Open Access
Add to BibSonomyAbstract
Three-dimensional (3D) photoacoustic imaging can be achieved by a two-dimensional (2D) ultrasound array matrix. However, the 2D matrix consisting of hundreds or thousands of transducer elements makes it not only expensive, but also a big technical challenge for both probe manufacturing and parallel data acquisition. In this study, we performed the photoacoustic imaging by scanning an unfocused linear ultrasound array probe over a planar geometry, resulting in an equivalent 2D matrix probe. The phantom study demonstrated that this method substantially increased imaging quality, which has great potential for animal and clinical photoacoustic imaging.
© 2017 Optical Society of America
1. Introduction
Photoacoustic tomography (PAT) is a hybrid imaging modality that combines rich optical contrast and high ultrasonic resolution. This modality breaks through the imaging barrier for pure high-resolution optical imaging in tissues caused by the strong optical scattering. Over the past decades of fast developing, PAT has made significant progresses to demonstrate its great potential for wide implementations [1–5].
Three-dimensional (3D) PAT imaging is important for its applications. In principle, the 3D PAT needs detection of PA signals over a 2D surface, such as sphere, cylinder, and plane. For instance, by scanning an optical point detector over a plane can provide high-quality 3D images deep in tissue. However, the low repetition rate by laser safety regulation makes it too time-cost by scanning single detector for clinical application. Therefore, 2D array systems, such as the planar matrix array [6–10] and the hemisphere ultrasound array [11], were developed for 3D PAT. But a 2D array consists of hundreds to thousands ultrasonic transducer elements, significantly increasing the system’s complexity and cost.
As an alternative, one can scan a linear array, and the 3D PAT imaging can be achieved by stacking multiple cross-section images. Various linear arrays have been used, including hand-held linear probe [12, 13], arc-shaped [14] or circular ring array [15, 16]. However, the linear array generally weakly focuses in the elevation direction, and the non-uniform resolution, especially low elevational resolution makes it challenging for high-quality 3D rendering. Therefore, although many 3D PA imaging have been explored by scanning a hand-held linear probe, most of these works provide multiple image slices and only a few that has high-frequency probe (>30MHz, thus better elevational resolution) presented 3D-rendered results [17, 18]. But high-frequency probe is not suitable for deep tissue imaging. Besides, another important limitation caused by focusing in elevation direction is that it only allows the array to detect PA signal from a very limited acquisition angle, as shown in Fig. 1. For instance, it will not effectively detect signals coming from an object whose PA waves transport in the direction out of the acquisition angle, as shown in Fig. 1(a). In the next section, a simulation study would demonstrate this limitation. Although scanning an unfocused arc-shaped linear array to form a spherical detection was explored for small animal imaging [19], for the purpose of clinical application, the planar 2D array is preferred. In this study, we explored the 3D PAT by scanning an unfocused linear array probe over a planar surface. The unfocused transducer has a much wider acquisition angle, as shown in Fig. 1(b). Moreover, it is much easier to manufacture a linear array and its parallel data acquisition system. Scanning such a linear array is equivalent to a 2D array matrix with thousands or even more elements, as discussed in the following.
Fig. 1 The detected angle of the focused transducer and the unfocused transducer. (a) and (b) represented the focused and unfocused transducer elements, respectively.
2. Simulation study
We first did a simulation study to compare imaging results by scanning focused and unfocused US probes for 3D PA imaging. Here, we assumed two linear US probe (A, B), and each contains the same 48 point elements. The difference is that probe A is a focused array with an acquisition angle of 5 degrees along the elevation direction, and probe B is an unfocused array. Both arrays have the same pitch of 1.35 mm. As shown in Fig. 2, we scanned these two probes to perform 3D PA imaging of a uniform sphere with the diameter of 5 mm. The sphere locates 15 mm under the center of the scanning plane. The probe is parallel to y-axis, scanning in z direction, and the scanning plan is parallel to the yoz plane.
Fig. 2 (a) Simulation schematic diagram; (b) Cross-section of reconstructed images by two probes, respectively.
The total scanning length is 40 mm with a step size of 0.1 mm. Since a uniform sphere absorber generate PA waves radially outward, the incident angle of the PA wave on each point detector can be feasibly determined. Figure 2(b) showed the cross-sections of the reconstructed images: the first row showed results by probe A (focused) and the bottom row showed the results by probe B (unfocused); the first column showed cross-section image in xoy plane, and the second column showed results in xoz plane. According to the results, it is clearly to see that only the result in xoz plane by probe A has serious image distortions. It is because the very limited acquisition angle along the elevation direction for a focused probe cannot effectively received PA signals that travels obliquely onto the probe. This simulation demonstrated that scanning traditional US probe with focusing lens has a risk of losing important target features.
3. Methods
3.1 Imaging System Setup
In this study, we used a customized unfocused linear probe (by TomoWave. Inc, Houston, USA), which has 48 elements and each element has a size of 1.0 mm × 1.0 mm, and the element pitch is 1.35 mm. we parallelly scanned the probe along a planar surface over the target with the step size of 1.0 mm, as shown in Fig. 3.
Fig. 3 Schematic diagram of our method: scanning a linear ultrasound probe is equivalent to a 2D ultrasound matrix for PAT. (a) Scanning of a linear ultrasound probe (48 elements); (b) an equivalent 2D matrix with n × 48 elements.
During the imaging process, the target was immersed in a water tank, and the probe was mounted on a one-dimensional translation stage that was controlled by LabVIEW. Data acquisition was done by a customized LOIS system (TomoWave Laboratories, Houston, USA) at 25 MHz sampling rate and a gain of 64 dB. A Q-switched Nd:YAG laser (LS-2137/2, LOTIS TII, Minsk, Belarus) generating 1064 nm, 10 Hz laser was employed to excite photoacoustic signals. The laser was delivered to the target through a one-to-two optical fiber bundle and illuminated the object from the bottom or sides of the container. Figure 4 showed the experimental scheme.
3.2 System calibration
The bandwidth of the system was measured by detecting an ultra-short PA signal. We uniformly painted a thin layer of black ink on the surface of a 2-cm-thick glass plate, and immersed this glass in a water tank with its painted surface facing the array. Then we illuminated a focused laser on the dark layer, generating a very short PA wave. A detected PA signal by one element was demonstrated in Fig. 5(a), and its frequency spectrum was shown in Fig. 5(b). According to the spectrum, the transducer has a wide bandwidth from hundreds of kHz to about 6.0 MHz. We also changed the relative position of the PA source with the element, and measured the FWHM of acquisition angle to be about 60 degrees (primarily limited by the finite element size).
The resolution of the system was measured by imaging a thin human hair (~100 μm in diameter). The hair was horizontally placed in the tank at 8 mm under the probe surface. We scanned the linear probe over 38 mm with the step size of 1.0 mm (equivalent to a 38 × 48 elements planar 2D matrix), and the hair is placed in the middle of the scanning area. Figure 6(a) showed the cross section of the reconstructed image, and Fig. 6(b) showed the normalized reconstructed values along two white lines. According to the results, the resolution along the horizontal direction is about 0.89 mm, and the elevation resolution is about 0.49 mm. Since we used the intensity in the image reconstruction (Eq. (2), instead of the amplitude, the resolution is a bit better than the theoretical analysis [16].
Fig. 6 (a) The cross section of the PAT result of hair; (b) (c) the normalized reconstructed PA values along horizontal and vertical white lines, respectively.
3.3 Image reconstruction and 3D rendering
In this study, we used the image reconstruction algorithm described in [20] for 2D planar detection configuration, whose time-domain formulation is
where c0 is the sound speed, p is the detected signal, is the infinitesimal solid angle at with respect to the reconstruction point , and . In addition, to improve the image contrast, we used the intensity of the reconstructed value:To perform the 3D image rendering, we first reconstructed multiple slices, and then loaded them into the VolView software (kitware.com).
4. Phantom studies
4.1 Carbon rods in agar phantom
We embedded two carbon rods (0.5 mm in diameter) in agar for photoacoustic imaging. Two rods were aligned at the same horizontal plane, and separating about 30°, as shown in Fig. 7(a). In order to mimic tissue optical scattering, 1% intralipid was added in the agar phantom. Figure 7(a) was the photograph of the phantom when the two carbon rods was embedded onto a 1.4 cm-thick solidified agar, then another 2.0-cm-thick agar layer was filled on top of the rods before solidification. The laser illuminated upwardly onto the bottom of the phantom. We scanned a length of total 40 mm at an interval of 1.0 mm, and the equivalent 40 × 48 elements matrix array provided the 3D reconstruction result in Fig. 7(b) (see Visualization 1). The dimension of the box in Fig. 7(b) is 20 mm × 6.0 mm × 39 mm. The reconstructed result is highly consistent with the carbon phantom.
Fig. 7 (a) The photograph of carbon rods in scattering medium during phantom process; (b) the 3D PAT of two carbon rods (see Visualization 1).
4.2 Double helixes in agar phantom
The second phantom study used a double-helix object made from the black strip. As shown in Fig. 8(a), the double helixes immersed in the agar phantom with 1% intralipid. The procedure to prepare the phantom is very similar to Fig. 7(a). Unlike the horizontally aligned straight rods in Fig. 7, which generated PA waves cylindrically outward, the helix generated PA waves have preferred direction depending on the local orientation of the helix. According to the discussion in the introduction part, the limited acquisition angle of traditional focused array cannot effectively image the part on the helix when its normal direction is out of the acquisition angle. We scanned a length of total 47 mm, and the equivalent 47 × 48 elements matrix array provided the 3D reconstruction result in Fig. 8(b). The dimension of the box in Fig. 8(b) is 20 mm × 13 mm × 46 mm. In our methods, with significantly increased acquisition angle, the reconstructed image accurately showed most of the helix, except for portions where the normal direction is close to parallel to x-z plane. A 3D animation was also provided online (see Fig. 8, Visualization 2).
Fig. 8 (a) The photograph of double helixes sample; (b) the 3D PAT of double helixes (see Visualization 2).
5. Discussion and conclusion
In this study, we demonstrated that 3D photoacoustic imaging can be achieved by scanning a linear unfocused ultrasound array probe. In this way, an equivalent 2D ultrasonic matrix array, with large amount of elements and large matrix dimension than traditional 2D matrix, can be built at a much less cost and system complexity. This method has great potential clinical use where a large 2D matrix is desired, such as the photoacoustic breast imaging. However, this method also has prominent limitations. First, the imaging speed would be much slower than real 2D array matrix; second, using the unfocused array will make it challenging for traditional ultrasound imaging. Due to the first limitation, this method is more suitable for imaging targets that can be hold stable during imaging, such as the fixed breast or arms. For instance, a mammography exam can take up to ten minutes while the patient’s breast was hold. For the latter one, a partial solution is scanning a traditional probe in the same imaging region after PA imaging.
Recently, a novel method that used a “slit” for traditional US probe can substantially improve the elevational resolution for 3D PA imaging [21]. This method placed a thin metal “slit” on the focal zone to diffract the incoming PA waves to the probe. However, PA signals generally have a broad bandwidth, and according to the classical Kirchhoff diffraction theory, wave diffraction depends on the frequency. The frequency-dependent detection in this slit-based mechanism would cause extra difficulties for the accurate PA image reconstruction. Besides, only a fraction of the diffracted energy can be detected by the probe (limited by the acquisition angle), which also hinders the signal to noise ratio. Compared with this slit-based method, our method used well-developed image reconstruction algorithms and has higher sensitivity, but at the expense of resolution. However, our method aims to image deep in tissue for clinical use, and millimeter-scale resolution is a typical value for several widely used clinical imaging modalities, such as the magnetic resonance imaging (MRI) and positron emission tomography (PET).
Funding
Ministry of Education of the People's Republic of China (20130001110035); the National Key Instrumentation Development Project (2011YQ030114, 2013YQ030651); National Nature Science Foundation of China (NSFC) (81421004).
Acknowledgments
Authors thank Mr. Wenzhao Li for helping analyzing and reducing electrical noise of the system.
References and links
1. S. Zackrisson, S. M. W. Y. van de Ven, and S. S. Gambhir, “Light in and sound out: emerging translational strategies for photoacoustic imaging,” Cancer Res. 74(4), 979–1004 (2014). [CrossRef] [PubMed]
2. L. V. Wang and S. Hu, “Photoacoustic tomography: in vivo imaging from organelles to organs,” Science 335(6075), 1458–1462 (2012). [CrossRef] [PubMed]
3. P. Beard, “Biomedical photoacoustic imaging,” Interface Focus 1(4), 602–631 (2011). [CrossRef] [PubMed]
4. L. V. Wang, “Multiscale photoacoustic microscopy and computed tomography,” Nat. Photonics 3(9), 503–509 (2009). [CrossRef] [PubMed]
5. C. Li and L. V. Wang, “Photoacoustic tomography and sensing in biomedicine,” Phys. Med. Biol. 54(19), R59–R97 (2009). [CrossRef] [PubMed]
6. S. Vaithilingam, T. J. Ma, Y. Furukawa, I. O. Wygant, X. Zhuang, A. De La Zerda, O. Oralkan, A. Kamaya, S. S. Gambhir, R. B. Jeffrey Jr, and B. T. Khuri-Yakub, “Three-dimensional photoacoustic imaging using a two-dimensional CMUT array,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 56(11), 2411–2419 (2009). [CrossRef] [PubMed]
7. Y. Wang, T. N. Erpelding, L. Jankovic, Z. Guo, J.-L. Robert, G. David, and L. V. Wang, “In vivo three-dimensional photoacoustic imaging based on a clinical matrix array ultrasound probe,” J. Biomed. Opt. 17(6), 061208 (2012). [CrossRef] [PubMed]
8. T. Kitai, M. Torii, T. Sugie, S. Kanao, Y. Mikami, T. Shiina, and M. Toi, “Photoacoustic mammography: initial clinical results,” Breast Cancer 21(2), 146–153 (2014). [CrossRef] [PubMed]
9. D. Piras, W. Wenfeng Xia, T. G. Steenbergen, van Leeuwen, and S. G. Manohar, “Photoacoustic Imaging of the Breast Using the Twente Photoacoustic Mammoscope: Present Status and Future Perspectives,” IEEE J. Sel. Top. Quantum Electron. 16(4), 730–739 (2010). [CrossRef]
10. Z. Xie, F. M. Hooi, J. B. Fowlkes, R. W. Pinsky, X. Wang, and P. L. Carson, “Combined photoacoustic and acoustic imaging of human breast specimens in the mammographic geometry,” Ultrasound Med. Biol. 39(11), 2176–2184 (2013). [CrossRef] [PubMed]
11. R. A. Kruger, C. M. Kuzmiak, R. B. Lam, D. R. Reinecke, S. P. Del Rio, and D. Steed, “Dedicated 3D photoacoustic breast imaging,” Med. Phys. 40(11), 113301 (2013). [CrossRef] [PubMed]
12. M. Kuniyil Ajith Singh, W. Steenbergen, and S. Manohar, “Handheld Probe-Based Dual Mode Ultrasound/Photoacoustics for Biomedical Imaging,” in Frontiers in Biophotonics for Translational Medicine: In the Celebration of Year of Light (2015), M. Olivo and U. S. Dinish, eds. (Springer Singapore, Singapore, 2016), pp. 209–247.
13. 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(4), 436–440 (2005). [CrossRef] [PubMed]
14. A. Dima, N. C. Burton, and V. Ntziachristos, “Multispectral optoacoustic tomography at 64, 128, and 256 channels,” J. Biomed. Opt. 19(3), 036021 (2014). [CrossRef] [PubMed]
15. C. Li, A. Aguirre, J. Gamelin, A. Maurudis, Q. Zhu, and L. V. Wang, “Real-time photoacoustic tomography of cortical hemodynamics in small animals,” J. Biomed. Opt. 15(1), 010509 (2010). [CrossRef] [PubMed]
16. J. Xia, M. R. Chatni, K. Maslov, Z. Guo, K. Wang, M. Anastasio, and L. V. Wang, “Whole-body ring-shaped confocal photoacoustic computed tomography of small animals in vivo,” J. Biomed. Opt. 17(5), 050506 (2012). [CrossRef] [PubMed]
17. L. Song, C. Kim, K. Maslov, K. K. Shung, and L. V. Wang, “High-speed dynamic 3D photoacoustic imaging of sentinel lymph node in a murine model using an ultrasound array,” Med. Phys. 36(8), 3724–3729 (2009). [CrossRef] [PubMed]
18. S. Y. Nam and S. Y. Emelianov, “Array-based real-time ultrasound and photoacoustic ocular imaging,” J. Opt. Soc. Korea 18(2), 151–155 (2014). [CrossRef]
19. H.-P. Brecht, R. Su, M. Fronheiser, S. A. Ermilov, A. Conjusteau, and A. A. Oraevsky, “Whole-body three-dimensional optoacoustic tomography system for small animals,” J. Biomed. Opt. 14(6), 064007 (2009). [CrossRef] [PubMed]
20. M. Xu and L. V. Wang, “Universal back-projection algorithm for photoacoustic computed tomography,” Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 71(1), 016706 (2005). [CrossRef] [PubMed]
21. Y. Wang, D. Wang, Y. Zhang, J. Geng, J. F. Lovell, and J. Xia, “Slit-enabled linear-array photoacoustic tomography with near isotropic spatial resolution in three dimensions,” Opt. Lett. 41(1), 127–130 (2016). [CrossRef] [PubMed]
