Abstract
In an ultrasound image-guided High Intensity Focused Ultrasound (HIFU) surgery, reflected HIFU waves received by an imaging transducer should be suppressed for real-time simultaneous imaging and therapy. In this paper, we investigate the feasibility of pulse compression scheme combined with notch filtering in order to minimize these HIFU interference signals. A chirp signal modulated by the Dolph-Chebyshev window with 3–9 MHz frequency sweep range is used for B-mode imaging and 4 MHz continuous wave is used for HIFU. The second order infinite impulse response notch filters are employed to suppress reflected HIFU waves whose center frequencies are 4 MHz and 8 MHz. The prototype integrated HIFU/imaging transducer that composed of three rectangular elements with a spherically con-focused aperture was fabricated. The center element has the ability to transmit and receive 6 MHz imaging signals and two outer elements are only used for transmitting 4 MHz continuous HIFU wave. When the chirp signal and 4 MHz HIFU wave are simultaneously transmitted to the target, the reflected chirp signals mixed with 4 MHz and 8 MHz HIFU waves are detected by the imaging transducer. After the application of notch filtering with pulse compression process, HIFU interference waves in this mixed signal are significantly reduced while maintaining original imaging signal. In the single scanline test using a strong reflector, the amplitude of the reflected HIFU wave is reduced to −45 dB. In vitro test, with a sliced porcine muscle shows that the speckle pattern of the restored B-mode image is close to that of the original image. These preliminary results demonstrate the potential for the pulse compression scheme with notch filtering to achieve real-time ultrasound image-guided HIFU surgery.
Keywords: Noninvasive HIFU surgery, Simultaneous therapy and imaging, Integrated HIFU/imaging transducer, Chirp signal, Notch filter
1. Introduction
In recent years, High Intensity Focused Ultrasound (HIFU) has become one of the potential surgical techniques for the treatment of benign or malignant tissues. It can generate necrotic lesions by increasing a temperature on the target to more than 70°C during a short time [1]. During HIFU surgery, not only precise targeting without damage to healthy tissues, but also monitoring of treatment response in real-time is important for a physician. Therefore, image-guided HIFU systems using Magnetic Resonance Imaging (MRI) and ultrasound imaging have become popular ways to achieve simultaneous imaging and therapy. Among them, the ultrasound image-guided HIFU (US-gHIFU) has been shown to be a promising tool for the treatment of localized malignant prostate tissues for more than a decade [2,3]. Several ultrasound imaging schemes, such as B-mode, elastography, Acoustic Radiation Force Impulse (ARFI), shear wave, and localized harmonic motion, have been applied to US-gHIFU [4–9].
In order to realize real-time therapy and imaging, a HIFU transducer should be combined with an ultrasound imaging transducer. However, in this configuration, the quality of an ultrasound image is severely corrupted by the reflected HIFU wave when HIFU and imaging transducers are simultaneously activated [10–12]. Several techniques, including the implementation of large frequency difference between HIFU and imaging transducers and interleaving the HIFU data with the imaging data have been developed. These were developed as an attempt to reduce these HIFU noises received by an imaging transducer [13–15].
Along the same vein, we previously proposed the pulse compression of binary encoded pulses with notch filtering on reception [16]. By employing this technique, the HIFU transducer could be run in not only the pulsed wave (PW) mode but also in the continuous wave (CW) mode during imaging without a mode switching circuit. On the other hand, the combination of the short pulse excitation and the fixed notch filtering increased the sidelobe level due to remnant ripples in the backward direction of the mainlobe [16]. The performance of the proposed method was previously demonstrated by Field-II simulation [17] and simple experiments using three commercial imaging transducers. The basic limitation of previous investigation was that the amplitude of HIFU reflection in all simulations and experiments was numerically calculated. This was because an integrated HIFU/imaging transducer which was purposed [16] was not available at that time.
The present work extends the performance demonstration of the pulse compression with notch filtering to the frequency modulated chirp signal through in vitro test. This is done by using a prototype integrated HIFU/imaging transducer. It has been well known that a chirp signal which is one of the coded excitation techniques can provide high Signal to Noise Ratio (SNR) and deep penetration depth. It can be implemented by relatively simple hardware. Additionally, after pulse compression, the range mainlobe width and the range sidelobe level of the chirp signal can be easily controlled by using window functions. A chirp signal modulated by the Dolph–Chebyshev window is employed as an imaging signal [18] and a CW was used for HIFU.
The general methodology of the proposed scheme is briefly described in Section 2. Computer simulations and A (Amplitude)-mode signal/B (Brightness)-mode image experiments using a prototype integrated HIFU/imaging transducer are carried out. All specifications for the signal processing and transducer configuration are chosen with an aim to treat malignant prostate tissues.
2. Methods
2.1. Principle of simultaneous therapy and imaging
Fig. 1a shows a schematic diagram of the proposed method. The integrated HIFU/imaging transducer is composed of 6 MHz imaging transducer and 4 MHz HIFU transducers. The acoustic stacks of the HIFU and imaging transducer are different, in order to maximize the performance of each transducer. The imaging transducer has a backing and a matching layer to obtain a broad bandwidth. The HIFU transducer has no backing layer in order to maximize the transmitted energy. No matching layer is employed to protect the surface of the HIFU transducer by the minimization of heat absorption during high temperature operation.
Fig. 1.
(a) Schematic diagram shows the procedure of the proposed method using the integrated HIFU/imaging transducer and system. (b) HIFU and imaging data sequence for simultaneous therapy and imaging based on the proposed scheme.
A continuous HIFU wave is chosen in this study to achieve higher heating effect compared to a pulsed HIFU wave. When a continuous HIFU beam and a frequency modulated chirp signal are simultaneously transmitted to the target, the imaging transducer receives the reflected chirp signal contaminated by the reflected strong HIFU wave. After the application of the notch filtering and pulse compression, the final output signal containing significantly reduced HIFU interference can be obtained. Fig. 1b shows HIFU and imaging data sequence for simultaneous therapy and imaging. Note that, during imaging HIFU transducer can transmit a continuous HIFU beam.
2.2. Frequency modulated chirp signal
The chirp signal whose frequency linearly changed with time is given below [19]:
| (1) |
where ωc is the angular center frequency of the chirp signal, Δw is the angular bandwidth, and T0 is a time period. Typically, the range mainlobe width and the range sidelobe level of the chirp signal after matched/mismatched filtering depend on the usage of windowing functions. The performances of windows such as Blackman, Hamming, Hanning, Kaiser, and Dolph-Chebyshev are summarized well by Harris [20]. Among them, the Dolph–Chebyshev windowing function is chosen in this study as its ripple ratio and range mainlobe width can be easily controlled. A −60 dB range sidelobe level is chosen for the Dolph–Chebyshev window. We also consider the relationship between the range mainlobe width and the range sidelobe level of the output signal.
The typical frequency range for HIFU treatment of malignant prostate tissues is 3–4 MHz considering the target location (4–5 cm from the rectal wall), the required intensity level (Ispta > 1000 W/cm2), −6 dB focal spot size, and the aperture size of the transducer [21]. The center frequency of the imaging transducer in the current approach is chosen to be between the fundamental and the second harmonic HIFU signals in order to maximize the available bandwidth for an optimized image. After notch filtering, not only the HIFU wave but also the image signal at the notch frequency can be minimized. Thus, the notch frequencies must be far from the imaging center frequency in order to prevent the loss of meaningful image data.
In this research, 4 MHz and 6 MHz center frequencies are used for HIFU and imaging, respectively. The imaging transducer has a maximum available bandwidth of 67% between the 4 MHz fundamental and 8 MHz harmonic HIFU signals. This bandwidth becomes a criterion for designing the transducer and selecting the input waveform, i.e., frequency sweep range. In our prototype transducer, the −6 dB bandwidth is 54% and the chirp signal with 3–9 MHz frequency sweep range is used for imaging. Note that this frequency sweep range of the chirp signal can be increased, depending on the transducer bandwidth. Although the bandwidth of the prototype transducer is 13% lesser than the theoretically maximum value, it can still show the feasibility of the proposed scheme.
2.3. Notch filtering
A notch filter widely used for communication system and audio processing can remove specific frequency components while it minimally affects other frequency components [22]. In the US-gHI-FU, notch filtering can reduce HIFU interference with specific frequency from the imaging signal as the frequencies of the reflected HIFU waves are known. In general, the frequency range of pulse echo data containing meaningful information may be broad enough to cover from very low frequency to high frequency. In this study, low frequency limitation is lower than 4 MHz and high frequency limitation is higher than 8 MHz. Thus, it is desirable that the reflected HIFU wave with 4 MHz and 8 MHz center frequencies should be only removed without causing distortion to other frequency components in the imaging signal.
As the notch attenuation value affects the shape of the restored signal, proper values must be chosen. Shallow notch attenuation can maintain the original shape of the mainlobe but it generates high remnant ripples after the mainlobe. Deep notch attenuation may significantly suppress interference signal but it also distorts final output signal [23]. Pulse compression of a chirp signal may yield greater robustness in suppressing remnant ripples compared to a normal short pulse signal. Thus, considering the amplitude of HIFU interference a shallow notch attenuation value is selected to maintain the shape of the pre-HIFU signal as closely as possible.
Fig. 2 shows the frequency responses of the 2nd order infinite impulse response notch filters with attenuation frequencies at 4 MHz and 8 MHz. The notch attenuation values are −37 dB and −31 dB at 4 MHz and 8 MHz, respectively. The 6 dB attenuation difference between the two notch filter responses may be compensated by the amplitude difference between the fundamental and harmonic components of the HIFU waves. The third harmonic signal is not considered in this study as it may not affect the imaging quality due to limited transducer bandwidth. It can be easily removed by using a broad bandpass filter.
Fig. 2.
Simulated frequency responses of two notch filters with 4 MHz and 8 MHz notch attenuation.
3. Simulations
3.1. Simulation procedure
A numerical simulation for the proposed method was performed with Matlab (The MathWorks Inc., Natick, MA). The chirp signal modulated by the Dolph–Chebyshev window with 20 µs time duration and 3–9 MHz frequency sweep range provided approximately −73 dB of the range sidelobe level after a pulse compression. The 4th order Butterworth bandpass filter was designed to model the realistic transducer transfer function for imaging as shown in Fig. 3. The simulated −6 dB bandwidth of the imaging transducer was 54% and this was similar to that of the prototype transducer. Peak amplitude of 4 MHz HIFU wave was assumed to be identical to that of 6 MHz imaging signal. Note that the amplitude of the HIFU wave can be controlled by using notch filtering with adjusted notch attenuation value, i.e., the higher amplitude of HIFU wave can be reduced by notch filtering with a deeper notch attenuation value. Additionally, the amplitude of 8 MHz HIFU wave was 6 dB lower than that of 4 MHz signal because typically a harmonic signal has lower amplitude than the fundamental frequency component [24].
Fig. 3.
Simulated impulse response of the imaging transducer.
3.2. Simulation results
Fig. 4a and c shows the time and frequency domain responses of the chirp signal without the HIFU wave. The 4 MHz and 8 MHz HIFU wave were added onto the imaging signal as shown in Fig. 4b and d. After pulse compression without notch filtering, this imaging signal had −21 dB sidelobe level (Fig. 4f) while the pure chirp signal had −73 dB (Fig. 4e). After pulse compression with filtering using the designed notch filters in Fig. 2, the sidelobe level was reduced by −73 dB (Fig. 4h). This was almost identical to the output of notch filtering for the pure chirp signal (Fig. 4g).
Fig. 4.
Simulated time and frequency responses for the windowed chirp signal. Time responses of the chirp signal: (a) without and (b) with HIFU wave. Frequency responses of the chirp signal: (c) without and (d) with HIFU wave. Envelope signals of the chirp signal before notch filtering: (e) without and (f) with HIFU wave. Envelope signals of the chirp signal after notch filtering: (g) without and (h) with HIFU wave. Note that all responses were obtained after convolving with the impulse response in Fig. 3.
4. Experiments
4.1. Transducer design and fabrication
In this study, we designed and fabricated a spherically focused HIFU/imaging transducer. Its dimension was 14.4 mm × 28 mm as shown in Fig. 5. Briefly, the 6 MHz imaging transducer was built with 1–3 composite of 69% volume fraction ratio. As a piezoelectric material, PZT-5H (TFT L-145N, TFT Corporation, Japan) was selected and a regular epoxy (EPO-TEK301, EPOXY Technology, Billerica, MA) was used for a matching layer. A conductive silver epoxy (E-SOLDER 3022, Von Roll Isola, Inc., New Haven CT) with an impedance of 5.92 MRayl was used as a backing layer. We controlled the driving conditions for the imaging transducer in order to satisfy the Food and Drug Administration (FDA) guideline [25], i.e., the mechanical index (MI) of the imaging transducer was 0.6, and the spatial peak temporal average intensity (ISPTA) was 104 mW/cm2 at 4 cm focal depth.
Fig. 5.
Photograph of the prototype integrated HIFU/imaging transducer.
In the case of HIFU transducer, PZT4 (840, APC International Ltd., Mackeyville, PA) was chosen as it has a high mechanical Q(500), low dielectric/mechanical loss which results in a lower internal loss, and high Curie temperature (328°C) [26]. Each element had 1–3 composite structure with 79% volume fractional ratio for aperture conformation and a reduced impedance mismatch between a transducer and a medium. As kerf filler, unloaded epoxy (EPO-TEK314, EPOXY Technology, Billerica, MA) with high glass transition temperature was used. No backing layer was used to maximize the forward traveling acoustic energy and no matching layer was used to minimize the heat absorption during high temperature operation. When the driving voltage was 141 Vpp, ISPTA measured in water was equal to 1170 W/cm2 at 4 cm focal depth and HIFU transducer efficiency was approximately 53%. This was relatively low but it was sufficient enough to generate necrotic region (Fig. 7c).
Fig. 7.
Schematic diagram of (a) experimental setup, (b) measured maximum temperature, and (c) photograph of the slice of porcine muscle with an ablated region by using the prototype transducer.
4.2. Experimental results
4.2.1. Transducer performance test
The measured center frequencies of HIFU and imaging transducers were 4.3 MHz and 6.1 MHz, respectively. Additionally, their measured −6 dB bandwidths were 16% and 54% as shown in Fig. 6. The performance of the prototype transducer was compared to the KLM (Krimholtz, Leedom and Matthaei) model [27] simulation result.
Fig. 6.
KLM modeling and experimental pulse echo responses of (a) HIFU and (b) imaging elements in the integrated HIFU/imaging transducer.
As shown in Fig. 7a, the therapeutic performance of the prototype transducer was tested on a sliced porcine muscle with 5 cm × 5 cm × 3 cm obtained from a butcher. The goal of this test was to verify whether the prototype transducer can generate enough HIFU signal or not. The target was mounted in the degassed/ deionized water at 27°C room temperature. Note that the degassing process for the target did not affect the experimental results in this study. A thermocouple (TMTSS-020G-6, OMEGA Engineering Inc., Stamford, CT) was placed inside the target to monitor the temperature. A HIFU transducer was driven by a 55 dB power amplifier (A300, ENI Co., Santa Clara, CA) connected to a function generator (33250A, Agilent, Santa Clara, CA). When the spherically focused beam was transmitted to the target, the temperature on the target increased to more than 80°C, after approximately 10 s (Fig. 7b). The total ISPTA of two HIFU elements was 1170 W/cm2 and the total acoustic power (TAP) was approximately 20 W. Thus, this intensity could generate a coagulated region in the porcine muscle specimen as shown in Fig. 7c. Note that this power was also used for the following A-mode signal and the B-mode imaging experiments.
In order to implement simultaneous therapy and imaging using the integrated HIFU/imaging transducer, an array imaging transducer is desirable for imaging without the movement of a transducer. However, the coded array imaging system capable of transmitting and receiving multiple coded signals were not yet available. So, we used a mechanical scanning system by fabricating a single element imaging transducer combined with two single element HIFU transducers. Under this experimental setup, it was difficult to observe the formation of the lesion due to the linear movement of the HIFU transducer. This results in a low HIFU sonication energy at a fixed target. However, its power was sufficient enough to generate a strong HIFU noise. As the goal of this study was to eliminate the reflected strong HIFU waves, we focused on the generation of high-amplitude HIFU wave that contains the fundamental and harmonics by using a prototype integrated HIFU/ imaging transducer.
4.2.2. A-mode signal test
Fig. 8 shows the experimental setup to acquire A-mode signals. Two identical function generators (33250A, Agilent, Santa Clara, CA) capable of producing continuous HIFU waves and windowed chirp signals were connected to two different RF power amplifiers (325LA and A300, ENI Co., Santa Clara, CA) in order to boost their amplitude. Subsequently, they were used to excite the transducers. The echo signal was amplified and recorded by using a receiver (5900PR, Panametrics Inc., Waltham, MA) and a digital oscilloscope (LC534, LeCroy, Chestnut Ridge, NY) with 8-bit Analog to Digital Converter (ADC) card. As a target, an aluminum plate was immersed in a degassed/deionized water tank and a thin rubber was placed beneath the aluminum plate to minimize the coming back of the reflected signals from the bottom of the water tank. Note that 4 MHz CW and chirp signals modulated by the Dolph–Chebyshev window with 3–9 MHz frequency sweep range were used for HIFU and imaging, respectively. Additionally, two notch filters whose specifications were identical to Fig. 2 were used to minimize the effect of HIFU interference.
Fig. 8.
Experimental setup for the measurement of A-mode signal.
When the HIFU and imaging transducers were simultaneously activated, high-amplitude 4 MHz and 8 MHz signals were detected by the 6 MHz imaging transducer as shown in Fig. 9a. Before notch filtering, the pulse compressed the chirp signal displayed −27 dB sidelobe level as shown in Fig. 9b (dashed line). After notch filtering, the sidelobe level was decreased to lower than −45 dB as shown in Fig. 9b (solid line). Note that the second largest peak at approximately 20 µs was the second reflected signal that came from the bottom of the target. The measured −6 dB mainlobe beamwidths of two curves were very close to each other, i.e., 0.52 µs (before notch filtering) and 0.58 µs (after notch filtering).
Fig. 9.
Frequency domain response (a) and envelope signal (b) of the chirp signal after pulse compression before (dashed line) and after (solid line) notch filtering.
Table 1 shows the performance comparison of the different notch filters. As the notch attenuation value was decreased, the sidelobe level slightly increased while maintaining the mainlobe width. In the case of deep notch attenuation value, sidelobe level improved but the mainlobe range increased. The effect of different chirp excitations was investigated as described in Table 2. The frequency sweep range was fixed. We consider the fixed −6 dB band- width of the prototype transducer. The different windows were employed and found that the Dolph–Chebyshev and Hanning windows can provide both reasonable mainlobe width and sidelobe level compared to other windows. When the Dolph-Chebyshev window was used and the frequency sweep range was only decreased, the range mainlobe width increased but the range sidelobe level reduced as shown in Table 3. This is due to the frequency sweep range of the chirp signal which determines the axial resolution [18].
Table 1.
Performance comparison of the different notch filters.
| Notch frequency (MHz) |
Notch attenuation value (dB) |
Mainlobe range (µs) |
Sidelobe level (dB) |
|
|---|---|---|---|---|
| Case I | 4 | −48 | 0.60 | −46 |
| 8 | −42 | |||
| Case II | 4 | −37 | 0.52 | −41 |
| 8 | −31 | |||
| Case III | 4 | −28 | 0.53 | −38 |
| 8 | −22 | |||
| Case IV | 4 | −22 | 0.53 | −35 |
| 8 | −16 |
Table 2.
Performance comparison of the different chirp excitations for various windows after notch filtering and pulse compression.
| Window | Frequency sweep range (MHz) |
Mainlobe range (µs) |
Sidelobe level (dB) |
|---|---|---|---|
| Blackman | 3–9 | 0.60 | −48 |
| Dolph–Chebyshev | 3–9 | 0.55 | −41 |
| Hanning | 3–9 | 0.54 | −40 |
| Kaiser | 3–9 | 0.60 | −38 |
Table 3.
Performance comparison of the different chirp excitations for different frequency sweep ranges after notch filtering and pulse compression.
| Window | Frequency sweep range (MHz) |
Mainlobe range (µs) |
Sidelobe level (dB) |
|---|---|---|---|
| Dolph–Chebyshev | 3–9 | 0.51 | −41 |
| 4–8 | 0.74 | −46 | |
| 4.5–7.5 | 0.98 | −49 |
4.2.3. B-mode imaging test
Fig. 10 shows the experimental setup for B-mode imaging. The integrated HIFU/imaging transducer was mounted on a linear mechanical positioner. The imaging and the HIFU elements were connected to the equipments in the same manner as shown in Fig. 8. All B-mode images in this experiment were obtained by using the pulse compression scheme with the chirp signal modulated by the Dolph-Chebyshev window. A fresh porcine muscle which was obtained from a butcher was sliced with 4 cm × 4 cm × 4 cm in dimensions and it was then, put in the degassed/deionized water.
Fig. 10.
Schematic diagram of mechanical scanning system to obtain B-mode image during HIFU sonication. The integrated HIFU/imaging transducer was linearly moved during the activation of the HIFU/imaging transducer.
Note that there was no coagulative necrosis in the target as HIFU and imaging elements of the integrated transducer were scanned together. This results in insufficient time for lesion formation during HIFU sonication. Fig. 11a shows the conventional B-mode image before HIFU sonication. When HIFU elements were activated, the image quality was severely interfered by high-amplitude HIFU noise as shown in Fig. 11b. The stripe pattern of the image came due to the un-uniform moving velocity of the linear motor during imaging. After using notch filters, the HIFU noise could be minimized as shown in Fig. 11c. Although there was a slightly different speckle pattern between Fig. 11a and c, the whole speckle pattern was similar to each other. Fig. 12 shows A-mode envelope signals obtained from the ROI (Region of Interest) in Fig. 11a and c. In this sample, the amplitude difference between the control signal (solid line) and the restored signal (dashed line) was lower than 0.1.
Fig. 11.
B-mode images of a sliced porcine muscle. All B-mode images were logarithmically compressed with a dynamic range of 40 dB. (a) Control image of the Dolph–Chebyshev-windowed chirp with pulse compression. (b) When the HIFU transducer was activated high-amplitude HIFU wave was observed after pulse compression. (c) After notch filtering with pulse compression, HIFU noise was significantly reduced. Note that the ROI in the (a) and (c) was chosen for the A-mode signal plots in the Fig. 12.
Fig. 12.
Comparison of two A-mode signals obtained from Fig. 11a and c.
5. Discussion and conclusion
In order to achieve real-time simultaneous HIFU therapy and ultrasonic imaging, the reflected HIFU wave received by an imaging transducer should be effectively minimized while maintaining ultrasonic image quality. Previously, we proposed the pulse compression of binary code combined with the fixed notch filtering. In the previous investigation, its feasibility was demonstrated using commercial imaging transducers. However, the amplitudes of HIFU reflections in all simulations and experiments were numerically calculated as an integrated HIFU/imaging transducer purposed [16] was not available at that time. In this article, the feasibility of pulse compressed scheme combined with notch filtering was extensively investigated. We fabricated spherically focused integrated HIFU/imaging transducer and used different pulse compression method using a frequency modulated signal. The fabrication process of the prototype transducer was similar to the previous research except for the aperture shape, i.e., the aperture of the previous one was cylindrical and that of the current transducer was spherical, in expecting a highly focused intensity. The frequency modulated chirp signal is one of the coded excitation schemes. It has a high compression ratio compared to the binary code. Thus, it may provide a high durability against HIFU interference [18,19].
In simulation, the chirp signal modulated by the Dolph–Chebyshev window shows sidelobe level of about −73 dB after pulse compression. When the HIFU interference was added onto this signal and the pulse compression process was only employed, the sidelobe level increased from −73 dB to −21 dB. After applying two notch filters with pulse compression, 4 MHz and 8 MHz HIFU interferences were minimized. Thus, the restored signal was almost identical to the pre-HIFU imaging signal. In A-mode signal test using the strong reflector, the range sidelobe level of the reflected HIFU wave could be lower than −45 dB after employing the proposed method. This effect was also shown in the B-mode imaging test, i.e., the speckle pattern of the restored image after the removal of HIFU interference was almost identical to that of the pre-HIFU B-mode image.
There may be several factors that affect the performance of the proposed method such as the frequency sweep range and notch attenuation value as described in Tables 3 and 1. The frequency sweep range of the chirp signal should be chosen by giving careful consideration to the transducer impulse response and notch frequencies. Within the limited transducer impulse response, if the chirp signal has too wide frequency sweep range then, the axial resolution will be improved. On the other hand, the frequency response of the mainlobe containing important information of the chirp signal will be severely distorted. This is due to the HIFU wave with fundamental and harmonic frequencies. On the contrary, if the frequency sweep range is too narrow, although most frequency information of the mainlobe will be protected from HIFU wave, the tradeoff is that the range mainlobe width will be increased and this results in a poor axial resolution (Table 3). In this study, the 3–9 MHz frequency sweep range was chosen and taking into consideration the 54% of bandwidth of the prototype transducer and the −6 dB axial resolution of the chirp signal, i.e. the −6 dB bandwidth of the chirp signal was approximately matched to the transducer bandwidth. If the transducer bandwidth is higher then, a chirp signal with broader frequency sweep range can be employed which results in a more improved axial resolution [18]. Table 2 shows that the performance of our scheme is slightly affected by the kinds of windows. The proper window can be chosen by considering the desired axial resolution and sidelobe level.
The notch attenuation is another important factor for the proposed method as described in Table 1. If the notch attenuation value is too shallow when compared to the amplitude of HIFU interference then, the remnant HIFU interference increases the range of the sidelobe level. If it is too deep then, the restored signal is severely distorted [23]. In this case, adaptive notch filtering scheme may be one of the alternative solutions for determining the proper notch attenuation. However, coded excitation has low correlation to CW compared to the short pulse excitation. For example, when the amplitude of HIFU wave was identical to that of the imaging signal, the range sidelobe level after pulse compression was approximately −20 dB (Fig. 4f). This implies that the notch attenuation values of the designed notch filters can be shallower than the HIFU interference within the 20 dB margin. This makes the fixed notch filter handle most of the HIFU noise instead of the adaptive filter scheme based on experimental results (Table 1). These shallow notch filters combined with pulse compression will allow the restoration of images as closely as possible to the pre-HIFU images. Note that the normal short pulse signal does not have this notch attenuation margin. In this case, if the notch attenuation is shallower than HIFU interference then, high amplitude ripples will appear in the envelope of the short pulse signal [23].
In this article, we experimentally evaluated the performance of the pulse compression combined with notch filtering in order to implement real-time ultrasonic imaging and HIFU surgery. By using a prototype integrated HIFU/imaging transducer and the chirp signal, the performance of the proposed method was extensively demonstrated. In this study, a mechanically moving system was employed to obtain a B-mode image during HIFU therapy as a coded array system was not available yet. Although this experimental system cannot realize real-time lesion formation during imaging, our preliminary experimental results show that the high amplitude HIFU interference mixed with imaging signals can be significantly reduced. Thus, pulse compression scheme combined with notch filtering may be useful for simultaneous B-mode ultrasonic imaging during HIFU treatment.
Acknowledgments
The authors would like to thank Dr. Jonathan Cannata and Mr. Jay Williams for their valuable advice and discussions. This work was supported by the Dongguk University Research Fund of 2011.
References
- 1.Azzouz H, Rosette JJMCH. HIFU: local treatment of prostate cancer. EAU-EBU Update Series. 2006;4:62–70. [Google Scholar]
- 2.Blana A, Walter B, Rogenhofer S, Wieland WF. High-intensity focused ultrasound for the treatment of localized prostate cancer 5-year experience. J. Urology. 2004;63:297–300. doi: 10.1016/j.urology.2003.09.020. [DOI] [PubMed] [Google Scholar]
- 3.Sanghvi NT, Foster RS, Bihrle R, Casey R, Uchida T, Phillips MH, Syrus J, Zaitsev AV, Marich KW, Fry FJ. Noninvasive surgery of prostate tissue by high intensity focused ultrasound an updated report. Eur. J. Ultrasound. 1999;9:119–129. doi: 10.1016/s0929-8266(99)00010-5. [DOI] [PubMed] [Google Scholar]
- 4.Chan AH, Fujimoto VY, Moor DE, Martin RW, Vaezy S. An image-guided high intensity focused ultrasound device for uterine fibroids treatment. Med. Phys. 2002;29:2611–2620. doi: 10.1118/1.1513990. [DOI] [PubMed] [Google Scholar]
- 5.Owen NR, Bailey MR, Hossack J, Crum LA. A method to synchronize highintensity focused ultrasound with an arbitrary ultrasound imager. IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 2006;53:645–650. doi: 10.1109/tuffc.2006.1610574. [DOI] [PubMed] [Google Scholar]
- 6.Curiel L, Souchon R, Rouvière O, Gelet A, Chapelon JY. Elastography for the follow-up of high-intensity focused ultrasound prostate cancer treatment initial comparison with MRI. Ultrasound Med. Biol. 2005;31:1461–1468. doi: 10.1016/j.ultrasmedbio.2005.06.013. [DOI] [PubMed] [Google Scholar]
- 7.Fahey BJ, Nightingale KR, Stutz DL, Trahey GE. Acoustic radiation force impulse imaging of thermally- chemically-induced lesions in soft tissues preliminary ex vivo results. Ultrasound Med. Biol. 2004;30:321–328. doi: 10.1016/j.ultrasmedbio.2003.11.012. [DOI] [PubMed] [Google Scholar]
- 8.Bercoff J, Pernot M, Tanter M, Fink M. Monitoring thermally-induced lesions with supersonic shear imaging. Ultrason. Imaging. 2004;26:71–84. doi: 10.1177/016173460402600201. [DOI] [PubMed] [Google Scholar]
- 9.Curiel L, Chopra R, Hynynen K. In vivo monitoring of focused ultrasound surgery using local harmonic motion. Ultrasound Med. Biol. 2009;35:65–78. doi: 10.1016/j.ultrasmedbio.2008.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Vaezy S, Shi X, Martin RW, Chi E, Nelson PI, Bailey MR, Crum LA. Realtime visualization of high-intensity focused ultrasound treatment using ultrasound imaging. Ultrasound Med. Biol. 2001;27:33–42. doi: 10.1016/s0301-5629(00)00279-9. [DOI] [PubMed] [Google Scholar]
- 11.Wu CC, Chen CN, Ho MC, Chen WS, Lee PH. Using the acoustic interference pattern to locate the focus of a high-intensity focused ultrasound (HIFU) transducer. Ultrasound in Med. Biol. vol. 2008;34:137–146. doi: 10.1016/j.ultrasmedbio.2007.07.001. [DOI] [PubMed] [Google Scholar]
- 12.Fleury G, Bouchoux G, Berriet R, Lafon C. Simultaneous imaging and therapeutic ultrasound. IEEE Ultrason. Symp. 2006:1045–1051. [Google Scholar]
- 13.Kluiwstra JUA, Tokano T, Davis J, Strickberger SA, Cain CA. Real time image guided high intensity focused ultrasound for myocardial ablation: in vivo study. IEEE Ultrason. Symp. 1997:1327–1330. [Google Scholar]
- 14.Pichardo S, Gelet A, Curiel L, Chesnais S, Chapelon JY. New integrated imaging high intensity focused ultrasound probe for transrectal prostate cancer treatment. Ultrasound Med. Biol. 2008;34:1105–1116. doi: 10.1016/j.ultrasmedbio.2007.12.005. [DOI] [PubMed] [Google Scholar]
- 15.Seip R, Tavakkoli J, Wunderlich A, Sanghvi NT, Dines KA, Crum LA. Real-time detection of multiple lesions during high intensity focused ultrasound (HIFU) treatments. in: 2nd Int Symp. on Therapeutic Ultrasound. 2002:168–175. [Google Scholar]
- 16.Jeong JS, Chang JH, Shung KK. Ultrasound transducer and system for realtime simultaneous therapy and diagnosis for noninvasive surgery of prostate tissue. IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 2009;9:1913–1922. doi: 10.1109/TUFFC.2009.1267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Jensen JA. Field: a program for simulating ultrasound systems. Med. Biol. Eng. Comput. 1996;34:351–353. [Google Scholar]
- 18.Misaridis T, Jensen JA. Use of modulated excitation signals in medical Part II ultrasound design and performance for medical imaging applications. IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 2005;52:177–191. doi: 10.1109/tuffc.2005.1406546. [DOI] [PubMed] [Google Scholar]
- 19.O’Donnell M. Coded excitation system for improving the penetration of realtime phased-array imaging systems. IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 1992;39:341–351. doi: 10.1109/58.143168. [DOI] [PubMed] [Google Scholar]
- 20.Harris FJ. On the use of windows for harmonic analysis with the discrete Fourier transform. Proc. IEEE. 1978;66:51–83. [Google Scholar]
- 21.ter Haar GR. Intervention and therapy. Ultrasound Med. Biol. 2000;26:S51–S54. doi: 10.1016/s0301-5629(00)00164-2. [DOI] [PubMed] [Google Scholar]
- 22.Carney R. Design of a digital notch filter with tracking requirements. IEEE Trans. Space Electron. Telem. 1963;9:109–114. [Google Scholar]
- 23.Jeong JS, Cannata JM, Shung KK. Adaptive HIFU noise cancellation for simultaneous therapy and imaging using an integrated HIFU/imaging transducer. Phys. Med. Biol. 2010;55:1889–1902. doi: 10.1088/0031-9155/55/7/007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Jensen JA, Fox PD, Wilhjelm JE, Taylor LK. Simulation of non-linear ultrasound fields. in: Proc. IEEE Ultrasonics Symp. 2002:1733–1736. [Google Scholar]
- 25.Barnett SB, ter Haar GR, Ziskin MC, Rott HD, Duck FA, Maeda K. International recommendations and guidelines for the safe use of diagnostic ultrasound in medicine. Ultrasound Med. Biol. 2000;26:355–366. doi: 10.1016/s0301-5629(00)00204-0. [DOI] [PubMed] [Google Scholar]
- 26.Zhang S, Xia R, Lebrun L, Anderson D, Shrout TR. Piezoelectric materials for high power high temperature applications. Mater. Lett. 2005;59:3471–3475. [Google Scholar]
- 27.Krimholtz R, Leedom DA, L Matthaei G. New equivalent circuits for elementary piezoelectric transducers. Electron. Lett. 1970;6:398–399. [Google Scholar]