Open Access
Add to BibSonomy
Abstract
Photoacoustic computed tomography (PACT) can ultrasonically image optical absorbers in biological tissues by using a linear piezoelectric transducer array, but some features can not be visualized as a result of the limited acceptance angle. The optical ultrasound sensors for photoacoustic imaging have received great interests, because of their compact sizes, comparable sensitivities to their electric counterparts, as well as the extended field/angle-of-view. In this work, we have developed a PACT system based on a fiber-laser based ultrasound sensor. Two-dimensional imaging was performed by horizontally scanning the sensor and image reconstruction via back projection, and three-dimensional imaging was further achieved by repeating such scanning process at multiple angles, based on inverse Radon transform. The axial and lateral resolutions are 93 and 220 µm in three-dimensional imaging. The fiber-based PACT can resolve more features than that with a piezoelectric transducer array, taking advantage of the dual-60-degree vision angles of the sensor.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Photoacoustic computed tomography (PACT) can image biological tissues at an optical diffusion regime, by ultrasonically detecting the optical absorbers, including blood vessels or synthetized extrinsic contrast agents [1,2]. Recently, the whole body of small animals and human organs have been imaged by PACT at high tempo-spatial resolutions towards medical studies and diagnostic applications [3–8]. The imaging capability relies on the detection of optically induced ultrasound waves by using a scanning piezoelectrical transducer or a transducer array [1,2,9]. For high imaging quality as well as frame rate, the ring- and spherical-shaped arrays can detect the ultrasound waves from all directions, which offers full detail of the fine absorber structures [8,10]. By contrast, the linear transducer array has found wide applications in hospital for ultrasound imaging in the side-viewing manner, which enables high flexibility for human and animal imaging. The linear array allows for a low-cost and practical strategy for PACT incorporation by using an additive nanosecond laser source for photoacoustic excitation [11–13]. However, due to the limited acceptance angle (usually 40 degrees or less), the linear array would lose many features of the absorptive tissues, resulting in a degradation of image quality [14,15].
Optical sensors provide alternative strategies for ultrasound detection in photoacoustic imaging. For example, high-finesse planar, on-tip, or in-fiber Fabry-Perot cavities, polymer/silicon ring resonators, as well as functionalized optical surfaces have been exploited for ultrasound sensing [16–20]. Notably, a light beam can naturally behave as a line-shaped detector, which translates the ultrasound waves travelling through the beam into a variation in optical phase. The acoustically induced optical response is a result of the phase integration throughout the optical path. The optical sensors bring great flexibility to photoacoustic imaging by enabling different imaging configurations. PACT by using such line detectors have been conceptually demonstrated in [21] and then with free-space and fiber-optic interferometers via rotary scanning [22,23]. We have developed a fiber-laser-based sensor by translating the ultrasound waves into a lasing frequency shift and incorporate it into fast-scanning photoacoustic microscopy [24,25]. In this work, we develop a side-viewing PACT by using the fiber-laser ultrasound sensor. The line-profiled sensor linearly scans to obtain a two-dimensional image via back projection [2,26,27]. Such linear scanning is repeatedly performed at multiple angles by rotating the imaging sample to have a three-dimensional image based on the inverse Radon transformation. The PACT system offers axial and lateral resolutions at 93 and 220 µm in three-dimensional imaging. Compared with the piezoelectric transducer array, this fiber optic sensor can resolve more features, as a result of the dual 60-degree vision.
2. Materials and methods
2.1 Methods
Figure 1 schematically shows the implementation of PACT performed by using a scanning fiber-optic sensor. The imaging contains the following two consecutive steps: Step one, as shown in Fig. 1(a), the fiber sensor horizontally scans to detect the photoacoustic signals at each position. The optically induced ultrasound waves can acoustically interact with the optical fiber, establish a stress field over it, and induce a change in optical phase, expressed as ${\Delta }\varphi = \mathop \smallint \nolimits_0^L k{\Delta }n{\; }dL$, where k, Δn, and L represent the optical wavenumber, acoustically induced refractive-index change, and the length of the fiber sensor. With a given acoustic speed, the absorbers can be reconstructed via back projection [2,26].
Fig. 1. (a) Schematic diagram of the optical fiber-based PACT. (b) Step one: the fiber optic sensor horizontally scans to detect the optically induced ultrasound waves to have a two-dimensional image; (c) Step two: Rotate the sample by an angle and repeat the linear scanning. Three-dimensional volumetric images can be obtained with multiple two-dimensional snapshots based on inverse Radon transform. EDFA: Erbium-doped fiber amplifier, WDM: wavelength division multiplexing. ISO: optical isolator, PC: Polarization controller.
Step two: The linear scanning, as well as the two-dimensional reconstruction, were repeated, before rotating the sample with a step θ, as demonstrated in Fig. 1(b). By repeating such a process multiple times to cover the full angular range and performing the inverse Radon transform, we can have a three-dimensional image. Here, we can equivalently consider each 2D image as a projection at a single angle. By viewing the sample at many different angles, a volumetric image can be formed.
2.2 Fiber optic ultrasound detection
To detect the photoacoustic signal, we have developed a sensor based on a dual-polarization fiber laser, as shown in Fig. 2(a). The laser contains two highly reflective Bragg gratings in a rare-earth-doped fiber (EY305, CorActive). Each grating has a length of 3 mm, and the grating separation is about 8 mm. Pumped by a 980 nm semiconductor laser diode, the laser cavity has single longitudinal mode output at both x- and y-polarizations. The two lasing frequencies are slightly different, creating a beat note at 2.7 GHz. The beat frequency determined by the intrinsic birefringence of the fiber due to the deviation from a perfect round geometry. The ultrasonically induced phase change is translated into a shift in beat frequency. The frequency modulation can be measured by using an I/Q demodulator at the radiofrequency regime. We have characterized the ultrasound sensitivity by measuring its response to planar ultrasound waves. The noise equivalent pressure is 46 Pa over a detection bandwidth 50 MHz, which is sufficient for photoacoustic imaging.
Fig. 2. (a) Fiber-laser based ultrasound sensor. Measured spatial sensitivity at (b) longitudinal and (c) azimuthal directions.
Figure 2(b) shows the longitudinal profile of the ultrasound sensitivity by measuring its response to a point acoustic source scanning along the fiber length. The point source was created by focusing a 532 nm nanosecond laser on a highly-absorptive black tape immersed in water. The spacing between the source and the fiber was 2 mm. The measured result presents an almost flat ultrasound response over the black region between the grating reflectors. The sensitivity rapidly decreases in the grating regions, because much fewer photons in the laser cavity were modulated. It is intuitive to increase the cavity length to have a larger field of view. However, with cavities longer than 1 cm, multiple longitudinal modes may simultaneously lase, which induces a mode competition as well as the instability of the beat signal. Moreover, the ultrasound sensitivity will also decrease, owing to the reduced ratio between the effective acoustic interaction length (Leff=200 µm at 20 MHz) over the cavity length.
Figure 2(c) shows the measured response at a different incidence angle, measured by rotating the fiber sensor with a step of 20 degrees. The sensor presents maximum sensitivities at incidences along the principal axis because the acoustic compression can be converted into the birefringence change most effectively. With the assumption of weak acoustic perturbation, the ultrasound response follows a cos(2θ) variation. The result suggests that the sensor has 4 angular acceptance ranges, with the full angle at half maximum 60 degrees.
To perform PACT, the sample is immersed in a water tank mounted on a motorized rotation stage. A 532-nm Nd:YAG nanosecond laser (DAWA-100, Beamtech) projects laser pulses onto the sample. The pulse duration is 5 ns, and the repetition rate is 10 Hz. The laser beam was expanded and homogenized by using a diffuser (DG120, Thorlabs), resulting in a 2 cm-diameter illumination area on the sample. With a single-pulse energy of 30 mJ, the optical fluence on the sample surface was 15 mJ/cm2, which is below the American National Standards Institute (ANSI) safety limit. The fiber sensor was scanned by using a linear stage (L-509, PI) at a speed of 100 µm/s with a step of 10 µm. The frequency modulated optical signal was amplified with an erbium-doped fiber amplifier to about 20 mW before launching on a photodetector to achieve a higher signal-to-noise ratio. The radio-frequency signal was subsequently digitalized by a data acquisition system (National Instruments, PXI7852R) at a sampling rate of 100 MHz. A computer-controlled the synchronization of laser pulses, ultrasound detection, and signal acquisition. The sample was rotated by a step of 3 degrees before the next cycle of scanning was performed.
3. Imaging results
The spatial resolutions in 2D PACT have been characterized by imaging three pieces of human hair. Each hair, 70 µm in diameter, was placed parallel to each other and vertical to the imaging plane. Figure 3(a) shows the reconstructed image of the samples. Extracted from the enlarged view of a single hair in Fig. 3(b), the photoacoustic amplitudes are plotted along the dashed horizontal and vertical lines in Fig. 3(c). The sensor has a maximum response at 22 MHz and a -3-dB bandwidth of 18 MHz, which corresponds to a theoretical axial resolution of 70 µm (Note that the frequency response is not an ideal Lorentze shape). The vertical resolution is 86 µm, by measuring the full widths at half maxima of the envelopes of the photoacoustic amplitudes, almost in accordance with the theoretical result determined by the working bandwidth of the fiber sensor. The horizontal resolution is 113 µm, which is determined by the sensor performance as well as the scanning distance.
Fig. 3. Resolution characterization of 2D imaging. (a) Schematic illustration. (b) Reconstructed image of the three pieces of human hair. (c) Enlarged view of a selected hair in the dashed box. Photoacoustic amplitude profiles along the vertical (d) and horizontal (e) dashed lines.
The receiving angle of the 2D PACT was tested by imaging a 5-mm-diameter metal ring, as shown in Figs. 4(a) and 4(b). A ring-shaped sample is selected here because it contains features at all angles. The result in Fig. 4(b) suggests the dual 60-degree vision angles (at one side), following the measured result in Fig. 2(c). Compared to the piezoelectric transducer linear array with a 40-degree vision, the fiber sensor has a greatly enhanced angle-of-view [14]. We then imaged an ink-stained leaf skeleton that has rich structural features and can mimic the vascular structures in biological tissues. The sample was fixed inside an agar, almost perpendicular to the imaging plane. Figures 4(c) and 4(d) exhibit the photo and the reconstructed photoacoustic image of the leaflet. Most of the structural features have been reconstructed, showing a good match to the photograph. Compared with the result obtained with a piezoelectric transducer array in which only a few horizontal structures can be visualized (Fig. 1(b) in Ref. [14]), more horizontal and some oblique structures have been well reconstructed by the fiber-based PACT, as a result of the extended angular vision range. Some features are still missing, which can be addressed by multiplexing more sensors to cover the entire visual range.
Fig. 4. Two-dimensional imaging results. (a) Photograph and (b) photoacoustic image of a metal ring. (c) photograph and (d) photoacoustic image of an inked leaflet.
In order to reconstruct a 3D image, an additional dimension of scanning is required. Inspired by the widely used X-ray computed tomography, the linear scanning was repeated at multiple angles and 3D images can be reconstructed based on the inverse Radon transformation [28]. This scanning manner can also maintain the side looking ability. Here each 2D image can be equivalently considered as a projection at a single angle. By viewing the sample at many different angles, a volumetric image can be formed. The 3D image capability was demonstrated by imaging human hairs that are assembled in different geometries. Figure 5(a) shows the photograph and the photoacoustic image, which is reconstructed by using 120 2D frames. The image contains a 500×500×500-pixel vortex, corresponding to a 10×10×14.8 mm3 volume. Figure 5(b) shows the axial resolution curve measured in the 3D reconstruction. The axial resolution is estimated as 93 µm, extracting from the reconstructed image of a point source and taking an envelope of the data. Figure 5(c) shows the lateral resolution curve with an FWHM of 220 µm. Ideally, the spatial resolution in 3D imaging should follow the 2D case. Here the resolution is degraded due to the under-sampling in the angular direction and the possible misalignment during the rotation scanning. This problem can be addressed by increasing the angular steps. However, this increases the measurement time.
Fig. 5. Characterization of the multi-angle fiber-scanning PACT system. (a) 3D reconstruction of hairs. Inset: photo of hairs. (b) photoacoustic amplitude distribution along the axial direction, (c) photoacoustic amplitude distribution along the lateral direction.
We then in vitro imaged a chick chorioallantois membrane (CAM) with the 3D PACT system. We opened a fertilized, 3-day-old chicken egg and transferred its content into a transparent plastic cup, covered with a breathable film, and incubated at 38 degrees for eight additional days [29]. On day 11, the CAM was almost fully developed, presenting the vascular structures and eyes embedded in a transparent background, as shown in Fig. 6(a). The 3D reconstructed image is exhibited in Figs. 6(b) and 6(c) at different view angles. The image has an 8.88×12×6.216 mm3 volume, corresponding to 370×500×175 pixels. The penetration depth is over 4 mm. An eye and the surrounding vascular branches (labeled from 1 to 5) can be visualized in three dimensions. The main vessels are covered by a capillary network, which has been imaged in Fig. 6(b). However, we are not able to tell apart each of them due to the limitation of spatial sensitivity. To see the main vessels more clearly, we have removed the surface layer which contains the capillaries in Fig. 6(c) (See more details in Visualization 1).
Fig. 6. In vitro 3D photoacoustic imaging result. (a) photograph of the CAM. (b) and (c) 3D reconstruction of CAM at different view angles (See Visualization 1 for more details).
Notably, the ultrasound waves induce equal but opposite frequency modulation along the fast and slow axis of the fiber-laser sensor. Ideally, the image reconstruction in PACT typically requires detection of the amplitude as well as the phase of the photoacoustic signal. However, in a practical case, the ultrasound transducer has a specific electrical impulse response (EIR), which can probably change the acoustic phase. Deconvolving of the raw channel data with the EIR is required, which demands the high signal-to-noise ratio and the accurate EIR measurement of each element. Here we used the envelope of the Hilbert transform of the bipolar signal for reconstruction, as indicated in [11]. It may introduce some image artifacts and the result is not as good as that by using universal back-projection reconstruction, as a result of the low-pass nature in spatial frequency of the Hilbert transformation. We found the image quality is acceptable despite these artifacts.
4. Conclusion
We have demonstrated 2D and 3D photoacoustic tomography by using a single fiber-laser based ultrasound sensor. Two-dimensional imaging was performed by linear scanning the sensor and three-dimensional imaging was further achieved by repeating this process at multiple angles via the inverse Radon transformation. Compared with the conventional piezoelectric transducer array, the fiber sensor offers a dual-60-degree angular vision, which enables visualizing more horizontal features. In vitro imaging of biological tissues suggests that the sensors can offer sufficient sensitivity. Notably, it is not sufficient to cover a full 180-degree angle, as a result of the cos(2θ) dependence. Also, the current scanning manner is time-consuming. We are currently working on the wavelength/beat-frequency multiplexing of such sensors to address these issues. By multiplexing an additional sensor whose principal axis is rotated by 45 degrees, we will be able to cover the full angular vision to resolve features at every orientation. Further, a multiplexed sensors array can help form a 2D image with a single laser pulse, which can significantly accelerate the imaging process.
Funding
National Natural Science Foundation of China (61675091, 61775083, 61805102, 81627805).
Disclosures
The authors declare no conflicts of interest.
References
1. L. V. Wang and S. Hu, “Photoacoustic tomography: in vivo imaging from organelles to organs”, Science 335(6075), 1458–1462 (2012). [CrossRef]
2. P. Beard, “Biomedical photoacoustic imaging,” Interface Focus 1(4), 602–631 (2011). [CrossRef]
3. J. Gamelin, A. Maurudis, A. Aguirre, F. Huang, P. Guo, L. V. Wang, and Q. Zhu, “A real-time photoacoustic tomography system for small animals,” Opt. Express 17(13), 10489–10498 (2009). [CrossRef]
4. J. Yao, J. Xia, K. I. Maslov, M. Nasiriavanaki, V. Tsytsarev, A. V. Demchenko, and L. V. Wang, “Noninvasive photoacoustic computed tomography of mouse brain metabolism in vivo,” NeuroImage 64, 257–266 (2013). [CrossRef]
5. A. P. Jathoul, J. Laufer, O. Ogunlade, B. Treeby, B. Cox, E. Zhang, P. Johnson, A. R. Pizzey, B. Philip, T. Marafioti, M. F. Lythgoe, R. B. Pedley, M. A. Pule, and P. Beard, “Deep in vivo photoacoustic imaging of mammalian tissues using a tyrosinase-based genetic reporter,” Nature Photonics 9, 239 (2015). [CrossRef]
6. L. Li, L. Zhu, C. Ma, L. Lin, J. Yao, L. Wang, K. Maslov, R. Zhang, W. Chen, and J. Shi, “Single-impulse panoramic photoacoustic computed tomography of small-animal whole-body dynamics at high spatiotemporal resolution,” Nat. Biomed. Eng. 1(5), 0071 (2017). [CrossRef]
7. P. Zhang, L. Li, L. Lin, J. Shi, and L. V. Wang, “In vivo super-resolution photoacoustic computed tomography by localization of single dyed droplets,” Light: Sci. Appl. 8(1), 36 (2019). [CrossRef]
8. A. Özbek, X. L Deán-Ben, and D. Razansky, “Optoacoustic imaging at kilohertz volumetric frame rates,” Optica 5(7), 857 (2018). [CrossRef]
9. Y. Qiu, J. V. Gigliotti, M. Wallace, F. Griggio, C. E. Demore, S. Cochran, and S. Trolier-McKinstry, “Piezoelectric micromachined ultrasound transducer (PMUT) arrays for integrated sensing, actuation and imaging,” Sensors 15(4), 8020–8041 (2015). [CrossRef]
10. J. Xia, M. R. Chatni, K. I. Maslov, Z. Guo, K. Wang, M. A. 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]
11. G. Li, L. Li, L. Zhu, J. Xia, and L. V. Wang, “Multiview Hilbert transformation for full-view photoacoustic computed tomography using a linear array,” J. Biomed. Opt. 20(6), 066010 (2015). [CrossRef]
12. D. Yang, D. Xing, S. Yang, and L. Xiang, “Fast full-view photoacoustic imaging by combined scanning with a linear transducer array,” Opt. Express 15(23), 15566–15575 (2007). [CrossRef]
13. P. Zhang, L. Li, L. Lin, P. Hu, J. Shi, Y. He, L. Zhu, Y. Zhou, and L. V. Wang, “High-resolution deep functional imaging of the whole mouse brain by photoacoustic computed tomography in vivo,” J. Biophotonics 11(1), e201700024 (2018). [CrossRef]
14. G. Li, J. Xia, K. Wang, K. Maslov, M. A. Anastasio, and L. V. Wang, “Tripling the detection view of high-frequency linear-array-based photoacoustic computed tomography by using two planar acoustic reflectors,” Quant Imaging Med Surg. 5(1), 57–62 (2015). [CrossRef]
15. W. Shu, M. Ai, T. Salcudean, R. Rohling, P. Abolmaesumi, and S. Tang, “Broadening the detection view of 2D photoacoustic tomography using two linear array transducers,” Opt. Express 24(12), 12755–12768 (2016). [CrossRef]
16. E. Zhang, J. Laufer, and P. Beard, “Backward-mode multiwavelength photoacoustic scanner using a planar Fabry-Perot polymer film ultrasound sensor for high-resolution three-dimensional imaging of biological tissues,” Appl. Opt. 47(4), 561–577 (2008). [CrossRef]
17. A. Maxwell, S.-W. Huang, T. Ling, J.-S. Kim, S. Ashkenazi, and L. J. Guo, “Polymer microring resonators for high-frequency ultrasound detection and imaging,” IEEE J. Sel. Top. Quantum Electron. 14(1), 191–197 (2008). [CrossRef]
18. C. Zhang, S.-L. Chen, T. Ling, and L. J. Guo, “Review of imprinted polymer microrings as ultrasound detectors: Design, fabrication, and characterization,” IEEE Sens. J. 15(6), 3241–3248 (2015). [CrossRef]
19. S. A. Carp and V. Venugopalan, “Optoacoustic imaging based on the interferometric measurement of surface displacement,” J. Biomed. Opt. 12(6), 064001 (2007). [CrossRef]
20. G. Wissmeyer, M. A. Pleitez, A. Rosenthal, and V. Ntziachristos, “Looking at sound: optoacoustics with all-optical ultrasound detection,” Light: Sci. Appl. 7(1), 53 (2018). [CrossRef]
21. P. Burgholzer, G. J. Matt, M. Haltmeier, and G. Paltauf, “Exact and approximative imaging methods for photoacoustic tomography using an arbitrary detection surface,” Phys. Rev. E 75(4), 046706 (2007). [CrossRef]
22. R. Nuster, M. Holotta, C. Kremser, H. Grossauer, P. Burgholzer, and G. Paltauf, “Photoacoustic microtomography using optical interferometric detection,” J. Biomed. Opt. 15(2), 021307 (2010). [CrossRef]
23. H. Grün, T. Berer, P. Burgholzer, R. Nuster, and G. Paltauf, “Three-dimensional photoacoustic imaging using fiber-based line detectors,” J. Biomed. Opt. 15(2), 021306 (2010). [CrossRef]
24. F. Zhou, L. Jin, Y. Liang, L. Cheng, and B.-O. Guan, “Spatial sensitivity characterization of dual-polarization fiber grating laser sensors,” J. Lightwave Technol. 33(19), 4151–4155 (2015). [CrossRef]
25. B.-O. Guan, L. Jin, L. Cheng, and Y. Liang, “Acoustic and ultrasonic detection with radio-frequency encoded fiber laser sensors,” IEEE J. Sel. Top. Quantum Electron. 23(2), 302–313 (2017). [CrossRef]
26. M. Xu and L. V. Wang, “Universal back-projection algorithm for photoacoustic computed tomography,” Phys. Rev. E 71(1), 016706 (2005). [CrossRef]
27. B. E. Treeby, E. Z. Zhang, and B. T. Cox, “Photoacoustic tomography in absorbing acoustic media using time reversal,” Inverse Probl. 26(11), 115003 (2010). [CrossRef]
28. H. P. Hiriyannaiah, “X-ray computed tomography for medical imaging,” IEEE Signal Process. Mag. 14(2), 42–59 (1997). [CrossRef]
29. M. Naito, K. Nirasawa, and T. Oishi, “Development in culture of the chick embryo from fertilized ovum to hatching,” J. Exp. Zool. 254(3), 322–326 (1990). [CrossRef]
