients maintained speaking and movement capacity during the procedure and in the 6 hours after treatment, verifying the absence of US-induced neuronal effects. None of the patients presented any clinical symptoms within the hours or days after the procedure, including the 11 epileptic patients who remained stable. In addition, no cerebral carboplatin-related toxicity was observed at any time, consistent with reports in the literature of intra-arterial carboplatin delivery (20). No treatment-related serious adverse events were reported. A few minor adverse effects were noted. Patient 14 complained of local pain (7 of 10 on a visual analog scale) during the 2.5-min connection of the intradermal needle in the scalp. Patient 17 had a transient (3 min) vagal episode consisting of brachycardia, followed by tachycardia, sweating, and feeling of weakness, without full loss of consciousness, during the intravenous arm puncture for the microbubble infusion. There were two non–treatment-related adverse effects. Patient 6 had a 4-mm-diameter cerebellar vascular stroke at a distance of 3.3 cm from the acoustic axis of the transducer, as detected in the immediate postsonication MRI. Numerical simulations demonstrated that reflections from the bone interfaces should not have produced any high-pressure regions at the location of the stroke; therefore, it was considered unrelated to the sonication procedure by an independent scientific monitoring committee and the investigators. Patient 14 had an episode of transient peritumoral cerebral edema detected 14 days after sonication, requiring hospitalization and intravenous steroids; the edema resolved after 48 hours. The event was considered related to tumor progression and not related to the US because the BBB closes at a maximum of 48 hours after sonication. For all patients, when patient status deteriorated, this was caused by tumor-related clinical symptoms and consistent with radiological tumor progression.


We report the minimally invasive, safe, reproducible, and transient BBB disruption by pulsed US in patients with recurrent GBM. BBB disruption was observed in 28 of 41 US treatments at acoustic pressure of 0.8 MPa or higher, concomitant with systemic bolus injection of microbubbles (SonoVue). In a subset of patients (68%), the degree of BBB disruption was sufficient to allow the 1-kD gadolinium MR contrast agent to enter into the brain. Extrapolating from the gadolinium enhancement observed in T1-weighted images, we infer that concentrations of the lower–molecular weight carboplatin (0.3 kD) were also increased in the brain, as previously observed in NHP studies (14).

Our study used carboplatin as the chemotherapeutic agent of choice for three main reasons: first, carboplatin normally has low BBB penetrance (2.6% brain/blood ratio) because of its hydrophilic properties and size (21); second, carboplatin is commonly used for GBM recurrence with some efficacy reported in phase 2 trials (17, 18); and third, carboplatin was tested in GBM patients in high-dose direct intra-arterial delivery studies, with improved tumor control and low neurotoxicity as reported in phase 2 studies (20). The low permeability of the BBB is generally considered to contribute to the modest efficacy of systemically delivered carboplatin in patients with GBM (19).

The trial (NCT02253212) is still ongoing because the maximum tolerated dose of pulsed US exposure has not been reached, but we have sufficient data to report that transient BBB disruption with the SonoCloud system, followed by systemic administration of carboplatin, was achieved, with a threshold pressure dose of 0.8 MPa. Additionally, we have seen an increased quality of BBB disruption, as seen on MRI, up to a dose of 1.1 MPa, without toxicity in GBM patients subsequently exposed to carboplatin chemotherapy. Although the primary objective of this phase 1/2a trial was not to assess efficacy against the tumor, in most of the patients (n = 9) who had confirmed BBB disruption (grade 2 or higher), the region encompassed by the US field had no detected tumor progression on MRI.

On the basis of our results for these first 15 patients, it appears safe to increase the pressure level higher than predicted in preclinical models. The differences between the observation in the clinical study and the preclinical experiments on the tolerated acoustic pressure before the appearance of adverse effects could be related to the size of the human skull, which limits standing waves relative to smaller animal models (22). Other possible explanations are related to the nature of the microvessels in humans, which may be more resistant to damage through microbubble/vessel interaction, or to different physiological conditions such as anesthesia administration, which was performed in animal models but not in the clinical study (23). Together, the safe escalation of the pressure doses in additional cohorts should increase the concentration of chemotherapeutics delivered to the brain tissue, as previously reported (10). These additional cohorts will help to definitively identify the recommended acoustic pressure dose for future phase 2/3 studies that will focus on demonstrating clinical activity in patients with relapsed GBM.

Our study had some limitations. First, the narrow sonication field (1-cm diameter) was not sufficient to cover the entire desired volume of the tumor and peritumoral region in the patients treated. However, because of induced glial and immune reactions observed in animal studies (24), a ~5-cm3 BBB disruption treatment volume with our device was considered an important achievement, especially if it acts as an “opened gate” for potential locoregional efficacy (14). In the absence of toxicity, a larger volume of BBB disruption could be envisioned with a future iteration of the device design to increase the volume of enhanced chemotherapy delivery. Second, our current study design did not allow the measurement of carboplatin concentrations in the brain through an additional posttreatment procedure for ethical reasons. In future studies, surgical resection could be performed after BBB disruption and chemotherapy administration to sample tissue and measure drug concentrations. Third, the dynamics of BBB closure were not measured because this would have required an additional MRI exam at 24 hours. The BBB was intact by the following sonication (+4 weeks), and preclinical studies have reported BBB closure in <8 hours after US-induced disruption in small animal models (25).

In addition to optimizing chemotherapy penetration in the brain, another potential benefit of US relies on induction of immunological changes in the brain. Preclinical and clinical studies in other tumor types have demonstrated that US can facilitate and amplify an immune antitumor response, up-regulation of heat shock proteins, enrichment of the population of tumor-infiltrating lymphocytes (TILs), augmentation of dendritic cell activity, and enhancement of cytokine delivery to the brain, resulting in an anticancer immune response (26, 27). As an example, in human breast cancer specimens collected 1 to 2 weeks after focused US treatment, an increase in TILs of both T and B cell subsets at the margin of the ablated region was shown (26). Further, in a mouse model of brain metastasis from breast cancer, specialized derived natural killer cells were systemically delivered and successfully targeted the brain tumor only when combined with BBB disruption by focused pulsed US (28). Together, engaging the immune system with US, in combination with chemotherapy or immunomodulatory therapeutics, is an exciting direction for the future development of treatments for GBM.

Beyond the potential therapeutic delivery application in the field of neuro-oncology, promising research in Alzheimer’s disease mouse models has shown that disrupting the BBB using scanning US facilitated the removal of amyloid-β plaques, elicited a targeted neuroimmune response, and restored functional memory (24, 29). In complementary studies, investigators demonstrated that US treatment promoted hippocampal neurogenesis in aged mice, indicating additional mechanisms of action regarding memory improvement and the overall effect of US on specialized brain regions in rodents (30). These studies underscore the potential of US technology as a therapeutic approach to modify the course of neurodegenerative diseases such as Alzheimer’s disease. In the near term, safe disruption of the BBB in humans should open avenues for the development of many innovative strategies for targeting cancer and neurodegenerative diseases.


Study design

This study is a prospective, open-label, single-center, single-arm, dose escalation, phase 1/2a clinical trial, enrolling recurrent de novo GBM patients. This investigator-driven study was developed, conceived, and performed at the Assistance Publique–Hôpitaux de Paris (AP-HP) University Hospital La Pitié-Salpêtriere. All patients provided written informed consent in accordance with institutional guidelines. Approval was obtained from the ANSM (French National Health Agency; ref. 140488A-12) in June 2014, and Ile-de-France VI ethical committee (ref. CPP/38-14). The study was conducted in accordance with good clinical practices of the Clinical Research department of AP-HP. The trial was registered as NCT02253212, EudraCT 2014-000393-19, and IDRCB: 2014-A00140-47. The study began in July 2014 and is still ongoing at the time of publication. Here, we report the data for the first 15 patients treated in the first five cohorts up to the end of January 2016.

Patient selection

Patients experiencing recurrence (first, second, or third) of a histologically proven de novo GBM, after at least a first-line standard of care (radiation with concurrent and adjuvant temozolomide) were recruited. Qualifying patients had to have a growing contrast-enhanced tumor of less than 35 mm in diameter and be eligible for carboplatin-based chemotherapy.

Trial design

The trial was designed as a dose escalation study (31), in which the US pressure was increased throughout the study, starting at 0.5 MPa and increasing to 1.1 MPa through five different “dose” levels (0.50, 0.65, 0.80, 0.95, and 1.1 MPa). The initial US pressure was selected to be 0.5 MPa because this corresponded to the threshold for BBB opening in small animal studies (8, 11). A minimum of three patients were included at each dose level.

DLT was predefined as occurrence of an adverse effect directly related to the US emission during the first cycle of treatment, which would include the following: a neurological deficit starting within 2 days after the procedure and persisting at day 15, localized brain edema not preexisting before the procedure, occurrence of cerebral midline shift not controlled by routine treatment or requiring a salvage surgical procedure, partial epilepsy induced or enhanced after the procedure and not controlled by routine therapy, irreversible focal encephalopathy in the area of the BBB opening, bleeding or ischemia of more than 1 cm in diameter in the area of the BBB opening occurring within 2 days of the procedure, and brain herniation requiring salvage surgery. If none of the patients experienced a DLT, the next dose level was pursued. If one patient experienced a DLT, three additional patients would have been enrolled at that dose level. If no additional patients had a DLT, the dose level was increased after approval by the independent scientific monitoring committee. The maximum tolerated dose definition used was the standard one for oncology phase 1/2a studies (highest dose level at which six patients are started and fewer than two experience first-course DLT).

Patients received US for BBB disruption every 4 weeks. The first treatment was performed at the initial dose level for the inclusion group, the second treatment was performed at the next highest dose level, and escalation was continued at subsequent cycles in the absence of toxicity. Patients were treated monthly for up to a maximum of six treatments or until there was evidence of tumor progression. Hematological parameters were assessed for all patients to monitor for potential toxicity of carboplatin. Patients were treated with carboplatin only when platelet counts reached >105 cells/μl.

Surgical procedure

The US device was implanted within the skull bone overlying the tumor area (contrast-enhancing region or high-signal FLAIR region). When the tumor was next to or within eloquent regions, the device was implanted in this critical area in an attempt to prevent tumor progression and to have the highest potential enhancement to patient’s quality of life. If patients were eligible for a debulking surgery under general anesthesia, the device was implanted during this surgical procedure within a burr hole after dura matter closing and before skin closure. If surgical resection was not indicated, the device was implanted during a dedicated surgical procedure in an ambulatory fashion under local anesthesia. This procedure consisted of a 3-cm skin opening, creation of a burr hole without dura matter opening, and, finally, implantation of the device and closure of the skin. In all cases, neuronavigation systems could be used to position the device in the desired location. As a result, the transducer was in contact with the external face of the dura matter with no residual bone in between to have no distortion and no attenuation of emitted US. The US output intensity was then known, so no MRI monitoring during sonications was required.

Monthly MRI before and after BBB opening

All patients were assessed monthly using MRI, blood sampling, and clinical neurological evaluation. Two days before each planned US treatment, tumor status was evaluated by MRI. Progression-free survival and overall survival were also assessed. Patients left the study if tumor progression was identified according to the Response Assessment in Neuro-Oncology criteria (32).

In the absence of tumor progression and toxicity, the BBB opening session followed by carboplatin infusion was scheduled 2 days after the pretreatment MRI. A subsequent MRI exam was performed immediately after US treatment (~10 min after US-induced BBB disruption). If tumor progression was observed, the patient exited the trial and received an alternative chemotherapy drug without BBB disruption.

A 3.0T GE Signa MRI (GE Medical Systems) was used for the imaging exams. At each exam, standard FLAIR, T1-weighted contrast-enhanced (0.2 ml/kg, Dotarem), SWAN, and diffusion sequences were obtained. T1-weighted MR images were analyzed to grade the type of BBB opening observed. Four different grading stages were defined as follows: grade 0, no BBB opening; grade 1, contrast enhancement in subarachnoid space; grade 2, contrast enhancement in subarachnoid space and gray matter; and grade 3, contrast enhancement in subarachnoid space, gray matter, and white matter.

Quantification of MR contrast enhancement after BBB disruption

T1-weighted gadolinium contrast enhancement was quantified by calculating the signal change in a 1-cm circular ROI in images obtained at the presonication MRI (2 days before treatment) and in the images obtained immediately after US sonications using Osirix software (Pixmeo SARL). The ROI was placed at the site of the highest apparent T1 signal of the US region. The signal change in this ROI was corrected for changes in background signal by taking the average change in three separate ROIs outside of the tumor and sonication region. The pixel values were analyzed for significance before and after BBB disruption using a paired t test analysis in statistical software (R Project, R Foundation for Statistical Computing).

Ultrasound parameters

T he SonoCloud US implant (CarThera) consisted of a 10-mm-diameter US transducer that had a resonance frequency of 1.05 MHz and was encased in an 11.5-mm-diameter biocompatible housing. The transducer was operated with a burst length of 25,000 cycles (23.8 ms) at a pulse repetition frequency of 1 Hz (2.38% duty cycle) for a total duration of 150 s. The implantable US transducer was passive and contained no internal power supply. To activate the device for 2.5 min, a transdermal needle connection device was connected to the implant and to an external radiofrequency generator. The external generator was custom-designed and contained a graphical user interface that guided the practitioner through the treatment protocol. The US pressure level was defined as the pressure at the near-field to far-field transition in water (12-mm distance from the source) and was calibrated using a previously described procedure (8). A representative pressure field measured before implantation and simulated for the implant used in patient 2 is shown in fig. S1. Simulations were performed using the Fast Object-Oriented C++ Ultrasound Simulator (FOCUS) toolbox in MATLAB (MathWorks) for comparison to measurements (33). Although the device had a nominal diameter of 10 mm, the simulation showed that the measurements best matched simulations assuming a piston diameter of 8.3 mm. Additional measurements of this transducer after removal from the patient showed that the acoustic output was unchanged when compared with initial measurements.

The sonication was initiated at the beginning of a bolus injection of SonoVue microbubbles (Bracco). The initial clinical protocol was designed with a dose of SonoVue corresponding to 0.1 ml/kg with a maximum dose of 4.8 ml (one vial of SonoVue). After 18 treatments in nine patients who showed limited BBB opening, the ANSM authorized an increase in SonoVue corresponding to the weight of the patient at 0.1 ml/kg, and, thus, subsequent patients received doses of SonoVue that were dependent on their weight, with a maximum dose of up to 8.7 ml (87-kg patient) and a mean dose of 7.2 ml.

Carboplatin chemotherapy administration

Carboplatin was started no later than 60 min after BBB opening and was intravenously infused for 60 to 90 min. Carboplatin dose was calculated on the basis of AUC with the Calvert formula taking the renal function into account. The starting dose was AUC5, further adapted (AUC4 or AUC6) on the basis of clinical and biological monitoring. The AUC levels and milligrams of drug calculated with the Calvert formula (34) administered to each patient are shown in Table 1. The elimination half-life for carboplatin is between 2 and 6 hours (35).

Sample size and statistics

This was a phase 1 dose escalation study that used a standard 3 + 3 design (31). A minimum of three patients per US dose pressure cohort were enrolled to assess safety and tolerability. Patient enrollment was anticipated to vary from 2 to 30 patients for five cohorts to determine the maximum tolerated US pressure dose.


Fig. S1. Acoustic field of the SonoCloud device.


  1. Acknowledgments: We thank AP-HP for promoting P120905 and funding the clinical trial with the support of CarThera. We would like to thank the members of the data monitoring committee, which included the clinical research unit Groupe Hospitalier Pitié-Salpêtrière consisting of L. Anemet, N. Mansour, A. Bissery, L. Gambotti, and A. Mallet; members of the Direction de la Recherche Clinique consisting of C. Hoffart-Jourdin, A. Bergera, E. Pougoue, and P.-A. Jolivot; and members of the Independent Scientific Committee consisting of S. Katshian (methodologist and president), J.-S. Guillamo (neurologist and vice president), S. Pujet (neurosurgeon), J. Savatovsky (neuroradiologist), C. Fernandez (pharmacist), and A. Grandvuillemin (pharmacist and pharmacovigilance). We are also very thankful to I. Tabah-Fisch and G. Bouchoux for fruitful discussions during the preparation of the manuscript. Funding: The device was developed by CarThera. The clinical trial was sponsored by AP-HP (clinical research contract grant #12109) with support from CarThera. The research leading to these results has received funding from the program “Investissements d’avenir” ANR-10-IAIHU-06. Author contributions: A.C. invented the SonoCloud device; A.C., V.R., C.K., L.C., and P.C. performed the device implantations and assisted in BBB disruption treatments; A.C., M.C., and A.I. wrote the manuscript; M.C., A.V., A.C., J.-Y.C., and C.L. developed the technology; D.L. assisted in MR data acquisition; J.-Y.D., M.S., A.I., and K.H.-X. assisted in patient recruitment; A.V. prepared technical and regulatory documentation of the medical device; K.B. and C.H. prepared preclinical data for ethical committee submission of the trial; and A.I. served as principal investigator for the trial. Competing interests: A.C., C.L., M.C., and J.-Y.C. have ownership in CarThera SAS, a spin-off start-up company from Paris VI Sorbonne University and AP-HP. A.C., K.B., M.C., C.L., and J.-Y.C. have submitted a patent application on the technology via Sorbonne Universités, UPMC Univ Paris 06, and AP-HP, which licensed it to CarThera SAS. M.C. is a paid consultant to CarThera SAS. K.B. performed research supported by CarThera SAS.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article