Characterization and Ex Vivo evaluation of an extracorporeal high‐intensity focused ultrasound (HIFU) system

Abstract Background High‐intensity focused ultrasound (HIFU) has been in clinical use for a variety of solid tumors and cancers. Accurate and reliable calibration is in a great need for clinical applications. An extracorporeal clinical HIFU system applied for the investigational device exemption (IDE) to the Food and Drug Administration (FDA) so that evaluation of its characteristics, performance, and safety was required. Methods The acoustic pressure and power output was characterized by a fiber optic probe and a radiation force balance, respectively, with the electrical power up to 2000 W. An in situ acoustic energy was established as the clinical protocol at the electrical power up to 500 W. Temperature elevation inside the tissue sample was measured by a thermocouple array. Generated lesion volume at different in situ acoustic energies and pathological examination of the lesions was evaluated ex vivo. Results Acoustic pressure mapping showed the insignificant presence of side/grating lobes and pre‐ or post‐focal peaks (≤−12 dB). Although distorted acoustic pressure waveform was found in the free field, the nonlinearity was reduced significantly after the beam propagating through tissue samples (i.e., the second harmonic of −11.8 dB at 500 W). Temperature elevation was <10°C at a distance of 10 mm away from a 20‐mm target, which suggests the well‐controlled HIFU energy deposition and no damage to the surrounding tissue. An acoustic energy in the range of 750–1250 J resulted in discrete lesions with an interval space of 5 mm between the treatment spots. Histology confirmed that the lesions represented a region of permanently damaged cells by heat fixation, without causing cell lysis by either cavitation or boiling. Conclusions Our characterization and ex vivo evaluation protocol met the IDE requirement. The in‐situ acoustic energy model will be used in clinical trials to deliver almost consistent energy to the various targets.


| INTRODUCTION
Ultrasound is an effective, low-cost, and non-ionizing diagnostic modality used for several decades in clinics. Recent years have seen a dramatic interest in its application as a surgical and therapeutic tool, especially high-intensity focused ultrasound (HIFU) for tissue ablation in the treatment of cancers and solid tumors. 1 High-intensity focused ultrasound energy can be focused inside the human body to raise the local temperature above 65°C within seconds. 2 Since the mid-1990s, advancements in clinical instrumentation have brought HIFU from the lab to the medical mainstream. In China and Europe, more than 100 000 patients have been involved for treatment of uterine fibroids, and cancers of the prostate, liver, kidney, breast, pancreas, brain, and bone. [2][3][4] In comparison to traditional cancer treatment methods (i.e., open surgery, radiotherapy, or chemotherapy) and other physical methods for tissue ablation (i.e., lasers, microwave, or radiofrequency ablation), HIFU has advantages of being a non-invasive and local treatment, deep penetration, better selectiveness without damaging adjacent vital structures, easier power control, and non-ionizing radiation. 1 International Electrotechnical Commission (IEC) has published several standards on ultrasound diagnostic, continuous-wave or pulsed Doppler, physiotherapy system, and pressure pulse lithotripter. [5][6][7][8] Medical ultrasound fields are typically characterized in water by measuring the spatial and temporal distribution of pressure using a piezoelectric (PVDF membrane or needle) hydrophone and the acoustic power (up to 30 W, i.e., ultrasound power meter from Ohmic Instruments) using an air-backed metal cone that intercepts the entire field. The acoustic intensity is derived from the measured pressure waveform assuming that the local pressure and particle velocity are in-phase ("plane-wave assumption"). However, the existing techniques may not be appropriate for characterizing HIFU fields due to the strong nonlinearity and potential cavitation damage to the hydrophone. 9 A fiber optic probe hydrophone (FOPH) was used in measuring the acoustic field with high intensity and focusing gain. [10][11][12] The FOPH has a small sensing element (100 μm), broad bandwidth (50 MHz or higher after de-convolution), and a large half aperture angle of 30°and a new fiber tip can be easily prepared and self-calibrated if cavitation damage occurs. In addition, a radiation force balance with the acoustic absorbing target placed between the source and the focus was designed for calibrating the HIFU transducers with no damage, and a minimal temperature rise was found at the electrical power up to 230 W. 13 Until now, there is no consensus and standard on calibration and methods for describing the HIFU field [i.e., American Institute of Ultrasound in Medicine (AIUM), FDA, National Institute of Standards and Technology (NIST) guideline], 9 except a national one in China. 14 Therefore, accurate and reliable HIFU calibration protocols are in a great need for both product development and clinical application.
One of the most important aspects of oncology therapy is delivering an appropriate dose, which is dependent on the type and stage of cancer, radiation method, whether administered before or after surgery, and the degree of surgery success. 15 However, the definition of dose varies with the treatment modality. In chemotherapy, the dose is determined by the patient's body surface area and severity of the disease; radiotherapy dose is usually measured in gray (Gy), exposure per unit mass of a medium to be treated; a temporal temperature relationship is used as the thermal dose in RFA. In HIFU literature, authors commonly reported the acoustic output (intensity and power) and exposure parameters in describing the HIFU field and exposure energy. 1 Although the term of treatment "dose" was sometimes used, its definition is not consistent with various targets (i.e., tissue type and propagation distance through biological tissue), the HIFU transducers (i.e., geometries and working frequency), and operation parameters (i.e., burst duration, duty cycle, and the total exposure time). The thermal dose of a 240-min exposure at 43°C is required to create thermally irreversible damage in most tissue types. 16 However, accurate in situ measurement of temperature profile or calculation of thermal dose is still challenging due to the high heating rate of HIFU and low temporal resolution of thermometry. 9 Therefore, a definition of dose in HIFU therapy, the amount of acoustic energy deposited in the tissue similar to the "absorbed dose" used in x-irradiation, is vital for guidance in HIFU application. An extracorporeal HIFU system applied for the investigational device exemption (IDE, #100169) to the Food and Drug Administration (FDA) for clinical trials. The characterization of the produced acoustic field, the safety of HIFU ablation (i.e., temperature elevation to the surrounding tissue away from the target), estimation of the acoustic energy delivered to the target, and evaluation of ex vivo performance were required. In this study, the acoustic field and power were measured by a fiber optic probe hydrophone and a radiation force balance system, respectively, with the electrical power up to 2000 W. A method of delivering an in-situ acoustic energy was established. Subsequently, the performance of the extracorporeal HIFU system was evaluated ex vivo. The relationship between the size and volume of HIFU-generated lesions with the acoustic energy was investigated. In addition, the temperature elevation in the focal region was measured by a thermocouple array.

2.A | HIFU system
A clinical extracorporeal HIFU system (FEP-BY02, Beijing Yuande Biomedical Engineering Inc., China) was used in this study. It has two identical HIFU transducers (upper and lower one for the treatment at the prone or the supine position, respectively), each consisting of 251 individual lead zirconate titanate (PZT) elements (frequency of~1 MHz and diameter of 16 mm) driven all in phase and positioned on a spherical surface. The HIFU transducer has an outer diameter of 37 cm, an inner diameter of 12 cm, and a radius of curvature of 25.5 cm. An ultrasound imaging probe (S3, GE, Seongnam, Korea) aligned coaxially with the HIFU transducer was connected to an ultrasound imaging system (Logiq 5, GE) to identify the region of interest (ROI). The operator specified the location, size, and shape of the treatment target, electrical power to the HIFU transducer, HIFU on and off time, the number of pulses per treatment spot, and the interval spacing between treatment spots in the control software. The ROI was treated in a raster scanning pathway.
Before each treatment, the water was degassed with an oxygen concentration of <4 mg/L. The fresh bovine livers were obtained from a local slaughterhouse (Schenk Packing, Stanwood, WA), immediately immersed in phosphate-buffered saline (PBS) solution, and chilled on ice. Within 2 h they were cut into the size of~20 × 20 × 6 cm (L × W × H) with the capsule left intact as the acoustic wave entry site, and attention was paid to exclude major vessels for consistent lesion production. After being degassed, the liver sample was positioned about 10 mm proximally to FOPH which was aligned to the HIFU focus in the free field. Then the focus of the HIFU beam after propagating through tissue was found using the acoustic field mapping method mentioned above. FOPH was realigned for the acoustic pressure waveform measurement with the presence of tissue samples. to absorb the ultrasound energy. The absorber in a diameter of 250 mm was suspended from a load cell (SML-100, Interface, Scottsdale, AZ) whose signal was digitized by a data acquisition (DAQ) board (SCB-68, National Instruments) at a sampling frequency of 1000 Hz. The sensitivity of the radiation force balance was calibrated using a weight set (VWR, West Chester, PA) each time before measurement. Then the position and air being trapped on the surface of the acoustic absorber were examined in the sonography. The discrepancy of the stabilized responses between HIFU on (3 s) and off (5 s) stage was used to calculate the radiation force, F, and the subsequent acoustic power output, P A , 14

2.C | Acoustic power measurement
where N is the number of PZT elements, ϑ i is the angle between the acoustic beam axis of the i-th element and the main axis of the HIFU array, α is the absorption of the propagation media, L is the distance between the center of the transducer and the absorbing target (usually L ≤0:7 Á D.), D is the focal length, corr is the planar wave correction factor of a piston element, 14 where k is the wave number, a is the activating sensor size of a single piston, J n (*) is the n-order Bessel function. 14 The electrical power to the HIFU transducer was determined by where V rms is the RMS voltage to the HIFU transducer measured by a high-voltage probe (PPE-2kV, LeCroy), G is the conductance of the HIFU transducer measured to be 17.56 ms using an impedance analyzer (4192A, Hewlett-Packard, Palo Alto, CA). Therefore, the electrical-to-acoustic energy conversion ratio was calculated as follows:

2.D | Ex vivo lesion production
The attenuation of tissue was determined using an "acoustic caliper." A pair of transducers was mounted on a digital caliper for precise measurement of the transmission path length. A chirped pulse containing frequencies ranging from 1 to 10 MHz was transmitted first through the tissue and then through a water path (reference signal); the received tissue and reference signals were acquired and compared in order to calculate sound speed and attenuation. 18 Measurements were repeated three times for each tissue sample. The electrical power achieving a desired and reproducible in-situ acoustic energy, E A , was calculated to be where α T is the whole attenuation through the propagation path in dB, T 1 is the HIFU pulse duration in ms, n P is the number of pulses per treatment spot. 19 The electrical power was set no higher than 500 W according to the clinical experience for the high safety.
Sections of the bovine liver in the size of 4.5 × 4.5 × 6.0 cm (L × W × H) were prepared using the method described above and then transferred to a tissue-mimicking phantom that contains 6.5% Alginate impression material (Jeltrate, Dentsply International, York, PA), taking care not to reintroduce gas. 20 Then the center of the sample was aligned to the HIFU focus under the guidance of sonography.
The relationship of HIFU-generated lesion size and in-situ acoustic energy within the range of 500-2000 J was investigated with the interval distance between the treatment spots of 7 mm, which results in a single lesion production according to preliminary observation. Furthermore, lesion interaction effects were studied by reducing the interval distance to 3-5 mm.
Immediately following HIFU ablation, the liver samples were placed in a custom-built tissue cassette with proper maintenance of orientation and no disruption on the tissue. The samples were then frozen at −10°C overnight, ensuring sufficient stiffness to facilitate slicing while minimizing the amount of tissue expansion. 21 Afterward, the samples were cut to record the lesion photographically in a step size of~1 mm until no lesions were observed for at least 3 consecutive slices. The image files were then processed by a Matlab (MathWorks, Natick, MA) program. Each lesion area was fitted with an ellipse so that the area, centroid coordinates, major axis, and minor axis could be determined, from which the lesion volume was calculated as follows: where N is the number of image files, a i is the cross-sectional area of the lesion in the i-th image, z i is the incremental distance of the ith image taken in the tissue sample. x-z plane, and (d) axial direction. Capsule hydrophone and fiber optic probe hydrophone were used in the field mapping at low and high power output, respectively. White arrow and arrowhead show the first side lobe and grating lobe in the focal plane, respectively.

2.F | Temperature measurement
The temperature elevation and distribution in the focal region were measured by an array of 6 type T thermocouples in a diameter of 0.3 mm (Omega Engineering, Stamford, CT). 5 thermocouples (#1-#5) were aligned in a line with 10 mm away from each other, and the sixth thermocouple was 5 mm distal to the third one that was aligned to the HIFU focus under the guidance of sonography as required by FDA for the IDE of this system (Fig. 2). To ensure a straight insertion into the tissue, the thermocouple was mounted  data were acquired through the SCB-68 DAQ board concurrently with HIFU treatment. The corresponding thermal dose was calculated using the measured temperatures, T(t), with R = 0.25 if T(t) < 43°C and 0.5 otherwise. 23 A 240-min exposure at 43°C (240 CEM) could create thermally irreversible damage in most tissue types. [23][24][25] After the ablation, the tissue was sliced into approximately 3 mm-thick pieces to identify the presence of lesions.

2.G | Statistics
At each experimental condition, at least six data were collected. Sig-maPlot (Systat, San Jose, CA) was used to calculate the average and standard deviation and determine the data regression.

3.A | Acoustic field mapping
The acoustic fields of the HIFU transducer in the focal (x-y) and axial (x-z) plane were measured by the capsule hydrophone at the low power (Fig. 3). It is demonstrated that HIFU energy concentrated in

| 351
The generation of harmonics is found to occur usually in the main lobe [ Fig. 5(a)].

3.C | Acoustic power
There was a linear relationship (R 2 = 0.99) between the electrical and acoustic power for both the upper and lower transducers with the electrical-to-acoustic energy conversion efficiencies of 42.5% and 47.8%, respectively (Fig. 6). Electrical power to the HIFU transducer was stable with a variation of only 4%. In comparison, the variation in the corresponding acoustic power was a little higher. At the low-power level (electrical power <100 W) the variation in measured acoustic power was up to 45%, which is due to the low signal-to-noise ratio of the load cell for small signals. However, with the increase of electrical power, the variation decreased (~10% and 5% for the upper and lower transducer, respectively). Although the upper and lower HIFU transducers were manufactured identically, the upper HIFU transducer has a little lower electric-to-acoustic conversion ratio and higher variation of acoustic power output, which is due to the use of a silicon rubber cushion as the degassed water reservoir instead of a large water cavity as used in the lower transducer. Although the silicon rubber is thin (~0.4 mm) and no significant effect on the pressure waveforms at the focal point was found, the vibration of the silicon cushion with a response to the HIFU pulse-generated radiation force may dissipate some acoustic energy and introduce additional noise.

3.D | Single lesion dosimetry
Dosimetry studies in ex vivo bovine liver yielded consistent results in 3-month experiments (n > 15) despite some variations due to unavoidable inhomogeneities of the tissue, such as the existence of small vessels. Viability (NADH-d) staining revealed distinct lesions that corresponded well to areas of discoloration/blanching in the tissue (Fig. 7). However, a blue-black ring was frequently observed, which is due to an extracellular deposition and an artifact of the staining by examination with higher magnification. Furthermore, lesions generated with low (750 J) and high (1000 J) acoustic energies were evaluated histologically (Fig. 8).

3.E | Lesion interaction
A single layer of lesions with different spacing between spots and in situ acoustic energies was generated (Fig. 10). Because of the thermal diffusion effect, the ambient temperature increased with the ongoing of HIFU treatment. As a result, the lesion became progressively larger. 20 When the lesion size was larger than the interval spacing between nearby spots, the lesion coalescence may occur. Although single lesions were produced in the initial stage with interval spacing of 3 mm (lesion size at 750 and 1000 J were 1.65 AE 0.19 mm and 2.3 AE 0.23 mm, respectively, in Fig. 9(b), a contiguous plane of ablated tissue was found at the end of ablation. With

3.F | Temperature elevation
Since the major mechanism of HIFU-induced cell damage is heat fixation, temperature measurement by the thermocouple array would provide direct evidence of the treatment safety. Representative lesions produced by the extracorporeal HIFU system with in situ acoustic energy of 1000 J for each spot are shown in Fig. 11. The initial temperatures were 33-35°C and the maximum temperatures determined from the positions where the temporal average pressure falls to 6 dB below the peak value in the focal plane. In shock wave lithotripter, the peak positive pressure at the focus varies from 40 to 120 MPa, depending on the output voltage, focusing gain, and the methods of shock wave generation. As a result, at the edge of beam width the pressure could be up to 60 MPa, which is much higher than the stone fragmentation threshold (~10 MPa). Therefore, the coagulation necrosis (i.e., pyknotic nuclei, nuclear disruption, and disappearance). Along the margin of the ablation, a narrow cellular band of fibrous tissue could be identified, with the presence of fibroblasts, inflammatory cells, collagen fibrin, and capillary network. Therefore, NADH-diaphorase stain is more accurate and objective than H&E staining in assessing acute cell death because it is based on the presence or absence of enzyme function instead of changes in the cellular structure. Thermal fixation dominates cell damage in HIFU ablation, which is based on the lack of histologic evidence of cavitation, the linear changes of temperature with power, and no sharp elevation of measured temperatures. In this study and future clinical trials, the electrical power output is only up to 500 W. Heat fixation is assumed as the dominant mechanism so that in-situ acoustic energy was developed for delivering almost the same absorbed acoustic energy to different patients. Although there are variations of acoustic attenuation in the same type of tissue, such protocol will allow us a more reasonable comparison of HIFU outcome rather than using only the operation parameters (i.e., power and effective exposure time).

| CONCLUSION
A clinical extracorporeal HIFU system was characterized using our protocols for the acoustic field (i.e., acoustic pressure waveform,