Open Access
Add to BibSonomyAbstract
We present a fast full-view photoacoustic imaging system for visualizing tissue structures using a linear transducer array with combined scan. In this system, a 128-element linear transducer array was used to detect photoacoustic signals by combined scanning of electronic scan and mechanical scan. An improved limited-field filtered back projection algorithm with directivity factors was applied to reconstruct the optical absorption distribution. The experiments on phantoms and in vivo blood vessels in a rat brain were performed with this system. And a clear view of the curve boundaries of objects and the network of blood vessels of rat’s brain were acquired. The experimental results demonstrate the multielement photoacoustic imaging system has the ability of imaging complicated structures of objects.
©2007 Optical Society of America
1. Introduction
Photoacoustic imaging (PAI) is a rapidly developing imaging modality of medical diagnostics that can noninvasively measure optical absorption heterogeneity of tissues by detecting photoacoustic signals generated by absorbed radiation energy of short laser pulses. Because PAI combines the merits of high tissue contrast of pure optical imaging and high spatial resolution of pure ultrasonic imaging, it has been developed as a promising tool for early detection of breast tumors or cancer [1–3], imaging of vasculature structure [4–6]. Photoacoustic technique has also been applied to measurement of epidermal melanin [7, 8], monitoring of oxygenation [9, 10] and glucose [11–13] in blood vessels, monitoring of vascular damage during photodynamic therapy [14], by time-resolved detection of laser-induced pressure profiles.
To be accepted as a diagnostic medical imaging technique, PAI method should provide images in a short time. Fast imaging can not only reduce motion artifacts by respiration or heartbeat but also allow the doctor to find crucial regions of interest and to instantly diagnose the investigated region. To realize fast photoacoustic imaging, transducer array with multi-channel parallel acquisition system, and high repetition excitation sources with high-peak power are usually required. In the past decades, many researchers have focused on multielement PAI. Oraevsky et al. applied a 128-channel photoacoustic imaging system for breast cancer diagnostics [15]. Niederhauser et al. used a transducer array for vascular imaging [5]. Wang et al. used a high- frequency ultrasound array transducer for photoacoustic imaging of microvasculature [16]. Xing et al. applied a 320-element array transducer to obtain the images of phantom and in vivo blood vessels [17–19]. Tang et al. used a 4f acoustic lens imaging system with a linear transducer array to achieve the photoacoustic images of phantoms [20]. The multi-element PAI has the potential to address some needs of diagnosing and monitoring the diseases, but the imaging quality need to be further improved.
In our previous works[17, 18], we applied the limited-field filtered back projection (LFBP) algorithm to reconstruct the photoacoustic images of the multi-element PAI system, which greatly improved the lateral resolution of the multi-element imaging system. The imaging system provided B-mode photoacoustic images with high soft-tissue contrast and good spatial resolution within a few seconds, but there was difficulty for this system to image complicated structures of objects. In fact, when the transducer array is placed at one position to receive signals, only the boundaries of the targets that are nearly perpendicular to the axis of the transducer array can be clearly imaged, because most of the photoacoustic waves travel in a small angle around the normals of boundaries and can be detected by the transducer array; spherical or oblique boundaries of targets can not be imaged clearly, because the induced photoacoustic waves travel in a large angle with the axis of the transducer array and the transducer array receives little signals from the boundaries [21]. In theory, a piece of boundary of object can be stably reconstructed as soon as at any position on the boundary at least one of two normal directions passes through a transducer position [22].
To improve the ability of imaging complicated structures of targets with the multi-element photoacoustic imaging system, we use the linear transducer array with combined scanning mode to detect photoacoustic signals at multiple locations on a circle around the sample. The LFBP algorithm is also modified to reconstruct the photoacoustic images. Weights are assigned to the signals according to the directivity pattern functions of the individual transducer and the total transducer array to further improve the lateral resolution. Experiments on phantoms and blood vessels in rat’s brain were performed and clear images were obtained. The experimental results demonstrate that the multi-element PAI system has the ability of imaging complicated structures of objects.
2. Materials and methods
2.1 Imaging principle
When a linear transducer array is used to detect laser-induced photoacoustic signals, because individual transducer of the transducer array is of finite size, the directivity of transducer must be taken into account. Let the transducer array consist of N rectangular transducers with spacing d between the piezo-elements, and the dimension of individual piezo-element is a × b. L is the total length of the linear transducer, L = N × d (see in Fig. 1). Here we only consider the directivity perpendicular to the short side of the transducer array.
The directivity pattern function of individual transducer of the transducer array can be expressed as [23]
where λ is the ultrasound wavelength, and θ is the incidence angle of acoustic waves approaching to the transducer. And the directivity pattern function of the transducer array D(θ) can be expressed
when the linear transducer array receives the photoacoustic signals by circularly scanning the tissues, the improved LFBP algorithm uses the directivity pattern functions of individual transducer and the total transducer array as weighting factors assigned to signals to reconstruct two-dimensional photoacoustic images. In this algorithm, the image value Sf in tissue pixel f is determined as
where N is the number of elements of the linear transducer array; and K is the total number of the transducer array scanning positions. Smn(t) is the recorded signal of the transducer n of the linear transducer array at the scanning position m. dfmn and Dfm are the directivity pattern functions taking into account the angular sensitivity of the transducer n and the total transducer array at position m for signals generated in pixel f, which can be calculated according to Eq(1) and Eq(2). τfmn is the arrival time delay corresponding to the distance from the transducer n at scanning position m to the pixel f
2.2 Imaging system
A new B-mode system (Model CTS-200, SIUI, China) was modified as a multi-element linear transducer array system (MLTAS) for PAI. The resonance frequency of the linear transducer array is 7.5 MHz, and the spanning width is 49 mm. The sensitivity of the transducer array is approximately 1 mV/ Pa. Each transducer of the transducer array has a thin cylinder ultrasonic lens that produces a geometric focus approximately 35mm in front of the transducer array to select 2D image plane and suppress the out-of-plane signals. The signals from the transducers, after pre-amplification and phase adjustment, are acquired with the data acquisition (DAS) card (compuscope 12100, Gage Applied Co., Montreal, Quebec, Canada). The card features a high-speed 12-bit analogue-to-digital converter with a sampling rate of 100 MHz. The system operation and data acquisition are controlled by a personal computer.
The schematic of the experimental setup is shown in Fig. 2. An Nd: YAG pumped laser (Ultra, USA, wavelength λ =532nm, pulse width τ =7 ns, pulse-energy E up to 200mJ, and repetition rate r=20) provides laser pulses to irradiate the sample for generating photoacoustic signals. A piece of ground glass was used to homogenize the laser beam. The incident energy density was controlled below 20mJ/cm 2. A 20Hz clock signal, provided by control circuit, was used for triggering the pulse laser and controlling the individual transducer of the linear transducer array to be dynamically selected to perform the linear scanning. And the linear transducer array was driven by a stepper motor to circularly rotate around the sample along a 40 mm diameter circle to perform the circular scanning. Thus this combined scanning is the combination of electronic scanning and mechanical scanning. The linear transducer array and the samples were both immersed in a tank of water for better coupling. The induced photoacoustic waves were captured every 18 degrees, and total 20 positions photoacoustic waves were recorded for a full view of 2π circularly scanning angle. After the 128×20 series data were transferred to a personal computer for further data processing using MATLAB (version 7.0, Mathworks), projections were calculated with the improved LFBP algorithm.
3. Results
To demonstrate that the improved LFBP algorithm with weighting factors can improve the lateral resolution of the imaging system, the first experiment was performed. A pencil rod (the length=7mm, the diameter=0.7mm) was embedded in a cylindrical phantom at a depth of 3 mm. The phantom made with 13% gelatin, 12.5% milk and 74.5% water, was used to simulate the optical properties of human breast. The phantom had an effective attenuation coefficient 1.2cm-1.The transducer array was placed parallel to the rod to detect photoacoustic signals in this experiment. Fig. 3(a) and Fig. 3(b) were reconstructed images with the LFBP algorithm and the improved LFBP algorithm, respectively. As it was expected, the lateral resolution of the reconstructed image with the improved LFBP algorithm was better than that with the LFBP algorithm. And the artifacts in Fig. 3(b) are obviously less than that in Fig. 3(a). Fig. 3(c) and Fig. 3(d) are the pixel profiles of the reconstruction images shown in Fig. 3(a) and Fig. 3(b) with y=30mm. The spatial resolution of the imaging system is estimated according to the resolution criterion defined in the literature [24]; the lateral resolutions are calculated as 0.24mm in Fig. 3(c) and 0.18 mm in Fig. 3(d), respectively.
Fig. 3. (a) Reconstructed image of the phantom with the LFBP algorithm. (b) Reconstructed image of the phantom with the improved LFBP algorithm. (c) The line normal to object axis profile of the reconstructed image shown in Fig. 3 (a) with y=30mm. (d) The line normal to object axis profile of the reconstructed image shown in Fig. 3 (b) with y=30mm.
To test the ability of computed tomography of the PAI system, the second experiment was made. A point source was fabricated by painting a small dot of black paint on a polyethylene thread with diameter of 0.2mm (shown in Fig. 4 (a)). The thread was fixed on positioner, and the laser beam vertically illuminated on the dot. The point source was positioned 35mm in front of the transducer array and the z-axis position of the linear transducer array was adjusted until the amplitude of the photoacoustic signal was a maximum. When so positioned, we assumed the point source was located within the imaging plane of the transducer. Then we recorded the amplitude photoacoustic signal as a function of the z position. We interpreted this response as the slice-width profile shown in Fig. 4(b). The full width at half maximum (FWHM) of the slice-width profile was measured about 1.6mm.
Fig. 4. (a) The schematic of measuring the slice width profile of the linear transducer array. (b) The amplitude slice-width profile of the linear transducer array to a point source placed at the position of 35mm in front of the linear transducer array.
In order to verify the ability of the imaging system for reconstructing the curved boundaries of sample, the third experiment was performed. Two thin elliptical absorbers with diameters of 2.5 mm and 4 mm, and a height of 0.5 mm, which were made of the same mixture as that used in the first experiment and a drop of 0.1% dark ink solution, which had an effective attenuation coefficient 4.2cm-1, were used to simulate breast tumors with higher optical absorption. Two absorbers were embedded in cylindrical phantom at a depth of 8 mm. Fig. 5 is the reconstructed photoacoustic image of two thin elliptical absorbers, and the inserted is the photograph of the cross section of the phantom. Fig. 5(a) is the reconstructed image using the improved LFBP algorithm with the directivity pattern functions of individual transducer d(θ) and the total transducer array D(θ) ; Fig. 5(b) is the reconstructed image applying the improved LFBP algorithm only with the directivity pattern function of individual transducer d(θ) . As we can see the artifacts in Fig. 5(a) is obviously less than that in Fig .5(b). It demonstrates that using directivity pattern function of the total transducer array as the weighting factor can reduce the artifacts when image is reconstructed. Because of the highly scattering effect of light in turbid media, the intensity of light and the photoacoustic signal-to-noise ratio (SNR), decrease with depth exponentially with a decay constant of a few millimeters. So the contrast of the reconstructed image is not good. The reconstructed images can agree well with the original sample. The experiment demonstrates that the photoacoustic computed tomography system has the ability of reconstructing the curved boundaries of targets with signals collected by linear transducer array circularly scanning sample.
Fig. 5. Reconstructed image of the two elliptical absorbers embedded in phantom (inset: cross section of the phantom). (a) is the reconstructed photoacoustic image using the improved LFBP algorithm with the directivity pattern functions of individual transducer and the total transducer array. (b) is the reconstructed image using the improved LFBP algorithm with the directivity pattern function of individual transducer.
The fourth experiment was performed to further test the ability of the imaging system to image complicated structure of targets. Five black cotton threads which were embedded in cylindrical phantom (made of the same mixture as that was used in the first experiment) at a depth of 10 mm were used to simulate the network of blood vessels. Fig. 6 is the reconstructed image of the simulated blood vessels, and the inserted is photograph of the cross section of the sample. The reconstructed image of the simulated blood vessels is in good agreement with the sample, but the contrast of the reconstruction image is not good and the image resolution is only in mm due to the attenuations of light intensity and photoacoustic signals. This experiment verifies that the imaging system with transducer array circularly scanning sample is of the ability of imaging complicated structural tissues, such as the network of blood vessels.
Fig. 6. Reconstructed image of the network of simulated blood vessels (inset: cross section of the phantom).
We also conducted in vivo experiment with rat to further evaluate the ability of imaging the network of blood vessels. A male rat weighting about 120g was used for the experiment. Before experiment, the hair on the head of the rat was removed with commercial hair remover lotion. General anesthesia was administered on the rat by intraperitoneal injection of sodium pentobarbital with the dose of 40 mg/kg. As required, supplemental injections of sodium pentobarbital (10 mg/kg/h) were performed to keep the rat motionless throughout the experiment. The rat was protruded into the water tank through a hole in the bottom of tank while a piece of clear membrane between the water and the rat seals the hole. A thin layer of ultrasonic coupling gel was applied on the surface of the animal head to couple the head and the membrane. A temperature controller is used to keep the temperature of the water and the mouse head stable at ~37°. After experiment, the rat was sacrificed by pentobarbital.
Figure 7 shows an in situ photoacoustic image of the vascular distribution in a rat brain. Figure 7(a) is the open-skull photograph of rat brain surface acquired after experiment, and Fig. 7(b) is photoacoustic imaging of the superficial layer of a rat brain acquired with the skin and skull intact. A satisfactory match between the photograph and the reconstructed image can be seen, the main vessels and the vascular ramifications in the rat brain can be imaged clearly and accurately with our current system. Due to the limitation of the spatial resolution, the subtle vessels are not distinct, and because the orientations of the histological cross section and the imaged layer are not identical, the shapes of the vessels in the histology slightly differ from those in the photoacoustic image.
Fig. 7. Photoacoustic imaging of the rat brain in vivo. (a) Open-skull photograph of rat brain surface acquired after experiment. (b) Photoacoustic imaging of the superficial layer of a rat brain acquired with the skin and skull intact.
4. Discussion
Since in this study a transducer array was used rather than a single transducer, mechanical step scanning was performed, leading a measurement time of 6.4seconds per position, and a total scan duration of 150 seconds. Compared with the acquisition time of 5 seconds in our previous works, the time resolution of this imaging system is reduced, but the ability of imaging complicated structures of samples is obviously improved. At present the data acquisition time is acceptable compared with that of a single transducer scanning the sample. By utilizing laser system with higher repetition rate and using a 128-channel parallel data acquisition system, the multi-element photoacoustic imaging system holds the potential to provide a clear and accurate photoacoustic image of the object with complicated structure within a few seconds. This is our next work.
The arrangement of the ultrasonic transducer determines the imaging mode and the regions that can be detected. For in vivo detection of tissue structures, the imaging mode that the transducer array circularly scanning the sample determines that this imaging system can be used to detect the protrudent parts of the body, such as breast. So the photoacoustic imaging system has the potential of detection of the breast cancer at the early stage. The imaging system also has the ability of imaging of the network of blood vessels, so it may be used to detect and monitor the tumor angiogenesis. The transducer array was made from piezoelectronic materials, which has a high sensitivity but a narrow band frequency response. Due to the limitation of the bandwidth of the linear transducer array, the spatial resolution of the imaging system is not high to image the subtle vessels.
5. Conclusion
In this paper, we successfully used a linear transducer array by combined scanning in a full view to form photoacoustic images of phantoms and the network of blood vessels in a rat brain. The improved LFBP algorithm with directivity pattern functions of individual transducer and the total transducer array can improve the lateral resolution of the imaging system. This photoacoustic imaging system may have the potential for detection of tumors at the early stage, monitoring the angiogenesis around the tumors.
Acknowledgments
This research is supported by the National Natural Science Foundation of China (30470494; 30627003), and the Natural Science Foundation of Guangdong Province (7117864).
References and links
1. R. O. Esenaliev, A. A. Karabutov, and A. A Oraevsky, “Sensitivity of laser opto-acoustic imaging in detection of small deeply embedded tumors,” IEEE J.Sel.Top.Quantum Electron. 5, 981–988 (1999). [CrossRef]
2. A. A. Oraevsky, V. A. Andreev, A .A . Karabutov, R. D. Fleming, Z. Gatalica, H. Singh, and R. O. Esenaliev,“Laser optoacoustic imaging of breast: detection of cancer angiogenesis,” Proc.SPIE 3597, 352–363 (1999). [CrossRef]
3. V. G . Andreev, A .A. Karabutov, and A .A . Oraevsky, “Detection of ultrawide-band ultrasound pulses in optoacoustic tomography,” IEEE Trans. Ultrason. Ferroelectr. Freq. control. 50, 1383–1390 (2003). [CrossRef] [PubMed]
4. X. Wang, Y. Pang, G. Ku, X. Xie, G. Stoica, and L.V. Wang, “Noninvasive laser-induced photoacoustic tomography for structural and functional in vivo imaging of the brain,” Nat.Biotechnol. 21, 803–806 (2003). [CrossRef] [PubMed]
5. J. J. Niederhauser, M. Jaeger, R. Lemor, P. Weber, and M. Frenz, “Combined ultrasound and optoacoustic system for real-time high contrast vascular imaging in vivo,” IEEE Trans. Med. Imaging 24, 436–440 (2005). [CrossRef] [PubMed]
6. R. I. Siphanto, K. K. Thumma, T. G. Van Leewen, F. F. M. de Mul, J. W. van Neck, L. N. A. van Adrichem, and W. Steenbergen, “Serial noninvasive photoacoustic imaging of neovascularization in tumor angiogenesis,” Opt. Express 13,89–95 (2004). [CrossRef]
7. E. V. Savateeva, A. A. Karabutov, B. Bell, R. Johnigan, M. Montamedi, and A. A. Oraevsky, “Noninvasive detection and staging of oral cancer in vivo with confocal optoacoustic tomography,” Proc.SPIE 3597, 55–67 (2000). [CrossRef]
8. J. A. Viator, L. O. Svaasand, G. Aguilar, B. Choi, and J. S. Nelson, “Photoacoustic measurement of epidermal melanin,” Proc. SPIE 4960,14–21 (2003). [CrossRef]
9. E. V. Savateenva, A. A. Karabutov, S. V. Solomatin, and A. A. oraevsky, “Optical properties of blood at various level of oxygenation studied by time detection of laser-induced pressure profiles,” Proc. SPIE 4618, 63–75 (2002). [CrossRef]
10. R. O. Esenaliev, I. V. larina, K. V. Larin, D. J. Deyo, M. Motamedi, and D. S. Prough, “Optoacoustic technique for noninvasive monitoring of blood oxygenation: a feasibility study,” Appl. Opt. 43, 3401 1–7 (2002).
11. H. A Mackenzie, H. S. Ashton, S. Spiers, Y. Shen, S. S. Freeborn, J. Hannigan, J. Lindberg, and P. Rare, “Advance in photoacoustic noninvasive glucose testing,” Clin. Chem. 45, 1587–1595 (1999). [PubMed]
12. Z. Zhao and R. Myllyla, “Photoacoustic determination of glucose concentration in whole blood by a near-infrared laser diode,” Proc.SPIE 4256, 77–83 (2001). [CrossRef]
13. M. Kinnunen and R. Myllyla, “Effect of glucose on photoacoustic signals at wavelengths of 1064 and 532 nm in pig blood and Intralipid,” J. Phys D: Appl. Phys. 38, 2654–2661 (2005). [CrossRef]
14. L.Zh. Xiang, D. Xing, H.M Gu, D.W. Yang, S.H. Yang, L. M. Zeng, and W. R. Chen, “Real-time optoacoustic monitoring of vascular damage during photodynamic therapy treatment of tumor,” J. Biomed Opt. 12, 014001 1–8 (2007). [CrossRef]
15. S. A. Ermilov, A. Coniusteau, K. Mehta, R. Lacewell, P. M. Henrichs, and A. A. Oraevsky, “128-channel laser optoacoustic imaging system for breast cancer diagnostics,” Proc. SPIE. 6086, 608609 1–12 (2006).
16. R. J. Zemp, R. Bitton, M. L. Li, K. K. Shung, G. Stoca, and L. V. Wang, “Photoacoustic imaging of the microvasculature with a high-frequency ultrasound array transducer,” J. Biomed Opt. Lett. 12, 010501 1–3 (2007). [CrossRef]
17. D. W. Yang, D. Xing, H. M. Gu, Y. Tan, and L. M. Zeng, “Fast multielement phase-controlled photoacoustic imaging based on limited-field-filtered back-projection algorithm,” Appl. Phys. Lett. 87, 194101 1–3 (2005). [CrossRef]
18. D. W. Yang, D. Xing, Y. Tan, H. M. Gu, and S. H. Yang, “Integrative prototype B-scan photoacoustic tomography system based on a novel hybridized scanning head,” Appl. Phys. Lett. 88, 174101 1–3 (2006). [CrossRef]
19. B.Zh. Yin, D. Xing, Y. Wang, Y. G. Zeng, Y. Tan, and Q. Chen, “Fast photoacoustic imaging system based on a 320-element linear transducer array,” Phys. Med. Biol. 49, 1339–1346 (2004). [CrossRef] [PubMed]
20. Zh. X. Chen, Zh. L. Tang, and W. Wan, “Photoacoustic tomography imaging based on a 4f acoustic lens imaging system,” Opt. Express , 15, 4966–4976 (2007). [CrossRef] [PubMed]
21. D. Feng, Y. Xu, G. Ku, and L.V. Wang, “Microwave-induced thermoacoustic tomography: reconstruction by synthetic aperture,” Med. Phys. 28, 2427–2431 (2001). [CrossRef]
22. Y. Xu and L. V. Wang, “Reconstructions in limited-view thermoacoustic tomography,” Med. Phys. 31, 724–733 (2004). [CrossRef] [PubMed]
23. A. A. oraevsky, V. G. Andreev, A. A. Karabutov, and R. O. Esenaliev, “Two-dimensional opto-acoustic tomography transducer array and image reconstruction algorithm,” Proc. SPIE 3601, 256–267 (1999). [CrossRef]
24. M. H. Xu and L. V. Wang, “Time-domain reconstruction for thermoacoustic tomography in a spherical geometry,” IEEE Trans. Med. Imaging. 21, 814–822 (2002). [CrossRef] [PubMed]
