We present a 3D-visual laser-diode-based photoacoustic imaging (LD-PAI) system with a pulsed semiconductor laser source, which has the properties of being inexpensive, portable, and durable. The laser source was operated at a wavelength of 905 nm with a repetition rate of 0.8 KHz. The energy density on the sample surface is about 2.35 mJ/cm2 with a pulse energy as low as 5.6 μJ. By raster-scanning, preliminary 3D volumetric renderings of the knotted and helical blood vessel phantoms have been visualized integrally with an axial resolution of 1.1 mm and a lateral resolution of 0.5 mm, and typical 2D photoacoustic image slices with different thickness and orientation were produced with clarity for detailed comparison and analysis in 3D diagnostic visualization. In addition, the pulsed laser source was integrated with the optical lens group and the 3D adjustable rotational stage, with the result that the compact volume of the total radiation source is only 10 × 3 × 3 cm3. Our goal is to significantly reduce the costs and sizes of the deep 3D-visual PAI system for future producibility.
©2012 Optical Society of America
Photoacoustic imaging (PAI) is a hybrid biomedical technique for visualizing blood vessels and blood-containing structures with high optical contrast and high ultrasound resolution [1,2]. In the past decade, PAI has attracted considerable attention and made great progress in the fields of detecting brain structure and function [2–7], vasculature networks [1,8–11], organic cells , heterogeneous media [13,14], chemical traces [15–18], and tumor tissue [19,20]. In order to generate the photoacoustic effect efficiently, a bulky and expensive solid-state laser is generally preferred with a pulse duration of tens of nanoseconds and a pulse energy of several millijoules, such as a Q-switched Nd:YAG laser. In order to realize functional PAI, wavelength-tunable laser sources are usually employed, such as Q-switched Nd:YAG pumped OPO, Ti:sapphire, and dye laser systems. Although widely used as laboratory tools, such laser systems have limitations in photoacoustic applications because of their high cost, large size, and cooling requirements. Furthermore, the low pulse repetition frequency (PRF) of tens of hertz is unsuited for developing real-time PAI systems for clinical medical diagnosis.
As an alternative laser source, laser diodes with pulsed or continuous-wave (CW) excitation could overcome those limitations for PAI. They are relatively inexpensive, compact, simple, and high PRF. Moreover, they are readily available in a wide range of visible and NIR wavelengths without additional instruments. The feasibility of the use of laser diodes in two-dimensional (2D) PAI has been explored in phantom studies [21,22]. Unfortunately, the main disadvantage of the laser diode is its low peak output power, which (for nanosecond pulse duration) limits the pulse energy available for an effective photoacoustic effect to several microjoules over a large region. One compensating method, at the cost of system complexity, is to combine the outputs of several diodes with a fiber bundle to increase the total energy several-fold; this technique has been applied with sideward-mode detection for visualizing 2D superficial vascular anatomy, and it has the potential to develop multiwavelength functional applications when different wavelength laser diodes are added . Modulation and coding techniques may be another potential concept for increasing the stimulus laser energy [24–26]. For instance, a photoacoustic microscope (PAM) using an intensity-modulated CW laser diode source has been built to image 2D rabbit ear vasculature, but it is difficult to capture depth-direction information about the photoacoustic time of flight for three-dimensional (3D) photoacoustic detection .
Here, we report an inexpensive, portable, and durable 3D-visual LD-PAI system using a high-power pulsed laser diode source. As a first attempt to verify the possibility, preliminary clear 2D and 3D photoacoustic visualizations of the blood vessel phantom were performed. The experimental results demonstrated the potential utility of semiconductor laser sources for miniaturizing deep 3D PAI systems.
2. Experimental setup
The laser source comprises a high-power pulsed laser diode array, whose structure is a 2D stacked chip with high-reliability AIGaAs quantum wells. The emitting area of the stack array is 400 × 340 μm2 with four active elements, which provide a peak optical output power of 140 W. With a custom-built driver circuit, the pulsed laser source operates at a 905 ± 15 nm wavelength when driven with a max peak forward current of 30 A. Its duty cycle was 0.08‰, allowing the laser diode array to be operated at a PRF of 0.8 KHz for a 100 ns pulse duration. A focusing lens group consisting of collimation and focusing modules was used to focus and project the laser pulses onto the phantom. The clear aperture (CA) of the lens group was 10 mm with a full-angle divergence of 3.0 mrad. Figure 1(a) shows the front photograph of the focusing lens group, which has a long focus length of 15 cm with a compact volume of 2.5 × 3.0 × 3.0 cm3. When the laser beam passes through the focusing lens group, the energy loss is 60% lower than that of a fiber-focused system, resulting in a pulse energy on the tissue phantom of approximately 5.6 μJ. As shown in Fig. 1(d), the effective spot size is determined to be 0.55 mm, according to the photoacoustic amplitude profile plotted as a function of the y-axes with a point absorber (diameter 0.5 mm). Therefore the pulse energy density on the phantom surface is ~2.35 mJ/cm2, which is much lower than the ANSI 2000 safety limit of 20 mJ/cm2. For accurate focus, the laser source integrated with the optical lens group is fixed on a 3D adjustable rotational stage, and the compact volume of the total radiation source is only 10 × 3 × 3 cm3 as shown in Fig. 1(b).
The schematic of the 3D-visual LD-PAI system is shown in Fig. 1(c). A focused ultrasonic transducer (I2P10NF40, Doppler, China) was used as a forward-mode photoacoustic sensor with a central frequency of 1.95 MHz and a focal length 37.5 mm. The diameter of the active element is 10 mm, and the relative pulse-echo sensitivity (Srel) is −48 dB. The tissue phantom was mounted on a C-scan translation stage (MTS101, Boif, China), which was moved in raster-scanning order in the perpendicular plane (y-z plane) for 3D photoacoustic acquisition. The phantom and transducer were immersed in the fluid medium of water. After a 20 minute system warm-up, the pulsed laser was generated and projected onto the tissue phantom for photoacoustic generation. The excited photoacoustic signal captured by the transducer was subsequently amplified by a preamplifier (5678, Olympus, Japan), then synchronously recorded by a mixed-signal oscilloscope (54642D, Agilent, USA), and finally transferred to a personal computer by a GPIB bus card (GPIB-USB-HS, National Instruments, USA). The preamplifier provides a high voltage gain (40 dB) and low input noise (20 μV p-p) at a 50 KHz to 20 MHz bandwidth, and the oscilloscope features a high-speed 8-bit analogue-to-digital converter with a maximal sampling rate of 500 MHz. The total system operation and 3D-visual data procession are executed on LabVIEW Full Dev. System (version 8.0, National Instruments, USA) and MATLAB (version 7.0, Mathworks, USA). In the experiments, the ultrasound velocity is assumed to be 1500 m/s for photoacoustic reconstruction.
The spatial resolution of the 3D-visual LD-PAI system was tested by imaging two carbon rods (diameter, 0.6 mm; distance between the rods, 3.3 mm), which were fixed in an artificial phantom (13% gelatine, 12.5% milk, and 74.5% water). The effective attenuation coefficient of the phantom is 1.2 cm−1, used to simulate the optical properties of human breast . Figure 2(a) shows the photoacoustic reconstructed image of the phantom with the two inserted carbon rods. The photoacoustic signal amplitude was plotted as a function of the x- and y-axes in Fig. 2(b), and the axial resolution was determined to be ~1.1 mm, according to the full width at half-maximum (FWHM) of the pressure sensitivity distribution along the x-axis. Figure 2(c) is the normalization line profile of the reconstructed image shown in Fig. 2(a) with x = 44.75 mm. The profile includes the two absorption rods. The 40.5%-amplitude line intercepts the profile at points A, B, C, and D, and it also intercepts the respective centerlines of the two absorption peaks at points E and F. The lateral resolution is estimated according to the Rayleigh criterion. The two sources can no longer be clearly distinguished when point B touches point C in Fig. 2(c). Therefore, the minimum distinguishable distance, R, between the two sources is approximately R = |EB| + |CF| − 2r, where r is the radius of the absorption source. From Fig. 2(c), R is measured to be ~0.50 mm.
To demonstrate the feasibility of the 3D-visual LD-PAI system for biomedical applications, a knotted blood vessel was embedded in the same cylindrical gelatin phantom at a depth of 5 mm as in the inset in Fig. 3(a) . The artificial blood vessel filled with Chinese ink dye water solution (μa = 73 cm−1 at 905 nm) was made of polytetrafluoroethene (PTFE) tubing with an inner diameter of 0.6 mm, and its non-planar structure has a volume of 24 × 10 × 3 mm3 approximately. Raster scanning of the phantom was performed along the x-axis in 130 0.2 mm steps and along the z-axis in 24 0.5 mm steps. Meanwhile, a 40 μs acquisition time of each A-scan line was recorded by signal averaging 128 pulses. Figure 3(a) shows the 3D rendering of the photoacoustic volumetric data set with the viewpoint of view(7, 15), which specifies a viewing angle for a 3D space in terms of azimuth and elevation. The knotted structure has been reconstructed well and matches the phantom photograph.
A typical 2D photoacoustic cross-sectional image of Fig. 3(b) in the x-y plane was obtained at z = 7.0 mm. It can be seen that three orders of vessel section (A–C) can be observed clearly, and especially section C has two blood vessels at the intersecting point. Their horizontal and depth-resolved intervals are approximately 0.75 mm and 1.35 mm, respectively. Meanwhile, the trigger signal has a peak voltage of 13 mV and a duration time of 2.2 μs, which may result in an ileocecal region of approximately 3.5 mm as shown in Fig. 3(b). The duration time of the trigger signal is dependent mainly on the current pulse duration, which is caused by the drive current pulse of the laser diode excitation source. Figure 3(c) shows the 3D photoacoustic rendering with the viewpoint of view(0, 0), and the random background noise can be identified in each x-y plane along the z-axis. It has better visual effect than the 2D photoacoustic projection image of Fig. 3(d), and the recorded photoacoustic signal-to-noise ratio (SNR) of 20.6 dB could be further improved by additional improvement, such as increasing the number of pulses averaged.
Biological tissue is relatively transparent to light in the NIR region (700–1000 nm) and gives better tissue transmission, allowing a deeply penetrating 3D PAI of up to several centimeters without an optical contrast agent. In order to validate the potential of the 3D-visual LD-PAI system, a helical blood vessel was embedded in the same gelatin phantom as shown in Fig. 4(a) . On the vessel, there was a pinch point, a mark made by hand vice. It had an inner diameter of 0.3 mm and a straight depth of approximately12 mm. The phantom was raster scanned along the y-axis in 120 0.2 mm steps and the z-axis in 100 0.2 mm steps. Meanwhile, a 20 μs acquisition time of each A-scan line was recorded by signal averaging 256 pulses. Figure 4(b) shows a typical 3D rendering for view(170, 22), which only projects the partial 3D volumetric data with a signal amplitude of 2.0 mV. It can be seen that the helical structure of the blood vessel phantom was reconstructed clearly, but there was a lot of background noise. Figure 4(c) shows the 3D rendering of the 3D photoacoustic volumetric data set for view(0, 0). It matches the photograph of Fig. 4(a), but the pinch point (D) is too blurry to be discriminated. With the viewpoint of view(−135, 12), the 3D volumetric structure of the blood vessel phantom has been visualized integrally. Compared with the foregoing experiment, the reconstructed 3D photoacoustic image has a lower level background noise with increased data averaging. By looking carefully at Fig. 4(c), it can be seen that the reconstruction of the middle section of the helical blood vessel blurry, because the focused transducer can receive only a limited signal from the boundaries of the phantom that are nearly perpendicular to the x-axis of the transducer.
In clinical diagnosis, reconstructed image slices with different thicknesses and orientations are usually needed in order to perform detailed comparison and analysis in 3D visualization. Figure 5(a) shows the 2D photoacoustic projection image of the helical blood vessel, which matches well with the 3D rendering of Fig. 4(c). The pinch point (D) can be discriminated clearly, especially in the photoacoustic projection image slices on the x-axis of 46.5 mm with a thickness of 0.5 mm [Fig. 5(b)]. Figures 5(c) and 5(d) show two typical 2D image slices in the y-z plane, whose thicknesses are 0.5 mm and 5.0 mm on the x-axis of 37.5 mm. The thinner the slice, the more image detail is revealed. In the procession of image reconstruction, the raw data was applied without any signal processing, such as filtering and smoothing. From the linear color bar, it can be seen that the maximal absorption of the reconstructed vessel in Figs. 5(c) and 5(d) is reduced by approximately one-third of that in Fig. 5(b), resulting mostly from the optical attenuation of the pulsed laser beam when it traverses through the gelatin background along the x-axis (in the depth direction). In addition, the intuitive depth-direction information along the x-axis can be accurately provided by the cross-sectional photoacoustic image in the x-y plane. Figures 5(e) and 5(f) show two typical 2D cross-sectional (B-scan) images in the x-y plane along the dotted lines in Fig. 5(a), whose positions on the z-axis are 4.4 mm and 8.4 mm, respectively. It can be seen that the different sections (E–K) of the helical blood vessel are reconstructed clearly, and the relative positions along the depth direction (x-axis) are located accurately. The brightness of the blood vessel section represents the amplitude of the captured photoacoustic signal, which is related to the intersection angle between the boundary of the phantom and the transducer axis.
4. Discussion and conclusion
After a careful observation of Fig. 4(c), it can be seen that the inner ring of the blood vessel is distinctly broadened relative to the outer ring, because the lateral resolution will degrade as a function of distance away from the transducer focus point. This can be improved effectively with a phased or line-focused transducer, and the axis resolution could be increased when the transducer is made of polyvinylidene fluoride (PVDF) film or composite piezoelectric material. Certainly, the penetrating depth in a living animal will be reduced to a certain extent because of the tissue’s attenuation of the weak diode laser, so this method would be developed mostly for imaging superficial vascular network and pathological tissue.
In medical applications, 3D-visual reconstruction is essential for observing the organ contour and internal structure. The marching cubes (MC) algorithm is one of the most popular methods for extracting a polygonal mesh (sometimes called a voxel) of an iso-surface from a 3D scalar field [27,28]. In the experiments, the voxel sizes in Fig. 3 and Fig. 4 are approximately 0.06 × 0.2 × 0.5 mm3 and 0.1 × 0.2 × 0.2 mm3, respectively. Rendering a 2D photoacoustic projection of the 3D data set requires every sample value to have a defined opacity and color in the volume. This is done with a simple ramp function to project an RGBA (for red, green, blue, alpha) value in the 3D-visual photoacoustic rendering, which is reconstructed by the following procedure: 1) defining the 3D PAI region; 2) drawing an iso-surface using histogram or specified value of the 3D data set; 3) calculating the normals of the iso-surface vertices with the gradient of the 3D data set; 4) setting the face and edge color; 5) adding lights; 6) specifying the transparency and viewpoint.
To avoid catastrophic optical damage at the facet, the peak output power of the laser diode (typically <200 W) is several orders of magnitude less than a conventional Nd:YAG laser. Therefore, the LD-PAI system is unsuitable for directly irradiating a relatively large tissue volume with a comparatively large-area incident beam; otherwise signal averaging needs thousands to tens of thousands of instances (i.e., 5,000–50,000) [21–23]. Early studies tended to rely on focusing the laser output to a small spot on the surface of the target to achieve sufficient energy density for chemical trace detection, such as blood oxygen and glucose . But with PAM development, the portable PAM based on a semiconductor laser source may be an attractive prospect for integration with raster-scanning of the focused spot by a stepper motor or optical galvanometer scanner .
The current 3D-visual LD-PAI system employed impractical forward-mode detection, and the next step will focus on developing backward-mode sensing with a hollow annular array transducer . Like a tunable laser source, the LD-PAI technique can provide in vivo functional information of the biological organism as well as if multiple laser diodes of different wavelengths had been used for spectral measurements . Because of the compact volume of LD-PAI system, it has possible applications for endoscopic PAI of internal organs , intravascular plaque , and pathological wall of intestine .
In this feasibility study, we demonstrated a 3D-visual LD-PAI system with a pulsed semiconductor laser source. Preliminary 3D volumetric renderings of the knotted and helical blood vessel phantoms have been visualized integrally and matched well with photographs. Meanwhile, typical clear 2D photoacoustic image slices with different thickness and orientation were obtained for detailed comparison and analysis in 3D diagnostic visualization. The laser diode excitation source costs no more than twenty thousand Yuan (RMB), including material and design charges for the laser diode and its driver circuit, which is far below the typical price of Q-switched Nd:YAG lasers. Compared with the conventional photoacoustic system with a solid-state laser, the proposed system is more suitable for future producibility with the properties of being inexpensive, portable, and durable.
This research is supported in part by the National Natural Scientific Foundation of China (61068002), the National Torch Plan Project of China (2010GH041570), the Science and Technology Pillar Program of Jiangxi Province (2009BSA12700), the Comprehensive Strategic Cooperation Project between Guangdong Province and Chinese Academy (2010B090300028), the Natural Science Foundation of Jiangxi Province (2008GQW0013), the Scientific Research Foundation of Jiangxi and Hunan Provincial Education Bureau (GJJ10243, 09C314), and the Open Foundation of the Key Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong University (ZD201029010).
References and links
1. T. Harrison, J. C. Ranasinghesagara, H. Lu, K. Mathewson, A. Walsh, and R. J. Zemp, “Combined photoacoustic and ultrasound biomicroscopy,” Opt. Express 17(24), 22041–22046 (2009). [CrossRef] [PubMed]
2. R. Olafsson, D. R. Bauer, L. G. Montilla, and R. S. Witte, “Real-time, contrast enhanced photoacoustic imaging of cancer in a mouse window chamber,” Opt. Express 18(18), 18625–18632 (2010). [CrossRef] [PubMed]
3. 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(19), 194101 (2005). [CrossRef]
4. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Functional photoacoustic microscopy for high-resolution and noninvasive in vivo imaging,” Nat. Biotechnol. 24(7), 848–851 (2006). [CrossRef] [PubMed]
5. L. Zeng, D. Xing, H. Gu, D. Yang, S. Yang, and L. Xiang, “High antinoise photoacoustic tomography based on a modified filtered backprojection algorithm with combination wavelet,” Med. Phys. 34(2), 556–563 (2007). [CrossRef] [PubMed]
6. D. W. Yang, D. Xing, S. H. Yang, and L. Z. Xiang, “Fast full-view photoacoustic imaging by combined scanning with a linear transducer array,” Opt. Express 15(23), 15566–15575 (2007). [CrossRef] [PubMed]
7. J. Gamelin, A. Aguirre, A. Maurudis, F. Huang, D. Castillo, L. V. Wang, and Q. Zhu, “Curved array photoacoustic tomographic system for small animal imaging,” J. Biomed. Opt. 13(2), 024007 (2008). [CrossRef] [PubMed]
9. M. P. Fronheiser, S. A. Ermilov, H. P. Brecht, A. Conjusteau, R. Su, K. Mehta, and A. A. Oraevsky, “Real-time optoacoustic monitoring and three-dimensional mapping of a human arm vasculature,” J. Biomed. Opt. 15(2), 021305 (2010). [CrossRef] [PubMed]
11. K. Jansen, A. F. van der Steen, H. M. van Beusekom, J. W. Oosterhuis, and G. van Soest, “Intravascular photoacoustic imaging of human coronary atherosclerosis,” Opt. Lett. 36(5), 597–599 (2011). [CrossRef] [PubMed]
12. E. I. Galanzha, E. V. Shashkov, T. Kelly, J. W. Kim, L. Yang, and V. P. Zharov, “In vivo magnetic enrichment and multiplex photoacoustic detection of circulating tumour cells,” Nat. Nanotechnol. 4(12), 855–860 (2009). [CrossRef] [PubMed]
13. Z. Yuan and H. Jiang, “Quantitative photoacoustic tomography: Recovery of optical absorption coefficient maps of heterogeneous media,” Appl. Phys. Lett. 88(23), 231101 (2006). [CrossRef]
14. Z. Yuan, Q. Wang, and H. Jiang, “Reconstruction of optical absorption coefficient maps of heterogeneous medium by photoacoustic tomography coupled with diffusion equation based regularized Newton Method,” Opt. Express 15(26), 18076–18081 (2007). [CrossRef]
17. A. de la Zerda, Z. Liu, S. Bodapati, R. Teed, S. Vaithilingam, B. T. Khuri-Yakub, X. Chen, H. Dai, and S. S. Gambhir, “Ultrahigh sensitivity carbon nanotube agents for photoacoustic molecular imaging in living mice,” Nano Lett. 10(6), 2168–2172 (2010). [CrossRef] [PubMed]
18. Y. Wang, S. Hu, K. Maslov, Y. Zhang, Y. N. Xia, and L. V. Wang, “In vivo integrated photoacoustic and confocal microscopy of hemoglobin oxygen saturation and oxygen partial pressure,” Opt. Lett. 36(7), 1029–1031 (2011). [CrossRef] [PubMed]
19. P. Ephrat, G. C. Albert, M. B. Roumeliotis, M. Belton, F. S. Prato, and J. J. Carson, “Localization of spherical lesions in tumor-mimicking phantoms by 3D sparse array photoacoustic imaging,” Med. Phys. 37(4), 1619–1628 (2010). [CrossRef] [PubMed]
20. J. Staley, P. Grogan, A. K. Samadi, H. Z. Cui, M. S. Cohen, and X. M. Yang, “Growth of melanoma brain tumors monitored by photoacoustic microscopy,” J. Biomed. Opt. 15(4), 040510 (2010). [CrossRef] [PubMed]
22. T. J. Allen and P. C. Beard, “Dual wavelength laser diode excitation source for 2D photoacoustic imaging,” Proc. SPIE 6437, 64371U, 64371U-9 (2007). [CrossRef]
25. M. P. Mienkina, C. S. Friedrich, N. C. Gerhardt, M. F. Beckmann, M. F. Schiffner, M. R. Hofmann, and G. Schmitz, “Multispectral photoacoustic coded excitation imaging using unipolar orthogonal Golay codes,” Opt. Express 18(9), 9076–9087 (2010). [CrossRef] [PubMed]
27. W. E. Lorensen and H. E. Cline, “Marching cubes: a high resolution 3D surface construction algorithm,” Comput. Graph. 21(4), 163–169 (1987). [CrossRef]
28. S. Raman and R. Wenger, “Quality isosurface mesh generation using an extended marching cubes lookup table,” Comput. Graph. Forum 27(3), 791–798 (2008). [CrossRef]
29. K. M. Quan, G. B. Christison, H. A. MacKenzie, and P. Hodgson, “Glucose determination by a pulsed photoacoustic technique: an experimental study using a gelatin-based tissue phantom,” Phys. Med. Biol. 38(12), 1911–1922 (1993). [CrossRef] [PubMed]
30. L. M. Zeng, G. D. Liu, D. W. Yang, Z. Ren, and Z. Huang, “Design of a portable noninvasive photoacoustic glucose monitoring system integrated laser diode excitation with annular array detection,” Proc. SPIE 7280, 72802F (2009).
32. S. Sethuraman, J. H. Amirian, S. H. Litovsky, R. W. Smalling, and S. Y. Emelianov, “Spectroscopic intravascular photoacoustic imaging to differentiate atherosclerotic plaques,” Opt. Express 16(5), 3362–3367 (2008). [CrossRef] [PubMed]
33. Y. Yuan, S. H. Yang, and D. Xing, “Preclinical photoacoustic imaging endoscope based on acousto-optic coaxial system using ring transducer array,” Opt. Lett. 35(13), 2266–2268 (2010). [CrossRef] [PubMed]