-
光声信号的产生是由于短时脉冲激光照射生物组织,组织中的吸收体吸收一部分能量使得局部温度升高,导致发生热弹性膨胀,从而产生超声波。在满足热力限制和应力限制的条件下,声压p(r, t)满足的关系[22]:
$ \begin{array}{l} \frac{{{\partial ^2}p(\mathit{\boldsymbol{r}}, t)}}{{\partial {t^2}}} - {c^2}\left( \mathit{\boldsymbol{r}} \right)\rho \left( \mathit{\boldsymbol{r}} \right)\nabla \cdot \\ \left[ {\frac{1}{{\rho \left( \mathit{\boldsymbol{r}} \right)}}\nabla p(\mathit{\boldsymbol{r}}, t)} \right] = \mathit{\Gamma }\frac{{\partial H(\mathit{\boldsymbol{r}}, t)}}{{\partial t}} \end{array} $
(1) 式中,r为3维空间内的位置坐标; t表示时间; c(r)和ρ(r)分别为组织的声速和密度; Γ为无量纲格鲁内森参量; H(r, t)是热源函数,代表单位时间、单位体积内的热量。假设生物组织的密度是均匀的,即在成像区域内ρ(r)是常数,并假设声速是均匀的,即c(r)为常数c,另外热源函数还可以表示为H(r, t)=H(r)×H(t),其中H(r)表示单位体积内沉积的热能量,H(t)表示脉冲激光光强随时间的分布函数。在实际成像过程中脉冲激光脉宽很短,理论上光强函数可假设为一个脉冲函数,即H(t)=δ(t),因此(1)式可表示为:
$ \frac{{{\partial ^2}p(\mathit{\boldsymbol{r}}, t)}}{{\partial {t^2}}} - {c^2}{\nabla ^2}p(\mathit{\boldsymbol{r}}, t) = \mathit{\Gamma }H\left( \mathit{\boldsymbol{r}} \right)\frac{{\partial \delta \left( t \right)}}{{\partial t}} $
(2) (2) 式可以等价地表示为一个初始值问题:
$ \frac{{{\partial ^2}p(\mathit{\boldsymbol{r}}, t)}}{{\partial {t^2}}} - {c^2}{\nabla ^2}p(\mathit{\boldsymbol{r}}, t) = 0 $
(3) 初始条件为:
$ \left\{ \begin{array}{l} p(\mathit{\boldsymbol{r}}, t)\left| {_{t = 0} = \mathit{\Gamma }H\left( \mathit{\boldsymbol{r}} \right)} \right.\\ \frac{{\partial p(\mathit{\boldsymbol{r}}, t)}}{{\partial t}}\left| {_{t = 0} = 0} \right. \end{array} \right. $
(4) 上述初始值问题可以通过求解一个泊松类型的积分[23]而得到一个解析解:
$ p(\mathit{\boldsymbol{r}}, t) = \frac{\mathit{\Gamma }}{{4{\rm{ \mathit{ π} }}c}}\;\frac{\partial }{{\partial t}}\int_{S'\left( {\mathit{\boldsymbol{r}}, t} \right)} {\frac{{H\left( {\mathit{\boldsymbol{r'}}} \right)}}{{\left| {\mathit{\boldsymbol{r}} - \mathit{\boldsymbol{r'}}} \right|}}} {\rm{d}}S'(\mathit{\boldsymbol{r}}, t) $
(5) 式中,积分的对象是一个半径为|r-r′|=ct的球形表面S′(r, t),在2维平面即断层平面,所有的光声信号源和测量点位于同一平面,此时积分是沿着半径为|r-r′|=ct的圆周L′(t)开展,忽略(5)式中的常数,光声断层成像的前向模型可表示为:
$ p(\mathit{\boldsymbol{r}}, t) = \frac{\partial }{{\partial t}}\int_{L'\left( t \right)} {\frac{{H\left( {\mathit{\boldsymbol{r'}}} \right)}}{{\left| {\mathit{\boldsymbol{r}} - \mathit{\boldsymbol{r'}}} \right|}}} {\rm{d}}L'(t) $
(6) -
光声断层重建算法主要分为3种:第1种是基于雷登变换的滤波反投影重建算法; 第2种是基于时间反转方法的重建算法; 第3种是基于模型的重建算法。滤波反投影算法虽然实现容易且重建速度很快,但是其重建图像含有条状伪影而影响图像质量。时间反转方法虽然通过反向模拟超声波传播来得到更好质量的重建图像,但是这个过程需要复杂的数值仿真,不适合实时成像的要求。而基于模型的重建算法是在采集的声压信号数据和组织的吸收分布之间建立一种线性映射关系,继而通过最优化方法去最小化采集的声压信号与利用模型计算的声压信号之间的误差。因而具有很强的灵活性,且模型矩阵只与所使用的图像网格和信号采集系统的参量有关,而与实际的成像对象无关。
在基于模型的光声断层重建算法中,第1步是计算模型矩阵,需要用到导数的数值近似表示,因此(6)式可近似表示为:
$ p(\mathit{\boldsymbol{r}}, t) \approx \frac{{I\left( {\mathit{\boldsymbol{r}}, t\Delta t} \right) - I\left( {\mathit{\boldsymbol{r}}, t - \Delta t} \right)}}{{2\Delta t}} $
(7) 式中, 采用的是导数的中间差分近似。I(r, t)为:
$ I\left( {\mathit{\boldsymbol{r}}, t\Delta t} \right) = \int_{L'\left( t \right)} {\frac{{H\left( {\mathit{\boldsymbol{r'}}} \right)}}{{\left| {\mathit{\boldsymbol{r}} - \mathit{\boldsymbol{r'}}} \right|}}} {\rm{d}}L'(t) $
(8) 第2步就是计算(8)式,方法有很多[24-25]。最后(8)式和(9)式可以表示为:
$ \mathit{\boldsymbol{p}} = \mathit{\boldsymbol{Ax}} $
(9) 式中,p∈Rm为向量化表示的超声换能器阵元采集到的声压信号, x∈Rn为向量化表示的吸收分布,也即初始声压分布,A∈Rm×n是模型矩阵或系统矩阵,表示一个线性算子描述组织的光学吸收分布与换能器探测的声压信号数据之间的关系。
基于模型的重建算法可以分为两类,第1类是通过求解(10)式最小二乘问题的算法,成为朴素算法,其解称为朴素解:
$ {\mathit{\boldsymbol{x}}_{{\rm{native}}}} = \mathop {{\rm{argmin}}}\limits_x \left\| {\mathit{\boldsymbol{p}} - \mathit{\boldsymbol{Ax}}} \right\|_2^2 $
(10) 式中,xnaive为最终求解的朴素解。
第2类算法为基于正则化的算法,由于基于模型的光声断层图像重建问题通常是病态的,因此朴素解与精确解之间通常有较大的偏离,为了获得更好的近似解,需要使用各种正则化项,比如Tikhonov正则化、稀疏正则化、全变分正则化等。
高速128通道小动物多光谱光声断层成像系统
High-speed and 128-channel multi-spectral photoacoustic tomography system for small animal
-
摘要: 为了实现小动物光声断层信号的高速采集和实时高质量图像的重建,采用了覆盖角度为270°的128阵元弧形聚焦超声换能器、4个32通道的NI公司数据采集模块和可调谐脉冲激光器以及正则化优化的基于模型的光声断层重建算法。结果表明,系统的空间分辨率可以达到180μm;此系统可以在1ms内完成光声断层数据的采集,在40s以内获得高质量的重建图像。该系统可以用于开展小动物在体的多光谱光声断层成像实验研究。Abstract: In order to achieve high-speed acquisition and real-time high-quality image reconstruction of photoacoustic tomography signals for small animal, a 128 element arc-shaped focused ultrasound transducer with 270° coverage angle, four 32 channel NI data acquisition modules, a tunable pulsed laser and model-based photoacoustic tomography reconstruction algorithm with regularized optimization were used. The spatial resolution of the system can be up to 180μm. The in vitro and in vivo imaging experiments show that the system can complete photoacoustic tomography data acquisition within 1ms, and obtain high quality reconstructed images within 40s. The system can be used to carry out in vivo multi-spectral photoacoustic tomography of small animals.
-
Figure 6. a—characteristic curve of horizontal pixel point in the reconstruction result of pen core b—partial enlargement of the second peak signals of Fig. 6a
-
[1] KARABUTOV A, SAVATEEVA E, ORAEVSKY A. Optoacoustic tomography:New modality of laser di agnostic systems[J]. Laser Physics, 2003, 13(5):711-723. [2] WANG L V, HU S. Photoacoustic tomography:in vivo imaging from organelles to organs[J]. Science, 2012, 335(6705):1458-1462. [3] WANG X, PANG Y, KU G, et al. Noninvasive laser-induced photoacoustic tomography for structural and functional in vivo imaging of the brain[J]. Nature Biotechnology, 2003, 21(7):803-806. doi: 10.1038/nbt839 [4] LI C, WANG L V. Photoacoustic tomography and sensing in biomedicine[J]. Physics in Medicine and Biology, 2009, 54(19):R59-R97. doi: 10.1088/0031-9155/54/19/R01 [5] WANG L V. Prospects of photoacoustic tomography[J].Medical Physics, 2008, 35(12):5758-5767. doi: 10.1118/1.3013698 [6] NTZIACHRISTOS V, RAZANSKY D. Molecular imaging by means of multispectral optoacoustic tomography (MSOT)[J]. Chemical Reviews, 2010, 110(5):2783-2794. doi: 10.1021/cr9002566 [7] KOLKMAN R G M, BRANDS P J, STEENBERGEN W, et al. Real-time in vivo photoacoustic and ultrasound imaging[J]. Journal of Biomedical Optics, 2008, 13(5):050510. doi: 10.1117/1.3005421 [8] NIEDERHAUSER J J, JAEGER M, LEMOR R, et al. Combined ultrasound and optoacoustic system for real-time high-contrast vascular imaging in vivo[J]. IEEE Transactions on Medical Imaging, 2005, 24(4):436-440. doi: 10.1109/TMI.2004.843199 [9] OLAFSSON R, BAUER D R, MONTILLA L G, et al. Real-time, contrast enhanced photoacoustic imaging of cancer in a mouse window chamber[J]. Optics Express, 2010, 18(18):18625-18632. doi: 10.1364/OE.18.018625 [10] XU D, LIANG Zh X, JI X. Photoacoustic imaging system based on multi-channel parallel acquisition[J]. Chinese Journal of Lasers, 2011, 38(2):0204002(in Chinese). doi: 10.3788/CJL [11] XIANG L, WANG B, JI L, et al. 4-D photoacoustic tomography[J]. Scientific Reports, 2013, 3:1113. doi: 10.1038/srep01113 [12] TAN Y, HE J F, REN Y J, et al. Fast photoacoustic imaging systems based on rotating multi-element linear transducer array[J]. Laser Technology, 2009, 33(5):300-302(in Chinese). [13] CHEN Z, TANG Z, WAN W. Photoacoustic tomography imaging based on a 4f acoustic lens imaging system[J]. Optics Express, 2007, 15(8):4966-4976. doi: 10.1364/OE.15.004966 [14] KOZHUSHKO V, KHOKHLOVA T, ZHARINOV A, et al. Focused array transducer for two-dimensional optoacoustic tomography[J]. The Journal of the Acoustical Society of America, 2004, 116(3):1498-1506. doi: 10.1121/1.1781710 [15] KRUGER R A, KISER W L, Jr, REINECKE D R, et al. Thermoacoustic computed tomography using a conventional linear transducer array[J]. Medical Physics, 2003, 30(5):856-860. doi: 10.1118/1.1565340 [16] PARK S, MALLIDI S, KARPIOUK A B, et al. Photoacoustic imaging using array transducer[C]//International Society for Optics and Photonics. San Jose, USA: SPIE, 2007: 643714. [17] YIN B, XING D, WANG Y, et al. Fast photoacoustic imaging system based on 320-element linear transducer array[J]. Physics in Medicine and Biology, 2004, 49(7):1339-1346. doi: 10.1088/0031-9155/49/7/019 [18] ZEMP R J, BITTON R, LI M L, et al. Photoacoustic imaging of the microvasculature with a high-frequency ultrasound array transducer[J]. Journal of Biomedical Optics, 2007, 12(1):010501. doi: 10.1117/1.2709850 [19] KRUGER R A, KISER W L, REINECKE D R, et al. Thermoacoustic molecular imaging of small animals[J]. Molecular Imaging, 2003, 2(2):113-123. doi: 10.1162/153535003322331993 [20] GAMELIN J, AGUIRRE A, MAURUDIS A, et al. Curved array photoacoustic tomographic system for small animal imaging[J]. Journal of Biomedical Optics, 2008, 13(2):024007. doi: 10.1117/1.2907157 [21] YANG S, XING D, ZHOU Q, et al. Functional imaging of cerebrovascular activities in small animals using high-resolution photoacoustic tomography[J]. Medical Physics, 2007, 34(8):3294-3301. doi: 10.1118/1.2757088 [22] COX B T, KARA S, ARRIDGE S R, et al. k-space propagation models for acoustically heterogeneous media:Application to biomedical photoacoustics[J]. The Journal of the Acoustical Society of America, 2007, 121(6):3453-3464. doi: 10.1121/1.2717409 [23] DEAN-Ben X L, BUEHLER A, NTZIACHRISTOS V, et al. Accurate model-based reconstruction algorithm for three-dimensional optoacoustic tomography[J]. IEEE Transactions on Medical Imaging, 2012, 31(10):1922-1928. doi: 10.1109/TMI.2012.2208471 [24] DEAN-BEN X L, NTZIACHRISTOS V, RAZANSKY D. Acceleration of optoacoustic model-based reconstruction using angular image discretization[J]. IEEE Transactions on Medical Imaging, 2012, 31(5):1154-1162. doi: 10.1109/TMI.2012.2187460 [25] LIU H, WANG K, PENG D, et al. Curve-driven-based acoustic inversion for photoacoustic tomography[J]. IEEE Transactions on Medical Imaging, 2016, 35(12):2546-2557. doi: 10.1109/TMI.2016.2584120 [26] ZENG L, XING D, GU H, et al. High antinoise photoacoustic tomography based on a modified filtered backprojection algorithm with combination wavelet[J]. Medical Physics, 2007, 34(2):556-563. doi: 10.1118/1.2426406