International Society for Therapeutic Ultrasound Conference 2016

ORAL PRESENTATIONS

O1 The role of bubbles and cavitation in therapy ultrasound

Brian Fowlkes

Basic Radiological Sciences Division, Department of Radiology, University of Michigan, Ann Arbor, Michigan, USA

When exposed to sufficiently high ultrasound pressures, microbubbles can be generated spontaneously in tissue and undergo inertial cavitation where collapses result in physical effects. These effects range from petechial haemorrhage to complete cellular disruption, termed Histotripsy, depending on ultrasound parameters. This presentation will explore the mechanisms associated with histotripsy along with the tissue effects and the wide range of potential applications for this mechanical disruption method.

O2 Challenges for clinical trials in therapeutic ultrasound, the need for an evidence base, & trial design

Pejman Ghanouni

Stanford School of Medicine, Stanford, California, USA

In this lecture, we will compare the design of clinical trials that led to approval of MR guided focused ultrasound for the treatment of uterine fibroids and osseous metastases. The impact of these trials on the evidence base, and thus on adoption by users and coverage by insurers will be compared. We will also review the process of expanding approved FUS applications, either via investigator- or industry-initiated studies or through off-label clinical use.

O3 Prostate HIFU – current status

Narendra Sanghvi

SonaCare Medical, Indianapolis, Indiana, USA

In this lecture, present status of focused ultrasound for the treatment of localized prostate cancer ablation will be discussed. High intensity focused ultrasound (HIFU) has been used for the ablation of prostate over two decades and has treated thousands of prostate cancer patients. Meanwhile, prostate cancer management is undergoing significant improvements as molecular markers, targeted biopsy and advanced multi-parametric MRI are routinely used to accurately localize the prostate cancer. These advances offer a unique opportunity for focal ablation of localized prostate cancer with HIFU as it plays a significant role in reducing morbidity and treatment cost. This presentation will focus on hardware design, software architecture and HIFU features of the devices. Presentation will demonstrate localization of prostate with ultrasound imaging, treatment planning with 3D volumetric rendering of the prostate with ultrasound and MRI fusion techniques for focal treatment and finally HIFU dose setting with guidance using real time Tissue Change Monitoring (TCM) with the Sonablate device. The presenter will encourage exchange of ideas and discussion for research topics.

O4 Enhancement of drug delivery - clinical challenges and solutions

Constantin Coussios¹, Paul C. Lyon¹, Michael Gray¹, Christophoros Mannaris¹, Marie de Saint Victor¹, Eleanor Stride¹, Robin Cleveland¹, Robert Carlisle¹, Feng Wu⁴, Mark Middleton³, Fergus Gleeson²

¹Institute of Biomedical Engineering, University of Oxford, Oxford, United Kingdom; ²Department of Radiology, Churchill Hospital, Oxford, United Kingdom; ³Department of Oncology, University of Oxford, Oxford, United Kingdom; ⁴HIFU Unit, Churchill Hospital, Oxford, UK

There are four key clinical challenges in optimizing drug based strategies: (i) achieving prolonged blood circulation of the therapeutic to enable active or passive accumulation in the target tissue; (ii) mediating triggered release or activation, or active accumulation of the therapeutic, to maximize its concentration at the target site whilst reducing off target side effects; (iii) enabling successful transport of the therapeutic from the blood stream into the target tissue, and achieving a homogenous distribution in that target tissue and (iv) where necessary, further enabling penetration of the therapeutic into the cell. The potential of therapeutic ultrasound with, or without, sonosensitive microparticle or nanoparticle formulations to address these challenges will be explored.

O5 Neuromodulation with ultrasound for beginners

Jean-Franҫois Aubry

Institut Langevin, Paris, France

In this lecture, the use of transcranial low intensity focused ultrasound for neuromodulation will be discussed. A historical review will be presented, with an emphasis on the experimental setups and the acoustical parameters. Models ranging from slice cultures to intact rodents and primates will be presented, together with recent trials on humans. Potential mechanisms will be described. Based on our own experience, exciting successful neuromodulation as well as disappointing failures will be presented.

O6 Thermometry in ultrasound fields, challenges & solutions in vivo, ex vivo & everywhere else!

Kim Butts Pauly

Stanford School of Medicine, Stanford, California, USA

This talk will cover image-based thermometry methods for guiding focused ultrasound. MR Thermometry based on the proton resonance frequency shift with temperature is linear and reversible in aqueous tissues and is utilized in clinical practice with common temperature resolutions of 1°C. Ultrasound based thermometry based on the speed of sound change with temperature can be used to 45°C in aqueous tissue. Both methods are sensitive to the presence of fat within the aqueous tissue, as well as motion. This talk will cover these basic concepts as well as their use in clinical practice.

O7 Motion compensation

Chrit Moonen

Center for Imaging Sciences, Imaging Division, University Medical Center, Utrecht, the Netherlands

Motion leads to several challenges for HIFU treatment. In this lecture, the effects of respiratory, cardiac, and peristalsis related motion on MRI thermometry will be discussed. In addition, methods for tracking the moving target with the HIFU will be described, as well as gating strategies.

O8 MRgFUS crossing the chasm from proof of concept to mainstream treatment alternative

Jacob Vortman

InSighTec, Haifa, Israel

MRgFUS treatment is a disruptive, non-invasive, outpatient treatment alternative that is capable to treat tumours and functional disorders under real-time monitoring and control via MR thermometry.

Transforming this breakthrough technology from the lab to a mainstream treatment alternative requires gaining the support and the agreement to change by a whole array of stakeholders in different areas some interrelated and some conflicting.

The stakeholders’ current position, changes they will need to go through and possible changes engines are mentioned below:

▪ Physicians (surgeons) need to transform into image guided surgeons where knowledge and understanding of the disease play a dominant role in the procedure outcome. The benefit is the confidence in the safety and efficacy coupled with income that wouldn’t decline. The change engine could be the patients and the payers.

▪ Payers should benefit from covering MRgFUS by saving cost and addressing patients’ demands. The change engine in this case should be the patients, governments and physicians.

▪ Governments should see the benefit of very fast recovery, very low level of adverse events and productivity enhancement. In this case patients and physicians should drive the change.

▪ Patients should benefit from safer treatment, fast recovery next day back to your life, minimal trauma and morbidity. The significant benefit to them should transform them to the dominant driver of this change. They will need to influence physicians, payers and providers to adopt this new treatment.

▪ Providers should adopt the technology and provide this treatment since the data exist proving safety and efficacy, proven cost savings and physicians and patients demand.

The current Medical ecosystem is biased against new technologies since the incumbent system/treatments are reimbursed while the new technologies are not. Could governments perform economic analysis and if found beneficial (example: saving money and improving productivity) decide on limited 2 years reimbursement during which RCT data will be collected based on which private insurance will decide to cover. This model should incentivize the physician, payers and providers to try the new technology.

Fig. 1 (abstract O8).

Overcoming the resistant to change - is there a strategy that could bring all the different stakeholders to combine and align efforts?

O9 MR guided focused ultrasound treatment of soft tissue tumours of the extremities

Pejman Ghanouni

Stanford School of Medicine, Stanford, California, USA

In this lecture, the use of MR guided focused ultrasound for the treatment of soft tissue and osseous tumours will be discussed. Results of treatment of these different types of tumours will be reviewed, including lessons learned from challenging patient treatments. Technical aspects of all parts of a treatment, including patient preparation, positioning, imaging, planning, thermometry, and methods of evaluation, will be described. The talk will also focus on methods developed to address these current challenges and opportunities for future development.

O10 Non-invasive, non-destructive FUS-induced neuro-modulation assessed by recording auditory evoked potentials – initial experience in small/large animals

Shirley Sharabi^1,2, Dianne Daniels^1,2, David Last¹, David Guez¹, Yoav Levy³, Alexander Volovick³, Javier Grinfeld³, Itay Rachmilevich³, Talia Amar³, Zion Zibly¹, Yael Mardor^1,2, Sagi Harnof^1,2

¹Sheba Medical Center, Ramat-Gan, Israel; ²Tel-Aviv University, Tel-Aviv, Israel, ³InSighTec, Haifa, Israel

Objectives

MR guided focused ultrasound (MRgFUS) has been extensively studied in recent years as a non-invasive treatment modality. Initial clinical trials have indicated promising treatment response to ablative FUS treatments of patients with brain tumours, neuropathic pain, essential tremor, obsessive compulsive disorder, and Parkinson’s disease. Apart from the ablative applications of FUS, this technology has been extensively evaluated for less destructive applications such as thrombolysis, blood–brain barrier disruption for increased drug delivery, and recently also neuro-modulation.

The objective of the presented study was to demonstrate non-invasive, non-destructive, reversible FUS-induced neuro-modulation in small (rats) and large (pigs) animals by inducing temporary suppression of auditory evoked potentials.

Methods

All animal experiments were performed under full anaesthesia. Rats were anesthetized by Xylazine/Ktalar and pigs were anesthetized by Propofol. Rat’s heads were shaved prior to treatment but they did not undergo craniotomy. Pigs underwent craniotomy to avoid FUS reflection/aberration by the skull. EEG was recorded using 3 small cup-shaped electrodes attached to the skin/dura with metal particles-containing gel for optimal sound conductivity (Fig. 2).

The audio stimulation system consisted of a pulse generator connected to the EEG trigger input and to speakers placed near the animal ears, producing a square-wave form at 10 KHz, resulting in 150 “click” sounds per min (Fig. 2). Each measurement consisted of 200 repetitions enabling acquisition of a full measurement in 1’20” min.

The ExAblate Neuro system (InSightec, Tel Aviv, Israel) is a combination of a standard MRI scanner and a FUS delivery system. The FUS device is in the shape of a helmet consisting of 1024 transducers which deliver US energy in the form of “sonications”. The system is designed to provide real-time therapy, planning, thermal dosimetry, and closed loop therapy control. Treatment starts with conventional MRI scans, displayed on the ExAblate computer, used to determine regions of interest of the target volume. During the procedure, the beam path is periodically reviewed to confirm the planned direction through the tissue. The set of sonication volumes is sequentially applied to cover the entire planned volume. The current experiments were performed with a modified ExAblate version developed for neuro-modulation as part of the MAGNET programs supported by the Israeli Ministry of Commerce.

Baseline auditory evoked potentials were recorded by EEG prior to FUS treatment, with the animals in the prone position. The animals were then placed in the supine position, with the skull dipped in degassed water at the centre of the FUS system, for localization MRI scanning followed by FUS treatment. The animals were then returned to the prone position for continuous post-treatment EEG recordings. Rats which did not show recovery of the auditory evoked potentials 30–60 min post treatment were monitored again 48 hours or 1 week post treatment.

The animals were treated by FUS for 52 sec using the Exablate Neuro system at 220 KHz, 12 W, and 100 ms on/ 2900 ms off pulses. Two rats were treated in the thalamus region (targeted at deep auditory tracks) and another two in the frontal cortex region (targeted at peripheral auditory tracks). Two sham rats underwent a similar procedure without activation of the FUS system. One pig was treated in the thalamus region and another in the right motor cortex region.

Results

Auditory evoked potential EEG signals shapes varied from one animal to the other but all were detected 2–10 ms after the trigger. The maximal peak-to-peak height was calculated for each measurement.

The sham rats showed no significant change in the auditory evoked potentials EEG signal.

The rats treated in the thalamus regions showed 50% and 65% suppression of the baseline auditory evoked potentials EEG signal. The first showed no recovery 2 hours post treatment with full recovery measured 1 week post treatment. The second showed no recovery for 1 hour post treatment and full recovery 48 hours post treatment. The rats treated in the cortex regions showed 50% and 67% suppression of the baseline EEG signal. The first showed no recovery 30 min post treatment with full recovery measured 1 week post treatment. The second showed initial recovery 14 min post treatment reaching full recovery within 28 min post treatment.

The pig treated in the thalamus region showed 90% suppression of the baseline signal with no recovery 30 min post treatment. The second pig, treated in the cortex region, showed complete suppression of the baseline signal immediately post treatment with initial recovery noted 18 min post treatment, reaching full recovery 63 min post treatment (Fig. 3).

Conclusions

Our preliminary results suggest that reversible neuro-modulation by non-invasive FUS is feasible. Full recovery was noted in all 4 treated rats and in 1 of the 2 treated pigs. Unfortunately we were not able to monitor the first pig for more than 30 min post treatment.

Fig. 2 (abstract O10).

See text for description

Fig. 3 (abstract O10).

See text for description

O11 Biophysical dissection of ultrasonic neuromodulation mechanisms

Michael Plaksin, Yoni Weissler, Shy Shoham, Eitan Kimmel

Faculty of Biomedical Engineering & Russell Berrie Nanotechnology Institute, Technion – Israel Institute of Technology, Haifa, Israel

Objectives

Low intensity US can noninvasively suppress or excite central nervous system (CNS) activity using different combinations of stimulation parameters. While applications are already emerging, the underlying biophysics remains unclear regarding the relative contribution of possible mechanisms: extracellular bubble cavitation, thermal effects, acoustic radiation pressure and US-induced intramembrane cavitation within the bilayer membrane (the bilayer sonophore or BLS model). Interestingly, both radiation pressure and intramembrane cavitation can induce plasma membrane capacitance changes. Here, we use detailed predictive modelling and find that only intramembrane cavitation can explain all the observed aspects of ultrasonic neuromodulation.

Methods

We analyzed the relevant experimental literature using modified Rayleigh–Plesset intramembrane cavitation BLS biomechanics and acoustic radiation pressure gradients (RPG) - induced membrane dynamics. By coupling these biomechanical models to biophysical membrane models we predict dynamical biophysical responses of artificial bilayer membranes, and of three common neocortical single cell Hodgkin-Huxley type models: i) Regular Spiking (RS) cortical pyramidal neuron, ii) Fast Spiking (FS) cortical inhibitory neuron and iii) Low Threshold Spiking (LTS) cortical inhibitory neuron, RS-FS-LTS Hodgkin-Huxley based network model and CNS axon model. In addition, live brain tissue RPG subjected areal strains were evaluated in a viscoelastic brain model.

Results

Only the Neuronal Intramembrane Cavitation Excitation (NICE) models were able to explain US-induced action potential generation through BLS-type pulsating nano-bubbles inside the bilayer plasma membrane: the leaflets' periodic vibrations induce US-frequency membrane capacitance and potential oscillations, leading to slow charge accumulation across the membrane (on a time scale of tens of milliseconds), until action potentials are generated. In contrast, the analysis of RPG-induced membrane capacitance variations associated with membrane area changes explain artificial membrane results, but were found to be highly unlikely sources for neural excitation, when considering the areal strains expected to form in brain tissue during normal sonication. Further, the NICE-LTS inhibitory neurons show a much higher relative sensitivity to sparse ultrasonic stimulation compared to the other neurons, resulting from their T-type voltage gated calcium channels. This model-based prediction was found to explain the results of a significant body of suppression and excitation experimental studies, including in humans.

Conclusions

These results provide a unified theoretical framework for a large body of experiments in multiple preparations across the field of US neuromodulation, lending further support to the hypothesis that intramembrane cavitation is responsible for ultrasonic neuromodulation. They could thus pave the way towards new CNS therapeutic protocols, using the only method that currently allows targeted non-invasive neuromodulation with millimetre spatial resolution essentially anywhere in the brain.

O12 Ultrasonic stimulation of mammalian retina in-vitro

Omer Naor^1,2, Nairouz Farah³, Shy Shoham²

¹ELSC Center for Brain Sciences, Hebrew University, Jerusalem, Israel; ²Biomedical Engineering, Technion- Israel Institute of Technology, Haifa, Israel; ³Faculty of Life Sciences, Bar Ilan University, Ramat Gan, Israel

Objectives

Following previous in vivo stimulation of the retina, we aimed to achieve a first direct measurement of the response of mammalian retinal neurons to ultrasonic (US) stimuli, and to study and characterize this response.

Methods

We coupled a high-density phased array (986 elements on a 25x35 mm² area) to a system for multi-electrode-array (MEA) recording with 256 contacts. Mouse retinas were dissected and placed on the MEA, and sonicated at 2.3 MHz, applying varying durations and intensities, as well as stimulated by light. The acquired data were processed to detect action potentials (spikes) elicited by retinal ganglion cells, and analysed to reveal the relations between the stimuli and the responses.

Results

We found prominent spike responses for stimuli in the range of 4.3-7.3 W/cm² and 0.5-1 s, which disappeared when the focus was steered 1.5 mm away. Furthermore, we found that the relation between the response strength and the stimulation intensity, or duration, followed a logistic sigmoid curve, while the response latency was described by a decreasing exponent. Lastly, we found indications that the observed responses to US stimuli are related to the "2^nd OFF" component in the responses to light stimuli.

Conclusions

These findings are the first direct demonstration of the response of the mammalian retina to US stimulation. The properties of the US transducer and the stimulation frequency indicate that non-invasive US stimulation of human retina is feasible, and may potentially evolve as an important tool for diagnosis and treatment of retinal diseases.

O13 Motor response elicitation and pupil dilation using megahertz-range focused ultrasound neuromodulation

Christian Aurup¹, Hermes Kamimura^2,1, Shutao Wang¹, Hong Chen¹, Camilo Acosta¹, Antonio A. Carneiro², Elisa E. Konofagou¹

¹Columbia University, New York, New York, USA; ²Universidade de São Paulo, São Paulo, Brasil

Objectives

Using transcranial focused ultrasound for the modulation of brain activity has been identified as a possible non-invasive means of treating neurological disorders. Most studies involving sedate rodents use frequencies in the kilohertz range, which allow for optimal transmission of acoustic power through the skull. The trade-off of using lower frequencies involves a lack of target specificity. Higher frequencies must be used in order to modulate activity in a more highly-specified manner. This study demonstrates that focused ultrasound in the megahertz range can be used to evoke motor- and cognitive-related responses in mice under deep anaesthesia by targeting specific brain structures. Contralateral-paired hind limb movements were observed when stimulating cortical regions, demonstrating the ability of MHz-range FUS to stimulate activity in highly-localized brain regions. Additionally, pupil dilation was observed when deep-seated anxiety-related structures were targeted, demonstrating the ability of FUS to modulate cognitive activity in a highly-specified manner.

Methods

For this study, wild-type adult male mice were anesthetized with intraperitoneal injections of sodium pentobarbital (65 mg/kg) and fixed in a stereotaxic frame. A single-element FUS transducer with fundamental frequency of 1.94 MHz was fixed to a 3D positioning system for accurate navigation through the brain. A 6x6 mm grid centred +2 mm rostral of the lambda skull suture was sonicated in a random order using a centre frequency of 1.9 MHz, pulse repetition frequency of 1 kHz, 50% duty cycle, 1 second pulse duration, 1 second inter-pulse interval for a total of 10 pulse repetitions. The acoustic pressure applied was varied in order to evaluate thresholds for eliciting physiological responses like motor movement, eye movement, or pupil dilation. Motor movements were validated using video recordings and electromyography via needle electrodes implanted into the biceps femoris of both hind limbs. Videos were recorded using a high-resolution camera focused at the right eye and processed to measure eye movements or changes in pupil size.

Results

The minimum acoustic pressure required to elicit motor movements was 1.45 MPa when targeting the somatosensory cortex, calibrated using an excised mouse skull. Higher pressures increased the success rate from 20% (at the 1.45 MPa threshold) to 70% (1.79 MPa). Targeting eye-motor and anxiety related regions of the brain elicited eye movements and pupil dilations up to 20%. Sonicating the superior colliculus resulted in both eye movement and pupil dilation at a lower threshold pressure (1.20 MPa) than the hippocampus and locus coeruleus which required pressures greater than 1.80 MPa.

Conclusions

This study successfully demonstrated that MHz-range transcranial focused ultrasound can be used to elicit motor- and cognitive-related physiological responses with high specificity in mice in vivo. It was also shown that the success rate of stimulation increased with acoustic pressure for motor movements associated with cortical activity modulation but highly depends on the region of the brain targeted. These findings emphasize the complex and yet to be determined mechanism of action involved in ultrasonic neuromodulation.

Fig. 4 (abstract O13).

Evaluation of the pressure threshold and success rate associated with applying FUS to location within the somatosensory cortex. This location resulted in contralateral hind-limb movement relative to the sonication site. Moving the transducer symmetrically about the midline resulted again in contralateral movement relative to the new sonication site

Fig. 5 (abstract O13).

Superior colliculus (top) threshold determined to be approximately 1.2 MPa while the locus coeruleus (bottom) was evaluated to be greater than 1.8 MPa

O14 Thermal dose effects by MR-guided focused ultrasound on the pig brain tissue - preliminary results

Dong-Guk Paeng^1,3, Zhiyuan Xu², John Snell¹, Anders H. Quigg¹, Matthew Eames¹, Changzhu Jin³, Ashli C. Everstine⁴, Jason P. Sheehan², Beatriz S. Lopes⁵, Neal Kassell¹

¹Focused Ultrasound Foundation, Charlottesville, Virginia, USA; ²Neurosurgery, University of Virginia, Charlottesville, Virginia, USA; ³Ocean System Engineering, Jeju National University, Jeju, Korea (the Republic of); ⁴Biology, University of Virginia, Charlottesville, Virginia, USA; ⁵Pathology, University of Virginia, Charlottesville, Virginia, USA

Objectives

The objective of this research is to investigate the effects of thermal dose (TD) delivered by magnetic resonance-guided focused ultrasound (MRgFUS) on in vivo pig brain tissue. In current clinical applications of transcranial MRgFUS systems, continuous acoustic wave emission is used to heat brain tissue to peak temperatures over 58°C. However, there are some situations where it has proven difficult to reach the desired peak temperature due to high absorption of acoustic energy by skull bone. There are reports that thermal effects on tissue are well correlated with thermal dose, which suggest that treatment delivery could be prescribed in terms of thermal dose rather than peak temperature or electric/acoustic power. It is also been demonstrated that the thermal dose threshold for permanent tissue damage is about 240 cumulative equivalent minutes (CEM) at 43°C for most of tissue. Currently available transcranial MRgFUS systems only allow the prescription of acoustic power and duration. In order to investigate the effects of thermal dose on in vivo brain tissue, we have developed a closed-loop control system to allow prescription thermal dose. This system monitors tissue heating via MR thermometry and provides pulse width modulation of output acoustic power in order to hold target tissue at a fixed temperature, and hence receives a nearly constant dose rate.

Methods

A FUS system (ExAblate 4000 Neuro 650 kHz system, InSightec) was used for sonication and an MRI system (Discovery MR75-3.0T, GE Medical systems) was used for thermometry and pre- and post-imaging. A closed-loop control system was implemented on a personal computer to control pulse width modulation of the FUS system acoustic power in order to maintain a specified temperature based on the MR thermometry. Accumulated thermal dose was calculated in real time and used to stop the sonication so that a prescribed thermal dose was delivered to the targeted tissue. Phantom studies were performed to test the control system to prepare for animal experiments. One acute and six chronic experiments (with three day survival) were conducted to observe the effects of TD on pig brain by behaviour observation and post MR imaging of the brain (1 hour and 70 hours post procedure). Craniotomy was performed to create an acoustic access window, and sonication was applied on 4 spots in the thalamus of each pig. Histology was also performed to compare it with MR imagery. Temperature in the pig brain tissue was estimated by rectal temperature for the MR thermometry baseline. TD was varied from 7 to 200 CEM while the target temperature was changed from 46 to 52 °C with appropriate acoustic power depending on target position and individual pig. This study was approved by the University of Virginia Institutional Animal Care and Use Committee.

Results

From the acute experiment, we could observe the lesions on MR images after 1 hour of sonication and histology subsequently confirmed the lesions. For the chronic experiments, no obvious problem was observed in the behavior of any of the six animals. Eighteen sonication spots in 5 pigs were analyzed through MR images. One pig experiment failed to control temperature due to introduction of air bubbles between the brain and scalp during surgery procedure, and 2 sonication spots were excluded due to technical problems. Large tissue changes were observed in MR images in all 6 spots over 100 CEM.

The diameter of those tissue changes in MR T2-weighted axial images were measured and averaged to 2.9 ± 0. 4 mm. There is inconsistency in generating lesions for TD below 100 CEM. No lesion was shown in some lower TD from 7 CEM and 61 CEM, while some smaller lesions (<2 mm in lesion diameter) were shown in TD from 18 CEM to 85 CEM except one large tissue change of 3.5 mm in diameter at 31 CEM. Some tissue changes were shown in both post MR images after 1 hour and 70 hours of sonication, while some were visible only at the 70 hour time point. Histology of 3 pig experiments is now available and the histology reports support the tissue changes and lesions in MR images. Lesion diameters in MR T2-weighted axial images versus TD in CEM are shown in Fig. 6 for all the results from the chronic pig study.

Conclusions

These preliminary results from pig brain tissue generally confirmed the previous results from rabbit brain tissue in generating tissue changes over a certain TD, even though there are some differences in the FUS systems and the experimental procedures and analysis. For lower thermal dose below 61 CEM, there is significant variability in generating of tissue changes, while large tissue changes whose average diameter is 2.9 mm were observed in MR T2-weighted axial images for higher TD over 100 CEM, which were reported with similar tendency but a little difference in TD from the rabbit brain study. These results may contribute to open the way to prescribe the thermal dose rather than peak temperature or acoustic power for brain treatments, and expand the treatment envelope beyond the current limitations in selecting targets and patients. This project is ongoing and will be further pursued with additional experiments for consolidation of the results and analysis.

Fig. 6 (abstract O14).

Pig chronic study results showing the relations of lesion diameter (mm) based on enhancing region appearing in T2-weighted axial MR images with applied thermal dose in CEM.

O15 In vivo feasibility study of boiling histotripsy with clinical Sonalleve system in a neurological porcine model

Thomas Looi^1,4, Vera Khokhlova², Charles Mougenot⁵, Kullervo Hynynen³, James Drake¹

¹Hospital for Sick Children, Toronto, Ontario, Canada; ²University of Washington, Seattle, Washington, USA; ³Sunnybrook Research Institute, Toronto, Ontario, Canada; ⁴University of Toronto, Toronto, Ontario, Canada; ⁵Philips Healthcare, Markham, Ontario, Canada

Objectives

To determine if a clinical focused ultrasound system (Philips Sonalleve) can be used to perform mechanical liquefaction of brain tissue for neurological lesioning through a simulated fontanelle in a porcine model (simulating neo-natal patients). This work will determine the power of the system required to induce lesions using a boiling histotripsy (BH) pulsing protocol. Post-treatment, the lesion volume and border will be measured with MRI imaging and histological examination.

Methods

A porcine model was used as the in vivo model with a maximum weight < 6.8 kg (4.9 - 6.8kg). A horse-shoe incision and blunt dissection was used to expose the skull. A craniotomy was performed to create a 4–5 cm² opening in the skull simulating the fontanelle in a neonatal patient. A degassed mixture of ultrasound gel and water (ratio 10:1) was poured on top of the dura to ensure good acoustic coupling. The scalp was sutured closed with 2–0 Vicryl cutting needle. The animal was placed supine feet first with the craniotomy centred about the Sonalleve V2 system with Flex-M surface coils. Pre-treatment T1-weighted (T1-w), T2-weighted (T2-w) and T2*-weighted (T2*-w) MRI imaging was conducted as a baseline. Each animal was treated at four cluster locations where each cluster consisted of seven sonication points; one point in the centre and six points uniformly distributed over a 4-mm diameter circle. The clusters were located approximately 15 mm deep in the brain, 7 mm off the midline, and separated by 14 mm in a rectangular geometry. In initial treatment on the first animal, the power was increased from 100 to 500 W for each cluster. After initial analysis, the treatment was repeated on second animal with refined power levels of 325, 350, 375, and 425 W. The treatment sequence consisted of 12000 pulses of 1.2 MHz frequency, 1 and 10 ms pulse duration, and 1% duty cycle for both 1 and 10 ms pulse duration. These protocols have been shown to generate BH lesions in ex vivo bovine liver in another Sonalleve system. During treatment, MR thermometry was used to monitor for surface, focal, and far field heating. A dedicated MATLAB-based interface was connected to the Sonalleve cavitation sensor to detect the signal generated during treatment points. After treatment, post T1-w, T2-w and T2*-w MRI scans were completed for comparison. The animals were euthanized, perfusion fixated and their brains were removed for histology. The brain specimens were cut at the centre for the treatment clusters to get a cross-sectional coronal view where each slice was 5 microns. The slides were stained with haematoxylin and eosin (H&E) and examined for lesion presence, blood and border definition.

Results

A total of 4 piglets were sonicated with the following configuration: 1 piglet with 5 clusters (100 – 500 W), 2 piglets with 4 clusters (325-425 W) using 10-ms long pulses and 1 piglet with 4 clusters and 1-ms long pulses. For all power levels, the MR-measured temperature in the near or far field of the treatment was below the noise level. For 10-ms long pulses, 100 and 200 W acoustic powers, no noticeable imaging change was observed during sonication and post-treatment MR evaluation. As power was increased from 300 to 400 W, a temperature increase of up to 5C was measured at the focus. With more discrete power levels, it appeared at 375W that the lesion was more contained whereas higher power levels created wider areas of tissue change. It was observed the timing of MR magnitude of the target cluster changed as the power level increased where the tissue change occurred at 15 s at 300 W, 12 s at 350 W, 10 s at 375 W and 5 s at 425 W. After sonication was completed, the detected temperature rise decreased immediately versus dissipating over time. This would indicate the detected thermometry was due to a phase change of the tissue rather than temperature increase. During treatment, a high amount of lower broad band emissions at frequency < 1.2MHz were detected by a cavitation sensor. Post-treatment MR imaging showed that at power levels between 300 and 400 W, there were areas of hypointensity indicating the lesion. H&E staining confirmed the presence of the mechanical lesion where various anatomical targets were fractionated. Power levels were sufficient to rupture vessels and cause a focused area of haemorrhage at the treatment cluster. It showed that BH treatment dissolved the anterior ventricle wall with presence of elements of blood. H&E staining also showed that maximum lesion diameter was approximate 7 mm in coronal plane therefore the treatment borders matched the treatment plan. For the piglet treated with 1-ms long pulses, the post-treatment imaging change was not noticeable. However, at 375 and 400 W power, H&E slides showed two areas where there was a perforation of the anterior ventricle wall with a lesion size of up to 2 mm. It appeared that the shorter pulse duration generated smaller but more focused lesions.

Conclusions

This pilot study shows that the clinical Sonalleve system is capable of generating mechanical ablation of a brain tissue in an in vivo porcine model using boiling histotripsy pulsing scheme. The power threshold to initiate lesions in brain using 10 ms pulses (375 W) was found to be similar the power levels used in BH studies in ex vivo bovine liver and porcine kidney tissue at a similar depth in tissue (250 – 300 W) in a Sonalleve V2 system at the University of Washington. The treatments can be accelerated by using higher power outputs and shorter pulses. H&E histological evaluation showed that BH treatment caused rupture of vessels focally while also creating wider well defined areas of mechanical ablation with no damage to surrounding tissue. Additional work is underway to characterize the pressure levels generated by the Sonalleve to correlate the power and pressure for treatment.

Fig. 7 (abstract O15).

Top Row: Left to Right: Treatment area, Pre T2-w (100–500 W), Post T2-w (100–500 W), Pre- T2-w (325–425) and Post T2-w (325–425); Bottom Row: Histology Left to Right: 10 ms treatment (375W), 10 ms treatment - zoomed (375W), 1ms treatment (375W and 425W), 1 ms treatment – zoomed.

O16 Musculoskeletal clinical applications of intense therapy ultrasound (ITU): part 1. Clinical study for plantar fasciitis

Michael Slayton¹, Richard C. Amodei¹, Keegan Compton¹, Ashley McNelly², Daniel Latt²

¹Guided Therapy Systems, LLC, Mesa, Arizona, USA; ²School of Medicine, University of Arizona, Tucson, Arizona, USA

Objectives

Chronic Plantar fasciitis (CPF) is a common cause of plantar heel pain that is a result of a degenerative process of the plantar fascia and its surrounding perifascial structures [1]. It is the most common cause of heel pain, affecting 10% of the U.S. population, and one of the most common foot and ankle problems.

More than twenty different treatments have been used for plantar fasciitis. Conservative treatments (rest, ice, stretching and NSAIDs) have been shown to effectively treat symptoms but 10% of patients fail conservative management and continue to have symptoms within 12 months and beyond. Surgery consisting of partial PF release is often considered with 50% of patients having residual symptoms, in addition to surgical risk exposure.

High frequency ITU, a novel potential approach to treating Plantar Fasciosis, was studied for the creation of small thermal injuries noninvasively inside symptomatic Plantar Fascia (PF). It has been shown to initiate a tissue repair cascade and promote collagen generation in musculoskeletal tissue [2, 3]. A double blinded, randomized, sham controlled clinical study for ITU treatment of chronic Plantar Fasciitis has been conducted by IRB approved clinical protocol to access clinical efficacy of the procedure.

Methods

Custom 3.2 MHz high intensity (10 kW/cm²) ultrasound therapy system was designed and fabricated (GTS, Mesa, AZ, USA). Field simulations, testing and Schlieren images verified intensity, high focal pressure (17.3 MPa) and focal distance of 13–15 mm.

Each treatment consisted of 250–320 100 ms pulses creating matrices of small ablative thermal lesions of 4–5 joules at pre-programmed pitch of 1.6 mm. Each patient underwent two treatment sessions in 2 weeks, each treatment time did not exceed 12 minutes. ITU placebo group consisted of the same treatment with energy set to 0.

Treatment effects were assessed with diagnostic imaging ultrasound at 12 MHz (Spark, Ardent Sound, Mesa, AZ, USA) by a certified sonographer. Ultrasound images were analysed to determine symptomatic hypoechoic lesion size with PF.

Patient reported outcomes consisted of PROMIS physical function computer adaptive test (PF-CAT), PROMIS global health, Foot Function Index pain subscale (FFIPS) [4, 5] and a non-validated heel pain specific questionnaire.

Clinical protocol included (35) patients diagnosed with chronic heel pain due to Plantar Fasciitis (more than 3 months) and failed conventional therapy treatments.

Patients were randomized to standard therapy (anti-inflammatory pills, stretching and gel heel cups) plus ITU (“Treatment” group, n=26) of standard therapy plus sham ITU (“Control” group, n=9).

Primary investigator, sonographer and study coordinator administering the study were blinded to group assignments. P-values were calculated via 2-tailed paired T-tests for both treatment and control groups.

Results

Patient-Reported Outcome Measures: Compared to the baseline assessment of Pain, the Treatment Group showed significantly improved pain scores compared to the Control (sham treatment) Group in follow-up visits including 12 weeks after the initial treatment.

Foot Function Index Pain Score: Compared to the baseline assessment, the treatment group pain scores also showed significant improvement compared to the sham group.

Diagnostic Ultrasound Imaging: During the 12 week follow-up period changes to the overall thickness of the PF were not statistically significant, while calculated volume size of hypoechoic lesions within the PF, just distal to the Calcaneus, showed significant change. For the experimental group (n=28) the average hypoechoic lesion volume reduction was followed and compared to the baseline measurements just before the first Treatment; 2 week follow-up and 2nd treatment date (−28%), 4 weeks (−50%), 6 weeks (−66%) and 12 weeks (−80%).

For the control group (n=10), the average hypoechoic lesion volume was followed and compared to the baseline just before the first Treatment; 2 weeks and 2nd sham treatment (+9%), at 4 weeks (+16%), 6 week (+29%) and 12 weeks (+31%). Unlike the experimental group, these lesions grew in size during the follow-up period.

P-values calculated for all outcome results discussed above for both treatment and control groups were below .01, showing the statistical significance of the results.

Conclusions

• (1) Results of the double blinded randomized, sham controlled study for the treatment of Plantar Fasciitis with ITU appeared to have statistically significant positive results within 12 weeks post-treatment in 80% of treated subjects.

• (2) Both quantitative measurements from diagnostic ultrasound imaging and applied standardized assessment protocols consisting of PROMIS PF-CAT, FFIPS along with Patient Reported Outcome Measures showed statistically significant coincidental improvements in treated subjects vs. control group.

• (3) Intense Therapeutic Ultrasound has shown potential for effective treatment of Chronic Plantar Fasciitis. Better designed studies with increased # of subjects will be considered to support ITU as an effective tool for the proposed clinical treatment.

References

[1] Neufeld S.K. and Cerrato R.; Plantar fasciitis: evaluation and treatment. J Am Acad Orthop Surg, June 2008, 16(6): p. 338–46.

[2] Slayton M. and Barton J.; Healing tissue response with ITU (Intense Therapy Ultrasound) in musculoskeletal tissue, feasibility study. Ultrasonics Symposium (IUS), 2014 IEEE International, Chicago, USA, p.1654–1657.

[3] Slayton M. H., Amodei R. C., McNelly A. and Latt D. L.; Intense Therapy Ultrasound (ITU) for the treatment of Chronic Plantar Fasciitis: Preliminary Results of Clinical Study. 37th International Conference of the IEEE Engineering in Medicine and Biology; Milan, Italy, August 2015.

[4] Rose, M., Bjorner, J.B., Becker J., Fries J.F. and Ware J.E.; Evaluation of a preliminary physical function item bank supported the expected advantages of the Patient-Reported Outcomes Measurement Information System (PROMIS). J. Clin. Epidemiology, January2008, 61(1): p. 17–33.

[5] Budiman-Mak E., Conrad K. J. and Roach K. E.; The Foot Function Index; a measure of foot pain and disability. J. Clin. Epidemiology, 1991, 44(6): p. 561–579.

Fig. 8 (abstract O16).

Treatment Group SROM 1

Fig. 9 (abstract O16).

Control Group SROM 1

Fig. 10 (abstract O16).

FFIP Score by Visit

Fig. 11 (abstract O16).

Average Lesion Volume change by Visit. SE applied.

O17 Musculoskeletal clinical applications of intense therapy ultrasound (ITU): part 2. Initial results of clinical study for lateral epicondylitis

Michael Slayton¹, Richard C. Amodei¹, Keegan Compton¹, John Kearney²

¹Guided Therapy Systems, Mesa, Arizona, USA; ²The CORE Institute, Phoenix, Arizona, USA

Objectives

Acute and Chronic pain of the Common Extensor Tendon (CET) region, lateral epicondylitis or tennis elbow is a common pathology of both athletes and non-athletes affecting up to 3% of the population at large [1], while the prevalence of chronic problems caused by overuse in tennis players can be as high as 40%. Elbow tendinopathy represents an important set of pathologies that account for lost recreation time, decreased quality of life, and work-related disability claims. Conservative treatment of the Epicondylitis or -osis is recommended as the initial strategy by most authors. This strategy includes identification and correction of possible etiological factors, and a symptom related approach. Generally, the initial treatment consists of a multifactorial approach that may include a combination of rest (complete or modified activity), medication (NSAIDs for Epicondylitis), stretching and strength training. More aggressive treatments for CET include: Cortisone injection, Plasma Rich Platelets (PRP), Tenotomy, ESWT.

High frequency ITU, a novel potential treatment for CET, was studied for the creation of small thermal injuries noninvasively inside symptomatic Common Extensor Tendon (CET). It has been shown to initiate a tissue repair cascade and promote collagen generation in musculoskeletal tissue [2–5]. A blinded, randomized, clinical study for ITU treatment of chronic Lateral Epicondylitis has been conducted by IRB approved clinical protocol to assess clinical efficacy of the procedure.

Methods

Custom 4.5 MHz high intensity (47.9 KW/cm²) ultrasound therapy system was designed and fabricated (GTS, Mesa, AZ, USA). Field simulations, testing and Schlieren images verified intensity, high focal pressure (37.9 MPa) and focal distance of 6 mm.

Each treatment consisted of 80 14 ms pulses creating matrices of small ablative thermal lesions of 1 joule at manually targeted area set by diagnostic ultrasound imaging. Each subject underwent two treatment sessions 4 weeks apart. Each treatment time did not exceed 10 minutes. Treatment effects were assessed with diagnostic imaging ultrasound at 20 MHz (Spark, Ardent Sound, Mesa, AZ, USA) by a certified sonographer. Ultrasound images were analysed to determine changes in the peri-tendon region, including hypoechoic areas, calcifications and dependent free fluid.

Subject reported outcomes consisted of PRTEE survey [6], physical examination, Universal Analog Visual Pain Score17 and a Patient Reported Satisfaction Survey [7].

Clinical protocol includes 25 subjects diagnosed with chronic Tennis Elbow, or Lateral Epicondylosis (more than 3 months) and failed conventional therapy treatments.

Subjects were subjected to standard therapy (stretching and strength exercises, hot and cold compresses and compression support) plus ITU.

Primary investigator, sonographer and study coordinator administering the study were blinded to group assignments. P-values were calculated via 2-tailed paired T-tests at each visit of the clinical study.

Results

The results presented below are initial findings for the first 12 subjects currently being followed through the study.

PRTEE: Patient Reported Tennis Elbow Evaluation Final Score is a weighted Pain Score based on 15 questions grouped into 3 categories: Overall Pain, Functional Disability and Usual Activities. Subjects respond to each question with a Pain Score of 0–10. Each category is then summed and weighted with a maximum score of 100 (Overall Pain 100%, Functional Disability 50% and Usual Activities (50%), n=12, Fig. 12.

Self-Reported Outcome Measures Surveys show a significant improvement and treatment satisfaction with Subjects reporting improvements in elbow pain 100%, improvement in Daily function 83% (vs. 17% no improvement) and treatment satisfaction 83% (vs. 17% not satisfied), n=6, Fig. 13.

Universal Analog Pain Scores also show progressive reduction (−3 on a 10 point scale) throughout the same period, n=12, Fig. 14.

Diagnostic Ultrasound Images: Diagnostic Ultrasound Images show a consistent increase of free fluid 2 weeks after the first treatment, with a progressive reduction in free fluid at 8 weeks in subjects with no to mild peri-tendon calcifications. These subjects correlated well with improving PRTEE survey scores. Subjects with little or no improvement in PRTEE scores consistently presented with moderate to severe peri-tendon calcifications.

P-values calculated for the above reported outcomes were not statistically significant for Visits 2 and 3 (P>0.05) while results for Visits 4 and 5 demonstrated P< 0.05.

Conclusions

• 1. Feasibility of Intense Therapeutic Ultrasound treatments of chronic pain in CET region has been established with the initial results (n=12) of the ongoing clinical study.

• 2. Significant reduction of pain scores per activities (PRTEE) and Self-Reported Outcome Measures (83% improvement) with average Universal Pain Scores reduction from 5.0 to 2.0 were statistically significant (p<0.05) at 8 and 12 weeks post treatment.

References

[1] Hong QN, Durand MJ, Loisel P. Treatment of lateral epicondylitis: where is the evidence? Joint Bone Spine 2004; 71(5):369–373.

[2] White, W. M., I. R. Makin, et al. (2007). Selective creation of thermal injury zones in the superficial musculoaponeurotic system using intense ultrasound therapy: a new target for noninvasive facial rejuvenation." Arch Facial Plast Surg 9(1): 22–29.

[3] Gliklich R, White WM, Barthe PG, Slayton MH, Makin IRS. Clinical pilot study of intense ultrasound (IUS) therapy to deep dermal facial skin and subcutaneous tissues. Arch Facial Plast Surg 2007; 9:88–95.

[4] Slayton M., Barton J, Feasibility of Modulating Healing Tissue Response by ITU (Intense Therapy Ultrasound) in Musculoskeletal Tissue ASLMS 2014 Annual Conference.

[5] Slayton M., Barton J., Healing tissue response with ITU (Intense Therapy Ultrasound) in musculoskeletal tissue, feasibility study, Ultrasonics Symposium (IUS), 2014 IEEE International, pp. 1654–1657. DOI 10.1109/ULTSYM.2014.010

[6] Rompe JD1, Overend TJ, MacDermid JC. Validation of the Patient-rated Tennis Elbow Evaluation Questionnaire.J Hand Ther. 2007 Jan-Mar; 2007 (1):3–10: quiz 11.

[7] Rose, M., et al., Evaluation of a preliminary physical function item bank supported the expected advantages of the Patient-Reported Outcomes Measurement Information System (PROMIS). J Clin Epidemiol, 2008. 61(1): p. 17–33.

Fig. 12 (abstract O17).

See text for description

Fig. 13 (abstract O17).

See text for description

Fig. 14 (abstract O17).

See text for description

O18 Clinical experience of intra-operative high intensity focused ultrasound in patients with colorectal liver metastases. Results of a phase ii study.

David Melodelima^1,2, Aurelien Dupre², Yao Chen², David Perol², Jeremy Vincenot¹, Jean-Yves Chapelon¹, Michel Rivoire^2,1

¹LabTAU - U1032, INSERM, Lyon, France; ²Centre Leon Berard, Lyon, France

Objectives

Managing colorectal liver metastases (CLM) is a major clinical challenge, and surgery remains the only potentially curative treatment. Five-year survival rates of up to 51% have recently been reported. However, only 10–20% of patients are eligible for surgery, which is often precluded by the number, size and/or location of metastases, or because the necessary resection will leave insufficient volume of functional liver. Radiofrequency ablation (RFA) is the main technology that has been used in association with surgery as a tool to expand the number of patients who may be candidates for liver directed therapy. However, RFA has several limitations. There is a risk of inadequate treatment due to the heat sink effect of blood flow, RFA does not allow reliable real-time monitoring, and it require intra-parenchymal introduction of a probe. Moreover, only small hepatic volumes can be targeted. These limitations could explain the high rates of local recurrence seen after RFA.

High intensity focused ultrasound (HIFU) has been proven effective in a wide range of clinical applications, especially prostate cancer. The ablation achieved by conventional HIFU is small and ellipsoidal. The dimensions vary according to transducer characteristics but are typically 1–3 mm (transverse) and 8–15 mm (along beam axis). In clinical practice, hundreds of superimposed ablations are required; and the procedure may take up to two hours. Even so, HIFU has several potential advantages in the treatment of liver tumours: there is no need to puncture the parenchyma, the extent of the thermal lesions achieved is not reduced by hepatic perfusion, and it is possible to monitor the effects of therapy in real time. However, extra-corporeal treatment of the liver is difficult because presence of the ribcage may stop propagation of ultrasound waves and respiratory motion may cause targeting problems. HIFU treatment of CLM needs to be improved, and reducing the duration of surgical intervention by increasing the size of ablated fields is particularly important. A HIFU device enabling destruction of larger liver volumes has been developed based on toroidal transducers.

Preliminary in vitro and preclinical work demonstrated the potential, feasibility and safety of such HIFU ablations. During laparotomy in a porcine model we demonstrated that this HIFU device achieves reproducible ablations with an average volume of 7 cm3 (with 20 mm diameter and 25 mm long axis) in 40 seconds. Such preclinical work has to be translated into clinical practice through controlled trials, and the aim of this study was to assess the feasibility and safety of HIFU ablation in patients undergoing hepatectomy for CLM, as well to collect efficacy and accuracy data. This study is registered with Clinical-Trials.gov (NCT01489787).

Methods

This study was a prospective, single-centre phase I/II study evaluating the feasibility, safety and accuracy of HIFU during surgery in patients with CLM. The protocol was reviewed and validated by a national ethics committee (CPP Sud-Est IV) according to French and European directives. Since this study was the first use in man of intra-operative hepatic HIFU, ablations were made only in areas of liver scheduled for resection. This allowed real-time evaluation of HIFU ablation while protecting participating patients from any adverse effects related to this new technique. The transducer has a toroidal shape 70 mm in diameter and is divided into 32 ultrasound emitters of 0.13 cm² operating at 3 MHz. The radius of curvature is 70 mm. A 7.5 MHz ultrasound imaging probe was placed in the centre of the device and was used to guide the treatment. The imaging plane was aligned with the HIFU focal zone.

Six patients were included in the Phase I. Two single thermal ablations were created in each patient. Thirteen patients were included in Phase IIa and two HIFU ablations were to be placed precisely in a target previously identified in ultrasound images (step 1) and then at distance (step 2) from a target. Five patients were included in Phase IIb until now. HIFU ablations were created to ablate metastases (20 mm maximal diameter) with safety margins in all directions. The exposure time varied from 40 s to 370 s according to the diameter of the metastases to be treated.

Results

In agreement with preclinical studies, the demarcation between ablated and non-ablated tissue was clearly apparent in ultrasound images and histology. The dimensions measured on ultrasound imaging were correlated (r=0.88, p<0.0001) with dimensions measured during histological analysis. All HIFU ablations were obtained in 40 seconds. The average dimensions obtained from each HIFU ablation were a diameter of 21.0 ± 3.9 mm and a depth of 27.5 ± 6.0 mm. The phase IIa study showed both that the area of ablation could be precisely targeted on a previously implanted metallic mark and that ablations could be created deliberately to avoid such a mark. Ablations were achieved with a precision of 1–2 mm. In Phase IIb, one metastasis of 10 mm in diameter was ablated in 40 seconds with safety margins. Using electronic focusing metastases of 2 cm in diameter were ablated with safety margins (>3 mm in all directions) in 370 seconds. The dimensions of HIFU ablations created in 370 s were a diameter of 48 mm and a long axis of 51 mm.

Conclusions

This new HIFU device safely achieved large volume liver ablations in 40 s, with a precision of one to two millimetres under real-time monitoring. HIFU ablations of small metastases (<20 mm) and peri-lesional healthy liver were successfully created with planned safety margins of at least 3 mm in all directions.

O19 Chemotherapy in oral cancer

Wei Guo, Guoxin Ren, Guofeng Shen

Department of Oral Maxillofacial and Head Neck Oncology, Shanghai 9th People's Hospital, Shanghai, China

Objectives

To evaluate the efficacy and the main side-effects during the clinical trial of this new ultrasound hyperthermia system combined with chemotherapy in oral cancer, meanwhile, to observe the preliminary clinical response of this combined therapeutic modality.

Methods

Thirty four cases of oral squamous cell carcinoma entered this clinical trial, from which 23 had advanced oral carcinoma and were treated with new ultrasound hyperthermia system combined plus docetaxel–cisplatin–fluorouracil regimen (test group). Eleven patients only received chemotherapy with docetaxel–cisplatin–fluorouracil regimen (control group). The thermo-index were detected during the course of hyperthermia, the chief-complain of the patients were also recorded. The systemic physiological, biochemical and immunological index were tested before and after the treatment respectively. The therapeutic response was estimated 1 month after 2 cycles of the treatment.

Results

Twenty three cases of oral squamous cell carcinoma enrolled the clinical trial of local ultrasound hyperthermia combined with chemotherapy. 230 times of ultrasound hyperthermia in total were performed. The ultrasound hyperthermia system operated smoothly, no malfunction was found. The main thermo-index were: the maximum heating temperature was108 F, the average heating temperature was 106 F, the minimum heating temperature was 104 F, the fraction of heating time more than 108 F was 0.46, the average treatment time was 37.74±8.88min. PR+CR was 71% (test group). The main local side-effects were low-grade pain (6/23). The incidence of adverse effects was similar between both study groups, no bone marrow suppression (over III).

Conclusions

The system combined with docetaxel–cisplatin–fluorouracil regimen is effective and safe in the treatment of advanced oral cancer. The main side-effects of local ultrasound hyperthermia combined with chemotherapy are low-grade pain or tolerable pain. There is no immune function and obtains satisfying short-term response.

O20 Non-thermal, non-cavitational, 20kHz Ultrasound applicators in wound healing

Michael Neidrauer¹, Leonid Zubkov¹, Michael S. Weingarten², David J. Margolis³, Peter A. Lewin¹

¹The School of Biomedical Engineering, Science and Health Systems, Drexel University, Philadelphia, Pennsylvania, USA; ²Department of Surgery, College of Medicine, Drexel University, Philadelphia, Pennsylvania, USA; ³Biostatistics and Epidemiology, University of Pennsylvania, Philadelphia, Pennsylvania, USA

Objectives

This talk examines the challenges associated with the design of clinically viable ultrasound applicators operating at the relatively low frequency (20 kHz) and intensity (<100 mW/cm², spatial peak, temporal peak) levels, and tailored to treatment of chronic wounds, such as venous or diabetic ulcers. These challenges were associated with the architecture and weight, and principle and efficiency of operation, including electrical power consumption. The ultimate goal of this work was to test the efficacy of the applicators in human subjects.

Methods

A fully wearable Band-Aid™-like, dial-in delivery, battery-operated ultrasound applicator was designed. The applicator included light weight (<25g) piezoelectric flexural transducer and was powered by 10-12V fully rechargeable lithium-ion batteries (total weight <200g); it was able to operate for up to 4 hours between re-charging. To emphasize the uniqueness of the design, it might be useful to note that typically, the thickness of the capacitive piezoelectric element is inversely proportional to the frequency, therefore a 20 kHz element would need to be 10 cm thick. Such element would be bulky and require hundreds of volts (demanding a large power amplifier), in excitation signal thus eliminating any chance of being a portable design. To overcome this, a mechanical displacement amplifier, which translates 2 MHz ultrasound waves into 20 kHz output at the desired pressure amplitude (55 kPa; i.e. 100 mW/cm²) with only 12 volts excitation was chosen as a preferred solution.

The applicators were extensively tested to ensure that the ultrasound field energy was below the level needed to generate inertial cavitation and any temperature elevation that would exceed 1°C. Also, the uniformity of the acoustic field distribution was verified. The pilot study included 32 individuals between ages of 18 and 80 having venous (n=23) or diabetic (n= 9) wounds that remained open for a minimum of 8 weeks. In compliance with the IRB study protocol the subjects were randomly assigned to either treatment or control group, with an equal chance of being assigned to receive active ultrasound treatment or sham (current standard care). Treatment sessions lasted 15 minutes and were administered once a week for a period of 12 treatments, or until the wound’s closure. Clinical efficacy was evaluated by measuring the reduction in wound area over time. For both etiologies, i.e. both venous and diabetic wounds the rate of closure was statistically faster (p<.05) in the treated group compared to the control group.

Results

The study findings show that the ultrasound treated venous ulcer group had statistically improved (p<0.04) rate of wound size change (reduction of 14.3%/week) compared to the rate of wound size change for the control group (increase of 3.6%/week on average). Diabetic wound closure was achieved typically after 4 sessions for treated wounds, as opposed to 7 sessions for the control group. Time to heal was also statistically faster (p< .05) for treated wounds (~5 weeks) when compared to non-treated wounds (~12 weeks).

Conclusions

Overall, the results from this study support the notion that low frequency ultrasound treatment can successfully improve healing outcomes in chronic wounds with different morphology and etiology. The evaluated device used safe levels (<100mW/cm² ISPTP) of ultrasound energy and featured unique portability, which opens possibility for personalized home treatment of chronic wounds in the future.

Acknowledgements

NIH Grant R01 EB9670, NSF 1064802, Wallace H. Coulter Foundation.

O21 Tests of thermal ablation with a 230 kHz transcranial MRI-guided focused ultrasound system in a large animal model

Nathan McDannold¹, Jonathan Sutton¹, Natalia Vykhodtseva¹, Margaret Livingstone²

¹Radiology, Brigham and Women's Hospital, Boston, Massachusetts, USA; ²Neurobiology, Harvard Medical School, Boston, Massachusetts, USA

Objectives

Thermal ablation with transcranial MRI-guided focused ultrasound (TcMRgFUS) is being tested clinically for as an alternative to surgery for functional neurosurgery and brain tumour resection. The current TcMRgFUS system, which operates at 650–670 kHz, is limited by skull heating to a small central region in the brain. Use of a lower acoustic frequency will reduce skull heating, but at the same time the focal heating will decrease and the risks of uncontrolled cavitation (the formation of microbubbles) increase. The purpose of this study was to evaluate the feasibility of thermal ablation in nonhuman primates using a system that operates at a lower acoustic frequency and to determine whether it can increase the “treatment envelope” for TcMRgFUS.

Methods

The experiments were approved by our institutional animal committee. Thermal ablation with the 230 kHz ExAblate Neuro system (InSighTec) was tested over five sessions in three rhesus macaques. In each session a target in the thalamus was sonicated transcranially at 40–50 s at acoustic power levels ranging from 90–560 W. The TcMRgFUS system software modulated the acoustic power in real time with a closed-loop controller that maintained a low-level of acoustic emissions, which are correlated with cavitation activity. MR temperature imaging (MRTI) was acquired at 3T (LX, GE) in a single plane using a 14 cm surface coil (TR/TE: 29/13 ms; flip angle: 30°). Measurements of the peak temperature rise at the focus and on the outer brain surface were compared for the different animals as a function of the applied acoustic energy. For the brain surface we measured the average temperature at the hottest 5% of the voxels and over a two voxel wide strip and we also normalized the measurement by the outer skull area. Temperature measurements were used to calculate the accumulated thermal dose, which was then compared to post-sonication T2-weighted, T2*-weighted, and contrast-enhanced T1-weighted MRI. The focal and skull-induced heating on the brain surface were compared to an earlier study performed in macaques with a 650 kHz version of this system.

Results

Focal heating sufficient to create an MRI-evident thermal lesion was achieved in 4/6 targets; the peak thermal dose exceeded 240 CEM43°C at these targets (Fig. 15). Heating at the focus was slightly higher than that measured on the brain surface. The focal heating increased linearly as a function of the applied energy at a rate of 3.2 ± 0.4°C per kJ (R2: 0.81). The surface area of the outer skull ranged from 47–55 cm2. For the hottest 5% of the voxels on the brain surface included in the MRTI imaging plane, the temperature rise increased linearly as a function of temperature at a rate of 126.6 ± 7.3°C per kJ/cm². For a two voxel wide strip over the entire brain surface, this increase was 62.7 ± 7.5°C per kJ/cm². The extent of MRI-evident changes (apparent oedema in T2-weighted MRI, BBB disruption post-contrast, no petechiae in T2*-weighted MRI) were consistent with 240 CEM43°C contours. One lesion imaged one week after FUS increased in size.

Conclusions

Analyses of the MRTI and post-sonication MRI suggest that the lesions were consistent with thermal mechanisms. The temperature rise increased linearly with the applied energy, and no evidence of cavitation-related petechiae were evident after sonication. The MRI-evident lesions were consistent with isodose contours drawn at 240 CEM43°C, a conservative threshold often used to guide thermal ablation. However, since it is known that thermal damage can take several hours to manifest in MRI and the lesion we imaged at one week increased in size, it is likely that the size of the lesion was underestimated by this dose value.

Similar tests in macaques with a version of this system operating at 670 kHz (Hynynen et al., Eur J Radiol 2006; 59: 149–56) measured skull-induced heating of 130°C per kJ/cm² of outer skull surface, more than twice of that measured here (63°C per kJ/cm²). While no or minimal focal heating was observed at 670 kHz, with this 230 kHz system we were able to reach ablation-level thermal dose values. Thus, these preliminary results thus suggest that this low frequency system can expand the area of the brain that can be targeted for thermal ablation without overheating the skull. The closed-loop feedback system successfully maintained a low level of acoustic emission (and presumably microbubble activity) and immediately stopped the sonication when excessive levels were detected. However, additional work is needed to understand whether low-level cavitation activity played a role in the focal heating, to characterize the lesions in histology, and to examine whether safe cavitation levels can be maintained in tumours where the cavitation threshold may vary.

Fig. 15 (abstract O21).

MRTI (top) and post-FUS imaging (bottom) obtained in two sessions in Monkey 1. Thermal dose contours at 30 (orange) and 240 (red) CEM43C were calculated from the MRTI. Immediately after each session, a small lesion was observed in contrast-enhanced T1-weighted MRI (CE-T1WI). The dimensions of this area were consistent with the 240 CEM43 contours. The lesion produced in session 1 was largely non-enhancing in in CE-T1WI at week 2. It was visible in T2-weighted imaging (T2WI) and increased in size

O22 Growth slowdown in a brain metastasis model by antibody delivery using focused ultrasound-mediated blood–brain barrier disruption

Thiele Kobus^1,2, Yong-Zhi Zhang², Natalia Vykhodtseva², Nathan McDannold²

¹Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands; ²Radiology, Brigham and Women's Hospital, Boston, Massachusetts, USA

Objectives

HER2-targeting antibodies prolong survival in patients with HER2-positive breast cancer metastases outside the brain. However, the response of brain metastases to these drugs is poor and it is hypothesized that the blood–brain barrier (BBB) limits drug delivery to the brain. We aim to improve delivery by temporary disruption of the BBB using focused ultrasound (FUS). Here we evaluate the treatment benefit of combining two antibody therapies that target the HER2-receptor with FUS-mediated BBB disruption in a breast cancer brain metastasis model.

Methods

MDA-MB-361 HER2-positive human cancer cells were injected in the right brain hemisphere of nude rats. The animals were divided in three treatment groups of 10 animals each: the control-group received no treatment; the antibody-only group received trastuzumab and pertuzumab (antibodies that target the HER2-receptor); and FUS+antibody-group received trastuzumab and pertuzumab in combination with FUS-mediated BBB disruption. The six weekly treatments started five weeks after tumour implantation, when the tumour diameter was around 2 mm. The FUS treatments took place in a 7T MR-scanner using a single-element, spherically-focused 690 kHz-transducer. Trastuzumab and pertuzumab were injected before the first sonication. At the start of each sonication (duration 60s, 10-ms bursts, burst repetition frequency 1 Hz), the ultrasound contrast agent Optison (100 μl/kg) was injected. The complete tumour was treated in 4 to 14 sonications that were separated 1 to 1.5 mm. Peak negative pressure amplitudes in water between 0.46 and 0.62 MPa were used.

Before and after the sonications, MR imaging was performed consisting of T2-weighted (T2w), T1w and T2*w imaging to locate the tumour, confirm BBB disruption and study the presence of hemorrhages. In two animals tumour leakiness was studied by comparing T1w imaging before and after gadolinium injection before the tumours were sonicated. The difference in signal intensity change in pre- and post-contrast T1w images was determined between the tumour and contralateral brain region (= SI%). In all FUS-treated animals BBB disruption was confirmed with contrast-enhanced T1w imaging and quantified using the same method as for the tumour leakiness. Pre- and post-sonication T2*w images were inspected for hypo-intense regions, which can indicate extravasated erythrocytes.

Every other week, high-resolution T2w imaging was performed to determine tumour volume. The growth rate (r) was determined by fitting the tumour volumes to the following formula: volume(t)=a*exp(r*t), in which t is the time in days. The growth rate of each tumour was determined for the treatment period (week 5 to 11) and the follow-up period (week 11 till sacrifice). An animal was classified as ‘responder’ if the growth rate was lower than the mean growth rate of the control animals minus two standard deviations.

The animals were euthanized if the tumour size exceeded 13 mm in diameter or if the condition of the animal was poor. From nine animals, histology was obtained (hematoxylin and eosin (H&E)) and the brains of five animals were stained for HER2.

Results

BBB disruption was successful in all sessions with an average SI% of 21.2% (range 4.5 – 77.6%). The mean SI% of two tumours before BBB disruption during the six treatment weeks were 0.4% and 0.6%, indicating that the tumours were not leaky before disruption (Fig. 16). In 33% (20/60) of the FUS-sessions, regions were present that were clearly more hypo-intense on post- than on pre-sonication T2*w images, suggesting hemorrhages. In the remaining 67% of the sessions, no or a small difference in hypo-intensity was observed.

In the FUS+antibody-group, 4/10 animals were classified as responders during the treatment period (week 5 to week 11) with an average growth rate of 0.010±0.007, compared to 0.043±0.013 for the non-responders. There was no difference in the average SI% of the responding rats (21.8%±16.7) and the non-responding rats (20.7%±9.7). None of the control or antibody-only animals were classified as responder. When the FUS+antibody-animals are grouped, no significant differences in mean growth rates were observed between the control, antibody-only and FUS+antibody animals for the treatment period, nor for the follow-up period.

High-resolution T2w imaging showed that the tumour was homogenous in almost all animals till week 13–15, when cystic and necrotic areas started to develop. The tumours showed also a heterogeneous appearance on H&E stained sections and the complete tumour was expressing the HER2-receptor in the examined animals.

Conclusions

In this study, we demonstrate that FUS-mediated BBB disruption in combination with antibody therapy can slow down the growth of brain metastasis from breast cancer. As the tumours were not leaky before BBB disruption and no difference in growth rates was observed in the antibody-only group, the disruption of the BBB is necessary for drug delivery to these brain metastasis. Interestingly, only part of the rats responded to the treatment, the other animals had the same growth rate as the control-group. This is in line with a previous study (Park et al. 2012, J. Control. Release), where antibody therapy was combined with FUS in a different brain metastasis model and in part of the animals a strong response was observed, while the other animals did not respond. We did not observe a difference in tumour volume at the start of the treatment, in HER2 expression on histopathology, or in contrast-enhancement on MR images between the responders and non-responders to explain this. Better understanding of why certain animals respond is needed and will help in translating this technique to the clinic.

Fig. 16 (abstract O22).

a T1weighted image before contrast administration. The red arrow indicates the tumour. b No difference in enhancement of the tumour is observed after contrast administration (SI=0.4%). c After focused ultrasound-mediated blood–brain barrier disruption, the tumour enhances after contrast administration (SI=30.1%)

O23 Long term follow up of 6 essential tremor patients treated with MR-guided focused ultrasound thalamotomy

Michael Schwartz^1,2, Yuexi Huang², Nir Lipsman¹, Jennifer Jain², Martin Chapman², Tejas Sankar³, Andres Lozano¹, Kullervo Hynynen²

¹Surgery (Neurosurgery), University of Toronto, Toronto, Ontario, Canada; ²Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada; ³Surgery (Neurosurgery), University of Alberta, Edmonton, Alberta, Canada

Objectives

To determine the factors influencing outcome after MR-guided focused ultrasound (MRgFUS) thalamotomy for essential tremor.

Methods

Between May 2012 and February 2013, 6 patients were treated with MRgFUS thalamotomy. The first 4 patients have been reported (Lipsman et al., Lancet Neurol. 2013 May; 12(5): 462–8). Prospective recording of data from preoperative screening to 3 month follow up according to an experimental protocol using the clinical rating scale for tremor (CRST), and then a 2 year follow up examination and assessment were done. Maximum temperature at the focus of sonication was determined and post-treatment day 1 MRI scans were reviewed for lesion size and location. A point derived by our current method of determining the probable location of the ventral intermediate nucleus (VIM) of the thalamus was taken as the starting point for each treated case and a deviation vector from that point to the centre of each lesion produced was plotted.

Results

For the 6 patients, the mean CRST A scale (rest, posture and action) for the treated hand and arm prior to treatment was 7.2. At 1 week post-treatment, the mean CRST A was 0.67. At 3 months, the mean CRST A was 0.83 and at 2 years approximately 3.33. (At 2 years, the CRST A for 2 patients was estimated from a narrative account). Mean lesion size, excluding one very large lesion with a volume 2.7 times the mean volume of the other 5 patients, was 90.2 mm³. Only 2 patients had no decline in thalamotomy effect. Of these, one had a relatively large lesion of 107.5 mm³, which was located at the predicted location of the VIM nucleus. The other patient’s lesion was located 2.4 mm lateral to the predicted VIM location. He had prominent ataxia following his treatment. This subsided by 3 months. The patient with the large lesion, treated for left hand tremor, had persistent tingling of the left side of his mouth, and his left thumb and 3 fingers but excellent early relief of tremor. Although there was some persisting reduction of tremor at 2 years, he could no longer write, nor drink without spilling. The centre of his lesion was 1.3 mm lateral and 0.3 mm posterior to the expected VIM location. In its superior-inferior dimension his lesion measured 9.4 mm. The maximum temperature achieved at treatment was 63°C, higher than the mean of 58°C. This decline may, in part, be due to the progress of his condition, with increasing tremor on his untreated side and the onset of tremor in both legs. Two of the patients with a decline in function had relatively small lesions.

Conclusions

Not all patients have lasting benefit from MRgFUS thalamotomy. The effect of relatively large lesions may be more durable, but lesioning temperatures of greater than 60°C should likely be avoided.

O24 Skull bone marrow injury caused by MR-guided focused ultrasound for cerebral functional procedures

Michael Schwartz^1,2, Robert Yeung¹, Yuexi Huang¹, Nir Lipsman², Jennifer Jain¹, Martin Chapman¹, Andres Lozano², Kullervo Hynynen¹

¹Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada; ²Surgery (Neurosurgery), University of Toronto, Toronto, Ontario, Canada

Objectives

To determine the factors that lead to bone marrow injury in the skulls of patients undergoing MRgFUS cerebral ablations for the treatment of movement disorders such as ventral intermediate (VIM) nucleus thalamotomy for tremor and pallidotomy for L-dopa induced dyskinesia (LID), and to follow the course of healing.

Methods

All patients undergoing functional cerebral ablations produced for the mitigation of movement disorders are followed with serial MRI scans according to study protocol to assess the evolution of the cerebral lesions. In one patient subjected to very high power sonication in the attempt to produce a pallidotomy for LID, skull lesions were noticed on a follow up MRI scan, produced 4 months after his MRgFUS procedure. A review of other treated patients is currently underway.

Results

Multiple ovoid lesions throughout the calvarium, new since the immediate post-treatment MRI scan done January 30, 2015, were seen on the MRI scan done May 12, 2015 (Fig. 17). Their appearance resembles that of bone infarcts. The MRI scan was repeated on October 2, 2015. Many of the ovoid lesions were still visible. This patient underwent sonication increasing to a maximum power of 1100 W for 31 seconds. During the procedure, the scalp and skull were constantly cooled with flowing degassed water at 13 °C. Despite this sonication, the target locus in the globus pallidus reached only 48 °C.

Conclusions

High power and duration sonication for functional cerebral lesions may cause injury to skull bone marrow. A review of all patients treated with MRgFUS for movement disorders is currently underway to determine whether there have been other cases, and to determine the threshold for bone marrow injury.

Fig. 17 (abstract O24).

See text for description

O25 In vitro study using MR-guided focused pulsed ultrasound for destroying clots using thrombolytic drugs

Christakis Damianou¹, Nikolaos Papadopoulos²

¹Cyprus University of Technology, Limassol, Cyprus; ²City University, London, United Kingdom

Objectives

In this paper an extensive study of using MR-guided focused pulsed ultrasound system is presented for the treatment of stroke using thrombolytic drugs in a model in vitro.

Methods

A single element spherically focused transducer of 4 cm diameter; focusing at 10 cm and operating at 1 MHz was used. The transducer was mounted in an MR compatible robotic system of 3 axes. The artery was modelled using a silicone tube. Tissue was modelled using agar-evaporation-silica gel. Coagulated blood was used to model thrombus. A thermocouple was placed in the thrombus in order to measure the thrombus temperature.

Results

The effect of power, presence of bubbles, temperature, presence of agar-evaporation milk-siligal gel, time of sonication, pulse repetition frequency, presence of standing waves, flow velocity were investigated. The goal was to maintain a temperature increase of less than 1°C during the application of pulsed ultrasound (called safe temperature). With the application of ultrasound alone or thrombolytic drug alone there was no notable destruction of the thrombus.

Conclusions

With the combination of ultrasound and thrombolytic drugs sufficient destruction occurred after 30 min, but the rate of destruction of thrombus (mL/min) is considered low. Thus, the clinical use of focused ultrasound for sonothrombolysis despite the full parametric study that we performed is considered pessimistic.

O26 Low frequency in-vivo cavitation mapping

Alexander Volovick¹, Javier Grinfeld¹, Yoav Levy¹, Omer Brokman¹, Eyal Zadicario¹, Ori Brenner², David Castel³

¹INSIGHTEC, Tirat Carmel, Israel, ²Weizmann Institute, Rehovot, Israel, ³Sheba Medical Center, Ramat Gan, Israel

Objectives

Current trans-cranial Magnetic Resonance guided Focused Ultrasound Surgery (tcMRgFUS) treatments are limited to the targets that are located at the centre of the skull. Superficial and peripheral targets are prone to lower acoustic intensity reaching the focal point due to the geometry of the skull and the ultrasound transversal through it. A possible solution for this problem is reduction of the transmitting frequency. Reducing the frequency is a trade-off between higher skull penetration and lower tissue absorbing. In addition, lower frequency has a lower threshold to induce cavitation. Previously it was reported ([1]) that the transient cavitation threshold for the muscle tissue acts linearly with frequency as where P_th is the cavitation pressure threshold and f is the transmitting frequency. However this threshold was not reported to cause any unintentional pathological damage. In the presented work, different levels of acoustic intensities resulting in various occurrences of cavitation were applied to an in-vivo pig brain model, reaching various temperature levels and causing various pathological effects.

Methods

36 subjects underwent craniotomy and were sonicated using various ultrasonic parameters (ExAblate Neuro low frequency system, 220 Khz, by INSIGHTEC, Ltd). The temperature rise was measured using MR thermometry (1.5T GE HDx MRI by GE Healthcare, with an integrated Head Coil, by InSighTec) and cavitation signal was measured and recorded using integrated hydrophones. 20 subjects underwent 2 weeks follow up with post procedural MR imaging one, five, seven and fourteen days after the procedure. The brains were harvested, fixed in formalin and sliced to 3 mm slices. Macro-pathological slices for randomly selected subjects were also performed.

Results

Figure 18 represents the graph of temperature rise as a function of applied energy. Figure 19 represents the graph of cavitation activity as a function of maximal acoustic power. Figure 20 shows a follow up imaging for 2 lesions performed on a single subject, Fig. 21 is a macro pathology slide for the subject presented in Fig. 20. In general the temperature rise grew linearly with the applied acoustic energy; the cavitation activity was linearly dependent to the maximum applied acoustic power. For sonications that were in the central cavitation activity area as presented on Fig. 19 the lesions were well defined and increased in size till the 5^th day follow up and then reduced in size becoming scarf tissue as presented on Figs. 20 and 21. Macro pathology revealed that tissue rapture is seen on the micro level; however it is well defined and limited within the lesion area.

Conclusions

Cavitation threshold levels that were observed in brain agree with the levels reported in literature. Additional levels of cavitation were observed and associated with effects on tissue as seen in MRI and histology. The data collected suggests cavitation levels that can be applied while keeping lesioning effect to the confined area and avoiding haemorrhages in tissue. Integrating a real time control over the level of cavitation and keeping the level below the safety threshold results in safe and effective tissue ablation in brain.

Fig. 18 (abstract O26).

Maximal average temperature as a function of effective energy. The graph is based on the data of 502 sonications

Fig. 19 (abstract O26).

Maximal cavitation activity as a function of effective power. The graph is based on the data from 253 sonications

Fig. 20 (abstract O26).

An example for two weeks follow up imaging for T2, T1 and T2 flair images presented. Two lesions are visible on the 1 and 5 days follow up images, single lesion is seen for the 7 days follow up and no lesion is visible at the 14 days follow up imaging. The lesions locations are marked by red circles at the 14 days follow up images in order not to hide the lesion and edemic tissue for the earlier follow up images. The lesion on the left was produced by the sonication reaching 58°C, whereas the lesion on the right reached temperature of 56°C

Fig. 21 (abstract O26).

Macro pathology of the slides presented on Fig. 20. Only one lesion was detected

O27 Real-time, transcranial passive cavitation mapping for monitoring of the focused ultrasound-induced blood–brain barrier opening in primates

Shih-Ying Wu¹, Julien Grondin¹, Wenlan Zheng¹, Marc Heidmann¹, Maria Eleni Karakatsani¹, Carlos J. Sierra Sánchez¹, Vincent Ferrera³, Elisa E. Konofagou^{1, 2}

¹Biomedical Engineering, Columbia University, New York, New York, USA; ²Radiology, Columbia University, New York, New York, USA; ³Neuroscience, Columbia University, New York, New York, USA

Objectives

Real-time cavitation monitoring during the focused ultrasound (FUS) induced blood–brain barrier (BBB) opening is crucial in assessing and controlling the BBB opening outcomes and safety. Currently, passive cavitation detection using a single-element detector for quantification of the stable and inertial cavitation doses has shown good correlation with the opening volume and the molecular delivery efficiency in nonhuman primates (NHPs). However, an off-line magnetic resonance imaging (MRI) is required for confirming the targeting or opening area in the brain after the FUS procedure. It is therefore essential to develop transcranial cavitation mapping in providing the spatial distribution of cavitation intensity in real time in order to precisely assess and control the BBB opening outcomes with accurate targeting during the procedure. The aim of this study is to develop real-time cavitation mapping using time-exposure acoustics and passive sparse matrix beamforming with the performance evaluated in both the in vitro primate skull (NHP and human) experiments and the in vivo NHP experiments during BBB opening.

Methods

Similar to array-based passive cavitation detectors, a linear probe (L7-4, Philips) with a programmable data acquisition system (Vantage, Verasonics) was used to acquire cavitation signal passively during the sonication (frequency: 0.5 MHz, pulse length: 10 ms, PRF: 5 Hz, pressure: 150–600 kPa) with in-house made microbubbles (lipid-shelled and monodisperse with a diameter of 4–5 μm) and a single-element FUS transducer with a coaxially aligned flat-band hydrophone as a single-element passive cavitation detector. Time-exposure acoustics for an integration of a series of passive cavitation images over the exposure time reconstructed by dynamic receive beamforming using sparse matrix calculation in graphic processing unit (GPU; Tesla K40, NVIDIA) were developed for the reconstruction of passive cavitation maps in real time. For the in vitro experimental setup, a phantom with a channel of 4 mm in diameter and an infusion pump was used to mimic the vessel with microbubble circulation (a concentration of 2×10⁵ bubbles/mL with a flow rate of 1 mL/min), and the FUS transducer and the linear array were both focused at the channel orthogonally. B-mode imaging with the linear array was performed before acquiring the passive cavitation signal in order to confirm the alignment of both the linear array and the FUS transducer to the channel. Three sets of passive cavitation maps were acquired using this in vitro setup: 1) without the skull, 2) with the NHP skull and 3) with the human skull placed between the phantom and the linear probe. The effects of acoustic pressure, exposure time and aperture size to the intensity and focal size of the cavitation maps were all evaluated, as well as the computational cost in GPU and the sensitivity through the skull. Furthermore, the in vivo cavitation maps were acquired during the sonication for BBB opening in NHPs (duration: 2 min) and compared with the BBB opening outcomes in the MRI.

Results

The results of the in vitro experiments showed that the cavitation location corresponded to the microbubble disruption in B-mode images, and both the intensity and the focal area of the cavitation maps increased with pressure. Increasing the exposure time eliminated the interference outside of the focus and enhanced the focalization by minimizing the focal area, while both the intensity and the focal area reached a plateau at the exposure time of 62.5 μs. The focal size especially in the axial direction increased with decreasing aperture size, suggesting an improved focalization by using a larger aperture size of the array-based PCD. The computational time for the exposure time of 62.5 μs was 9.5 s, which can be decreased to 0.27 s in achieving real-time monitoring by decreasing the exposure time to 1.44 μs. By placing the skull, it was found that the cavitation signals were still detectable through the NHP and human skulls at 300 kPa and 600 kPa with the mapping system, respectively. For the in vivo experiments, the BBB opening in NHP were successfully monitored with passive cavitation mapping targeting at the caudate and the hippocampus, the deep subcortical structure in the brain.

Conclusions

A real-time cavitation mapping technique using time-exposure acoustics and passive sparse matrix beamforming has been developed with the performance and the sensitivity through the primate skull evaluated, and was used for monitoring of the BBB opening in NHPs. This novel transcranial monitoring technique providing both the spatial and intensity information of cavitation in real time during the FUS procedure is promising in assessing and controlling the targeting, treatment efficacy, and safety precisely.

O28 Blood brain barrier opening using focused ultrasound for the reduction of amyloid beta plaques in synergy with antibodies in a rabbit model fed with high cholesterol diet

Christakis Damianou, Marinos Yiannakou

Cyprus University of Technology, Limassol, Cyprus

Objectives

An animal model that creates Amyloid beta (Ab) plaques in the brain was implemented by delivering high cholesterol diet in rabbits for 3 months. The goal was to reduce the plaques using focused ultrasound (FUS) in combination with external antibodies.

Methods

A single spherically focused transducer was used which operates at 1 MHz, has focal length of 10 cm and diameter of 3 cm. The rabbit was placed in a custom made MRI compatible positioning device. Theoflavin staining was used in order to measure the plaque load at the end of each experiment.

Results

Using pulse FUS the blood brain barrier (BBB) was opened repeatedly up to 5 times at three day intervals. The opening of the BBB disruption was imaged using contrast-enhanced T1-weighted fast spin echo. By increasing the number of sessions, the number of plaques decreases (both for internal and external antibodies). With the use of FUS only (internal antibodies) the drop of average number of plaques/mm² was reduced by 20% (in 5 sessions). The effect of external antibodies was more drastic. With 5 BBB sessions the average number of plaques/mm² was reduced by 60%.

Conclusions

This study demonstrated that by opening the BBB, it will be possible to deliver internal and external antibodies to the brain, which eliminates Alzheimer disease (AD) plaques. More important by opening the BBB frequently (up to 5 times in this study) the reduction in the number of plaques is enhanced. Therefore FUS has the potentials to be used non-invasively for the treatment of AD.

O29 Correlation between down-regulation of p-glycoprotein and blood–brain barrier disruption in rat brain by mri-guided focused ultrasound and microbubbles

HongSeok Cho¹, Hwayoun Lee¹, Mun Han², Jong-Ryul Choi¹, Taekwan Lee¹, Sanghyun Ahn¹, Yongmin Chang², Juyoung Park¹

¹Daegu-Gyeongbuk Medical Innovation Foundation, Daegu, Republic of Korea; ²Kyungpook National University, Daegu, Korea (the Republic of )

Objectives

Blood–brain barrier (BBB) is composed of both physical barrier with tight junctions and functional barrier with active efflux transporters. Mechanism of the functional barrier is mediated by P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP) in brain endothelial cells. Past studies have shown that focused ultrasound (FUS) combined with microbubbles can disrupt the physical barrier of the BBB by exerting mechanical stress on the tight junctions; however, no study was performed to investigate the impact of FUS and microbubbles on the functional barrier of the BBB. Therefore, this study investigated the impact of BBB disruption induced by FUS and microbubbles on the expression of P-gp. We also investigated correlation between magnitude of the BBB opening and the down-regulation of P-gp.

Methods

A single target region was sonicated (0.5~0.7 MPa) transracially in one hemisphere in 3 rats using a 1 MHz FUS transducer; the other hemisphere served as a control. For the BBB disruption, 10 ms bursts were applied at 1Hz pulse repetition frequency (PRF) for 120 s and combined with IV injection of a microbubble ultrasound contrast agent (Definity 0.1 ml/kg). An MR contrast agent (Magnevist 0.4 mM/kg) and Evans Blue (0.15 ml/kg) were injected immediately after the sonication to indicate area of the BBB disruption in MR image and fluorescence spectroscopy, respectively. In order to measure the P-gp expression using a confocal fluorescence microscopy, the brains were fixed after perfusion and then stained immunohistochemically with a monoclonal antibody (C219) which reacts with a P-gp epitope.

Results

A T1 contrast enhanced MR image and Evans Blue fluorescent intensity at the sonicated regions indicated localized BBB disruption (Fig. 22). Both the MR contrast intensity and the Evans Blue fluorescent intensity were significantly increased in the targeted regions compared to the control regions (p<0.001). The fluorescence intensity of the P-gp expression at the confirmed locations of the BBB disruption was reduced by an average of 63.2±18.4% compared to the control area in all three rats. From the three sonicated regions, a total of 31 locations were selected and the P-gp fluorescence intensities were measured to observe the correlation between the degree of the BBB opening and the P-gp expression. Both the Evans Blue intensity and the MR contrast intensity were significantly correlated with the P-gp expression intensity (r=−0.72, p<0.001; r=−0.62, p<0.001, respectively). Histologic analysis on the sonicated region of the brain tissue revealed no apparent damage in the endothelial cells, and no significant amount of extravasated red blood cells was observed.

Conclusions

This study demonstrates that the BBB disruption induced by FUS and microbubbles reduces the expression of P-gp, and the level of the down-regulation of P-gp is significantly correlated with the magnitude of the BBB opening. These results suggest FUS + microbubble as a promising mean for the brain drug delivery through the BBB by overcoming both the physical and the functional barrier of the BBB.

Fig. 22 (abstract O29).

a A T1 contrast MR Image of a rat brain at sonification region (the opposite hemisphere of the brain served as control). b The fluorescence intensity of the P-gp expression at the soicated locations were reduced by an average of 63.2±18.4% compared to the control locations

O30 MR-guided focused ultrasound blood–brain barrier disruption through an intact human skull in a rat model using a clinical body system

Nicholas Ellens¹, Ari Partanen^1,2, Keyvan Farahani^1,3, Raag Airan¹

¹Radiology, Johns Hopkins University, Baltimore, Maryland, USA; ²Philips, Andover, Massachusetts, USA; ³National Cancer Institute, Bethesda, Maryland, USA

Objectives

MR-guided focused ultrasound (MRgFUS) can be used for a large range of non-invasive therapies using mechanical and thermal mechanisms. For instance, it has been demonstrated that a combination of low intensity focused ultrasound and synthetic microbubble scan be used to safely, locally, and transiently disrupt the blood–brain barrier (BBB)[1].

Typically, transcranial MRgFUS thermal ablation requires large aperture phased arrays with many elements operating at lower frequencies and higher powers than body systems due to high transmission losses through the skull. However, in low pressure applications at low duty cycles, such as BBB disruption or neuromodulation, it may be feasible to use 'body' systems that have smaller aperture transducers with fewer elements [2]. The objectives of this study were to use a clinical body MRgFUS system to:

• 1) Quantify the ultrasound transmission and focal distortion through a human cranium at different orientations and in different focus locations inside the skull to assess the possible treatment envelope with such a device.

• 2) Demonstrate the ability to disrupt the blood–brain barrier (BBB) in rat neurovasculature through a human cranium.

Methods

A clean, degassed human cranium (gift of Dr. Quiñones-Hinojosa, JHU Neurosurgery) was mounted in degassed water above a 256-element phased-array transducer (14 cm focal length) of a clinical body MRgFUS system (Sonalleve V2, Philips, Vantaa, Finland). Hydrophone measurements were made with a lipstick hydrophone (Onda Corp., Sunnyvale, USA) mounted to a 3D stage, both in water and through the human skull in 11 locations (three orientations, 3–4 depths at each, ranging between the centre of the cranium and 2 cm from the skull surface). 40-cycle sonications at 1 MHz were applied at different acoustic powers ranging from 5 to 20 W (water) and 20 to 500 W (trans-skull). In each location, the full-width at half-maximum (FWHM) of the ultrasound focal point (in three dimensions) and its peak negative pressure were measured. This setup and the locations sampled are shown in Fig. 23.

Results

Hydrophone measurements demonstrated a great deal of variation in ultrasound transmission with changing transducer/skull incidence angle and skull thickness. The pressure attenuation ranged from −5.8 dB to −9.3 dB (mean +/−standard deviation, −8 dB +/− 1 dB), and the FWHM varied between 1.83 mm and 3.79 mm (2.2 mm +/− 0.3 mm in the anterior/posterior direction, 2.7 +/− 0.7 mm in the left/right direction). The insertion losses as a function of depth at different orientations are shown in Fig. 24. The dorsal midline skull (with the ultrasound beam passing through the sagittal and coronal sutures) was thicker than either of the angled approaches (passing through the parietal and temporal bones) and experienced a greater insertion loss for most sonication target depths. For both angled approaches, the insertion loss decreased as the focus was moved away from the skull surface. For the centered approach, the insertion loss increased slightly as the focus moved closer to the skull surface. It appears that the intervening skull thickness dominates the attenuation, though the angle of incidence seen by transducer elements also affects transmission.

Gadolinium enhancement in the brain on post-sonication T1-weighted MRI indicated successful BBB disruption. Disruption accuracy is shown in Figs. 25 and 26. For sonications of sufficient pressure to produce BBB disruption, the region of gadolinium enhancement did not appear to be shifted at all from the desired target. For higher estimated in situ pressures (>0.55 MPa), the disruption region was larger and less uniform, as evident in the 0.58 MPa point in Fig. 25. Even without element-by-element refocusing, the focusing quality and targeting precision appeared to be adequate. Post-experiment, there were no signs of gross damage to the animals to suggest significant off-target sonication.

Conclusions

Transcranial FUS focal point distortion was minimal despite the lack of element-by-element transducer refocusing. The clinical MRgFUS body transducer and driving electronics had sufficient power and aperture to generate the in situ pressures for BBB disruption through a human skull, using a variety of clinically practical approaches and patient orientations.

Though transcranial thermal ablation typically requires high acoustic powers, a lower operating frequency, and large aperture arrays with high number of elements, this study demonstrates that a clinical body MRgFUS system with a smaller transducer may be a safe and feasible alternative for non-invasive BBB disruption and other low-pressure therapeutic ultrasound transcranial applications, potentially offering a wide treatment envelope.

References

[1] Hynynen, K., et al. (2001). Radiology, 220(3), pp. 640–646.

[2] Airan, R. D., et al. (2015). 23^rd ISMRM.

Fig. 23 (abstract O30).

a and b: Coronal and sagittal representations of the locations examined. c: Picture of the hydrophone and skull arrangement for one of the 'Left 45°' orientations. The transducer is below the table

Fig. 24 (abstract O30).

Insertion losses for different sonication orientations and depths

Fig. 25 (abstract O30).

Baseline T1-weighted coronal image, left, and post-contrast post-BBB disruption image, right, showing two locations of BBB disruption with the estimated in situ pressures labelled

Fig. 26 (abstract O30).

Sagittal images showing columns of BBB disruption produced by the labelled estimated in situ pressures. Ultrasound propagation is from the right to the left of these images.

O31 Interim results from a phase 1 clinical trial to disrupt the blood–brain barrier by pulsed ultrasound

Alexandre Carpentier^3,4, Michael Canney¹, Alexandre Vignot¹, Cyril Lafon², Jean-Yves Chapelon², Jean-yves Delattre^4,5, Ahmed Idbaih⁵

¹CarThera, Lyon, France; ²INSERM, U1032, LabTau, Lyon, France; ³Assistance Publique Hopitaux de Paris, Hopital de la Pitie Salpetriere, Department of Neurosurgery, Paris, Paris, France; ⁴Universite Paris, UPMC, Paris, France; ⁵Assistance Publique Hopitaux de Paris, Hopital de la Pitie Salpetriere, Department of Neuro-Oncology, Paris, France

Objectives

Pulsed ultrasound, coupled with peripheral injection of microbubbles, has been shown in pre-clinical studies to be an effective method for enhancing the delivery of chemotherapy to the brain. In this work, an intra-skull implantable ultrasound device, SonoCloud®, was developed for temporarily disrupting the BBB. The device was implanted in patients with recurrent glioblastoma (GBM) undergoing systemic carboplatin chemotherapy in a Phase 1 clinical trial and the safety of repeated BBB disruption was assessed.

Methods

A Phase 1 clinical trial began in July 2014 at a single centre at the Hospital Pitie Salpetriere in Paris, France. Patients with recurrent GBM with an enhancing volume of less than 35 mm in diameter were included in the trial. Participants were implanted with a 11.5-mm diameter biocompatible 1 MHz ultrasound transducer, which was fixed into the skull bone thickness. The device was either implanted during a regular surgical resection of the enhancing region or during a unique ambulatory procedure under local anaesthesia. Once a month, the device was connected to an external radiofrequency generator using a transdermal needle, and patients received a two minute pulsed ultrasound exposure in combination with systemic administration of an ultrasound contrast agent (7 min mean total duration procedure). BBB disruption was assessed immediately after sonications using dynamic T1-weighted imaging with a gadolinium based MR contrast agent. Systemic intravenous injection of carboplatin chemotherapy was delivered immediately following MR imaging. Patients followed a progression of ultrasound dose in which the pressure was increased from 0.5 to 0.8 MPa throughout the course of the study.

Results

As of July 2015, eleven patients had been included in the study with a total of 25 BBB disruption procedures performed. No adverse effects were observed in patients treated. BBB opening was clearly observed in 12/25 treatments on contrast-enhanced T1w imaging. The procedure was safely tolerated in all patients. No evidence of acute haemorrhage, petechia, ischemia or oedema was observed in post-sonication SWAN T2*, Diffusion or FLAIR MRI sequences.

Conclusions

The BBB was safely opened by pulsed ultrasound using an implantable ultrasound device in patients with recurrent glioblastoma. Additional follow up and recruitment will be used to further evaluate the safety and potential efficacy of such an approach.

Acknowledgments

Work supported by CarThera and the Hospital Pitie Salpetriere.

O32 Investigation of temperature dependent changes in MR signal intensity, t1 and t2* in cortical bone

Henrik Odéen², Bradley Bolster¹, Eun Kee Jeong², Dennis L. Parker²

¹Siemens Healthcare, Salt Lake City, Utah, USA; ²Radiology, University of Utah, Salt Lake City, Utah, USA

Objectives

For MR guided focused ultrasound treatments in or close to bone, such as for transcranial applications focusing through the intact skull bone and treatments of bone metastases, significant heating can occur in the bone due to its high ultrasound absorption. MR imaging of bone is in general challenging due to the short T2 relaxation time of bone. For MR temperature imaging the short T2 also severely decreases the accuracy that can be achieved with the standard proton resonance frequency shift method. Instead researchers have investigated the temperature dependence of the MR signal intensity (SI) and T1 relaxation time for temperature monitoring (1–4). Miller (1) and Fielden (3) et al. showed that the SI from cortical bone decreases with increasing temperature using ultrashort echo time (UTE) pulse sequences, and Ramsay et al. (2) found that, contrary to what Miller and Fielden observed, the SI increases with increasing temperature using a short TE gradient recalled echo (GRE) pulse sequence. Han et al. (4) further showed that T1 increases with temperature, also using UTE.

In this work we investigate the temperature dependence of the SI (dSI/dT) and the T1 and T2* relaxation times (dT1/dT and dT2*/dT, respectively) using a 3D UTE pulse sequence to investigate which parameter has the highest sensitivity to temperature change.

Methods

All imaging was performed on a 3T MRI scanner (MAGNETOM PrismaFit, Siemens Healthcare, Erlangen, Germany) using a 3D UTE pulse sequence. The sequence utilizes radial, ramped sampling of k-space in 3D starting at the k-space centre after a 80 μs non-selective hard RF pulse, allowing TEs down to 50 μs. T2* was measured by performing an exponential fit to data acquired at TE = 50, 90, 130, 170, and 250 μs (other scan parameters are listed in Table 1). The 50 μs TE was also used for dSI/dT calculations. T1 was measured using the variable flip angle (VFA) method (5) with TR = 11 ms and FA = 8 and 36° (other scan parameters are listed in Table 1).

An approximately 4-cm long bovine femur (marrow and connective tissue removed) was placed in a phantom holder that allowed heated water to circulate around the bone, Fig. 27. One fiber optic probe measured the water temperature, and three probes were inserted in 1-mm diameter, 2-cm deep, holes drilled in the bone to measure the temperature of the bone. The water was warmed to ~22, 35, 50, and 65°C and data was collected when all 4 probes measured within 1°C. The whole setup was places in a 20-channel RF head coil for signal detection.

Results

Figure 28 shows 2D maps of SI and T1 and T2* relaxation times for the 4 different temperatures. Mean and standard error values from a 9x9 ROI close to each probe is shown in Fig. 29, together with calculated changes in %/°C. A decrease in SI of 0.3-0.5%/°C, and increases in T1 and T2* of 0.5-0.9%/°C and 0.6-0.9%/°C, respectively, was observed. Using the temperature dependent spoiled GRE signal equation (6) and the observed values for dT1/dT and dT2*/dT, dSI/dT can be closely predicted.

Conclusions

The decrease in SI is in accordance with previously published results by Miller and Fielden. The measured change in T1 using UTE agrees well with the 0.6%/°C reported by Han, although we measured higher absolute T1 values (~160-175 ms at 25°C, compared to ~115-125 ms as reported by Han).

The effect of T1 and T2* on SI are counter-acting each other (both increasing), which reduces the sensitivity of dSI/dT. This may suggest that dT1/dT and dT2*/dT are more suitable candidates for bone MR thermometry, and Fig. 29 also shows higher sensitivity for relaxation times that for SI. However, it should be noted that SI can be detected from a single image, whereas T1 and T2* measurements utilize two or more images, therefor resulting in longer scan times.

Future studies will acquire data with longer TEs (out to ~10 ms) to investigate the difference in temperature sensitivity for shorter and longer T2 components in the bone. These results will be compared to the increase in SI with temperature observed by Ramsay using a TE ≈ 1 ms. The temperature dependence of the long and short T2* components can be found by a multi-exponential fit. Flip angle mapping will also be implemented to improve the accuracy of the T1 measurements (7).

Table 1 (abstract O32). Scan parameters

Fig. 27 (abstract O32).

Scan setup. A ~4 cm long bone sample was placed in a phantom holder that allowed water circulation to homogenously heat the bone. 4 fiber optic probes were used; 1 in the water and 3 in the bone sample.

Fig. 28 (abstract O32).

2D maps of SI and relaxation times for the 4 different temperatures (~22, 35, 50, and 65 °C), a SI, b T1, and c T2*

Fig. 29 (abstract O32).

Changes versus temperature for a SI, b T1, and c T2*. Mean and standard error value from a 9x9 voxel ROI close to each probe is shown

O33 Spatially-segmented MRI brain and water bath reconstruction for undersampled transcranial mr-guided focused ultrasound thermometry

Pooja Gaur¹, Xue Feng², Samuel Fielden², Craig Meyer², Beat Werner³, William Grissom¹

¹Vanderbilt University, Nashville, Tennessee, USA; ²University of Virginia, Charlottesville, Virginia, USA; ³University Children's Hospital, Zurich, Switzerland

Objectives

MR-guided focused ultrasound (MRgFUS) brain systems deliver targeted thermal energy into the brain using a hemispheric array of transducers that surround the head with an intervening water bath (Fig. 30a). During treatment the localized heating (hot spot) is measured from a change in image phase between baseline (pre-treatment) and dynamic (during treatment) images. Accelerating temperature mapping by undersampling k-space is desirable to increase spatiotemporal resolution and coverage, but is difficult to do with parallel imaging since coils must be placed outside the transducer, far away from the head. Multiple groups have instead developed accelerated temperature mapping methods that exploit temporal correlations between baseline and dynamic images [1, 2]. However, circulation of the water bath to cool the skull causes dynamic signal changes that are not captured by baseline images (Fig. 30b), which breaks those correlations and results in artefacts throughout the temperature maps. We propose a spatially-segmented approach for reconstructing temperature maps in brain MRgFUS, in which we separately estimate a water bath image without a baseline, and a temperature map in the brain with a baseline. The method can estimate artefact-free temperature maps from undersampled data during brain MRgFUS treatments using a single receive coil.

Methods

Our iterative approach alternates between updating the parameters of a k-space hybrid signal model which is fit in the brain region of the image [1], and a baseline-free estimate of the water bath image. The fitting of k-space hybrid brain model results in a phase drift-corrected brain image without the temperature phase shift and a sparse temperature phase shift map. An algorithm to fit the model is described in [1]. The water bath is reconstructed using a POCS algorithm that alternately enforces data consistency, consistency with a water bath support mask (brain and water bath masks are obtained from a baseline image), and sparsity in the Coiflet domain using soft thresholding [3]. Figure 30c illustrates the overall undersampled dynamic image model.

To test the method, a gel-filled human skull phantom was sonicated by an InSighTec ExAblate Neuro 4000 transcranial MRgFUS system (InSighTec Ltd, Haifa, Israel) while imaging with a GE 3T MR750 scanner (GE Healthcare, Waukeshaw, WI). 27 single-slice 2DFT gradient echo images were collected with the body coil and 28 x 28 x 0.3 cm3 field of view, 256 x 128 acquisition matrix, 30° flip angle, 13 ms TE, and 28 ms TR. Images and maps were reconstructed to a 128 x 128 matrix and retrospectively randomly undersampled by 2x, with full sampling over 22 central k-space lines. Temperature maps were reconstructed by fitting the k-space hybrid model to the entire image, or to the brain only with keyhole or POCS methods used to reconstruct the water bath image.

Results

Figure 30d shows the temperature reconstruction results. When the k-space hybrid model is fit to the entire image without distinguishing between the brain and water bath, phase artefacts obscure the hot spot in the reconstructed temperature map and (in this case) lead to an overestimation of the temperature rise in the sonicated region across image dynamics (RMSE across dynamics: 0.0121°C). Restricting the temperature reconstruction to within the brain, in combination with keyhole reconstruction of the water bath image (using the baseline image’s k-space to fill in missing k-space lines), produces temperature maps with lower errors in the hot spot but still large errors outside (RMSE across dynamics: 0.0039°C). The proposed k-space brain/POCS bath approach yields a more accurate estimate of the water bath image (not shown), resulting in much lower in-brain temperature artefacts (RMSE across dynamics: 0.0029°C).

Conclusions

Unpredictable water bath motion confounds model-based approaches to accelerated MR temperature mapping, resulting in large temperature artefacts due to aliased water bath signal. We demonstrated that a spatially-segmented reconstruction that applies a model-based reconstruction in the brain and a POCS reconstruction in the water bath can reconstruct temperature maps without undersampling artefacts at a moderate acceleration factor using a single receive coil. Future work will focus on integrating the approach with other accelerated temperature mapping methods [2] and extending it to non-Cartesian trajectories [4]. The method is compatible with multiple receive coils.

Acknowledgements

This work was supported by the Focused Ultrasound Foundation and NIBIB T32EB014841.

References

[1] Gaur P et al. Magn Reson Med 2015;73: pp. 1914–1925.

[2] Todd et al. Magn Reson Med. 2009;62: pp. 406–19.

[3] Lustig M et al. Magn Reson Med 2007;58: pp. 1182–1195.

[4] Fielden et al. Proc Intl Soc Mag Reson Med 23. 2015:1631.

Fig. 30 (abstract O33).

a During MRgFUS treatment, the patient’s head is immobilized in the transducer and circulating water bath. b The water bath signal varies significantly during a single focused ultrasound (FUS) sonication (arrow indicates sonication target). c In the proposed method, undersampled dynamic data are reconstructed using the k-space hybrid method in the brain and a POCS reconstruction in the bath. d Reconstructed temperature changes and maximum temperature errors in the brain with 2x undersampling. Temperature change averaged over the hot spot center is plotted at the bottom for each reconstruction. Circles on the x-axis indicate dynamics for which temperature maps are displayed above

O34 Efficient volumetric thermometry for MR-guided FUS brain treatment monitoring, using multiple-echo spirals and mixed update rates

Michael Marx, Pejman Ghanouni, Kim Butts Pauly

Radiology, Stanford University, Stanford, California, USA

Objectives

MR-guided focused ultrasound (MRgFUS) brain treatments are currently guided by one thermometry sequence: single-slice 2DFT MR thermometry. In this work, we divided treatment monitoring into two tasks, with different thermometry design goals for each, and developed sequences optimized for these goals: “Focal Spot Localization” and “Monitoring”. These sequences achieve greater imaging performance by utilizing multi-echo spiral thermometry, region-specific update rates, and MASTER slice interleaving.

Currently, focal spot targeting confirmation requires several low-power sonications to obtain high-resolution measurements in three dimensions. We developed a Focal Spot Localization sequence that obtains high-resolution measurements in-plane, at improved temperature precision compared to 2DFT, while also providing multiple-slices for through-plane characterization. This would reduce the number of sonications required, improving treatment time. Additionally, even lower-power sonications could be detected, improving patient safety.

During ablative treatment sonications, single-slice monitoring cannot detect through-plane shifts of the ultrasound focus, or unexpected out-of-plane heating. We developed a multi-rate thermometry Monitoring sequence that interleaves different sequences at different update rates to simultaneously achieve fast and precise focal monitoring, 3-dimensional focal spot measurement, and full brain monitoring. Fast and precise multi-slice monitoring of the focus ensures accurate thermal dose estimates for treatment feedback, while (slower) full-brain monitoring ensures patient safety.

Methods

All sequences were implemented using RTHawk (HeartVista, Menlo Park, CA) on a GE 3T 750 Signa scanner (GE Healthcare, Milwaukee, WI) equipped with the InSighTec Exablate Neuro (InSighTec, Haifa, Israel). All imaging was performed with the body coil, as is normally used with the Exablate system. Conventional 2DFT was implemented as a gold standard for comparison. The new sequences all use a 36 cm FOV to ensure that the transducer and water bath do not alias. Volunteer imaging was done with informed consent under IRB approval. Multi-frequency reconstruction was performed on all spiral data. Temperature uncertainty was measured as the per-voxel temporal standard deviation of temperature measurements, and averaged within manually segmented ROIs. Performance was also tested inside the transducer, using a gel phantom. Sequence parameters are compiled in Table 2.

The Focal Spot Localization sequence is a 5-slice 3-echo thermometry sequence, with doubled in-plane resolution as compared to conventional 2DFT (1.1x1.1 mm vs 1.09x2.18 mm), and acquisition time of 7 s. The Monitoring sequence interleaves 3 distinct sequences at different rates to monitor 29 total slices. The “Focus” is monitored using 3-slice 3-echo spiral imaging, for high-speed high-precision measurement of focal heating. Two adjacent slices, the “Boundary”, are monitored at half the temporal rate (also using 3-echo spiral) to fully characterize the focal spot. The remaining 24 slices, “Background”, were acquired using 8 blocks of 3-slice MASTER, with spiral readouts. Use of MASTER improves temperature uncertainty, compared to traditional slice interleaving, by increasing echo time. Limiting each MASTER block to 3 slices reduces inherent diffusion and motion-encoding artefacts. The three component sequences were interleaved such that Focus utilized 45% (15% per-slice) of the timeline, Boundary used 15% (7.5% per-slice), and Background the remaining 40% (1.7% per-slice).

Results

Figure 31a compares temperature uncertainty in vivo between 2DFT and Focal Spot Localization while Fig. 31b compares 2DFT with Monitoring for the same volunteer. Large images compare the centre slice, while stacks of images at the right show additional slices monitored by the new sequences. Figure 31c compares 2DFT and Monitoring uncertainty within the transducer. The new sequences obtained better uncertainty than 2DFT, with average values compiled in Table 2. Relative “Efficiency” is also listed in Table 2, which accounts for differences in speed and voxel volume. Each multi-echo spiral sequence is more than 150% as efficient as 2DFT. Background is 69% as efficient (but collects 3 slices per TR, for an effective 120% efficiency).

Conclusions

In this work, we have shown that multi-slice multi-echo spiral thermometry is an effective imaging approach for volumetric treatment monitoring. Improved imaging performance was successfully used to achieve imaging objectives for different aspects of ablative treatments. Focal spots may be localized faster and with reduced heating using the higher-resolution higher-precision Focal Spot Localization sequence. The mixed update rate Monitoring sequence successfully delivers high-speed high-precision monitoring of the targeted focus, while improving safety by simultaneously monitoring the full brain at a lower update rate. These sequences have also been validated within the transducer, to help ensure they will work in the clinical setting.

Table 2 (abstract O34). Implemented Sequence Performance Parameters. tSNR efficiency is proportional to (δ_xyz*√(T_seq)*σ_T)⁻¹ . For Monitoring - Background, median slice uncertainty reported

Fig. 31 (abstract O34).

Temperature uncertainty comparisons between 2DFT and a Focal Spot Localization, in vivo; b Monitoring, in vivo; c Monitoring, phantom within transducer. Large images compare centre slices, while stack of images at right show additional slices. Dotted outline in 1C delineates the phantom

O35 Towards MR-guided focused ultrasound treatments near metallic hardware

Hans Weber, Valentina Taviani, Kim Butts Pauly, Pejman Ghanouni, Brian Hargreaves

Radiology, Stanford University, Stanford, California, United States

Objectives

To demonstrate how MRI can be used in FUS treatments near metallic hardware for treatment planning, sonication monitoring and treatment assessment.

Methods

Using standard MRgFUS product sequences, we treated a 73-year-old patient with a metastasis in the right femur that was painful despite prior radiation and surgical stabilization with a metallic rod, demonstrating the inability of both echo-planar imaging (EPI) and gradient-recalled echo (GRE) imaging to be used for conventional proton resonance frequency (PRF) shift thermometry.

Based on the experience from this patient treatment, we have proposed an imaging strategy for MRgFUS near metallic hardware using multi-spectral imaging (MSI) techniques (Fig. 32a). Both MAVRIC-SL [Koch et al.; MRM 2011; pp. 65:71] and 2DMSI [Hargreaves et al.; ISMRM 2014, #615] are spin-echo-train-based imaging approaches that enable artefact-reduced imaging near metal. MAVRIC-SL compensates for distortions induced by field inhomogeneities by additional encoding along the slice dimension. Its ability to adjust the image contrast makes it a promising candidate both for treatment planning and assessment. 2DMSI limits the excitation to finite spectral and spatial regions (“frequency bins”) that can be imaged with minimal artefact quickly enough to be used for sonication monitoring.

We preliminarily tested the feasibility of MAVRIC-SL for treatment planning and assessment in a 65-year-old patient without metal hardware and with a metastasis in the right pelvis undergoing MRgFUS treatment for pain palliation. We used a GE 3T MRI system equipped with an InSightec ExAblate2000 FUS system. Proton density (PD) weighted MAVRIC-SL images were acquired prior to the treatment in addition to standard 2D fast-spin-echo (FSE) images. After treatment, T1-weighted MAVRIC-SL images were acquired before and after gadolinium injection in addition to standard 3D fast RF-spoiled GRE images with 2-point-Dixon fat suppression (LAVA-Flex, see Fig. 32c for details).

We tested the feasibility of 2DMSI to monitor a 25 s sonication in an acrylamide egg-white phantom containing the CoCr stem of a knee replacement and a pork loin sample containing a CoCr augment plate from the same hardware, each placed in a container filled with water. We acquired a time series of 10 single-slice 2DMSI images with a temporal resolution of 8 s/frame and the sonication starting after the third image. Each 2DMSI image was composed of 12 frequency bins ranging from −4.5 kHz to +5.4 kHz, and was acquired with TE = 30 ms, (bin) TR = 500 ms, 280 x 280 mm FOV, 5 mm slice thickness and 128 x 128 matrix size (effective number of phase encodings after half Fourier). To monitor the temperature-induced signal change, we subtracted the mean signal of the 3 baseline images from all 10 images voxel-by-voxel.

Results

Figure 32b shows the GRE and EPI image acquired in the patient with metallic hardware. With both techniques, the stabilized femur and its surrounding area are not visible due to distortions and signal dephasing. For comparison, the MAVRIC-SL image depicts the anatomy including the bone marrow surrounding the metallic rod.

Figure 32c presents the acquired images for treatment planning and assessment in the patient without metal. PD-weighted MAVRIC-SL achieves a contrast comparable that of the conventional FSE image and allows for localization of the metastasis in the lower part of the right pelvis. The T1-weighted MAVRIC-SL pre/post contrast difference image reveals the treatment area similar to the LAVA-Flex water difference image.

Figure 33 shows the thermometry results in the phantom and the tissue sample. In both cases, metal-induced field inhomogeneities of up to ± 4 kHz cause strong distortions and signal loss in the GRE image, whereas 2DMSI clearly depicts the area around the metal. In the phantom, the 2DMSI difference images yield a clear signal change at the focal spot next to the metallic stem and a noticeable change in signal at the focal spot next to the metallic plate in the tissue sample, despite a 70% reduction in SNR due to the lower water content. Averaging over the frames during sonication improves the localization of the focal spot.

Conclusions

We have presented initial results for our proposed imaging strategy for MRgFUS in the presence of metallic hardware.

MAVRIC-SL is an established technique for artefact-reduced imaging near metal. Here, we have shown that the image contrast can be adjusted to yield the relevant information for both planning and assessment of MRgFUS treatments. The reduced image resolution compared to the standard FSE protocol (to keep the scan duration at an acceptable length) did not noticeably reduce diagnostic image quality for the treating radiologist.

We have also demonstrated that 2DMSI enables the measurement of temperature-induced signal changes in close proximity to metallic hardware and thus in regions where conventional PRF shift thermometry fails. The bin-selective approach allows for a temporal resolution of 8 s/frame, which is sufficiently high to resolve the temperature evolution in sonications as short as 20 s, which is the typical duration in MRgFUS treatments. In case of less severe metal-induced artefacts, the number of frequency bins could be reduced to increase the temporal resolution. Averaging over time frames could facilitate the detection of the focal spot in lower SNR cases. For the given temporal resolution and the low temperature sensitivity of the T2 relaxation time of aqueous tissues, the measured signal change is expected to be highly dominated by the temperature sensitivity of the PD, whereas a higher temporal resolution is expected to increase the T1 weighting. While the latter provides higher temperature sensitivity and thus lowers the SNR requirements, PD weighting offers the benefit of a tissue-independent temperature mechanism that could facilitate quantitative thermometry.

In conclusion, the proposed imaging strategy has the potential to enable MRgFUS treatments near metallic hardware. The patient population at greater risk for cancers, and hence osseous metastases, overlaps with the elderly demographic more likely to have metallic hardware such as joint replacements. Further, orthopaedic bone stabilization is often used as a treatment for osseous metastases at risk for fracture. Overcoming these technical limitations is therefore important to allow the use of MRgFUS in a larger patient population.

Fig. 32 (abstract O35).

a Components of the proposed imaging strategy for MRgFUS near metallic hardware. b GRE and EPI images underlying conventional PRF thermometry and a MAVRIC-SL image, all acquired in a patient with a femur stabilized with a metallic rod. The dashed line in the x-ray image depicts the location of the MRI slices. c Comparison of both planning images and difference images (pre and post contrast injection) for treatment assessment, acquired in a patient without metallic hardware

Fig. 33 (abstract O35).

GRE and 2DMSI image and 2DMSI signal change during sonication in the phantom (a) and the ex vivo tissue sample (b). The 2DMSI difference images are cropped and masked to the dashed area. For the tissue sample, the 2DMSI difference image in the lower right shows the average over the three time frames during sonication

O36 Thermal monitoring of HIFU using thermal memory effect of phase-change nano droplet

Jun Tanaka, Kentaro Kikuchi, Ayumu Ishijima, Takashi Azuma, Kosuke Minamihata, Satoshi Yamaguchi, Teruyuki Nagamune, Ichiro Sakuma, Shu Takagi

The University of Tokyo, Tokyo, Japan

Objectives

Phase change nano droplets (PCND), whose diameters are 200–400 nm, are droplets of perfluorocarbon (PFC) covered with phospholipid layers. Since they can be vaporized by ultrasound and transformed to microbubbles, they will be utilized as ultrasound contrast agents and ultrasound therapy sensitizers. There are several types of PCND which are often used in research, whose internal compositions are different, such as perfluoropentane (PFP) and perfluorohexane (PFH).Its boiling point can be adjusted by changing the mixture ratio of PFP and PFH.

The lifetimes of generated microbubbles changed from PCND are different from type to type. In this study, the lifetime dependence to ambient temperature at the moment of its vaporization was investigated. If remaining efficiency depends on the ambient temperature, this drug has a potential to be used as indicator of thermal memory effects and temperature monitoring agents.

Methods

Experimental setup for ultrasound exposure and high-speed imaging are shown in Fig. 34.In this study, we used the two types of PCND, whose internal compositions were PFP and the mixture (MIX; PFP: PFH = 1:1), respectively. Their main differences are boiling points (PFP: 29°C, MIX: 40°C), and the boiling point of MIX is estimated by thermodynamic calculation.

For the observation of the vaporization of PCND, we used two-layer structure of polyacrylamide gel phantom (layer with PCND and layer without PCND), and set the layer with PCND at the focal point of the transducer. PCND were vaporized by ultrasound (5 MHz centre frequency, 5 cycle bursts, Peak Negative Pressure = 3.5 MPa), which were irradiated with an arbitrary ultrasound beam controller (Verasonics) and a linear array transducer (EUP-L73S, Hitachi Aloka Medical). Time-lapse behaviours of PCND through phase changes were recorded with the high-speed camera (HPV-1A, SHIMADZU), coupled with inverted microscope (NIKON Eclipse Ti-U). The ambient temperature (gel phantom temperature) conditions were controlled with a hot water bath and thermo plate (TOKAI HIT) on the stage of the microscope.

Results

First, we observed the vaporization of the two types of PCND at 37 °C. The high-speed images of vaporization at 37 °C are shown in Fig. 35.

As to MIX (B.P. = 40 °C), generated microbubbles disappeared soon (within 10 μs) after the ultrasound exposure. On the other hand, as to PFP (B.P.=29 °C ), generated microbubbles remained for a while (more than thirty seconds). Because the main difference of these two types of PCND is the boiling point, we assumed that temperature is the key factor, and tried to control the difference between the boiling point and the ambient temperature. Then, we did some vaporization experiment at various temperatures. Some of the high-speed images of vaporization at 26 °C, 48 °C are shown in Fig. 36.

As to PFP (B.P. = 29 °C), generated microbubbles disappeared soon after the sonication at 26 °C, although they remained for some time at 37 °C. As to MIX (B.P.=40 °C), generated microbubbles had long lifetimes at 48 °C, although they had very short lifetimes at 37 °C. It can be considered that the lifetimes of generated microbubbles are greatly affected by not only the internal composition, but also the ambient temperature. Besides, behaviours after vaporization at 26~48 °C are shown in Fig. 37. Whether generated microbubbles will remain or disappear was switched around the boiling point.

Conclusions

We found that the lifetime of microbubble highly depended on the ambient temperature at the moment of vaporization. This effect has a potential to be used as indicator of thermal memory effects and temperature monitoring agents.

Acknowledgements

Authors thanks to Dr. Kawabata and Mrs. Asami in Hitachi Central Research Laboratory.

Fig. 34 (abstract O36).

Experimental setup

Fig. 35 (abstract O36).

High Speed Images of Vaporization at 37 °C

Fig. 36 (abstract O36).

High speed images of vaporization at 26, 48 °C

Fig. 37 (abstract O36).

Behavior after vaporization at 26~48°C