You are on page 1of 33

Scanning microwave-induced thermoacoustic tomography: Signal, resolution, and contrast

Geng Ku and Lihong V. Wanga)


Optical Imaging Laboratory, Biomedical Engineering Program, Texas A&M University, 3120 TAMU, College Station, Texas 77843-3120

Received 12 July 2000; accepted for publication 25 October 2000 Scanning thermoacoustic tomography was explored in the microwave region of the electromagnetic spectrum. Short microwave pulses were used to induce acoustic waves by thermoelastic expansion in biological tissues. Cross sections of tissue samples were imaged by a linear scan of the samples while a focused ultrasonic transducer detected the time-resolved thermoacoustic signals. Based on the microwave-absorption properties of normal and cancerous breast tissues, the piezoelectric signals in response to the thermoacoustic contrast were investigated over a wide range of electromagnetic frequencies and depths of tumor locations. The axial resolution is related to the temporal prole of the microwave pulses and to the impulse response of the ultrasonic transducer. The lateral resolution is related to the numerical aperture of the ultrasonic transducer as well as to the frequency spectra of the piezoelectric signals in the time window corresponding to the axial resolution. Gain compensation, counteracting the microwave attenuation, was applied to enhance the image contrast. 2001 American Association of Physicists in Medicine. DOI: 10.1118/1.1333409 Key words: thermoacoustics, microwave, ultrasonics, tomography, photoacoustics

I. INTRODUCTION Purely microwave imaging of biological tissues has been investigated for a number of years.15 The advantages of microwave imaging include the use of nonionizing radiation and its relatively good imaging contrast. However, purely microwave imaging has had difculties in multichannel detection of microwaves without cross coupling, in reconstruction algorithms, and especially in achieving good spatial resolution because of the strong diffraction of microwaves determined by the long wavelength. Purely ultrasound imaging ultrasonography, an established medical imaging modality, can yield good spatial resolution but has poor contrast for early-stage tumors. Microwave-induced thermoacoustics can potentially bridge the gap and fuse the advantages of the two imaging modalities. In microwave-induced thermoacoustics, microwave pulses generate acoustic waves in electromagnetically lossy media. This phenomenon was known as the microwaveauditory or microwave-hearing effect in the early years.6 Microwave-induced thermoacoustics was used to quantify physical parameters in media such as the power density and the concentration of a given substance.79 Several investigators employed microwave-induced thermoacoustics in the 1980s for imaging of biological tissues.1015 However, these early works did not produce any tomographic or depthresolved images. Recently, images of biological tissues were computationally reconstructed based on microwave-induced thermoacoustics.16,17 This approach requires measurements of a large amount of data around the tissue and postprocessing computation. Our group explored scanning microwave-induced thermoacoustic tomography.18,19 Our approach is similar to the conventional ultrasonic B scan except that the ultrasound is produced internally inside the tissue by
4 Med. Phys. 28 1, January 2001

microwave pulses. This scanning approach can potentially provide real-time imaging and coregistration with ultrasonic B-scan images. Microwave-induced thermoacoustic imaging is based on the detection of the thermoacoustic signals generated by microwave radiation in the samples. Pulses of microwaves are used to irradiate the samples. Absorbed microwave energy causes thermoelastic expansion, which radiates acoustic waves, termed thermoacoustic waves. An ultrasonic transducer detects the time-resolved thermoacoustic signals. If optical radiation instead of microwave radiation is used, this thermoacoustic phenomenon is better known as photoacoustics. Microwave-induced thermoacoustic imaging shares similar principles with its optical counterpart.2024 However, microwave-induced thermoacoustic imaging may nd unique applications in medical imaging because microwave radiation provides a deeper penetration depth in biological tissues than does optical radiation and has different contrast mechanisms. We further investigated scanning microwave-induced thermoacoustic tomography toward biomedical applications. The piezoelectric signal, which is related to the thermoacoustic contrast, was simulated over a wide range of electromagnetic frequencies and depths of tumor locations based on the microwave-absorption properties of normal and cancerous breast tissues. The imaging resolutions were investigated in relation to the experimental parameters. Images of thick biological tissue samples were acquired experimentally, which clearly revealed both articially buried tissue objects and intrinsic structures as well. Gain compensation based on electromagnetic attenuation was applied to enhance the image contrast.
2001 Am. Assoc. Phys. Med. 4

0094-2405200128147$18.00

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

FIG. 1. Experimental setup for scanning thermoacoustic tomography.

II. METHODS The experimental setup for this study is shown in Fig. 1. A Cartesian coordinate system was set up for reference. The z axis was along the acoustic axis pointing upward. The x axis was perpendicular to the drawing plane and pointed outward. The y axis was in the drawing plane pointing to the right. A 3-GHz microwave generator transmitted microwave pulses toward the tissue samples. The pulse width was set to 0.5 s for most of the experiments unless it is stated otherwise. The energy of a single microwave pulse was estimated to be 5 mJ, causing a temperature rise of 1 mK in the tissue samples. A function generator DS345, Stanford Research System was employed to trigger the microwave generator, to control the pulse repetition rate of microwave radiation, and to synchronize the sampling by the oscilloscope. Microwave energy was delivered through two waveguides in tandem: a standard rectangular waveguide with a cross section of 72 mm34 mm and a tapered waveguide that narrowed the output cross section to 72 mm5 mm. The narrow output port helped concentrate the microwave energy into the imaging zone in the samples near the port; however, the imaging zone far from the port may receive less energy than it would otherwise from the original port as a result of increased diffraction. The object to be imaged was placed on a plastic holder, which was xed onto a two-dimensional x - y translation stage MD2, Arrick Robotics. A personal computer controlled the two stepper motors to drive the translation stage in the x and the y directions. An extension tube lled with acoustic-coupling water was mounted onto the ultrasonic transducer. The length of the extension tube was determined such that its focal zone was inside the region of imaging interest in the samples. The bottom surface of the extension tube was in contact with the samples for good acoustic coupling. The central frequency of the ultrasonic transducer V314, Panametrics was 1 MHz, the bandwidth was 0.6 MHz, the diameter was 1.9 cm, the focal length was 2.5 cm, and the focal diameter at 1 MHz was 0.2 cm. The piezoelectric output of the ultrasonic transducer was conMedical Physics, Vol. 28, No. 1, January 2001

nected to a pulse amplier. The amplied signal was averaged 10100 times by an oscilloscope TDS-640A, Tektronix and then transferred to the personal computer. Because the propagation speed of electromagnetic waves is much greater than the speed of sound, the microwave pulses stimulated the entire tissue sample essentially simultaneously. The induced thermoacoustic waves took different amounts of time to reach the ultrasonic transducer. The ultrasonic transducer measured the time of arrival of the thermoacoustic waves. The distance between the acoustic sources and the transducer was calculated by multiplying the time of arrival with the speed of sound in the medium, assuming the speed of sound is constant and known. Because the ultrasonic transducer was focused, it responded primarily to the acoustic sources along its focal column. Consequently, a time-domain signal can be converted into a onedimensional image along the acoustic axis z axis, which is similar to an ultrasonic A scan. Scanning the sample along the x or the y axis yielded a two-dimensional cross-sectional image of the sample. The narrow elongated output port of the tapered waveguide may be considered as a line source of electromagnetic radiation. Cylindrical wave propagation was assumed throughout the range of frequencies in our simulation of the thermoacoustic signals, although waveguides are rarely employed in practice for the low-frequency region of the considered electromagnetic spectrum. An electromagnetic wave was emitted from the tapered waveguide and attenuated in tissue approximately as I z I0 exp 2 z , 1

where I 0 is the intensity at the output port of the tapered waveguide, z is the distance from the output port to the point of observation along the vertical axis, I ( z ) is the intensity at z, and is the eld-absorption coefcient in the tissue samples.

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

The absorption coefcient can be expressed as

2 1

1 ,

where is the angular frequency, is the permeability, is the permittivity, and is the conductivity. In the frequency range of 0.110 GHz, the dielectric constant ratio of the permittivity in the material to that in vacuum has a value of 570 for soft tissues, and the conductivity has a value of 0.023 1 m 1 . 2527 The complex dielectric properties of tissues at various microwave frequencies determine the propagation and absorption distribution of microwaves. Consequently, the induced thermoacoustic pressure depends on the microwave intensity and the complex dielectric constant of the tissue samples. A model was used to estimate the microwave-induced thermoacoustic signals. A small breast tumor was embedded in normal breast tissue. The normal breast tissue predominantly determined the microwave attenuation. The thermoacoustic pressure p was proportional to the local absorbed microwave power: pI0 2 exp 2 z

The ultrasonic transducer responded to the components of the thermoacoustic pressure within its response bandwidth and rejected the components outside the bandwidth. The piezoelectric signal V p from the ultrasonic transducer was proportional to the thermoacoustic contrast: V p p t p n , 4

where p t and p n were, respectively, the thermoacoustic pressures in the tumor and the normal background tissue at depth z, where the tumor and the normal tissue interfaced. Based on Eqs. 3 and 4, we obtained V p I 0 2 t n exp 2 n z

FIG. 2. a Penetration depths of various biological tissues versus the microwave frequency; b simulated piezoelectric signal in response to a thermoacoustic wave from a tissue sample containing a buried tumor versus the microwave frequency.

where t and n represented the microwave-absorption coefcients of the tumor and the normal breast tissue, respectively.

III. RESULTS AND DISCUSSION Figure 2a shows the penetration depth of electromagnetic waves in various human tissues as a function of the electromagnetic frequency in the radio frequency rf region, where the penetration depth is the inverse of the absorption coefcient. The penetration depths of muscle and fat tissues are plotted in solid lines.28 At the frequency of our experimental setup, 3 GHz, the penetration depths for fat and muscle are 9 and 1.2 cm, respectively. The penetration depths of normal and malignant human breast tissues are plotted in dashed lines.29 At 3 GHz, the penetration depths for malignant and normal breast tissue are approximately 2.3
Medical Physics, Vol. 28, No. 1, January 2001

and 4.4 cm, respectively. Malignant breast tissues are more strongly absorbing than surrounding normal breast tissues.30,31 Water and ion concentrations are key factors in microwave absorption.32 Muscle and fat tissues have very high and very low water contents, respectively, and therefore represent the extreme microwave absorption properties. Most other soft tissues have an absorption coefcient in between those for muscle and fat tissues. This wide range of values among various tissues is desired to achieve a high image contrast. We calculated the piezoelectric signal as a function of the electromagnetic frequency and tumor location using Eq. 5 Fig. 2b. If the detection system is instrument-noise limited, the signal-to-noise ratio SNR of the system is different from this piezoelectric signal by a constant factor. The noise in our detection system was mainly from the pulse amplier, which was used to amplify the piezoelectric signal from the ultrasonic transducer. The noise remained almost constant in our experiment and was independent of the microwave frequency or the depth of the tumor. The SNR decreased as the tumor location increased because of the increasingly attenuated microwave intensity. When the tumor was located near

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

the tissue surface, e.g., at 1 cm depth, the SNR was better at higher frequencies. When the tumor was located more deeply, the choice of frequency for an optimal SNR was rather broad. This was because the decrease in thermoacoustic pressure was compensated by the increasing difference of the rf-absorption coefcients between the tumor and the normal breast tissues as the frequency increased. The axial resolution is limited by two factors: the temporal width of the microwave pulse and the temporal width of the impulse response of the ultrasonic transducer. The temporal width of the impulse response of the ultrasonic transducer is inversely proportional to the bandwidth of the ultrasonic transducer. To illustrate the effect of the microwavepulse width on the axial resolution, we simulated the thermoacoustic pressures and the corresponding piezoelectric signals from a microwave-absorbing slab of 5 mm in thickness using Eq. 3 and the convolution method that was described previously.19 The excitation microwave pulses had the same peak power but different pulse widths: 1.0, 0.5, and 0.1 s Fig. 3. There were two dipolar structures corresponding to the two boundaries of the slab. The width of the dipolar structures was determined by the width of the microwave pulses and the impulse response time of the ultrasonic transducer. The distance between the dipolar structures was determined by the acoustic-transit time across the slab. For the 5-mm thick slab, the acoustic-transit time was 3.3 s based on the speed of sound of 1.5 mm/s. Pumping with a narrower microwave pulse decreased the width of the dipolar structures and therefore improved the axial resolution because the dipolar structures dened the time window for the axial resolution. The narrower microwave pulses of the same peak pressure also produced smaller signals. Likewise, an ultrasonic transducer of a higher central frequency and a broader bandwidth produced narrower dipolar structures and therefore improved the axial resolution at the expense of signal strength. The lateral resolution at the focal plane was limited by the focal diameter of the ultrasonic transducer. Based on the ultrasound reciprocity, the focal diameter determined both the beam diameter when the ultrasonic transducer transmits ultrasound and the detection directivity factor when the ultrasonic transducer detects ultrasound. The focal diameter was determined approximately by

f a /NA c a / NA f a ,

6
FIG. 3. a Waveforms of the microwave pulses of various pulse widths that were tested in the experiment; b simulated thermoacoustic waves from a 5-mm-thick microwave-absorbing slab by the microwave pulses; c simulated piezoelectric signals of the thermoacoustic waves in b.

where a was the acoustic wavelength, NA was the numerical aperture of the ultrasonic transducer, c a was the speed of sound, and f a was the central frequency of the piezoelectric signal. The numerical aperture NA was solely determined by the geometric properties of the ultrasonic transducer. The speed of sound c a was relatively constant throughout the frequency range. The central frequency of the piezoelectric signal f a was determined by the frequency spectrum of the thermoacoustic signal in the dipolar structures and by the frequency response of the ultrasonic transducer. Therefore, the lateral resolution was not only related to the ultrasonic parameters, including the numerical aperture and the frequency response of the ultrasonic transducer, but also related
Medical Physics, Vol. 28, No. 1, January 2001

to the frequency content of the thermoacoustic signal in the dipolar structures. A high frequency was obviously desired for high-resolution imaging. The ultrasonic transducer functioned as a frequency lter to the thermoacoustic signals. The homogeneous tissue between interfaces produced slowly varying thermoacoustic signals. The slowly varying signals were outside the band-

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

FIG. 4. a Photograph of a y - z cross section of a lard sample containing a piece of muscle tissue, which was taken after the sample was imaged with the scanning thermoacoustic imaging system; b two-dimensional thermoacoustic image of the cross section; c typical temporal thermoacoustic signal corresponding to a vertical line in the two-dimensional thermoacoustic image.

width of the ultrasonic transducer and therefore were rejected. The variations in microwave absorption at tissue interfaces caused abrupt changes in the thermoacoustic pressure. The abruptly varying signals had frequency components falling into the response bandwidth of the ultrasonic transducer and made a primary contribution to the piezoelectric signals of the ultrasonic transducer. An ultrasonic transducer of a higher central frequency would provide better lateral resolution. Such a transducer usually has a broader bandwidth and would consequently yield better axial resolution as well. If the resolution were improved by varying the ultrasonic parameters, the SNR would be reduced because the volume of the thermoacoustic signal contributing to the piezoelectric signal is reduced. Therefore, there is a tradeoff between imaging resolution and SNR. Two samples were imaged with our microwave-induced thermoacoustic imaging system. The background of the rst sample was made intentionally as homogeneous as possible. A piece of muscle tissue was embedded into melted lard before it solidied, where the lard was used for its homogeneity. The sample was cut across to reveal the cross section after it was imaged with scanning thermoacoustic tomography Fig. 4. The time-resolved thermoacoustic signals were acquired at each step while the sample was scanned horizontally along the y axis with a step size of 1 mm. Each vertical line in this two-dimensional thermoacoustic image was obtained from a temporal thermoacoustic waveform. Figure 3c shows the temporal thermoacoustic signal for y equal to 23 mm as an example. The earliest arriving signal came from
Medical Physics, Vol. 28, No. 1, January 2001

the upper surface of the lard. The following signals corresponded to the two surfaces of the muscle and the bottom surface of the lard. Because the sample was in the near eld of microwave, a logical question was whether the heterogeneity of the electromagnetic eld would cause heterogeneity in thermoacoustic images. As shown in Fig. 4, the signal from the background lard is very weak, indicating the heterogeneity of the electromagnetic eld does not affect thermoacoustic imaging signicantly. The heterogeneity of the electromagnetic eld is of the scale of the wavelength, several centimeters in biological tissue at the 3-GHz frequency. Spatial variations of this scale correspond to low-frequency thermoacoustic signals and are therefore ltered out by the ultrasonic transducer. In the second sample the background was left with heterogeneity. A piece of swine muscle tissue of 5 mm in thickness was buried inside a piece of swine fat tissue. The fat tissue was naturally separated into several layers by thin 1 mm in thickness connective tissue that has greater microwave absorption than the adjacent fat tissue. A cross section of the sample was exposed and photographed after the sample was imaged with our scanning thermoacoustic imaging setup Fig. 5. Both the buried muscle and the connective tissue are clearly visible in the thermoacoustic image. Because of microwave attenuation, the deeper structures in the tissue samples received less microwave radiation and produced weaker thermoacoustic signals than did the shallower structures. Consequently, the deeper structures were

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

FIG. 5. a Photograph of a y - z cross section of a fattissue sample containing a piece of muscle tissue, which was taken after the sample was imaged with the scanning thermoacoustic imaging system; b two-dimensional thermoacoustic image of the cross section; c gain compensation to a typical piezoelectric signal; d twodimensional thermoacoustic image of the cross section after the gain compensation.

not imaged as clearly as the shallower ones. To enhance the image contrast, we compensated the piezoelectric signal with the following gain factor: g z z exp 2 z , 7

which is the inverse of microwave intensity attenuation. In Fig. 5c, the original piezoelectric signal for y equal to 20 mm was plotted with a solid line. The microwave-induced thermoacoustic signal decreased for increasing distance z. The inverse of the gain factor was shown in the gure with a dashed line for a comparison with the decay of the original piezoelectric signal. The deeper signals were compensated with greater gains. The compensated data, shown as a pointdashed line, had a nearly constant envelope throughout the imaged depth. The gain-compensated image Fig. 5d shows the deeper structures clearly as well. Some interference from the pulse amplier was also amplied by the gain compensation, shown as the artifacts in the image near the upper surface of the muscle tissue. This technique is similar to the time-gain compensation in conventional ultrasonography. Of course, the decay constant in Eq. 7 depends on the types of tissue being imaged, but it can be adjusted dynamically to optimize the image quality. It is worth noting that the tissue samples also function as a frequency lter to the thermoacoustic signals. Higherfrequency components experience greater attenuation than the lower-frequency ones. This attenuation is severe for the
Medical Physics, Vol. 28, No. 1, January 2001

high-frequency components that pass through a long path in the tissue samples. The preferential reduction in the highfrequency components would adversely affect the imaging resolution. We also assumed that the tissue samples were electromagnetically heterogeneous and acoustically homogeneous, and hence we revealed the electromagnetic contrast. However, if the tissue samples were also acoustically heterogeneous, additional measures would have to be taken to account for the effect of the acoustic heterogeneity. One straightforward method would be to co-register the thermoacoustic images with conventional ultrasonograms of the same imaging planes. Of course, the purpose of thermoacoustic tomography is to image objects that do not have acoustic contrast and cannot be imaged by conventional ultrasonography. IV. CONCLUSIONS We have investigated scanning microwave-induced thermoacoustic tomography both experimentally and theoretically. Our theoretical analysis of the SNR showed that for malignant tumors located at a 36 cm depth, the choice of the optimal microwave frequency was broad, ranging between 100 MHz and 3 GHz. The axial resolution was obtained by measuring the time-resolved microwave-induced thermoacoustic signals. Depth-resolved tomographic images were acquired directly without resorting to computational

10

G. Ku and L. V. Wang: Scanning microwave-induced thermoacoustic tomography

10

image reconstructioninverse problems. The key factors affecting the axial resolution include the pulse width of microwave radiation and the bandwidth of the ultrasonic transducer. The lateral resolution was achieved by the use of a focused ultrasonic transducer. The factors affecting the lateral resolution include the numerical aperture of the ultrasonic transducer and the central frequency of the piezoelectric signal in response to the thermoacoustic waves. The central frequency of the piezoelectric signal was determined by the frequency response of the ultrasonic transducer and the frequency spectrum of the microwave-induced thermoacoustic wave in the resolvable dipolar structures. Gain compensation can be used to counteract the microwave attenuation and improve the image contrast. ACKNOWLEDGMENTS We thank Dr. K. Chang for a helpful discussion. This project was sponsored in part by National Institutes of Health Grants No. R29 CA68562, No. R01 CA71980, and No. R21 CA83760, National Science Foudation Grant No. BES9734491, and Texas Higher Education Coordinating Board Grant No. ARP 000512-0123-1999.
a

Author to whom all correspondence should be addressed. Telephone: 979-847-9040; fax: 979-845-4450; electronic mail: LWang@tamu.edu URL: http://oilab.tamu.edu. 1 L. E. Larsen and J. H. Jacobi, in Medical Applications of Microwave Imaging IEEE, Piscataway, NJ, 1986. 2 J. C. Lin, Frequency optimization for microwave imaging of biological tissues, Proc. IEEE 73, 374375 1985. 3 S. Caorsi, A. Frattoni, G. L. Gragnani, E. Nortino, and M. Pastorino, Numerical algorithm for dielectric-permittivity microwave imaging of inhomogeneous biological bodies, Med. Biol. Eng. Comput. 29, NS3744 1991. 4 M. S. Hawley, A. Broquetas, L. Jofre, J. C. Bolomey, and G. Gaboriaud, Microwave imaging of tissue blood content changes, J. Biomed. Eng. 13, 197202 1991. 5 P. M. Meaney, K. D. Paulsen, and J. T. Chang, Near-eld microwave imaging of biologically-based materials using a monopole transceiver system, IEEE Trans. Microwave Theory Tech. 46, 3145 1998. 6 J. C. Lin, On microwave-induced hearing sensation, IEEE Trans. Microwave Theory Tech. MTT-25, 605613 1977. 7 F. Caspers and J. Conway, Measurement of power density in a lossy material by means of electromagnetically-induced acoustic signals for non-invasive determination of spatial thermal absorption in connection with pulsed hyperthermia, Proceedings of the 12th European Microwave Conference, 1982, pp. 565568. 8 J. L. Su and J. C. Lin, Thermoelastic signatures of tissue phantom absorption and thermal expansion, IEEE Trans. Biomed. Eng. 43, 178182 1987. 9 J. S. K. Wan, Microwaves and chemistry: the catalysis of an exciting marriage, Rev. Chem. Intermed. 19, 147158 1993. 10 T. Bowen, L. Nasoni, A. E. Pifer, and G. H. Sembrosk, Some experimental results on the thermoacoustic imaging of soft tissue-equivalent phantoms, Ultrason. Symp. Proc. 2, 823827 1981. 11 R. G. Olsen, Generation of acoustic images from the absorption of pulsed microwave energy, in Acoustic Imaging, edited by J. P. Powers Plenum, New York, 1982, pp. 5359. 12 R. G. Olsen and J. C. Lin, Acoustic imaging of a model of a human

hand using pulsed microwave irradiation, Bioelectromagnetics N.Y. 4, 397400 1983. 13 J. C. Lin and K. H. Chan, Microwave thermoelastic tissue imaging system design, IEEE Trans. Microwave Theory Tech. 32, 854860 1984. 14 R. L. Nasoni, G. A. Evanoff, Jr., P. G. Halverson, and T. Bowen, Thermoacoustic emission by deeply penetrating microwave radiation, Ultrason. Sym. Proc. 5, 633637 1984. 15 K. H. Chan and J. C. Lin, Microwave-induced thermoacoustic tissue imaging, Proceeding of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society New Orleans, Louisiana, November 47, 1988, pp. 445446 1988. 16 R. A. Kruger, D. R. Reinecke, and G. A. Kruger, Thermoacoustic computed tomographytechnical considerations, Med. Phys. 26, 18321837 1999. 17 R. A. Kruger, K. K. Kopecky, A. M. Aisen, D. R. Reinecke, G. A. Kruger, and W. L. Kiser, Jr., Thermoacoustic CT with radio waves: a medical imaging paradigm, Radiology 211, 275278 1999. 18 L. V. Wang, X. Zhao, H. Sun, and G. Ku, Microwave-induced acoustic imaging of biological tissues, Rev. Sci. Instrum. 70, 37443748 1999. 19 G. Ku and L. V. Wang, Scanning thermoacoustic tomography in biological tissue, Med. Phys. 27, 11951202 2000. 20 A. A. Oraevsky, S. L. Jacques, R. O. Esenaliev, and F. K. Tittel, Laserbased optoacoustic imaging in biological tissues, in Laser Tissue Interaction V, edited by S. L. Jacques, Proceedings of the Society of Photo-Instrument Engineering, 1994, Vol. 2134A, pp. 122128. 21 R. A. Kruger and P. Liu, Photoacoustic ultrasound: pulse production and detection of 0.5% Liposyn, Med. Phys. 21, 11791184 1994. 22 A. A. Oraevsky, R. Esenaliev, F. K. Tittel, M. R. Ostermeyer, L.-H. Wang, and S. L. Jacques, Laser opto-acoustic imaging of turbid media: determination of optical properties by comparison with diffusion theory and Monte Carlo simulation, Laser Tissue Interaction VII, 1996, Vol. 2681, pp. 277284. 23 C. G. A. Hoelen, F. F. M. Demul, R. Pongers, and A. Dekker, Threedimensional photoacoustic imaging of blood vessels in tissue, Opt. Lett. 23, 648650 1998. 24 A. A. Oraevsky, V. A. Andreev, A. A. Karabutov, and R. O. Esenaliev, Two-dimensional optoacoustic tomography: transducer array and image reconstruction algorithm, in Laser Tissue Interaction X: Photochemical, Photothermal, and Photomechanical, edited by S. L. Jacques, G. J. Mueller, A. Roggan, and D. H. Sliney, Proceedings of the Society of Photo-Instrument Engineering, 1999, Vol. 3601, pp. 256267. 25 C. Gabriel, S. Gabriel, and E. Corthout, The dielectric properties of biological tissues: I. literature survey, Phys. Med. Biol. 41, 22312249 1996. 26 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: II. measurements in the frequency range 10 Hz to 20 GHz, Phys. Med. Biol. 41, 22512269 1996. 27 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: III. parametric models for the dielectric spectrum of tissues, Phys. Med. Biol. 41, 22712293 1996. 28 C. C. Johnson and A. W. Guy, Nonionizing electromagnetic wave effects in biological and system, Proc. IEEE 60, 692718 1972. 29 S. Chaudhary, R. Mishra, A. Swarup, and J. Thomas, Dielectric properties of normal & malignant human breast tissues at radiowave and microwave frequencies, Indian J. Biochem. Biophys. 21, 7679 1984. 30 W. Joines, R. Jirtle, M. Rafal, and D. Schaeffer, Microwave power absorption differences between normal and malignant tissue, Int. J. Radiat. Oncol., Biol., Phys. 6, 681687 1980. 31 W. Joines, Y. Zhang, C. Li, and R. Jirtle, The measured electrical properties of normal and malignant human tissues from 50900 MHz, Med. Phys. 21, 547550 1994. 32 K. R. Foster and J. L. Schepps, Dielectric properties of tumor and normal tissues at radio through microwave frequencies, J. Microwave Power 16, 107119 1981.

Medical Physics, Vol. 28, No. 1, January 2001

Microwave-induced thermoacoustic tomography: Reconstruction by synthetic aperture


Dazi Feng, Yuan Xu, Geng Ku, and Lihong V. Wanga)
Optical Imaging Laboratory, Biomedical Engineering Program, Texas A&M University, 3120 TAMU, College Station, Texas 77843-3120

Received 14 May 2001; accepted for publication 20 September 2001 We have applied the synthetic-aperture method to linear-scanning microwave-induced thermoacoustic tomography in biological tissues. A nonfocused ultrasonic transducer was used to receive thermoacoustic signals, to which the delay-and-sum algorithm was applied for image reconstruction. We greatly improved the lateral resolution of images and acquired a clear view of the circular boundaries of buried cylindrical objects, which could not be obtained in conventional linearscanning microwave-induced thermoacoustic tomography based on focused transducers. Two microwave sources, which had frequencies of 9 and 3 GHz, respectively, were used in the experiments for comparison. The 3 GHz system had a much larger imaging depth but a lower signalnoise ratio than the 9 GHz system in near-surface imaging. 2001 American Association of Physicists in Medicine. DOI: 10.1118/1.1418015 Key words: thermoacoustics, tomography, synthetic aperture, microwave

I. INTRODUCTION Microwave-induced thermoacoustic tomography is emerging as a nonionizing imaging modality. When electromagnetic radiation is irradiated upon biological tissues, the resulting heat-related expansion of the tissues produces acoustic waves. From the acoustic signals, we can reconstruct the distribution of electromagnetic absorption in soft tissues. The large differences in electromagnetic absorption in various tissues, which are associated with their physiological and pathological status, provide signicant contrasts in imaging. For example, the absorption coefcient in cancerous breast tissue is two to ve times greater than that in normal breast tissue, due to the increment of water and sodium bounded within malignant cells.13 This large difference makes it promising to use microwave-induced thermoacoustic tomography to detect breast cancers. Microwave-induced thermoacoustic tomography combines the advantages of both pure ultrasound and pure microwave imaging. Traditional imaging technology with pure ultrasound ultrasonography offers satisfactory spatial resolution but poor soft-tissue contrast, while pure microwave imaging provides good imaging contrast but barren spatial resolution.4 7 Microwave-induced thermoacoustic tomography bridges the gap between them. By integrating ultrasound and microwave technology, microwave-induced thermoacoustic tomography has both satisfactory spatial resolution and good soft-tissue contrast. In conventional linear-scanning microwave-induced thermoacoustic tomography LMTT, a focused ultrasonic transducer is used to detect time-resolved acoustic signals. Since the focused transducer has a good response only along the transducer axis, each acoustic signal can be converted into a one-dimensional image. Linear scanning of the ultrasonic transducer yields multiple one-dimensional images, which can be combined to form a two-dimensional image of the
2427 Med. Phys. 28 12, December 2001

sample.8 10 In the two-dimensional images obtained with conventional LMTT, only boundaries that are nearly perpendicular to the axis of the ultrasonic transducer can be detected because most of the thermoacoustic waves travel in a small solid angle around the normals of boundaries; spherical or oblique boundaries of buried objects whose thermoacoustic waves have a large angle with the axis of the ultrasonic transducer cannot be imaged because the ultrasonic transducer receives little signal from these boundaries. To overcome this deciency of traditional LMTT, we have applied the synthetic-aperture method to LMTT. In this method, thermoacoustic signals were detected from multiple locations and the synthetic delay-and-sum algorithm was then used for the reconstruction of the images. The syntheticaperture method has been applied in PAT previously;1114 and weights were assigned to the signals according to the sensors directivity to improve the SNR at the expense of lateral resolution. In our experiment, the raw data were obtained by a 2.25 MHz nonfocused transducer instead of a focused one, as in traditional LMTT, and the syntheticaperture reconstruction method based on the delay-and-sum algorithm was applied to reconstruct the images. By applying the synthetic-aperture method, we have improved the lateral resolution of the images and enhanced our ability to image spherical or oblique boundaries in the samples. Images acquired from two microwave sources with different frequencies were compared; the 3 GHz system has a much larger imaging depth but a lower SNR than the 9 GHz system in near-surface imaging. II. METHODS
A. Reconstruction method

The image reconstruction method is illustrated in Fig. 1. For convenience, we converted both the signal-delay time
2001 Am. Assoc. Phys. Med. 2427

0094-24052001281224275$18.00

2428

Feng et al.: Microwave-induced thermoacoustic tomography

2428

FIG. 2. Experimental setup.

FIG. 1. Scanning and reconstruction method.

and the spatial distance into the number of pixels according to the speed of the acoustic wave traveling in the media. For the 50 MHz sampling rate, 1 s 50 pixels 1.5 mm. Because the transducer is nonfocused, it receives signals from a larger solid angle than does a focused counterpart. In the reconstruction, we evenly projected the signals to each point within the whole solid angle according to the time delay. The time delay corresponds to the distance from the transducer to the point to which we project the signal. This is the algorithm called delay and sum. In other words, the signal intensity of each point A ( i , j ) is the sum of the signals from the transducer at different positions delayed with the transit time from the transducer position to the point. So the signal intensity at any point, A ( i , j ), can be expressed as
M

tried applying weights to the signals according to the transducers directivity, as was done previously. It improved the SNR, but the lateral resolution deteriorated as a result. In our situation, we had already acquired a satisfactory SNR by averaging the thermoacoustic signals 100 to 200 times, indicating that the direct delay-and-sum algorithm works well in LMTT technology.
B. Experimental setup

A i, j

k0

B k,l ,

where B ( k , l ) is the signal intensity coming from the l th pixel point in the signals and from the k th transducer scanning position, M is the total number of steps that the transducer scanned, and l is the distance from the k th transducer scanning position to the point ( i , j ): l k i 2 j 2 . Let us consider a point ( i 1 , j 1 ) where there is a thermoacoustic source. During the data acquisition, all the detectors will receive signals from this particular point after time delays determined by the above equation. Conversely, in the reconstruction, all the detectors contribute signals to this particular point with the appropriate time delays. On the other hand, for a point ( i 2 , j 2 ) where there is no source, few detectors contribute signals to this point after time delays determined by the above equation. Consequently, the reconstructed intensity at point ( i 1 , j 1 ) will be much greater than that at point ( i 2 , j 2 ). In this reconstruction scheme, the detection can be articially focused onto any specied point ( i , j ), which is the basic concept of synthetic aperture. We attempted to add some corrections to the delay-andsum algorithm but found them unnecessary. For example, we
Medical Physics, Vol. 28, No. 12, December 2001

The experimental setup for this study is shown in Fig. 2. A Cartesian system was set up for reference. The z axis is along the ultrasonic axis pointing downward. The x axis is perpendicular to the drawing plane and pointed outward. The y axis is in the drawing plane and points to the right. In the experimental setup, microwave pulses of 9 or 3 GHz, with a width of 0.5 s were delivered into the samples. The sizes of the cross section of waveguides were 72 mm 34 mm in the 3 GHz system and 23 mm 10 mm in the 9 GHz system. A function generator DS345, Stanford Research System was employed to trigger the microwave pulses and synchronize the sampling of an oscilloscope. A linear translation stage MD2, Arrick Robotics, on which an ultrasonic transducer was mounted, was driven by a computer-controlled stepper motor. The transducer was scanned linearly. The nonfocused ultrasonic transducer V323, Panametrics mounted on the translation stage had a central frequency of 2.25 MHz and a 6 mm diam of an active element. A low-noise pulse preamplier 500 PR, Panametrics amplied the acoustic signals coming from the transducer. Then the amplied signals were collected and averaged by an oscilloscope TDS-640A, Tektronix and subsequently transferred to a personal computer. The acoustic waves from the sample were coupled to the ultrasonic transducer by mineral oil. III. RESULTS AND DISCUSSION In this section, we will present and discuss the images acquired from the 3 GHz microwave system and the 9 GHz microwave system, respectively. The rst two sets of images were acquired from the 9 GHz system. In the rst set of images, we imaged a semicylindrical fat sample with a

2429

Feng et al.: Microwave-induced thermoacoustic tomography

2429

FIG. 3. a Cross-sectional picture of the sample with two curved boundaries marked with 1 and 2; b raw data of thermoacoustic signals acquired with the 9 GHz system, where the signals from the two curved boundaries are marked with 1 and 2, respectively; c image after the reconstruction, where the images of the two curved boundaries are marked with 1 and 2, respectively.

simple structure to verify the synthetic-aperture method. In the second set of images, we imaged a small fat cylinder containing several detailed structures. The third set of images was acquired, using the 3 GHz system, from two muscle cylinders buried in an ellipse of pork fat. Comparing the last two sets of images demonstrates the effects of different object shapes and the different frequencies of the microwave sources.
A. Results

FIG. 4. a Cross-sectional picture of the sample, a pork fat cylinder with one layer of connective tissue 1 and discrete layers of muscle; b raw data of thermoacoustic signals acquired with the 9 GHz system; c Image after the reconstruction, where the images of the connective tissue layer is marked with 1.

The sample in the rst set of images is a semicylinder of pork fat immersed in mineral oil. Figure 3a shows a cross section of the semicylinder with two curved boundaries marked by 1 and 2. The transducer is on the left side of the picture, pointing to the sample and moving along the ruler. The original signals from the transducer are shown in Fig. 3b. Because the thermoacoustic waves were propagating almost perpendicularly to the boundaries, the lateral signals 1 and 2 in Fig. 3b were from the corresponding curved boundaries 1 and 2 in Fig. 3a. After reconstruction, the original signals 1 and 2 formed the curved boundaries 1 and 2 of the semicylinder in Fig. 3c. The synthetic-aperture method was proved to be effective in imaging the curved boundaries. The at boundary of the semiMedical Physics, Vol. 28, No. 12, December 2001

cylinder is not visible in the image because this boundary was parallel to the axis of the transducer. The thermoacoustic waves from this at boundary traveled perpendicularly to the axis of the transducer and never reached the transducer. In the second set of images, Fig. 4a is the cross section of the sample, which was a pork fat cylinder with one layer of connective tissue 1 and several small pieces of muscle. The transducer was mounted on the left side of the picture, pointing to the sample and moving along the ruler. The images before and after reconstruction are shown in Fig. 4b and Fig. 4c, respectively. The reconstructed image describes the structure of the sample very well. The connective tissue across the cylinder has been imaged clearly, as marked by 1 in Fig. 4c. The muscles have been imaged as three slides parallel to each other because of the different distances between the muscles and the transducer. Because of the lateral convolution effect caused by the 6 mm diam of the transducer, the images of the muscles were stretched and overlapped along the y direction. In the 9 GHz system, because of the small cross-sectional area in the 9 GHz waveguide, the microwaves were incident

2430

Feng et al.: Microwave-induced thermoacoustic tomography

2430

sample from the left side and moved along the ruler. In the 3 GHz system, the microwaves were incident upon the sample perpendicularly to the picture and pointed outward, which made the microwaves evenly distributed in the cross section of the sample. In the original signal data in Fig. 5b, the signals from the muscle cylinder and the edge of the pork fat are spread into a hyperbola with similar shapes. The signals from the two muscle cylinders even cross each other in the center. After the reconstruction, the rebuilt image shown in Fig. 5c is in good agreement with the real sample shown in Fig. 5a.
B. Discussion

FIG. 5. a Cross-sectional picture of the sample, an ellipse of pork fat containing two muscle cylinders; b raw data of thermoacoustic signals acquired with the 3 GHz system; c image after the reconstruction.

upon the sample from the right to the left side along the transducer axis but in the opposite direction. In this case, the thermoacoustic waves emitted from the left curved edge were smaller than those emitted from the right curved edge. One reason is that the strong microwave absorption of the sample, especially the muscle layers in the sample in Fig. 4, decreased the intensity of the microwave eld on the left side. The other reason is that the distribution of the microwave eld, without considering sample absorption, also decreased with increasing distance from the sample to the outlet of the waveguide. Therefore, the left boundary of the sample in Fig. 4c is much weaker than the right one. In the above two images acquired with the 9 GHz system, objects in the right part of the image produce a mirror image due to reected thermoacoustic waves from the boundaries of the container. The reected signals could be eliminated by time gating, but for comparison they were kept to maintain the same image scale as the images from the 3 GHz system. The third set of images shown in Fig. 5 was gathered with the 3 GHz system. Shown in Fig. 5a is a cross section of the sample, which was an elliptic slab of pork fat with two muscle cylinders buried inside. With the sample being immersed in the mineral oil, the transducer pointed to the
Medical Physics, Vol. 28, No. 12, December 2001

From all of the images above, it is clear that the image resolution of curved boundaries of samples and of small cylinders is worse than that of horizontal boundaries, which emit thermoacoustic waves along the transducer axis. For example, the axial resolution of the slabs in Fig. 4c is much better than that of the two small cylinders in Fig. 5c. The reason is that when the thermoacoustic wave does not come from the center of the transducers receiving solid angle, it may reach different parts of the transducer surface at different times. In this case, the pulse signal is broadened and this broadening of the pulse is proportional to the dimensions of the transducer surface. If the thermoacoustic waves come from the center of the receiving solid angle of the transducer, the resolution is optimized. The propagating direction of most thermoacoustic waves from curved boundaries and small cylinders have large angles with the transducer axis. In other words, when the transducer detects these thermoacoustic waves, they are not from the center of the receiving solid angle. Therefore even after reconstruction, the axial resolution of the curved boundaries or small cylinders has been compromised. Also because of the 6 mm diam of the transducer, all of the images have been stretched along the y direction and the lateral resolution has deteriorated. Therefore, the diameter of the active element of the transducer is a key to both the axial and lateral resolutions. We can alleviate the effect of stretching and improve the axial resolution of the images by reducing the diameter of the transducer at the cost of losing the SNR. Comparing the above images before and after reconstruction, the SNR has been greatly improved by the reconstruction. In the delay-and-sum algorithm, the signal intensity of every point is the sum of the signals from different positions that the transducer scanned. From the perspective of the SNR and randomization of the noise, summing up signals from k different positions has the equivalent effect of averaging the signal k times and will increase the SNR by k times. We can take advantage of this property to greatly reduce our average time in data acquisition and the dose of microwaves. A comparison of the images in Fig. 4c and Fig. 5c shows that the 3 GHz system has a much larger image volume than the 9 GHz system due to the deeper penetration depth of the microwaves with a lower frequency and a larger cross section of the waveguide. For microwaves of 3 GHz, the penetration depths for muscle and fat are 1.2 and 9 cm,

2431

Feng et al.: Microwave-induced thermoacoustic tomography

2431

respectively, and for microwaves of 9 GHz, the penetration depths for muscle and fat are 0.27 and 2.6 cm. The depth we can image in the tissues is proportional to the penetration depth of the microwaves in the tissue. On the other hand, in near-surface imaging the 9 GHz setup has a better SNR than the 3 GHz system due to the larger attenuation of 9 GHz microwaves and the higher-power density of the microwave source. In traditional LMTT, the ghost caused by the relatively small diameter of the focused transducer affects the lateral resolution. In synthetic aperture, the scanning nonfocused transducer can be articially focused onto any specied point and the effect equals a focused transducer with a diameter of the scanning dimension. The large diameter, which is much larger than the diameter of the focused transducer used in conventional LMTT, greatly reduces the ghost and, therefore, improves the lateral resolution of the image. In our study we point out that, compared with traditional LMTT, the synthetic-aperture method is effective for improving the lateral resolution of images and imaging the curved boundaries in samples. The resolution of the images can be further improved by reducing the diameter of the transducer or applying deconvolution with respect to the nite size of the transducer surface. According to the IEEE standard, our case involves exposures under a controlled environment, which means the exposure is incurred by persons who are aware of the potential of exposure. For both the 3 and 9 GHz microwaves under a controlled environment, the upper limit of safe exposure is 10 mW/cm2. If it is used for partial body exposure, the upper limit is relaxed to 20 mW/cm2 for 3 GHz microwaves and 22.1 mW/cm2 for 9 GHz microwaves. The peak power of our 3 GHz microwave generator is 10 kW; the microwave pulse width is 0.5 s; the pulse repetition rate is less than 40 Hz; and the outlet of the microwave generator is 72 mm 34 mm. As a result, the power density of the 3 GHz microwave system is 8.2 mW/cm2. The peak power of our 9 GHz microwave generator is 25 kW; the microwave pulse width is 0.5 s; the pulse repetition rate is about 2 Hz; and the outlet of the microwave generator is 23 mm 10 mm. Consequently, the power density of the 9 GHz microwave system is 10.9 mW/cm2. Further, we assumed that the entire microwave has been coupled out of the waveguide without divergence. In practice, however, only part of the microwave is coupled out of the waveguide and diverged into a much larger area than the outlet of the waveguide. The power densities used in our experiments are below the limits of the IEEE standard and are safe to humans. IV. CONCLUSIONS The synthetic aperture, which has never been used in LMTT, is proved to be a powerful image reconstruction method. The reconstruction method based on the delay-andsum algorithm has been veried to work well in LMTT because of its ability to image curved boundaries in samples, to improve the lateral resolution, and to reduce the noise of the system. The large diameter of the transducer causes resoluMedical Physics, Vol. 28, No. 12, December 2001

tion deterioration; the diameter can, however, be reduced to improve the resolution at the expense of the SNR. The comparison of the images shows that the 3 GHz system has a larger imaging volume but a poorer SNR than the 9 GHz system in near-surface imaging.

ACKNOWLEDGMENTS This project was sponsored in part by the U.S. Army Medical Research and Material Command Grant No. DAMD17-00-1-0455, the National Institutes of Health Grants No. R01 CA71980 and No. R21 CA83760, the National Science Foundation Grant No. BES-9734491, and Texas Higher Education Coordinating Board Grant No. ARP 000512-0123-1999.
a

Author to whom correspondence should be addressed. Telephone: 979847-9040; fax: 979-845-4450; electronic mail: LWang@tamu.edu; URL: http://oilab.tamu.edu 1 W. Joines, R. Jirtle, M. Rafal, and D. Schaeffer, Microwave power absorption differences between normal and malignant tissue, Int. J. Radiat. Oncol., Biol., Phys. 6, 681 687 1980. 2 S. Chaudhary, R. Mishra, A. Swarup, and J. Thomas, Dielectric properties of normal human breast tissues at radiowave and microwave frequencies, Indian J. Biochem. Biophys. 21, 76 79 1984. 3 W. Joines, Y. Zhang, C. Li, and R. Jirtle, The measured electrical properties of normal and malignant human tissues from 50900 MHz, Med. Phys. 21, 547550 1994. 4 L. E. Larsen and J. H. Jacobi, in Medical Applications of Microwave Imaging IEEE, Piscataway, NJ, 1986. 5 S. Caorsi, A. Frattoni, G. L. Gragnani, E. Nortino, and M. Pastorino, Numerical algorithm for dielectric-permittivity microwave imaging of inhomogeneous biological bodies, Med. Biol. Eng. Comput. 29, NS37 44 1991. 6 M. S. Hawley, A. Broquetas, L. Jofre, J. C. Bolomey, and G. Gaboriaud, Microwave imaging of tissue blood content changes, J. Biomed. Eng. 13, 197202 1991. 7 P. M. Meaney, K. D. Paulsen, and J. T. Chang, Near-eld microwave imaging of biologically-based materials using a monopole transceiver system, IEEE Trans. Microwave Theory Tech. 46, 31 45 1998. 8 L.-H. V. Wang, X. Zhao, H. Sun, and G. Ku, Microwave-induced acoustic imaging of biological tissues, Rev. Sci. Instrum. 70, 3744 3748 1999. 9 G. Ku and L.-H. V. Wang, Scanning thermoacoustic tomography in biological tissue, Med. Phys. 27, 11951202 2000. 10 G. Ku and L.-H. V. Wang, Scanning microwave-induced thermoacoustic tomography: signal, resolution, and contrast, Med. Phys. 28, 4 10 2001. 11 C. G. A. Hoelen, F. F. M. Demul, R. Pongers, and A. Dekker, Threedimensional photoacoustic imaging of blood vessels in tissue, Opt. Lett. 23, 648 650 1998. 12 C. G. A. Hoelen, R. Pongers, G. Hamhuis, F. F. M. Demul, and J. Greve, Photoacoustic blood cell detection and imaging of blood vessels in phantom tissue, SPIE 3196, 142153 1998. 13 A. A. Karabutov, E. V. Savateeva, N. B. Podymova, and A. A. Oraevsky, Backward mode detection of laser-induced wide-band ultrasonic transients with optoacoustic transducer, J. Appl. Phys. 87, 20032014 2000. 14 R. A. Kruger, P. Liu, Y. R. Fang, and C. R. Appledorn, Photoacoustic ultrasound PAUSReconstruction tomography, Med. Phys. 22, 1605 1609 1995.

Scanning thermoacoustic tomography in biological tissue


Geng Ku and Lihong V. Wanga)
Optical Imaging Laboratory, Biomedical Engineering Program, Texas A&M University, 3120 Tamu, College Station, Texas 77843-3120

Received 5 August 1999; accepted for publication 2 March 2000 Microwave-induced thermoacoustic tomography was explored to image biological tissue. Short microwave pulses irradiated tissue to generate acoustic waves by thermoelastic expansion. The microwave-induced thermoacoustic waves were detected with a focused ultrasonic transducer. Each time-domain signal from the ultrasonic transducer represented a one-dimensional image along the acoustic axis of the ultrasonic transducer similar to an ultrasonic A-scan. Scanning the system perpendicularly to the acoustic axis of the ultrasonic transducer would generate multi-dimensional images. Two-dimensional tomographic images of biological tissue were obtained with 3-GHz microwaves. The axial and lateral resolutions were characterized. The time-domain piezo-electric signal from the ultrasonic transducer in response to the thermoacoustic signal was simulated theoretically, and the theoretical result agreed with the experimental result very well. 2000 American Association of Physicists in Medicine. S0094-24050003105-9 Key words: microwave, ultrasonics, thermoacoustics, photoacoustics, tomography

I. INTRODUCTION Purely microwave imaging of biological tissues has been investigated for a number of years.15 The advantages of microwave imaging include the use of nonionizing radiation and relatively good imaging contrast. However, purely microwave imaging has had difculties in multi-channel detection of microwave without cross coupling, in reconstruction algorithms, and especially in achieving good spatial resolution because of the large wavelength of microwaves. Purely ultrasound imaging ultrasonography, an established medical imaging modality, can yield good spatial resolution but has poor contrast. Microwave-induced thermoacoustics may bridge the gap and combine the advantages of the two types of radiation. In microwave-induced thermoacoustics, microwave pulses generate acoustic signals in lossy media. This phenomenon was known as microwave-auditory or microwavehearing effect in the early years.6 Microwave-induced thermoacoustics was used to quantify physical parameters in media such as the power density and the concentration of a given substance.79 Microwave-induced thermoacoustics was also employed by several investigators in the 1980s for imaging of biological tissues.1015 However, these early works did not produce any tomographic or depth-resolved images. Recently, images of biological tissues were reconstructed based on microwave-induced thermoacoustics.16,17 This approach requires measurements of a large amount of data around the tissue and expensive computation following the data acquisition. Microwave-induced thermoacoustic imaging is based on the detection of the thermoacoustic signals generated by microwaves in the samples. Pulsed microwave radiation is used to irradiate the samples. Absorbed microwave energy causes thermoelastic expansion that radiates acoustic waves. An ultrasonic transducer detects the time-resolved thermoacoustic
1195 Med. Phys. 27 5, May 2000

signals. If optical radiation instead of microwave radiation is used, this phenomenon is better known as photoacoustics. Although microwave-induced thermoacoustic imaging shares similar principles with the photoacoustic imaging in the optical-wavelength region,1828 it may have broader applications in medical imaging because microwave radiation provides a deeper penetration depth in biological tissues than light. The electric eld strength of a plane wave in a lossy media is attenuated exponentially as E E 0 exp z , 1

where E 0 is the electrical eld at the sample surface, E is the electrical eld at the depth z, and is the electric-eld absorption coefcient expressed as

2 1

1 ,

where is the angular frequency, is the permeability, is the permittivity, and is the conductivity. The induced thermoacoustic pressure depends on the intensity of microwave and the complex dielectric constant of the material. In the frequency range of 0.110 GHz, the dielectric constant ratio of the permittivity in material to that in vacuum has a value of 570 for soft tissues, and the conductivity has a value of 0.023 1 m1.2931 The dielectric properties of tissues determine the absorption of microwave at various microwave frequencies. At 3 GHz, the penetration depths for fat and muscle, which are the inverse of the absorption coefcients, are 9 cm and 1.2 cm, respectively; while at 500 MHz, the penetration depths for fat and muscle are 23.5 cm and 3.4 cm, respectively.32 Most of the other soft tissues have an absorption coefcient in between those for muscle and fat. This wide range of values among various tissues can provide a high imaging contrast for biological tissues. Cancerous
2000 Am. Assoc. Phys. Med. 1195

0094-2405200027511958$17.00

1196

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1196

FIG. 1. Experimental setup for scanning thermoacoustic tomography.

breast tissues are 25 times more strongly absorbing than surrounding normal breast tissues, which has been attributed to an increase in bound water and sodium within malignant cells.3335 Therefore, microwave-induced thermoacoustic imaging may potentially be used to detect early-stage cancers. We here present our studies on scanning microwaveinduced thermoacoustic tomography toward biomedical applications. The microwave-induced piezo-electric signal from the ultrasonic transducer was simulated theoretically to explain the experimental observation for the rst time to our knowledge. Cross sections of biological tissues were imaged in full view using our current 3-GHz imaging system, representing a signicant advance from the previous 9-GHz system that imaged cross sections only partially because of the limited penetrating power of the radiation at 9 GHz.36 Our imaging approach differs signicantly from the prior arts in microwave-induced thermoacoustic imaging. Lateral resolution was achieved by use of a focused ultrasonic transducer. Axial resolution was obtained by measuring the temporal proles of the microwave-induced acoustic signals. Depthresolved tomographic images were acquired directly without resorting to image reconstruction. II. METHODS The experimental setup used for this study is shown in Fig. 1. A Cartesian coordinate system was set up for reference. The z axis was along the acoustic axis pointing upward. The x axis was perpendicular to the drawing plane and pointed outward. The y axis was in the drawing plane pointing to the right. A 3-GHz microwave generator transmitted microwave pulses. The peak power was estimated to be 2 kW. The pulse width was modied to 0.5 s. A function generator DS345, Stanford Research System was employed to trigger the microwave generator, control its pulse repetition frequency, and synchronize the oscilloscope sampling. Microwave energy was delivered by a rectangular waveguide with a cross section of 72 mm34 mm. The object to be
Medical Physics, Vol. 27, No. 5, May 2000

imaged was placed on a plastic stand inside a plexiglass container lled with mineral oil. The container was xed on a two-dimensional x - y translation stage MD2, Arrick Robotics. A personal computer controlled the two stepper motors to drive the translation stage in the x and y directions. Both mineral oil and plexiglass have a small absorption coefcient for microwaves. Mineral oil also provides good acoustic coupling. An ultrasonic transducer was immersed in the mineral oil facing the microwave waveguide. The transducer was connected to a pulse amplier. The amplied signal was recorded and averaged 100 times by an oscilloscope TDS640A, Tektronix and then transferred to a personal computer for imaging. Two ultrasonic transducers were used in this study. For the rst one V314, Panametrics, the central frequency of the ultrasonic transducer was 1 MHz, the bandwidth was 0.6 MHz, the diameter was 1.9 cm, and the focal length at 1 MHz was 2.5 cm. For the second one V384, Panametrics, the central frequency was 3.5 MHz, the bandwidth was 2.5 MHz, the diameter was 0.64 cm, and the focal length at 3.5 MHz was 1.8 cm. Unless otherwise stated, the rst ultrasonic transducer was used to obtained the results presented here. In our scanning microwave-induced thermoacoustic tomography, the ultrasonic transducer measured the time-ofarrival signals of the thermoacoustic waves. The distance between the thermoacoustic source and the transducer was calculated by multiplying the time of arrival with the speed of sound in the medium. Therefore, a time-domain signal was converted into a one-dimensional image along the acoustic axis z axis, which is similar to an ultrasonic A-scan image. Scanning the sample along the x or y axis and combining the multiple one-dimensional images yielded a two-dimensional cross-sectional image of the sample in the x - z or y - z plane. III. RESULTS AND DISCUSSION The generation of thermoacoustic waves by deposition of microwave energy can be described by the following differential equation:

H 1 2 , 2 2 p r, t t C vs p t

where p ( r, t ) is the thermoacoustic pressure at position r and time t, v s is the speed of sound, is the isobaric volume expansion coefcient, C p is the heat capacity, and H is the heating function describing the microwave-energy deposition in the sample per unit volume per unit time. Thermalconnement condition is assumed, where the acoustic transit time across the acoustic source is less than the heat conduction time. The solution of the three-dimensional wave equation under the zero-initial-value conditions p (0,r) 0 and ( / t ) p (0,r) 0 can be expressed as an integral: p r, t

4Cp

1 H r , t d r . r r t

The integral is calculated inside a sphere with a radius of v s t centered at r, and r is the position inside the sphere where

1197

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1197

FIG. 2. Simulation of the piezo-electric signal in response to the microwave-induced thermoacoustic signal from a 4.8 mm thick gel slab. a Thermoacoustic signal in a slab induced by an ideal microwave impulse; b Temporal prole of the microwave pulses used in the experiment; c Thermoacoustic signal in a slab induced by the microwave pulses used in the experiment; d Piezo-electric impulse response of the ultrasonic transducer; e Experimental and simulated piezo-electric outputs of the ultrasonic transducer in response to the thermoacoustic signals.

microwave is absorbed and acoustic signal is generated. In the integration, the heating function is not taken at time t but at an earlier time t t r r / v s ; therefore, the integration function is also called retarded potential.37 Analytic solutions can be obtained for simple geometric structures such as an innite layer, a sphere, and a cylinder under delta heating, where the heating function is a delta function in time.38 For a slab with a thickness d under delta heating, the impulse-response pressure is
2 vs u z v st , p 1 z , t 2Cp

wave absorption coefcient, and s ( t ) is the temporal prole of the microwave pulse. Figure 2b shows the temporal prole of the microwave pulses used in our experiment. The thermoacoustic pressure induced by the microwave pulses can be derived by the following convolution: p z,t

p 1 z , H z , t d .

where u ( z v s t ) is dened as a function that is unity when 0 ( z v s t ) d and zero otherwise. The impulse response is a traveling square wave as shown in Fig. 2a for a 4.8 mm gel slab if the microwave attenuation across the slab is negligible. Because the propagation speed of electromagnetic wave is much greater than the speed of sound, the sample volume illuminated by microwave pulses radiates acoustic waves simultaneously. The heating function in one-dimensional lossy media may be expressed as H ( z , t ) e z s ( t ), where is the microMedical Physics, Vol. 27, No. 5, May 2000

Figure 2c illustrates the thermoacoustic pressure at the ultrasonic transducer generated from the slab, which was obtained by convolving the two temporal waveforms in Figs. 2a and 2b by use of Eq. 6. For more general nonthermal-conned cases when heat transfer in the medium cannot be neglected, the following heat conduction equation must be taken into account in the pressure calculation:

Cp

T r, t k 2 T r, t H r, t , t

where is the density of the medium, k is the thermal con-

1198

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1198

ductivity, and T ( r, t ) is the temperature distribution in the microwave-illuminated space. The thermoacoustic pressure is: p r, t

4Cp
k2

1 H r , t r r t

T r , t d r . t

The integration involves the heat conducted from the surrounding medium besides the heat due to the absorbed microwave energy. The experimentally measured piezo-electric impulse response q ( t ) of the ultrasonic transducer is shown in Fig. 2d. The piezo-electric output of the ultrasonic transducer in response to thermoacoustic pressure can be calculated by the following convolution between the thermoacoustic pressure at the transducer and the impulse response of the transducer: P 0 z , t

p z, q t d.

The piezo-electric signal from a 4.8 mm slab was obtained by the convolution of the two waveforms in Figs. 2c and 2d with Eq. 9 and is plotted in Fig. 2e as a dashed line. The experimental piezo-electric signal from a 4.8 mm gel slab is plotted in Fig. 2e as a solid line for comparison, where the slight offset was caused by the dc drift of the amplier in the experimental detection. There are two dipolar structures in each waveform in Fig. 2e, and the signal between the dipolar structures is weak. The dipolar structures from the upper and lower surfaces of the slab have opposite polarities, which could not be explained by any existing theories in the literature. The polarity, the width of each dipole, and the distance between the two dipolar structures are in good agreement between the results of our theory and experiment. Because the piezo-electric signal of the ultrasonic transducer can be simulated by the two convolutions as shown above, the dipole width is related to the width of the microwave pulses and the width of the impulse response of the ultrasonic transducer, which were 0.5 s and 1.7 s, respectively. The time intervals between the zero-crossing points of the two dipolar structures in the two waveforms are determined by the slab thickness and are equal to the acoustic transit time 3.2 s over the 4.8-mm-thick slab. Dipolar structures were also observed in photoacoustics by lasers, where a Q-switched laser with a pulse width of 10 ns and a wide-band ultrasonic transducer were employed. Our detected dipolar structures resemble the wellknown dipolar structures that originate from small spherical or cylindrical objects excited by laser pulses39 or from acoustic reection at soft acoustic interfaces.40 However, our detected dipolar structures in the thermoacoustic signals in slabs resulted from the limited bandwidth of the ultrasonic transducers. This phenomenon can also be explained in the frequency domain, where the ultrasonic transducer acts as a bandpass lter. A frequency spectrum of the ultrasonic transducer is
Medical Physics, Vol. 27, No. 5, May 2000

FIG. 3. Frequency-domain analysis of the microwave-induced thermoacoustic signals. a Spectrum of the 1 MHz ultrasonic transducer; b Spectrum of the microwave-induced thermoacoustic signal; c Spectrum of the piezoelectric signal, which was the ltered microwave-induced thermoacoustic signal.

shown in Fig. 3a. The temporal prole of the thermoacoustic pressure varies sharply near the slab boundaries and slowly inside the slab as shown in Fig. 2c. The corresponding spectrum is peaked at dc as shown in Fig. 3b. The ltered spectrum is peaked near 0.5 MHz as shown in Fig. 3c, where the dc is rejected and the low-frequency components are attenuated signicantly. In other words, the ultrasonic transducer cannot respond efciently to the thermoacoustic waves emitted between the sample boundaries, which have a lower-frequency spectrum. Therefore, the observed piezo-electric signal between the two sample boundaries is low.

1199

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1199

This model includes only the key concept. Scattering and diffraction of microwaves becomes certainly important if strictly quantitative modeling is performed. The detected thermoacoustic signals are also related to other factors including microwave attenuation, conversion efciency from microwave energy to acoustic energy, acoustic scattering and diffraction, and acoustic attenuation. A tissue sample with a fat-muscle-fat structure as shown in Fig. 4a was placed on the plastic stand inside the container as shown in Fig. 1. Figure 4b shows a 2D image of the sample obtained with the scanning thermoacoustic tomography technique. Thermoacoustic signals were acquired in the time domain while the fat-muscle-fat sample was scanned horizontally along the y axis with a step size of 1 mm. The 2D image of the sample was formed by combining these temporal waveforms taken successively at the scanning stops along the y axis. Each vertical line in this 2D image was from a temporal waveform. The muscle inside the fat is clearly visible with a good contrast. The fat-tissue interface to the left of the muscle is also visible, which was possibly caused by the slight difference in the microwave properties between the two fat sections. Figure 4c illustrates a time-domain waveform that was measured above the center of the sample at y equal to 54 mm. The strongest dipole near 27 s was from the bottom boundary of the sample where microwave experienced the least attenuation, whereas the weakest dipole near 14 s was from the top boundary of the sample where microwave experienced the most attenuation. The two dipolar structures corresponding to the boundaries of the muscle layer are also clearly distinguishable. The time intervals between the adjacent dipolar structures agree with the thickness values of the tissue layers very well. However, the vertical boundaries of the muscle slab are not visible in the image because the thermoacoustic waves from these boundaries propagate perpendicularly to the acoustic axis of the ultrasonic transducer and therefore cannot be received by the transducer. The axial resolution along the acoustic axis z axis is determined by the width of the thermoacoustic dipolar structures, which is related to the width of the microwave pulse and the width of the impulse response of the transducer the inverse of the bandwidth of the ultrasonic detector. When the 1 MHz ultrasonic transducer was used, the width of the thermoacoustic signal was estimated to be 2.2 s, which was the sum of the width of the microwave pulses 0.5 s and the width of the impulse response of the transducer 1.7 s. Because the speed of sound in tissue is 1.5 mm/s, the corresponding axial resolution should be approximately 3.3 mm along the z axis. For the 3.5 MHz ultrasonic transducer, the axial resolution should be improved to 1.4 mm theoretically. Figure 5 shows the thermoacoustic signals from slab samples of various thickness values measured by the 1 MHz ultrasonic transducer. The slab samples were made of 5% gelatin and 5% NaCl, where NaCl controlled the microwave absorption. As the thickness of the samples decreased, the temporal distance between the adjacent dipolar structures corresponding to the two boundaries of the slabs decreased
Medical Physics, Vol. 27, No. 5, May 2000

FIG. 4. a Cross section of a fat-muscle-fat sample on the y - z plane; b Two-dimensional image of the y - z cross section of the sample obtained by scanning thermoacoustic tomography; c Temporal microwave-induced thermoacoustic signal along the vertical center line of the sample. The 1 MHz ultrasonic transducer was used in this experiment.

as well. The two dipolar structures became barely distinguishable when the thickness was reduced to 3.8 mm and completely inseparable when the thickness was reduced to 2 mm. Therefore, the experimentally measured axial resolution was 3.8 mm, close to the above calculated resolution of 3.3 mm based on the dipole width. The discrepancy was caused by the long tail of the dipolar structures. The relative variation in intensity between the two dipolar structures was caused by microwave attenuation in the slabs. Similarly, Fig. 6 shows the thermoacoustic signals from slab samples of various thickness values measured by the 3.5

1200

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1200

FIG. 5. Microwave-induced thermoacoustic signals in gel slabs of various thickness values: a 4.5 mm, b 3.8 mm, and c 2 mm. The 1 MHz ultrasonic transducer was used.

FIG. 6. Microwave-induced thermoacoustic signals in gel slabs of various thickness values: a 3.4 mm, b 1.9 mm, and c 1 mm. The 3.5 MHz ultrasonic transducer was used.

MHz ultrasonic transducer. The experimentally measured axial resolution was 1.9 mm, close to the above calculated resolution of 1.4 mm based on the dipole width. As expected from the theoretical consideration, the wider-bandwidth transducer produced better axial resolution. Potentially, shorter microwave pulses and deconvolution may be used to improve the axial resolution further.
Medical Physics, Vol. 27, No. 5, May 2000

The lateral resolution is determined by the numerical aperture of the ultrasonic transducer. The ultrasonic transducer responds to the thermoacoustic signals along its acoustic axis. The detected source area is related to the numerical aperture of the ultrasonic transducer and the distance between the thermoacoustic source and the ultrasonic transducer. The minimum detected source area is at the focal

1201

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1201

from the thermoacoustic waves propagating directly upward toward the ultrasonic transducer, whereas the dark lower band near z 9 mm is the ghost image caused by acoustic reection from the plastic stand. The gaps of greater than 4 mm can be easily recognized. The gap of 2.5 mm can be barely identied, which denes the lateral resolution and is comparable with the focal diameter of the ultrasonic transducer. The measured thermoacoustic signal is a convolution between the thermoacoustic signal in the sample and the detection-sensitivity distribution of the ultrasonic transducer over the detected area. The convolution reduces the lateral resolution, which is worsened when the ultrasonic transducer is out of focus. Figure 6c was acquired when the ultrasonic transducer was deliberately moved far away from the sample so as to create defocusing. The gaps are not distinguishable in the 2D image due to reduced resolution.

IV. CONCLUSIONS The simulation of the piezo-electric signal, by convolving the thermoacoustic signal of delta heating with the microwave-pulse waveform and then with the impulse response of the ultrasonic transducer, can be used to explain the bipolar experimental data. Our studies show that scanning microwave-induced thermoacoustic tomography is a promising imaging tool for biological tissue. The boundaries of different tissue constituents can be imaged clearly. The width of the microwave pulses and the bandwidth of the ultrasonic transducer determine the axial resolution along the acoustic axis of the ultrasonic transducer, which is 1.9 mm under our current experimental conditions. The axial resolution can be improved by employing broader-band ultrasonic transducers and shorter microwave pulses. The lateral resolution is determined by the numerical aperture of ultrasonic transducers and is 2.5 mm in our current setup.

FIG. 7. Two-dimensional tomographic images of a linear array of gel slabs obtained by scanning thermoacoustic tomography. a y - z cross section of the sample units in mm; b Image of the y - z cross section when the sample was placed at the focal plane of the ultrasonic transducer; c Image of the y - z cross section when the sample was placed far from the focal plane of the ultrasonic transducer.

ACKNOWLEDGMENTS Thanks to X. Zhao and G. Yao for their assistance in electronics and computer programming, respectively. This project was sponsored in part by the National Institutes of Health Grants Nos. R29 CA68562, R01 CA71980, and R21 CA83760 and by the National Science Foundation Grant no. BES-9734491.
a

plane of the ultrasonic transducer. Therefore, a better lateral resolution is expected when the sample is located within the focal column. We examined the lateral resolution with the 1 MHz ultrasonic transducer. The ultrasonic transducer has a 3 dB focal diameter of 2.1 mm and a focal zone of 17.6 mm along the acoustic axis. Several pieces of rectangular gel slabs were arranged linearly along the y direction as shown in Fig. 7a. The y - z cross section was imaged with a step size of 1 mm when the sample was scanned along the y axis. Figure 7b shows the two-dimensional image when the sample was on the focal plane of the ultrasonic transducer. The bright upper band near z 14 mm is the primary image
Medical Physics, Vol. 27, No. 5, May 2000

To whom correspondence should be addressed: Tel: 979 847-9040; Fax: 979 845-4450; electronic mail: lwang@tamu.edu; URL: http:// oilab.tamu.edu 1 L. E. Larsen and J. H. Jacobi, Eds., Medical Applications of Microwave Imaging IEEE Press, Piscataway, NJ, 1986. 2 J. C. Lin, Frequency optimization for microwave imaging of biological tissues, Proc. IEEE 73, 374375 1985. 3 S. Caorsi, A. Frattoni, G. L. Gragnani, E. Nortino, and M. Pastorino, Numerical algorithm for dielectric-permittivity microwave imaging of inhomogeneous biological bodies, Med. Biol. Eng. Comput. 29, NS3744 1991. 4 M. S. Hawley, A. Broquetas, L. Jofre, J. C. Bolomey, and G. Gaboriaud, Microwave imaging of tissue blood content changes, J. Biomed. Eng. 13, 197202 1991.

1202
5

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1202

P. M. Meaney, K. D. Paulsen, and J. T. Chang, Near-eld microwave imaging of biologically-based materials using a monopole transceiver system, IEEE Trans. Microwave Theory Tech. 46, 3145 1998. 6 J. C. Lin, On microwave-induced hearing sensation, IRE Trans. Microwave Theory Tech. MTT-25, 605613 1977. 7 F. Caspers and J. Conway, Measurement of power density in a lossy material by means of electromagnetically-induced acoustic signals for noninvasive determination of spatial thermal absorption in connection with pulsed hyperthermia, Proc. of the 12th European Microwave Conference, 565568 1982. 8 J. L. Su and J. C. Lin, Thermoelastic signatures of tissue phantom absorption and thermal expansion, IEEE Trans. Biomed. Eng. 43, 178 182 1987. 9 J. S. K. Wan, Microwaves and chemistry: The catalysis of an exciting marriage, Res. Chem. Intermed. 19, 147158 1993. 10 T. Bowen, L. Nasoni, A. E. Pifer, and G. H. Sembrosk, Some experimental results on the thermoacoustic imaging of soft tissue-equivalent phantoms, Ultrason. Symp. Proc. 2, 823827 1981. 11 R. G. Olsen, Generation of acoustic images from the absorption of pulsed microwave energy, in Acoustic Imaging, edited by J. P. Powers Plenum, New York, 1982, pp. 5359. 12 R. G. Olsen and J. C. Lin, Acoustic imaging of a model of a human hand using pulsed microwave irradiation, Bioelectromagnetics N.Y. 4, 397400 1983. 13 J. C. Lin and K. H. Chan, Microwave thermoelastic tissue imagingsystem design, IEEE Trans. Microwave Theory Tech. 32, 854860 1984. 14 R. L. Nasoni, G. A. Evanoff, Jr., P. G. Halverson, and T. Bowen, Thermoacoustic emission by deeply penetrating microwave radiation, Proc.IEEE Ultrason. Symp. 5, 633637 1984. 15 K. H. Chan and J. C. Lin, Microwave-induced thermoacoustic tissue imaging, Proc. Engineering in Medicine and Biology Society 10th Annual International Conference, 445446 1988. 16 R. A. Kruger, D. R. Reinecke, and G. A. Kruger, Thermoacoustic computed tomographyTechnical considerations, Med. Phys. 26, 1832 1837 1999. 17 R. A. Kruger, K. K. Kopecky, A. M. Aisen, D. R. Reinecke, G. A. Kruger, and W. L. Kiser, Jr., Thermoacoustic CT with radio waves: a medical imaging paradigm, Radiology 211, 275278 1999. 18 V. E. Gusev and A. A. Karabutov, Laser Optoacoustics American Institute of Physics, New York, 1993. 19 A. A. Oraevsky, S. L. Jacques, R. O. Esenaliev, and F. K. Tittel, Laserbased optoacoustic imaging in biological tissues, in Laser-Tissue Interaction V , edited by S. L. Jacques, Proc. Soc. Photo-Instrum. Eng. 2134A, 122128 1994. 20 R. A. Kruger and P. Liu, Photoacoustic ultrasound: Pulse production and detection of 0.5% Liposyn, Med. Phys. 21, 11791184 1994. 21 A. A. Oraevsky, S. L. Jacques, R. O. Esenaliev, and F. K. Tittel, Direct measurement of laser uence distribution and optoacoustic imaging in heterogeneous tissues, in Laser Interaction with Hard and Soft Tissue II, edited by H. J. Albrecht, G. P. Delacretaz, T. H. Meier, R. W. Steiner, L. O. Svaasand, and M. J. van Gemert, Proc. Soc. Photo-Instrum. Eng. 2323, 3746 1995. 22 R. A. Kruger, P. Liu, Y. R. Fang, and C. R. Appledorn, Photoacoustic ultrasound PAUSreconstruction tomography, Med. Phys. 22, 1605 1609 1995. 23 A. A. Oraevsky, R. Esenaliev, F. K. Tittel, M. R. Ostermeyer, L.-H. Wang, and S. L. Jacques, Laser opto-acoustic imaging of turbid media:

determination of optical properties by comparison with diffusion theory and Monte Carlo simulation, Proc. SPIE 2681, 277284 1996. 24 A. A. Oraevsky, R. O. Esenaliev, and A. A. Karabutov, Laser optoacoustic tomography of layered tissues: signal processing, in Optical Tomography and Spectroscopy of Tissue: Theory, Instrumentation, Model, and Human Studies II, edited by B. Chance and R. R. Alfano, Proc. Soc. Photo-Instrum. Eng. 2979, 5970 1997. 25 R. O. Esenaliev, H. Alma, F. K. Tittel, and A. A. Oraevsky, Axial resolution of laser opto-acoustic imaging: inuence of acoustic attenuation and diffraction, in Laser-Tissue Interaction IX, edited by S. L. Jacques, Proc. Soc. Photo-Instrum. Eng. 3254, 294306 1998. 26 C. G. A. Hoelen, F. F. M. Demul, R. Pongers, and A. Dekker, Threedimensional photoacoustic imaging of blood vessels in tissue, Opt. Lett. 23, 648650 1998. 27 A. A. Oraevsky, V. A. Andreev, A. A. Karabutov, and R. O. Esenaliev, Two-dimensional optoacoustic tomography: Transducer array and image reconstruction algorithm, in Laser-Tissue Interaction X: Photochemical, Photothermal, and Photomechanical, edited by S. L. Jacques, G. J. Mueller, A. Roggan, and D. H. Sliney, Proc. Soc. Photo-Instrum. Eng. 3601, 256267 1999. 28 A. A. Karabutov, E. B. Savateeva, and A. A. Oraevsky, Imaging of layered structures in biological tissues with optoacoustic front surface transducer, in Laser-Tissue Interaction X: Photochemical, Photothermal, and Photomechanical, edited by S. L. Jacques, G. J. Mueller, A. Roggan, and D. H. Sliney, Proc. Soc. Photo-Instrum. Eng. 3601, 284 295 1999. 29 C. Gabriel, S. Gabriel, and E. Corthout, The dielectric properties of biological tissues: I. literature survey, Phys. Med. Biol. 41, 22312249 1996. 30 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: II. Measurements in the frequency range 10 Hz to 20 GHz, Phys. Med. Biol. 41, 22512269 1996. 31 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues, Phys. Med. Biol. 41, 22712293 1996. 32 F. A. Duck, Physical Properties of Tissue: A Comprehensive Reference Book Academic, London, 1990. 33 W. Joines, R. Jirtle, M. Rafal, and D. Schaeffer, Microwave power absorption differences between normal and malignant tissue, Int. J. Radiat. Oncol. Biol. Phys. 6, 681687 1980. 34 S. Chaudhary, R. Mishra, A. Swarup, and J. Thomas, Dielectric properties of normal human breast tissues at radiowave and microwave frequencies, Indian J. Biochem. Biophys. 21, 7679 1984. 35 W. Joines, Y. Zhang, C. Li, and R. Jirtle, The measured electrical properties of normal and malignant human tissues from 50900 MHz, Med. Phys. 21, 547550 1994. 36 L.-H. Wang, X. Zhao, H. Sun, and G. Ku, Microwave-induced acoustic imaging of biological tissues, Rev. Sci. Instrum. 70, 37443748 1999. 37 I. N. Bronshtein and K. A. Semendyayev, Handbook of Mathematics Van Nostrand Reinhold, New York, 1978. 38 G. J. Diebold and T. Sun, Properties of photoacoustic wave in one two and three dimensions, Acustica 80, 339351 1994. 39 C. G. A. Hoelen, F. F. M. de Mul, and J. Greve, Non-destructive photoacoustic subsurface tissue imaging: a feasibility study, Proc. SPIE 2628, 304314 1995. 40 A. O. Oraevsky, S. L. Jacques, and F. K. Tittel, Measurement of tissue optical properties by time-resolved detection of laser-induced transient stress, Appl. Opt. 36, 402415 1997.

Medical Physics, Vol. 27, No. 5, May 2000

Scanning thermoacoustic tomography in biological tissue


Geng Ku and Lihong V. Wanga)
Optical Imaging Laboratory, Biomedical Engineering Program, Texas A&M University, 3120 Tamu, College Station, Texas 77843-3120

Received 5 August 1999; accepted for publication 2 March 2000 Microwave-induced thermoacoustic tomography was explored to image biological tissue. Short microwave pulses irradiated tissue to generate acoustic waves by thermoelastic expansion. The microwave-induced thermoacoustic waves were detected with a focused ultrasonic transducer. Each time-domain signal from the ultrasonic transducer represented a one-dimensional image along the acoustic axis of the ultrasonic transducer similar to an ultrasonic A-scan. Scanning the system perpendicularly to the acoustic axis of the ultrasonic transducer would generate multi-dimensional images. Two-dimensional tomographic images of biological tissue were obtained with 3-GHz microwaves. The axial and lateral resolutions were characterized. The time-domain piezo-electric signal from the ultrasonic transducer in response to the thermoacoustic signal was simulated theoretically, and the theoretical result agreed with the experimental result very well. 2000 American Association of Physicists in Medicine. S0094-24050003105-9 Key words: microwave, ultrasonics, thermoacoustics, photoacoustics, tomography

I. INTRODUCTION Purely microwave imaging of biological tissues has been investigated for a number of years.15 The advantages of microwave imaging include the use of nonionizing radiation and relatively good imaging contrast. However, purely microwave imaging has had difculties in multi-channel detection of microwave without cross coupling, in reconstruction algorithms, and especially in achieving good spatial resolution because of the large wavelength of microwaves. Purely ultrasound imaging ultrasonography, an established medical imaging modality, can yield good spatial resolution but has poor contrast. Microwave-induced thermoacoustics may bridge the gap and combine the advantages of the two types of radiation. In microwave-induced thermoacoustics, microwave pulses generate acoustic signals in lossy media. This phenomenon was known as microwave-auditory or microwavehearing effect in the early years.6 Microwave-induced thermoacoustics was used to quantify physical parameters in media such as the power density and the concentration of a given substance.79 Microwave-induced thermoacoustics was also employed by several investigators in the 1980s for imaging of biological tissues.1015 However, these early works did not produce any tomographic or depth-resolved images. Recently, images of biological tissues were reconstructed based on microwave-induced thermoacoustics.16,17 This approach requires measurements of a large amount of data around the tissue and expensive computation following the data acquisition. Microwave-induced thermoacoustic imaging is based on the detection of the thermoacoustic signals generated by microwaves in the samples. Pulsed microwave radiation is used to irradiate the samples. Absorbed microwave energy causes thermoelastic expansion that radiates acoustic waves. An ultrasonic transducer detects the time-resolved thermoacoustic
1195 Med. Phys. 27 5, May 2000

signals. If optical radiation instead of microwave radiation is used, this phenomenon is better known as photoacoustics. Although microwave-induced thermoacoustic imaging shares similar principles with the photoacoustic imaging in the optical-wavelength region,1828 it may have broader applications in medical imaging because microwave radiation provides a deeper penetration depth in biological tissues than light. The electric eld strength of a plane wave in a lossy media is attenuated exponentially as E E 0 exp z , 1

where E 0 is the electrical eld at the sample surface, E is the electrical eld at the depth z, and is the electric-eld absorption coefcient expressed as

2 1

1 ,

where is the angular frequency, is the permeability, is the permittivity, and is the conductivity. The induced thermoacoustic pressure depends on the intensity of microwave and the complex dielectric constant of the material. In the frequency range of 0.110 GHz, the dielectric constant ratio of the permittivity in material to that in vacuum has a value of 570 for soft tissues, and the conductivity has a value of 0.023 1 m1.2931 The dielectric properties of tissues determine the absorption of microwave at various microwave frequencies. At 3 GHz, the penetration depths for fat and muscle, which are the inverse of the absorption coefcients, are 9 cm and 1.2 cm, respectively; while at 500 MHz, the penetration depths for fat and muscle are 23.5 cm and 3.4 cm, respectively.32 Most of the other soft tissues have an absorption coefcient in between those for muscle and fat. This wide range of values among various tissues can provide a high imaging contrast for biological tissues. Cancerous
2000 Am. Assoc. Phys. Med. 1195

0094-2405200027511958$17.00

1196

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1196

FIG. 1. Experimental setup for scanning thermoacoustic tomography.

breast tissues are 25 times more strongly absorbing than surrounding normal breast tissues, which has been attributed to an increase in bound water and sodium within malignant cells.3335 Therefore, microwave-induced thermoacoustic imaging may potentially be used to detect early-stage cancers. We here present our studies on scanning microwaveinduced thermoacoustic tomography toward biomedical applications. The microwave-induced piezo-electric signal from the ultrasonic transducer was simulated theoretically to explain the experimental observation for the rst time to our knowledge. Cross sections of biological tissues were imaged in full view using our current 3-GHz imaging system, representing a signicant advance from the previous 9-GHz system that imaged cross sections only partially because of the limited penetrating power of the radiation at 9 GHz.36 Our imaging approach differs signicantly from the prior arts in microwave-induced thermoacoustic imaging. Lateral resolution was achieved by use of a focused ultrasonic transducer. Axial resolution was obtained by measuring the temporal proles of the microwave-induced acoustic signals. Depthresolved tomographic images were acquired directly without resorting to image reconstruction. II. METHODS The experimental setup used for this study is shown in Fig. 1. A Cartesian coordinate system was set up for reference. The z axis was along the acoustic axis pointing upward. The x axis was perpendicular to the drawing plane and pointed outward. The y axis was in the drawing plane pointing to the right. A 3-GHz microwave generator transmitted microwave pulses. The peak power was estimated to be 2 kW. The pulse width was modied to 0.5 s. A function generator DS345, Stanford Research System was employed to trigger the microwave generator, control its pulse repetition frequency, and synchronize the oscilloscope sampling. Microwave energy was delivered by a rectangular waveguide with a cross section of 72 mm34 mm. The object to be
Medical Physics, Vol. 27, No. 5, May 2000

imaged was placed on a plastic stand inside a plexiglass container lled with mineral oil. The container was xed on a two-dimensional x - y translation stage MD2, Arrick Robotics. A personal computer controlled the two stepper motors to drive the translation stage in the x and y directions. Both mineral oil and plexiglass have a small absorption coefcient for microwaves. Mineral oil also provides good acoustic coupling. An ultrasonic transducer was immersed in the mineral oil facing the microwave waveguide. The transducer was connected to a pulse amplier. The amplied signal was recorded and averaged 100 times by an oscilloscope TDS640A, Tektronix and then transferred to a personal computer for imaging. Two ultrasonic transducers were used in this study. For the rst one V314, Panametrics, the central frequency of the ultrasonic transducer was 1 MHz, the bandwidth was 0.6 MHz, the diameter was 1.9 cm, and the focal length at 1 MHz was 2.5 cm. For the second one V384, Panametrics, the central frequency was 3.5 MHz, the bandwidth was 2.5 MHz, the diameter was 0.64 cm, and the focal length at 3.5 MHz was 1.8 cm. Unless otherwise stated, the rst ultrasonic transducer was used to obtained the results presented here. In our scanning microwave-induced thermoacoustic tomography, the ultrasonic transducer measured the time-ofarrival signals of the thermoacoustic waves. The distance between the thermoacoustic source and the transducer was calculated by multiplying the time of arrival with the speed of sound in the medium. Therefore, a time-domain signal was converted into a one-dimensional image along the acoustic axis z axis, which is similar to an ultrasonic A-scan image. Scanning the sample along the x or y axis and combining the multiple one-dimensional images yielded a two-dimensional cross-sectional image of the sample in the x - z or y - z plane. III. RESULTS AND DISCUSSION The generation of thermoacoustic waves by deposition of microwave energy can be described by the following differential equation:

H 1 2 , 2 2 p r, t t C vs p t

where p ( r, t ) is the thermoacoustic pressure at position r and time t, v s is the speed of sound, is the isobaric volume expansion coefcient, C p is the heat capacity, and H is the heating function describing the microwave-energy deposition in the sample per unit volume per unit time. Thermalconnement condition is assumed, where the acoustic transit time across the acoustic source is less than the heat conduction time. The solution of the three-dimensional wave equation under the zero-initial-value conditions p (0,r) 0 and ( / t ) p (0,r) 0 can be expressed as an integral: p r, t

4Cp

1 H r , t d r . r r t

The integral is calculated inside a sphere with a radius of v s t centered at r, and r is the position inside the sphere where

1197

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1197

FIG. 2. Simulation of the piezo-electric signal in response to the microwave-induced thermoacoustic signal from a 4.8 mm thick gel slab. a Thermoacoustic signal in a slab induced by an ideal microwave impulse; b Temporal prole of the microwave pulses used in the experiment; c Thermoacoustic signal in a slab induced by the microwave pulses used in the experiment; d Piezo-electric impulse response of the ultrasonic transducer; e Experimental and simulated piezo-electric outputs of the ultrasonic transducer in response to the thermoacoustic signals.

microwave is absorbed and acoustic signal is generated. In the integration, the heating function is not taken at time t but at an earlier time t t r r / v s ; therefore, the integration function is also called retarded potential.37 Analytic solutions can be obtained for simple geometric structures such as an innite layer, a sphere, and a cylinder under delta heating, where the heating function is a delta function in time.38 For a slab with a thickness d under delta heating, the impulse-response pressure is
2 vs u z v st , p 1 z , t 2Cp

wave absorption coefcient, and s ( t ) is the temporal prole of the microwave pulse. Figure 2b shows the temporal prole of the microwave pulses used in our experiment. The thermoacoustic pressure induced by the microwave pulses can be derived by the following convolution: p z,t

p 1 z , H z , t d .

where u ( z v s t ) is dened as a function that is unity when 0 ( z v s t ) d and zero otherwise. The impulse response is a traveling square wave as shown in Fig. 2a for a 4.8 mm gel slab if the microwave attenuation across the slab is negligible. Because the propagation speed of electromagnetic wave is much greater than the speed of sound, the sample volume illuminated by microwave pulses radiates acoustic waves simultaneously. The heating function in one-dimensional lossy media may be expressed as H ( z , t ) e z s ( t ), where is the microMedical Physics, Vol. 27, No. 5, May 2000

Figure 2c illustrates the thermoacoustic pressure at the ultrasonic transducer generated from the slab, which was obtained by convolving the two temporal waveforms in Figs. 2a and 2b by use of Eq. 6. For more general nonthermal-conned cases when heat transfer in the medium cannot be neglected, the following heat conduction equation must be taken into account in the pressure calculation:

Cp

T r, t k 2 T r, t H r, t , t

where is the density of the medium, k is the thermal con-

1198

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1198

ductivity, and T ( r, t ) is the temperature distribution in the microwave-illuminated space. The thermoacoustic pressure is: p r, t

4Cp
k2

1 H r , t r r t

T r , t d r . t

The integration involves the heat conducted from the surrounding medium besides the heat due to the absorbed microwave energy. The experimentally measured piezo-electric impulse response q ( t ) of the ultrasonic transducer is shown in Fig. 2d. The piezo-electric output of the ultrasonic transducer in response to thermoacoustic pressure can be calculated by the following convolution between the thermoacoustic pressure at the transducer and the impulse response of the transducer: P 0 z , t

p z, q t d.

The piezo-electric signal from a 4.8 mm slab was obtained by the convolution of the two waveforms in Figs. 2c and 2d with Eq. 9 and is plotted in Fig. 2e as a dashed line. The experimental piezo-electric signal from a 4.8 mm gel slab is plotted in Fig. 2e as a solid line for comparison, where the slight offset was caused by the dc drift of the amplier in the experimental detection. There are two dipolar structures in each waveform in Fig. 2e, and the signal between the dipolar structures is weak. The dipolar structures from the upper and lower surfaces of the slab have opposite polarities, which could not be explained by any existing theories in the literature. The polarity, the width of each dipole, and the distance between the two dipolar structures are in good agreement between the results of our theory and experiment. Because the piezo-electric signal of the ultrasonic transducer can be simulated by the two convolutions as shown above, the dipole width is related to the width of the microwave pulses and the width of the impulse response of the ultrasonic transducer, which were 0.5 s and 1.7 s, respectively. The time intervals between the zero-crossing points of the two dipolar structures in the two waveforms are determined by the slab thickness and are equal to the acoustic transit time 3.2 s over the 4.8-mm-thick slab. Dipolar structures were also observed in photoacoustics by lasers, where a Q-switched laser with a pulse width of 10 ns and a wide-band ultrasonic transducer were employed. Our detected dipolar structures resemble the wellknown dipolar structures that originate from small spherical or cylindrical objects excited by laser pulses39 or from acoustic reection at soft acoustic interfaces.40 However, our detected dipolar structures in the thermoacoustic signals in slabs resulted from the limited bandwidth of the ultrasonic transducers. This phenomenon can also be explained in the frequency domain, where the ultrasonic transducer acts as a bandpass lter. A frequency spectrum of the ultrasonic transducer is
Medical Physics, Vol. 27, No. 5, May 2000

FIG. 3. Frequency-domain analysis of the microwave-induced thermoacoustic signals. a Spectrum of the 1 MHz ultrasonic transducer; b Spectrum of the microwave-induced thermoacoustic signal; c Spectrum of the piezoelectric signal, which was the ltered microwave-induced thermoacoustic signal.

shown in Fig. 3a. The temporal prole of the thermoacoustic pressure varies sharply near the slab boundaries and slowly inside the slab as shown in Fig. 2c. The corresponding spectrum is peaked at dc as shown in Fig. 3b. The ltered spectrum is peaked near 0.5 MHz as shown in Fig. 3c, where the dc is rejected and the low-frequency components are attenuated signicantly. In other words, the ultrasonic transducer cannot respond efciently to the thermoacoustic waves emitted between the sample boundaries, which have a lower-frequency spectrum. Therefore, the observed piezo-electric signal between the two sample boundaries is low.

1199

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1199

This model includes only the key concept. Scattering and diffraction of microwaves becomes certainly important if strictly quantitative modeling is performed. The detected thermoacoustic signals are also related to other factors including microwave attenuation, conversion efciency from microwave energy to acoustic energy, acoustic scattering and diffraction, and acoustic attenuation. A tissue sample with a fat-muscle-fat structure as shown in Fig. 4a was placed on the plastic stand inside the container as shown in Fig. 1. Figure 4b shows a 2D image of the sample obtained with the scanning thermoacoustic tomography technique. Thermoacoustic signals were acquired in the time domain while the fat-muscle-fat sample was scanned horizontally along the y axis with a step size of 1 mm. The 2D image of the sample was formed by combining these temporal waveforms taken successively at the scanning stops along the y axis. Each vertical line in this 2D image was from a temporal waveform. The muscle inside the fat is clearly visible with a good contrast. The fat-tissue interface to the left of the muscle is also visible, which was possibly caused by the slight difference in the microwave properties between the two fat sections. Figure 4c illustrates a time-domain waveform that was measured above the center of the sample at y equal to 54 mm. The strongest dipole near 27 s was from the bottom boundary of the sample where microwave experienced the least attenuation, whereas the weakest dipole near 14 s was from the top boundary of the sample where microwave experienced the most attenuation. The two dipolar structures corresponding to the boundaries of the muscle layer are also clearly distinguishable. The time intervals between the adjacent dipolar structures agree with the thickness values of the tissue layers very well. However, the vertical boundaries of the muscle slab are not visible in the image because the thermoacoustic waves from these boundaries propagate perpendicularly to the acoustic axis of the ultrasonic transducer and therefore cannot be received by the transducer. The axial resolution along the acoustic axis z axis is determined by the width of the thermoacoustic dipolar structures, which is related to the width of the microwave pulse and the width of the impulse response of the transducer the inverse of the bandwidth of the ultrasonic detector. When the 1 MHz ultrasonic transducer was used, the width of the thermoacoustic signal was estimated to be 2.2 s, which was the sum of the width of the microwave pulses 0.5 s and the width of the impulse response of the transducer 1.7 s. Because the speed of sound in tissue is 1.5 mm/s, the corresponding axial resolution should be approximately 3.3 mm along the z axis. For the 3.5 MHz ultrasonic transducer, the axial resolution should be improved to 1.4 mm theoretically. Figure 5 shows the thermoacoustic signals from slab samples of various thickness values measured by the 1 MHz ultrasonic transducer. The slab samples were made of 5% gelatin and 5% NaCl, where NaCl controlled the microwave absorption. As the thickness of the samples decreased, the temporal distance between the adjacent dipolar structures corresponding to the two boundaries of the slabs decreased
Medical Physics, Vol. 27, No. 5, May 2000

FIG. 4. a Cross section of a fat-muscle-fat sample on the y - z plane; b Two-dimensional image of the y - z cross section of the sample obtained by scanning thermoacoustic tomography; c Temporal microwave-induced thermoacoustic signal along the vertical center line of the sample. The 1 MHz ultrasonic transducer was used in this experiment.

as well. The two dipolar structures became barely distinguishable when the thickness was reduced to 3.8 mm and completely inseparable when the thickness was reduced to 2 mm. Therefore, the experimentally measured axial resolution was 3.8 mm, close to the above calculated resolution of 3.3 mm based on the dipole width. The discrepancy was caused by the long tail of the dipolar structures. The relative variation in intensity between the two dipolar structures was caused by microwave attenuation in the slabs. Similarly, Fig. 6 shows the thermoacoustic signals from slab samples of various thickness values measured by the 3.5

1200

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1200

FIG. 5. Microwave-induced thermoacoustic signals in gel slabs of various thickness values: a 4.5 mm, b 3.8 mm, and c 2 mm. The 1 MHz ultrasonic transducer was used.

FIG. 6. Microwave-induced thermoacoustic signals in gel slabs of various thickness values: a 3.4 mm, b 1.9 mm, and c 1 mm. The 3.5 MHz ultrasonic transducer was used.

MHz ultrasonic transducer. The experimentally measured axial resolution was 1.9 mm, close to the above calculated resolution of 1.4 mm based on the dipole width. As expected from the theoretical consideration, the wider-bandwidth transducer produced better axial resolution. Potentially, shorter microwave pulses and deconvolution may be used to improve the axial resolution further.
Medical Physics, Vol. 27, No. 5, May 2000

The lateral resolution is determined by the numerical aperture of the ultrasonic transducer. The ultrasonic transducer responds to the thermoacoustic signals along its acoustic axis. The detected source area is related to the numerical aperture of the ultrasonic transducer and the distance between the thermoacoustic source and the ultrasonic transducer. The minimum detected source area is at the focal

1201

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1201

from the thermoacoustic waves propagating directly upward toward the ultrasonic transducer, whereas the dark lower band near z 9 mm is the ghost image caused by acoustic reection from the plastic stand. The gaps of greater than 4 mm can be easily recognized. The gap of 2.5 mm can be barely identied, which denes the lateral resolution and is comparable with the focal diameter of the ultrasonic transducer. The measured thermoacoustic signal is a convolution between the thermoacoustic signal in the sample and the detection-sensitivity distribution of the ultrasonic transducer over the detected area. The convolution reduces the lateral resolution, which is worsened when the ultrasonic transducer is out of focus. Figure 6c was acquired when the ultrasonic transducer was deliberately moved far away from the sample so as to create defocusing. The gaps are not distinguishable in the 2D image due to reduced resolution.

IV. CONCLUSIONS The simulation of the piezo-electric signal, by convolving the thermoacoustic signal of delta heating with the microwave-pulse waveform and then with the impulse response of the ultrasonic transducer, can be used to explain the bipolar experimental data. Our studies show that scanning microwave-induced thermoacoustic tomography is a promising imaging tool for biological tissue. The boundaries of different tissue constituents can be imaged clearly. The width of the microwave pulses and the bandwidth of the ultrasonic transducer determine the axial resolution along the acoustic axis of the ultrasonic transducer, which is 1.9 mm under our current experimental conditions. The axial resolution can be improved by employing broader-band ultrasonic transducers and shorter microwave pulses. The lateral resolution is determined by the numerical aperture of ultrasonic transducers and is 2.5 mm in our current setup.

FIG. 7. Two-dimensional tomographic images of a linear array of gel slabs obtained by scanning thermoacoustic tomography. a y - z cross section of the sample units in mm; b Image of the y - z cross section when the sample was placed at the focal plane of the ultrasonic transducer; c Image of the y - z cross section when the sample was placed far from the focal plane of the ultrasonic transducer.

ACKNOWLEDGMENTS Thanks to X. Zhao and G. Yao for their assistance in electronics and computer programming, respectively. This project was sponsored in part by the National Institutes of Health Grants Nos. R29 CA68562, R01 CA71980, and R21 CA83760 and by the National Science Foundation Grant no. BES-9734491.
a

plane of the ultrasonic transducer. Therefore, a better lateral resolution is expected when the sample is located within the focal column. We examined the lateral resolution with the 1 MHz ultrasonic transducer. The ultrasonic transducer has a 3 dB focal diameter of 2.1 mm and a focal zone of 17.6 mm along the acoustic axis. Several pieces of rectangular gel slabs were arranged linearly along the y direction as shown in Fig. 7a. The y - z cross section was imaged with a step size of 1 mm when the sample was scanned along the y axis. Figure 7b shows the two-dimensional image when the sample was on the focal plane of the ultrasonic transducer. The bright upper band near z 14 mm is the primary image
Medical Physics, Vol. 27, No. 5, May 2000

To whom correspondence should be addressed: Tel: 979 847-9040; Fax: 979 845-4450; electronic mail: lwang@tamu.edu; URL: http:// oilab.tamu.edu 1 L. E. Larsen and J. H. Jacobi, Eds., Medical Applications of Microwave Imaging IEEE Press, Piscataway, NJ, 1986. 2 J. C. Lin, Frequency optimization for microwave imaging of biological tissues, Proc. IEEE 73, 374375 1985. 3 S. Caorsi, A. Frattoni, G. L. Gragnani, E. Nortino, and M. Pastorino, Numerical algorithm for dielectric-permittivity microwave imaging of inhomogeneous biological bodies, Med. Biol. Eng. Comput. 29, NS3744 1991. 4 M. S. Hawley, A. Broquetas, L. Jofre, J. C. Bolomey, and G. Gaboriaud, Microwave imaging of tissue blood content changes, J. Biomed. Eng. 13, 197202 1991.

1202
5

G. Ku and L. V. Wang: Scanning thermoacoustic tomography

1202

P. M. Meaney, K. D. Paulsen, and J. T. Chang, Near-eld microwave imaging of biologically-based materials using a monopole transceiver system, IEEE Trans. Microwave Theory Tech. 46, 3145 1998. 6 J. C. Lin, On microwave-induced hearing sensation, IRE Trans. Microwave Theory Tech. MTT-25, 605613 1977. 7 F. Caspers and J. Conway, Measurement of power density in a lossy material by means of electromagnetically-induced acoustic signals for noninvasive determination of spatial thermal absorption in connection with pulsed hyperthermia, Proc. of the 12th European Microwave Conference, 565568 1982. 8 J. L. Su and J. C. Lin, Thermoelastic signatures of tissue phantom absorption and thermal expansion, IEEE Trans. Biomed. Eng. 43, 178 182 1987. 9 J. S. K. Wan, Microwaves and chemistry: The catalysis of an exciting marriage, Res. Chem. Intermed. 19, 147158 1993. 10 T. Bowen, L. Nasoni, A. E. Pifer, and G. H. Sembrosk, Some experimental results on the thermoacoustic imaging of soft tissue-equivalent phantoms, Ultrason. Symp. Proc. 2, 823827 1981. 11 R. G. Olsen, Generation of acoustic images from the absorption of pulsed microwave energy, in Acoustic Imaging, edited by J. P. Powers Plenum, New York, 1982, pp. 5359. 12 R. G. Olsen and J. C. Lin, Acoustic imaging of a model of a human hand using pulsed microwave irradiation, Bioelectromagnetics N.Y. 4, 397400 1983. 13 J. C. Lin and K. H. Chan, Microwave thermoelastic tissue imagingsystem design, IEEE Trans. Microwave Theory Tech. 32, 854860 1984. 14 R. L. Nasoni, G. A. Evanoff, Jr., P. G. Halverson, and T. Bowen, Thermoacoustic emission by deeply penetrating microwave radiation, Proc.IEEE Ultrason. Symp. 5, 633637 1984. 15 K. H. Chan and J. C. Lin, Microwave-induced thermoacoustic tissue imaging, Proc. Engineering in Medicine and Biology Society 10th Annual International Conference, 445446 1988. 16 R. A. Kruger, D. R. Reinecke, and G. A. Kruger, Thermoacoustic computed tomographyTechnical considerations, Med. Phys. 26, 1832 1837 1999. 17 R. A. Kruger, K. K. Kopecky, A. M. Aisen, D. R. Reinecke, G. A. Kruger, and W. L. Kiser, Jr., Thermoacoustic CT with radio waves: a medical imaging paradigm, Radiology 211, 275278 1999. 18 V. E. Gusev and A. A. Karabutov, Laser Optoacoustics American Institute of Physics, New York, 1993. 19 A. A. Oraevsky, S. L. Jacques, R. O. Esenaliev, and F. K. Tittel, Laserbased optoacoustic imaging in biological tissues, in Laser-Tissue Interaction V , edited by S. L. Jacques, Proc. Soc. Photo-Instrum. Eng. 2134A, 122128 1994. 20 R. A. Kruger and P. Liu, Photoacoustic ultrasound: Pulse production and detection of 0.5% Liposyn, Med. Phys. 21, 11791184 1994. 21 A. A. Oraevsky, S. L. Jacques, R. O. Esenaliev, and F. K. Tittel, Direct measurement of laser uence distribution and optoacoustic imaging in heterogeneous tissues, in Laser Interaction with Hard and Soft Tissue II, edited by H. J. Albrecht, G. P. Delacretaz, T. H. Meier, R. W. Steiner, L. O. Svaasand, and M. J. van Gemert, Proc. Soc. Photo-Instrum. Eng. 2323, 3746 1995. 22 R. A. Kruger, P. Liu, Y. R. Fang, and C. R. Appledorn, Photoacoustic ultrasound PAUSreconstruction tomography, Med. Phys. 22, 1605 1609 1995. 23 A. A. Oraevsky, R. Esenaliev, F. K. Tittel, M. R. Ostermeyer, L.-H. Wang, and S. L. Jacques, Laser opto-acoustic imaging of turbid media:

determination of optical properties by comparison with diffusion theory and Monte Carlo simulation, Proc. SPIE 2681, 277284 1996. 24 A. A. Oraevsky, R. O. Esenaliev, and A. A. Karabutov, Laser optoacoustic tomography of layered tissues: signal processing, in Optical Tomography and Spectroscopy of Tissue: Theory, Instrumentation, Model, and Human Studies II, edited by B. Chance and R. R. Alfano, Proc. Soc. Photo-Instrum. Eng. 2979, 5970 1997. 25 R. O. Esenaliev, H. Alma, F. K. Tittel, and A. A. Oraevsky, Axial resolution of laser opto-acoustic imaging: inuence of acoustic attenuation and diffraction, in Laser-Tissue Interaction IX, edited by S. L. Jacques, Proc. Soc. Photo-Instrum. Eng. 3254, 294306 1998. 26 C. G. A. Hoelen, F. F. M. Demul, R. Pongers, and A. Dekker, Threedimensional photoacoustic imaging of blood vessels in tissue, Opt. Lett. 23, 648650 1998. 27 A. A. Oraevsky, V. A. Andreev, A. A. Karabutov, and R. O. Esenaliev, Two-dimensional optoacoustic tomography: Transducer array and image reconstruction algorithm, in Laser-Tissue Interaction X: Photochemical, Photothermal, and Photomechanical, edited by S. L. Jacques, G. J. Mueller, A. Roggan, and D. H. Sliney, Proc. Soc. Photo-Instrum. Eng. 3601, 256267 1999. 28 A. A. Karabutov, E. B. Savateeva, and A. A. Oraevsky, Imaging of layered structures in biological tissues with optoacoustic front surface transducer, in Laser-Tissue Interaction X: Photochemical, Photothermal, and Photomechanical, edited by S. L. Jacques, G. J. Mueller, A. Roggan, and D. H. Sliney, Proc. Soc. Photo-Instrum. Eng. 3601, 284 295 1999. 29 C. Gabriel, S. Gabriel, and E. Corthout, The dielectric properties of biological tissues: I. literature survey, Phys. Med. Biol. 41, 22312249 1996. 30 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: II. Measurements in the frequency range 10 Hz to 20 GHz, Phys. Med. Biol. 41, 22512269 1996. 31 S. Gabriel, R. W. Lau, and C. Gabriel, The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues, Phys. Med. Biol. 41, 22712293 1996. 32 F. A. Duck, Physical Properties of Tissue: A Comprehensive Reference Book Academic, London, 1990. 33 W. Joines, R. Jirtle, M. Rafal, and D. Schaeffer, Microwave power absorption differences between normal and malignant tissue, Int. J. Radiat. Oncol. Biol. Phys. 6, 681687 1980. 34 S. Chaudhary, R. Mishra, A. Swarup, and J. Thomas, Dielectric properties of normal human breast tissues at radiowave and microwave frequencies, Indian J. Biochem. Biophys. 21, 7679 1984. 35 W. Joines, Y. Zhang, C. Li, and R. Jirtle, The measured electrical properties of normal and malignant human tissues from 50900 MHz, Med. Phys. 21, 547550 1994. 36 L.-H. Wang, X. Zhao, H. Sun, and G. Ku, Microwave-induced acoustic imaging of biological tissues, Rev. Sci. Instrum. 70, 37443748 1999. 37 I. N. Bronshtein and K. A. Semendyayev, Handbook of Mathematics Van Nostrand Reinhold, New York, 1978. 38 G. J. Diebold and T. Sun, Properties of photoacoustic wave in one two and three dimensions, Acustica 80, 339351 1994. 39 C. G. A. Hoelen, F. F. M. de Mul, and J. Greve, Non-destructive photoacoustic subsurface tissue imaging: a feasibility study, Proc. SPIE 2628, 304314 1995. 40 A. O. Oraevsky, S. L. Jacques, and F. K. Tittel, Measurement of tissue optical properties by time-resolved detection of laser-induced transient stress, Appl. Opt. 36, 402415 1997.

Medical Physics, Vol. 27, No. 5, May 2000

REVIEW OF SCIENTIFIC INSTRUMENTS

VOLUME 70, NUMBER 9

SEPTEMBER 1999

Microwave-induced acoustic imaging of biological tissues


Lihong V. Wang, Xuemei Zhao, Haitao Sun, and Geng Ku
Optical Imaging Laboratory, Biomedical Engineering Program, Texas A&M University, College Station, Texas 77843-3120

Received 19 March 1999; accepted for publication 15 June 1999 We present tomographic imaging of biological tissues by use of microwave-induced acoustic signal. It was demonstrated that the acoustic signal was proportional to the intensity of the incident microwave and was related to the absorption property of microwave in the medium. Pulsed microwave radiation was used to illuminate the samples. Absorbed microwave energy caused thermoelastic expansion that radiated acoustic waves. A focused ultrasonic transducer detected the time-resolved acoustic signals. Each acoustic signal was converted into a one-dimensional image. A linear scanning of the ultrasonic transducer yielded multiple one-dimensional images, which formed a two-dimensional image. The imaging contrast is based on the difference in the dielectric constants among biological tissues. Because of the large contrast in microwave absorption among different tissue types, microwave-induced acoustic tomography could potentially provide a new modality for detecting early-stage cancers. 1999 American Institute of Physics. S0034-67489904709-7

I. INTRODUCTION

The electric eld strength in a lossy media is attenuated exponentially as E E 0 exp z , 1

Microwave imaging of biological tissues has been investigated for a number of years.15 The advantages of the technique include the use of nonionizing radiation and high imaging contrast. However, the technique has had difculties in multichannel detection of microwave without cross coupling, in the reconstruction algorithms, and in achieving good spatial resolution. Ultrasound imaging ultrasonography, an established medical imaging modality, can yield good spatial resolution but has poor contrast. Microwave-induced acoustics MIA may bridge the gap and combine the advantages of the two types of radiation. The phenomenon that pulsed microwave generates acoustic signals in lossy media has been recognized since long ago. This phenomenon was called microwave auditory or MIA. MIA was used to quantify physical parameters in media such as the power density and the concentration of a given substance.68 MIA was also employed by several investigators in the 1980s for imaging of biological tissues or tissue phantoms.914 However, these early works did not produce any tomographic or depth-resolved images. Recently, images of biological tissue were reconstructed based on MIA.1518 This approach requires measurements of a large amount of data around the tissue and expensive computation following the data acquisition. Microwave-induced acoustic imaging MIAI is based on the detection of the acoustic signals generated by microwave in the samples. Pulsed microwave radiation is used to illuminate the samples. Absorbed microwave energy causes thermoelastic expansion that radiates acoustic waves. An ultrasonic transducer or an array detects the time-resolved acoustic signals. Although MIAI shares similar principles with the photoacoustic imaging in the optical-wavelength region,1922 MIAI may have broader applications in medical imaging because microwave has a greater contrast and a deeper penetration depth in biological tissues than light.
0034-6748/99/70(9)/3744/5/$15.00 3744

where E 0 is the electrical eld at the sample surface, E is the electrical eld at the depth z, and is the electric-eld absorption coefcient expressed as

2 1

1 ,

where is the angular frequency, is the permeability, is the permittivity, and is the conductivity. The induced acoustic pressure is expected to depend on the intensity of microwave and the complex dielectric constant of the material. In the frequency range of 0.110 GHz, the dielectric constant ratio of the permittivity in material to that in vacuum has a value of 570 for soft tissues, and the conductivity has a value of 0.023 s/m.2325 The dielectric properties of tissue determine the absorption of microwave at various frequencies of microwave. At 9.4 GHz, the penetration depths, which are the inverse of the absorption coefcients, are 3.4 and 0.35 cm for fat and muscle, respectively; while at 500 MHz, the penetration depths for fat and muscle are 23.5 and 3.4 cm, respectively.26 Most of the other soft tissues have an absorption coefcient in between those of muscle and fat. This wide range of values can provide an extremely high imaging contrast for biological tissues.2 Therefore, MIAI may potentially be used to detect earlystage cancers. We here present our study on MIAI toward biomedical applications. First, some basic properties of MIA were examined using various tissue phantoms. Then, twodimensional cross sections of biological tissues were imaged using MIAI. Our imaging approach differs signicantly from the prior arts in MIAI. Lateral resolution was achieved by use of a focused ultrasonic transducer. Axial resolution was
1999 American Institute of Physics

Rev. Sci. Instrum., Vol. 70, No. 9, September 1999

Acoustic imaging of biological tissues

3745

FIG. 1. Experimental setup.

obtained by measuring the temporal proles of the microwave-induced acoustic signals. Depth-resolved tomographic images were acquired directly without resorting to image reconstruction.
II. EXPERIMENTAL SETUP

FIG. 2. Dependence of the microwave-induced acoustic pressure on the microwave energy density per pulse. The test sample was an 8-mm-thick gel made of 5% gelatin and 95% water.

The experimental setup used for this study is shown in Fig. 1. A Cartesian coordinate system was set up for reference. The z axis was along the ultrasonic axis pointing upward. The x axis was perpendicular to the drawing plane and pointed outward. The y axis was in the drawing plane pointing to the right. A 9.4-GHz microwave generator transmitted microwave pulses of 10-kW peak power into the mineral oil bath through an X-band 16-dB horn antenna. The aperture of the horn antenna was 55 mm long and 74 mm wide. The pulse width of microwave was set to either 2.2 or 0.6 s in our experiments. A function generator DS345, Stanford Research System was employed to trigger the microwave generator and control its pulse repetition frequency. The mineral oil bath was contained in a Plexiglas container and was xed on a two-dimensional x-y translation stage MD2, Arrick Robotics. A personal computer controlled the two stepper motors to drive the translation stage in the x and y directions. The object to be scanned was put on a Plexiglas stand inside the bath. Mineral oil has a very small absorption coefcient for microwave and allowed good acoustic coupling. Plexiglas has low microwave absorption as well. An ultrasonic transducer was immersed in the mineral oil facing the microwave antenna. Two ultrasonic transducers were used in this study. For the rst one V314, Panametrics, the central frequency of the ultrasonic transducer was 1 MHz, the bandwidth was 0.65 MHz, the diameter was 1.9 cm, and the focal length at 1 MHz was 2.5 cm. For the second one V384, Panametrics, the central frequency was 3.5 MHz, the bandwidth was 2.5 MHz, the diameter was 0.64 cm, and the focal length at 3.5 MHz was 1.8 cm. The transducer was connected to a pulse amplier 500PR, Panametrics. The amplied signal was recorded by an oscilloscope TDS-640A, Tektronix and transferred to a personal computer.
III. RESULTS

bath. The acoustic pressure versus the input energy density is plotted in Fig. 2. The input energy density was varied by adjusting the cathode current of the magnetron M602, Toshiba in the microwave generator. The result shows that the acoustic pressure increased linearly with the input density as expected. In Fig. 3, the time-domain acoustic waveforms induced by the microwave in its full power and 12% of the full power are compared. The full-power microwave had a peak power of 10 kW and an energy density of 0.55 mJ/cm2 . The 12% reduced source power yielded a surface energy density equivalent to the energy density at the penetration depth when the sample was illuminated with the full power. The signal-to-noise ratio at the penetration depth was expected to be about 3:1 when the full-power microwave was used in our system based on this measurement. This type of experiments may be used for selecting necessary microwave energy density in MIAI applications. Three gel slabs were made of the same amount of gelatin 5% and different amounts of NaCl 0%, 3%, and 5%. The dimensions of the slabs were 30, 30, and 10 mm in the x, y, and z directions, respectively. The ultrasonic transducer was scanned across each of the slabs along the x direction while

In order to study the dependence of the acoustic pressure on the microwave power, we tested a gel made of 5% gelatin and 95% water. The gel was a slab with a thickness of 8 mm and an area much greater than the receiving surface of the ultrasonic transducer. The gel was buried in the mineral oil

FIG. 3. Waveforms of the microwave-induced acoustic pressure when the sample was exposed to the full power of the microwave generator 0.55 mJ/cm2 ) and to 12% of the full power.

3746

Rev. Sci. Instrum., Vol. 70, No. 9, September 1999

Wang et al.

FIG. 4. One-dimensional images of the gel slabs that contain 0%, 3%, and 5% NaCl, respectively. The slabs were 30 mm wide and 10 mm thick, located between 0 and 30 mm on the horizontal axis.

the microwave-induced acoustic signals were recorded. Figure 4 shows the scanned one-dimensional images based on the peak amplitude of the acoustic signals from the bottom surface of the slabs. The peak amplitude increased with increased NaCl concentration because increasing NaCl led to an increase in the conductivity of microwave. According to Eq. 2, the energy absorbed by the gel would increase with an increased conductivity and result in an increased acoustic amplitude. These three samples had similar values of acoustic impedance. This experiment demonstrated the contrast mechanism of MIAI compared with that of the conventional ultrasound imaging. MIAI depends on the difference in dielectric constant among different tissue types or materials, while ultrasound imaging depends on the difference in acoustic impedance. Therefore, MIAI has potential to image objects that are invisible with ultrasound imaging. To examine soft tissue contrast, we scanned in the x-y plane a tissue sample containing muscle and fat. The dimensions of the sample were 60, 60, and 15 mm along the x, y, and z directions, respectively. The fat at the center of the sample was a 15-mm tall cylinder with a diameter of 16 mm. Figure 5a shows a two-dimensional image based on the peak amplitude of the acoustic signals generated at the bottom surface of the sample. Figure 5b shows a onedimensional image along the horizontal line across the center of the two-dimensional image. A good soft tissue contrast was observed, where the peak acoustic amplitude from the muscle was approximately six times as much as that from the fat. Since the broad ( 2.2 s microwave pulses and the low-frequency 1 MHz ultrasonic transducer were used for the experiments shown in Figs. 4 and 5, the lateral resolution was not very good. If the microwave pulses are shortened and the bandwidth of the ultrasonic transducer is broadened, the frequency of the received acoustic signal would be increased leading to improved resolution. However, the key point of these two experiments was to study the contrast. When a microwave pulse illuminated a sample and deposited energy into the sample, acoustic signals were generated wherever microwave energy was absorbed. The ultrasonic transducer detected the time of arrival signal of the

FIG. 5. a Two-dimensional surface image of a tissue sample by use of MIAI. The dark center of the image corresponded to the fat, the bright portion in the image corresponded to the muscle, and the outer black portion corresponded to the mineral oil background. b One-dimensional image along the horizontal line across the center of the two-dimensional image in a.

acoustic sources. The distance between the acoustic sources and the transducer was calculated by multiplying the time of arrival and the speed of sound in the medium. Therefore, a time-domain signal could be converted into a onedimensional image along the transducer axis z axis. Scanning the transducer across the sample along the x or y axis would yield a two-dimensional cross-sectional image of the sample in the x-z or y-z plane. Figure 6 shows a typical acoustic signal from a slice of fat tissue as a function of the relative time of arrival. The rst pulse near 5 s came from the top surface of the fat tissue. The second pulse near 10 s originated from the bottom surface of the fat tissue. The third pulse near 15 s was from the downward propagating acoustic wave originated from the top surface but reected by the bottom surface of the fat

Rev. Sci. Instrum., Vol. 70, No. 9, September 1999

Acoustic imaging of biological tissues

3747

FIG. 6. Sample waveform of microwave-induced acoustic pressure in an 8-mm-thick fat tissue.

tissue. Because the thickness of the sample was 8 mm and the speed of sound was 1.5 mm/ s, the expected time interval between the adjacent pulses should be 5.3 s, which matched the observed values very well. The ultrasonic transducer had a central frequency of 3.5 MHz and a 6-dB bandwidth of 2.5 MHz. Therefore, the transducer acted as a band-pass lter and rejected the low-frequency acoustic signal originated between the two tissue boundaries. The sample was translated along the x axis with a step size of 2.6 mm. At each step, the time-domain signal similar to that in Fig. 6 was recorded, which revealed the tissue structure along the z axis. Two-dimensional x-z crosssectional images were then formed by combining the recorded signals at all of the steps, where the negative portion of the signals was set to zero. Three different tissue samples were imaged. Figure 7a shows the image of an 8-mm-thick fat slab. The top and bottom surfaces were clearly visible. Figure 7b shows the image of an 8-mm-thick fat slab with a U-shaped ditch in the middle. The ditch had a width of 20 mm and a depth of 5 mm. The observed maximum thickness and minimum thickness in the image were 8 and 3 mm, respectively, which agreed with the actual size of the tissue sample. Figure 7c shows the image of a similar 8-mm-thick fat slab with a U-shaped ditch in the middle lled with muscle. The bottom surface of the muscle was visible, but the top surface was invisible because of the high attenuation of microwave at 9.4 GHz in muscle. Microwave with lower frequencies would be necessary to image thicker tissue samples because of the deeper penetration depth of microwave at lower frequencies. The axial resolution along the z axis was determined by the pulse width of the microwave and the bandwidth of the ultrasonic transducer. We observed that the 3.5-MHz ultrasonic transducer yielded better axial resolution than the 1-MHz transducer. For the results shown in Figs. 6 and 7, the pulse width of the microwave signal was 0.6 s, and the temporal width of the acoustic signal was limited to 1 s. The corresponding axial resolution was approximately 1.5 mm along the z axis under the present experimental conditions. The lateral resolution perpendicularly to the z axis was determined by the numerical aperture and the bandwidth of

FIG. 7. Two-dimensional tomographic images of samples with different shapes by use of MIAI. a An 8-mm-thick, 30-mm-wide fat tissue slab. b An 8-mm-thick fat tissue slab with a U-shaped ditch in the middle, which was 20 mm long and 5 mm wide. c A similar fat tissue slab as in b but having a ditch lled with muscle.

the ultrasonic transducer and by the pulse width of the microwave as well. The lateral resolution of our current setup was 10 mm and was to be improved by optimizing the parameters of the ultrasonic transducer and the microwave transmitter. Synthetic aperture technique may be used to further improve the lateral resolution.22
IV. DISCUSSION

Our experimental results demonstrated that microwaveinduced acoustic pressure was proportional to the intensity of the incident microwave. The difference in dielectric constant between different tissue types should provide good imaging contrast in MIAI, which is a different contrast mechanism from that in the conventional ultrasound imaging. We showed that the MIAI technique was able to generate tomog-

3748

Rev. Sci. Instrum., Vol. 70, No. 9, September 1999


7 8 9

Wang et al. J. L. Su and J. C. Lin, IEEE Trans. Biomed. Eng. 43, 178 1987. J. S. K. Wan, Rev. Chem. Intermed. 19, 147 1993. T. Bowen, L. Nasoni, A. E. Pifer, and G. H. Sembrosk, Proc. IEEE Ultrasonics Symposium 2, 823 1981. 10 R. G. Olsen, in Acoustic Imaging, edited by J. P. Powers Plenum, New York, 1982, pp. 5359. 11 R. G. Olsen and J. C. Lin, Bioelectromagnetics N.Y. 4, 397 1983. 12 J. C. Lin and K. H. Chan, IEEE Trans. Microwave Theory Tech. 32, 854 1984. 13 R. L. Nasoni, G. A. Evanoff, Jr., P. G. Halverson, and T. Bowen, Proc. IEEE Ultrasonics Symposium 5, 633 1984. 14 K. H. Chan and J. C. Lin, Proceedings of Engineering in Medicine and Biology Society 10th Annual International Conference, 1988, pp. 445 446 . 15 R. A. Kruger, D. R. Reinecke, and G. A. Kruger, Medical Physics to be published. 16 R. A. Kruger, K. K. Kopecky, A. M. Aisen, D. R. Reinecke, G. A. Kruger, and W. L. Kiser, Jr., Radiology 211, 275 1999. 17 R. A. Kruger, W. L. Kiser, Jr., D. R. Reinecke, and G. A. Kruger, presented at SPIE Conference on Physics of Medical Imaging 3659-43, 2026 February, 1999. 18 W. L. Kiser, Jr. and R. A. Kruger, presented at SPIE Conference on Physics of Medical Imaging 3659-96, 2026 February, 1999. 19 R. A. Kruger and P. Liu, Med. Phys. 21, 1179 1994. 20 R. A. Kruger, P. Liu, Y. R. Fang, and C. R. Appledorn, Med. Phys. 22, 1605 1995. 21 A. A. Oraevsky, R. Esenaliev, F. K. Tittel, M. R. Ostermeyer, L.-H. Wang, and S. L. Jacques, Laser-Tissue Interaction VII 2681, 277 1996. 22 C. G. A. Hoelen, F. F. M. Demul, R. Pongers, and A. Dekker, Opt. Lett. 23, 648 1998. 23 C. Gabriel, S. Gabriel, and E. Corthout, Phys. Med. Biol. 41, 2231 1996. 24 S. Gabriel, R. W. Lau, and C. Gabriel, Phys. Med. Biol. 41, 2251 1996. 25 S. Gabriel, R. W. Lau, and C. Gabriel, Phys. Med. Biol. 41, 2271 1996. 26 F. A. Duck, Physical Properties of Tissue: A Comprehensive Reference Book Academic, London, New York, 1990.

raphic images of biological tissues. The axial resolution along the ultrasonic axis depended on the pulse width of the microwave signal and the bandwidth of the ultrasonic transducer. Combining microwave and ultrasound overcame the resolution problem and the reconstruction problem in the purely microwave imaging. In the future studies, we plan to lower the microwave frequency and shorten the pulse width to image thicker tissue samples with improved resolution. This technique may also be applied in measurements of the dielectric constants of other materials or of the SAR patterns of electromagnetic elds.
ACKNOWLEDGMENTS

This project was sponsored in part by the National Institutes of Health Grant Nos. R29 CA68562 and R01 CA71980 and by the National Science Foundation Grant No. BES9734491.
1

Medical Applications of Microwave Imaging, edited by L. E. Larsen and J. H. Jacobi IEEE, Piscataway, NJ, 1986. 2 J. C. Lin, Proc. IEEE 73, 374 1985. 3 S. Caorsi, A. Frattoni, G. L. Gragnani, E. Nortino, and M. Pastorino, Med. Biol. Eng. Comput. 29, NS37 1991. 4 M. S. Hawley, A. Broquetas, L. Jofre, J. C. Bolomey, and G. Gaboriaud, J. Biomed. Eng. 13, 197 1991. 5 P. M. Meaney, K. D. Paulsen, and J. T. Chang, IEEE Trans. Microwave Theory Tech. 46, 31 1998. 6 F. Caspers and J. Conway, Proceedings of the 12th European Microwave Conference, 1982, pp. 565568.