Focused ultrasound robotic system for very small bore magnetic resonance imaging

Abstract Background A magnetic resonance imaging (MRI) compatible robotic system for focused ultrasound was developed for small animal like mice or rats that fits into a 9.4 T MRI scanner (Bruker Biospec 9420, Bruker Biospin, Ettlingen, Germany). The robotic system includes two computer‐controlled linear stages. Materials and Methods The robotic system was evaluated in a mouse‐shaped, real‐size agar‐based mimicking material, which has similar acoustical properties as soft tissues. The agar content was 6% weight per volume (w/v), 4% w/v silica while the rest was degassed water. The transducer used has a diameter of 4 cm, operates with 2.6 MHz and focuses energy at 5 cm. Results The MRI compatibility of the robotic system was evaluated in a 9.4 T small animal scanner. The efficacy of the ultrasonic transducer was evaluated in the mimicking material using temperature measurements. Conclusions The proposed robotic system can be utilized in a 9.4 T small animal MRI scanner. The proposed system is functional, compact and simple thus providing a useful tool for preclinical research in mice and rats.


| INTRODUCTION
The technology of magnetic resonance guided focused ultrasound (MRgFUS) is directed recently mostly for brain applications after the major success of treating essential tremor non-invasively. 1 Another promising application for brain is the treatment of glioblastoma using MRgFUS. 2 Because of these two major and promising applications, a lot of efforts have been directed towards research regarding brain. In the last decade, a lot of emphasis was given in opening the blood brain barrier (BBB) in order to push therapeutic drugs in targets in the brain. Recently, MRgFUS was explored in treating Alzheimer by opening the BBB. 3 Experimental studies for BBB disruption using focused ultrasound started in 2001 by Hynynen et al. 4 by sonicating the brain of 18 rabbits. The most popular animal for experimentation is the mouse [5][6][7][8][9] because it provides useful data with minimal cost and effort.
Therefore, there is significant demand for exploring new concepts in the area of MRgFUS. The exploration of new concepts requires robotic systems for mice. One of the first positioning devices was reported by the team of Chopra 10 that developed an magnetic resonance imaging (MRI)-compatible three-axis robotic FUS system for small animals.
The company MEDSONIC developed several robotic systems for animal use. One such system was developed for use in the rabbit brain. 11,12 Other systems were dedicated for preclinical use for prostate, 13,14 gynaecological applications, 15 and small animals. 16 The French company Image Guided Therapy (Pessac, France) 17 developed also an MRgFUS system for small animal experiments using phased arrays, a rather complicated and expensive solution for small animals.
The main goal in this article was to develop an MRgFUS robotic system for mice that operates in the environment of a 9.4 T scanner.
The proposed system is reduced in size in order to fit in a 9.4 T preclinical system. With the use of a standard resonators coil, the bore diameter of MRI is reduced to 7 cm. The MRI-compatibility of the materials used (transducer, motors, plastics, encoders and screws) was evaluated extensively in other studies reported by this group [11][12][13][14][15] and therefore such results are not presented in this article. The proposed robotic system can accommodate an ultrasonic transducer with 4 cm diameter. Based on the size of the robot and size of mice, the maximum radius of curvature that can be accommodated is 5 cm. Due to the fact that mice are small in size, there is no need for stage in the X axis of the MRI. Therefore two linear axes are needed (Z and Y in MRI planes). The advantage of using the proposed system over the phased array (electronic steering of the focus) is that the focus is steered mechanically making the system less complex and thus inexpensive. Since mice are small, it is possible to achieve transcranial ablation with a single element transducer.

| Robotic system
The robotic system includes two computer-controlled axes (Z and Y in MRI). Figure 1 shows the drawing of the linear actuator for motion along the MRI Y axis. All the parts of the robotic system were 3D and manufactured with a 3D printer (FDM400, Stratasys) using ABS material. The Z-axis design was identical to the design of the Y axis. Figure 2 shows the front part of the robotic system (the working envelope of the transducer is highlighted in yellow). In this part, the small animal is placed in prone position (natural position). It is possible to place the animal in supine position by using a specially designed holder. The volume of the working envelope was calculated by the outer diameter of the transducer (22 mm), the motion range of 2.8 � 3 cm in Y-and Z-axis respectively and the thickness of the transducer (10 mm). The robotic system weights around 2.2 Kg. The flow rate of the degassed water was 10 mL/s. The circulating water was kept at room temperature. Figure 3 shows the interior of the robotic system highlighting the two linear axes, the transducer arm, ultrasonic transducer and water space. The transducer arm goes through a plastic bellow which isolates the water space from the mechanical part of the two axes.
Transducer was immersed in a container which was filled with degassed water.
Compared to previous small animal positioning devices developed by our group, 16 the main difference was the usability at 9.4 T and the size of the motor which is almost 50% smaller. With these smaller size motors, we were able to develop more compact robotic systems.
The extended arm of the robotic system is cylindrical with an external diameter of 70 mm and can be placed inside the standard 70 mm resonator of the 9.4 T MRI scanner. The Robotic system has a maximum height of 7 cm, a length of 105 cm and a width of 15 cm.
The cylindrical extension that connects the positioning device with the water enclosure, has a diameter of 7 cm and it is approximately 80 cm in length. Figure 4A shows the placement of the device inside the 9.4 T MRI scanner. Figure 4B shows a closer view of the water enclosure's drawing with the mouse model placed on the acoustic opening. Figure 5A shows the drawing of the robotic system, and Figure 5B shows the photo of the developed robotic system.

| FUS system
The   After every sonication, the heat conduction increases the temperature of the nearby tissue. A delay between each step can be added in the software to avoid the near field heating effect. Six different algorithms were developed to produce non-linear motion patterns to avoid sonications between adjacent targets to reduce the delay between steps, thus reducing the treatment time.

| Electronic system
The electronic system includes a DC supply (24 V, 2 A) which drives the piezoelectric Shinsei motors. Cables connect the motor drivers

| Agar-based mimicking material
The functionality of the transducer and the MRI compatibility of the robot was evaluated in an agar-based mimicking material. The mixture of the agar-based mimicking material consisted of 6% weight per volume (w/v) agar, 4% w/v silica dioxide and the rest were degassed water. This agar-based mimicking material was already used successfully in other studies. 19,20 This phantom was used to measure the temperature produced by the system's transducer. Additionally, the phantom was used during the MR compatibility of the robotic system as the imaging load. The shape and size of the phantom is very close to real mice. Specifically, the phantom has a length of 78 mm, width of 28 mm and height 26 mm. This was achieved by producing with 3D printer, a specially designed mould of a real mouse. The volume required to fill the mould was approximately 27 ml. The attenuation of the phantom was found to be equal to 1.1 dB/cm-MHz. 20

| Magnetic resonance imaging
The robotic system was tested in a 9.4 T MR system with PET insert  Figure 7A shows the difference of intended step (blue line) to actual distance measured (red line) on the Z axis. The blue line represents the ideal expected motion and the red line represents what was actually achieved. Therefore, the difference between the two represents the motion error. Figure 7B shows the same graph for the Y axis. Motion steps of 1, 5 and 10 mm were evaluated. The average error measured for 20 movements of 1 mm was 0.02 mm, for 5 mm was 0.0375 mm and for 10 mm was 0.024 mm. The step of 1 mm is typically used in high-frequency sonications and represents the minimum step that is normally used. The 5 mm step is used typically with low-frequency sonications or with high-power sonications. The 10 mm step is typically used to place the robot in home position.

| RESULTS
The ability of the transducer to create high temperature was tested in the laboratory setting using the agar-based phantom. Figure 8  increases the rate of temperature decreases due to conduction. When the ultrasound is turn OFF, then the temperature drops rapidly. Figure 9 shows the placement of the mimicking mouse in the front part of the robotic system. The robotic system is then inserted in the coil of the 9.4 T MRI scanner. Water was poured on the membrane film in order to provide good coupling between the robot and mouse-mimicking material. Figure 10A shows the coronal image in the 9.4 T MRI scanner highlighting the transducer, water enclosure and mouse-mimicking material. Figure 10B shows the axial image in the 9.4 T MRI scanner highlighting the transducer, water enclosure and mouse-mimicking material. Note that no major artefacts were caused due to the presence of the ultrasonic transducer. Some artefacts that appeared were due to the transducer cable and are located outside of the clinical MRI area of interest.

| DISCUSSION
An MRgFUS positioning device was developed that can be used with mice and rats in a commercial 9.4 T small animal MRI scanner with a bore diameter of 7 cm. Due to the size of the animal only two axes are needed (called Z and Y in this article in compliance with the MRI axes). The robotic system can be navigated in two axes with a good spatial accuracy. This range of motion is more than enough for the size of mice or rats. The MRI compatibility of the system was tested successfully in a 9.4 T MRI system, using high-resolution imaging. In another article, we will present extensive MR thermometry of ultrasonic ablation. In this robotic system, the animal is placed in supine position, but with a special holder it can be placed also in prone position. The main innovation of this device is that its size is small compared to the available devices. Currently, the only device to perform this 17 uses phased arrays which moves the beam electronically. The proposed robotic system uses a single element transducer, which makes the system simple and affordable and yet as effective as the other available systems. In our opinion for preclinical efforts in mice, the use of a robotic system with two axes and the use of a single element transducer can result into a functional system. In this preliminary design, a mouse-mimicking material was used based on water and agar. In a follow up study, we will use real mice or rats to acquire data for this robotic system. With these small animals, it will be possible to explore fully an attractive application (e.g., BBB using HIFU).

ACNOWLEDGEMENTS
The project has been funded by the Research Promotion Foundation of Cyprus under the project FUSROBOT (ENTERPRISES/0618/0016).