Small animal in vivo multi-parameter dynamics monitoring method based on photoacoustic imaging

Abstract

The application discloses a small animal in-vivo multi-parameter dynamics monitoring method based on photoacoustic imaging and relates to the technical field of biomedical imaging, which comprises the following steps: obtaining photoacoustic signals generated by a small animal under the excitation of pulsed laser output by an optical excitation module through an acoustic detection module; reconstructing photoacoustic images by filtering back projection algorithm on the photoacoustic signals; calculating the similarity between adjacent images in the photoacoustic images, identifying the images with similarity lower than a set frame threshold as cross-sectional slices contaminated by respiratory motion artifacts and capable of being removed, using interpolation of adjacent non-artifact images to replace the removed image frames, thereby obtaining updated photoacoustic images; selecting a region of interest in the photoacoustic images to monitor oxygenation dynamics, pharmacokinetics, perfusion dynamics or in-vivo dynamics of nanoparticles; the monitoring method effectively suppresses motion artifacts caused by small animal respiration and realizes in-vivo multi-parameter dynamics monitoring of small animals.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of biomedical imaging technology, and particularly to a method for monitoring multi-parameter dynamics of small animals in vivo based on photoacoustic imaging. BACKGROUND

[0002] The monitoring of multi-parameter dynamics of small animals in vivo, including oxygenation dynamics, pharmacokinetics, perfusion dynamics and nanoparticle dynamics in vivo, is crucial for revealing the physiological and pathological processes of small animals.

[0003] The method for monitoring using traditional biomedical imaging technology has inherent contradictions between spatial resolution, temporal resolution, imaging depth and contrast. For example, X-ray computed tomography (CT) has high resolution, but lacks functional information sensitivity and is accompanied by ionizing radiation; magnetic resonance imaging (MRI) and positron emission tomography (PET) have deep tissue and functional imaging capabilities, but have low spatial and temporal resolution and rely on exogenous agents or ionizing radiation; ultrasound (US) imaging can achieve real-time imaging, but lacks optical contrast; optical imaging has subcellular resolution and molecular sensitivity, but has limited penetration depth. In contrast, photoacoustic imaging (PA) combines the advantages of rich optical contrast and high ultrasonic resolution, and becomes an ideal tool for in vivo biological dynamic monitoring, with a penetration depth of centimeters. Among many PA technologies, photoacoustic computed tomography (PACT) can realize the imaging of tissue composition, vascular function and oxygenation state without the aid of contrast agents. The two-dimensional PACT monitoring method based on a ring array transducer has the advantages of compactness, low cost and dense planar angle sampling, and can realize high-fidelity image reconstruction without relying on complex geometric structures or sparse sampling strategies, thereby realizing in vivo dynamic monitoring.

[0004] Currently, typical methods for small animal in vivo multi-parameter dynamics monitoring using two-dimensional PACT with annular transducer array include single-pulse panoramic photoacoustic tomography (SIP-PACT), transmission-reflection photoacoustic ultrasonic computed tomography (TROPUS) and video rate full-ring ultrasonic and photoacoustic tomography (VF-USPACT). SIP-PACT can realize imaging and monitoring of small animals at the anatomical structure, functional state, cellular activity and molecular level, but the soft tissue contrast performance is poor; TROPUS combines photoacoustic tomography and transmission-reflection ultrasound to simultaneously present optical absorption, acoustic reflectivity, speed of sound (SOS) and acoustic attenuation information, and realize sub-millimeter anatomical structure imaging, but it is still limited to anatomical structure visualization and cannot realize precise imaging at the functional specificity and molecular level; VF-USPACT realizes dual contrast imaging of optical absorption and ultrasonic reflectivity, and has video real-time transmission capability, but it is also limited to anatomical structure visualization and lacks functional and molecular specificity analysis. In addition, these methods generally use nonlinear image post-processing techniques such as blood vessel filtering, contrast limited adaptive histogram equalization and non-local mean filtering for enhancement, but these methods increase the complexity of post-processing and may introduce artifacts.

[0005] In summary, these technical defects hinder the realization of high-precision multi-parameter imaging of multi-scale dynamic physiological and pathological processes in vivo, and at the same time highlight the urgent need for small animal in vivo multi-parameter dynamics monitoring methods. SUMMARY

[0006] Based on the technical problems existing in the background art, the present application proposes a small animal in vivo multi-parameter dynamics monitoring method based on photoacoustic imaging, which effectively suppresses the motion artifacts caused by small animal respiration and realizes the monitoring of small animal in vivo multi-parameter dynamics.

[0007] The small animal in vivo multi-parameter dynamics monitoring method based on photoacoustic imaging proposed by the present application comprises:

[0008] The acoustic detection module acquires the photoacoustic signal generated by the small animal under the output of the pulsed laser excitation of the optical excitation module;

[0009] The photoacoustic signal is reconstructed into a photoacoustic image by a filtered back-projection algorithm;

[0010] The similarity between adjacent images in the photoacoustic image is calculated, and the image with a similarity lower than a set frame threshold is identified as a transverse section contaminated by respiratory motion artifacts and can be removed, and the removed image frame is replaced by interpolation of adjacent non-artifact images to obtain an updated photoacoustic image;

[0011] The region of interest in the photoacoustic image is selected, and the oxygenation dynamics monitoring, pharmacokinetics monitoring, perfusion dynamics monitoring or in vivo dynamics monitoring of nanoparticles is performed.

[0012] Further, the acoustic detection module is fixed at a set height by the motion scanning module, so as to obtain photoacoustic signals of selected sections of the small animal, and a two-dimensional photoacoustic image frame sequence is reconstructed from the photoacoustic signals by a filtered back-projection algorithm.

[0013] Further, the acoustic detection module is controlled to move up and down on the outside of the small animal by the motion scanning module, so as to obtain photoacoustic signals of different sections of the small animal, and a three-dimensional photoacoustic image is reconstructed from the photoacoustic signals by a filtered back-projection algorithm.

[0014] Further, after the two-dimensional photoacoustic image frame sequence is reconstructed from the photoacoustic signals by the filtered back-projection algorithm, a similarity value between adjacent frames in the two-dimensional photoacoustic image is calculated, and an image frame with a similarity value lower than a set frame threshold is identified as a cross-sectional slice contaminated by a respiratory motion artifact and can be removed, an adjacent non-artifact image is used for interpolation replacement of the removed image frame, so as to obtain an updated two-dimensional photoacoustic image.

[0015] Further, after the three-dimensional photoacoustic image is reconstructed from the photoacoustic signals by the filtered back-projection algorithm, a similarity value of each adjacent two-dimensional image slice in the three-dimensional photoacoustic image is calculated, and a two-dimensional image slice with a similarity value lower than a set frame threshold is identified as a cross-sectional slice contaminated by a respiratory motion artifact and is removed, an adjacent non-artifact image is used for interpolation replacement of the removed image frame, so as to obtain an updated three-dimensional photoacoustic image.

[0016] Further, the adjacent non-artifact image is used for interpolation replacement of the removed image frame, wherein linear interpolation replacement is adopted.

[0017] Further, the adjacent non-artifact image is used for interpolation replacement of the removed image frame, wherein three-dimensional spline interpolation replacement is adopted.

[0018] Further, when monitoring oxygenation kinetics, first, an anoxic state and an oxygen-rich state of the small animal are established, and the anoxic state and the oxygen-rich state are alternately performed, in the process, the small animal is photoacoustically imaged by the in-vivo multi-parameter kinetic monitoring method of the small animal, so as to obtain an updated photoacoustic image, and the oxygenation kinetics of the small animal in different oxygenation states is monitored.

[0019] Further, when monitoring pharmacokinetics, after a drug is injected into the small animal, the small animal is photoacoustically imaged by the in-vivo multi-parameter kinetic monitoring method of the small animal, so as to quantitatively analyze and compare the metabolic processes of a region of interest in the small animal;

[0020] When monitoring perfusion kinetics, after injecting a drug into a small animal and perfusion phenomenon exists, the small animal is subjected to photoacoustic imaging by the small animal in-vivo multi-parameter kinetic monitoring method, and an updated photoacoustic image is obtained, based on which the perfusion dynamics of a region of interest in the small animal is quantitatively analyzed and compared.

[0021] Further, when monitoring nanoparticle in-vivo kinetics, after injecting a solution containing suspended nanoparticles into a small animal, the small animal is subjected to photoacoustic imaging by the small animal in-vivo multi-parameter kinetic monitoring method, and an updated photoacoustic image is obtained, based on which the nanoparticles in the image time sequence are detected, located and tracked, so that the trajectory of the nanoparticles in the blood vessels is obtained, and the flow rate of the nanoparticles in the blood vessels is quantified.

[0022] The small animal in-vivo multi-parameter kinetic monitoring method based on photoacoustic imaging provided by the application has the advantages that: the photoacoustic imaging technology is adopted, linear interpolation and smoothing filtering technology are combined, motion artifacts caused by small animal respiration are effectively suppressed, noise is reduced on the basis of retaining dynamic characteristics, only nonlinear processing is used to correct the respiration motion artifacts in the three-dimensional volume photoacoustic imaging process to reduce the dependence on image post-processing, so that the shortcomings of the prior art in monitoring fast, spatially heterogeneous and hierarchical physiological and pathological dynamic changes are made up. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 The figure is a flowchart of the application;

[0024] Figure 2 The figure is a schematic diagram of the detection equipment corresponding to the small animal in-vivo multi-parameter kinetic monitoring method of Example 1;

[0025] Figure 3 The figure is a schematic diagram of the correction process of respiration motion artifacts in the three-dimensional volume photoacoustic imaging based on three-dimensional spline interpolation in Example 1;

[0026] Figure 4 The figure is a schematic diagram of the monitoring results of the whole body oxygenation kinetics of a mouse in Example 1, wherein (a) is three-dimensional photoacoustic images of five typical oxygenation states of a mouse under the condition of continuous alternation of hypoxic and oxygen-rich environments, and (b) is a curve graph of photoacoustic signals of the mouse trunk and intestinal tract changing with time;

[0027] Figure 5 The figure is a schematic diagram of the monitoring of the oxygenation kinetics of the kidney-spleen joint part of a mouse, wherein (a) is two-dimensional photoacoustic images of the kidney-spleen joint part section under the condition of continuous alternation of hypoxic and oxygen-rich environments, (b) is a curve graph of photoacoustic signal intensity of different blood vessel corresponding regions of the kidney-spleen joint part section changing with time, and (c1) is a schematic diagram of the correction process of respiration motion artifacts in the two-dimensional volume photoacoustic imaging based on two-dimensional spline interpolation in Example 1; Figure 5, (c2) is the curve diagram obtained after the (b) is processed by linear interpolation of respiratory artifacts and smoothing filter, and (d) is the curve diagram obtained after the (b) is processed by linear interpolation of respiratory artifacts and smoothing filter. Figure 5 , (c2) is the curve diagram obtained after the (c2) is processed by Savitzky-Golay filter smoothing, and (d) is the curve diagram obtained after the (b) is processed by linear interpolation of respiratory artifacts and smoothing filter. Figure 5 , (c2) is the curve diagram obtained after the (b) is processed by linear interpolation of respiratory artifacts and smoothing filter.

[0028] Figure 6 is a schematic diagram of three-dimensional volume photoacoustic imaging results of a small animal in a baseline state (oxygen-rich) and a severe hypoxia state when photoacoustic imaging is performed at different wavelengths of laser;

[0029] Figure 7 is a schematic diagram of a detection device corresponding to the multi-parameter dynamics monitoring method of embodiment 2;

[0030] Figure 8 is a schematic diagram of monitoring of ICG pharmacokinetics in a small animal, wherein (a) is a three-dimensional volume photoacoustic image of the small animal at five key moments (baseline state, i.e., before injection, 10 minutes, 60 minutes, 150 minutes, and 280 minutes after injection), (b) is Figure 8 , (a) is a schematic diagram of four different sections (S1, S2, S3, and S4) of the three-dimensional volume photoacoustic image at the five moments, and (c) is a curve diagram of normalized photoacoustic intensity of four typical metabolic organs (liver, duodenum, small intestine, and large intestine) region versus time after injection;

[0031] Figure 9 is a schematic diagram of monitoring of ICG perfusion kinetics in a mouse, wherein (a) is a two-dimensional photoacoustic image of a liver cross-section at 0 seconds (baseline state), 64 seconds, 80 seconds, and 300 seconds, (b) is a curve diagram of photoacoustic intensity of different blood vessels or tissue regions of the liver versus time, (c1) is Figure 9 , (b) is an enlarged view of artifacts caused by respiratory motion, (c2) is a curve diagram obtained after linear interpolation processing of (c1), and (c3) is a curve diagram obtained after filter smoothing processing of (c2); Figure 9 , (c2) is a curve diagram obtained after linear interpolation processing of (c1), and (c3) is a curve diagram obtained after filter smoothing processing of (c2); Figure 9 , (c2) is a curve diagram obtained after linear interpolation processing of (c1), and (c3) is a curve diagram obtained after filter smoothing processing of (c2);

[0032] Figure 10 is a schematic diagram of monitoring of ICG nanoparticle kinetics in a mouse, wherein (a) is a photoacoustic image of a liver cross-section before injection, (b) is a photoacoustic image of a liver cross-section at the 14th second after injection of ICG preparation B (nanoparticles), and (c) is Figure 10 , (b) is a schematic diagram of a series of consecutive frames in the brown dashed box, with a frame interval of 0.1 seconds;

[0033] Wherein, 1-optical excitation module, 2-small animal, 3-acoustic detection module, 4-motion scanning module, 5-reconstruction processing module, 6-water tank, 7-fiber bundle, 8-gas mixer, 9-injector. DETAILED DESCRIPTION

[0034] Hereinafter, the technical solutions of the present application will be described in detail through specific embodiments. In the following description, a lot of specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the concept of the present application, so the present application is not limited to the specific implementation disclosed below.

[0035] As shown in Figures 1 to 10 The small animal in-vivo multi-parameter dynamics monitoring method based on photoacoustic imaging provided by the present application comprises:

[0036] Step one, obtaining the photoacoustic signal generated by the small animal under the excitation of the pulsed laser output by the optical excitation module through the acoustic detection module;

[0037] Step two, reconstructing the photoacoustic image through the filtered back-projection algorithm;

[0038] Step three, calculating the similarity between adjacent images in the photoacoustic image, identifying the cross-sectional slice contaminated by the respiratory motion artifact as the image whose similarity is lower than the set frame threshold and can be removed, using the interpolation of the adjacent non-artifact image to replace the removed image frame, and obtaining the updated photoacoustic image;

[0039] Step four, selecting the region of interest in the photoacoustic image to monitor the oxygenation dynamics, the pharmacokinetics, the perfusion dynamics or the in-vivo dynamics of nanoparticles.

[0040] The present embodiment is used to realize the accurate monitoring of the in-vivo multi-parameter dynamics of the living small animal. The multi-parameter dynamics includes oxygenation dynamics, pharmacokinetics, perfusion dynamics and in-vivo dynamics of nanoparticles. Different small animals are used for different dynamics detection.

[0041] The method uses the photoacoustic imaging technology, combines the linear interpolation and smoothing filtering technology, effectively suppresses the motion artifact caused by the respiration of the small animal, reduces the noise on the basis of retaining the dynamic characteristics, only uses the non-linear processing to correct the respiratory motion artifact in the three-dimensional volume photoacoustic imaging process to reduce the dependence on the image post-processing, and thus makes up for the deficiencies of the prior art in monitoring the rapid, spatially heterogeneous and hierarchical physiological and pathological dynamic changes.

[0042] In one embodiment, the acoustic detection module is controlled by the motion scanning module to be fixed or moved, so as to perform two-dimensional planar photoacoustic imaging or three-dimensional volumetric photoacoustic imaging on the small animal, specifically:

[0043] First, the acoustic detection module is arranged on the periphery of the small animal, and the motion scanning module controls the acoustic detection module to be fixed or moved.

[0044] The optical excitation module is used to generate pulsed laser, which excites the small animal to generate ultrasonic waves through the photoacoustic effect. Such ultrasonic waves are also called photoacoustic signals.

[0045] Two-dimensional planar photoacoustic imaging: the acoustic detection module is fixed at a set height by the motion scanning module, so as to obtain photoacoustic signals of the small animal in a selected cross section, and a two-dimensional photoacoustic image frame sequence is reconstructed from the photoacoustic signals by a filtered back-projection algorithm.

[0046] Three-dimensional volumetric photoacoustic imaging: the acoustic detection module is controlled by the motion scanning module to move up and down on the periphery of the small animal, so as to obtain photoacoustic signals of the small animal in different cross sections, and a three-dimensional photoacoustic image is reconstructed from the photoacoustic signals by a filtered back-projection algorithm.

[0047] In one embodiment, for two-dimensional planar photoacoustic imaging, in order to suppress the respiratory motion artifact of the small animal, the similarity (for example, structural similarity SSIM) value between adjacent frames in the two-dimensional photoacoustic image is calculated, the image frame with a similarity value lower than a set frame threshold is identified as a cross-sectional slice contaminated by the respiratory motion artifact and can be removed, the removed image frame is replaced by interpolation (for example, linear interpolation) of adjacent non-artifact images, and the interpolation result is further denoised by a Savitzky-Golay filter. The parameters of the filter, such as window width and polynomial order, are set to suppress noise in the signal as much as possible while retaining signal characteristics (i.e., the fluctuation characteristics of photoacoustic image intensity caused by fluctuations in blood oxygen saturation due to pulsation can be retained) to the greatest extent, so as to balance the noise suppression effect and the signal characteristic retention effect, so as to obtain an updated two-dimensional photoacoustic image.

[0048] In one embodiment, for three-dimensional volumetric photoacoustic imaging, in order to suppress the respiratory motion artifact of the small animal, the similarity (for example, structural similarity SSIM) value of each adjacent two-dimensional image slice in the three-dimensional photoacoustic image is calculated, the two-dimensional image slice with a similarity value lower than a set frame threshold is identified as a cross-sectional slice contaminated by the respiratory motion artifact and is removed, and the removed image frame is replaced by interpolation (for example, three-dimensional spline interpolation) of adjacent non-artifact images, so as to obtain an updated three-dimensional photoacoustic image.

[0049] The present embodiment can use the structural similarity index (SSIM) to identify cross-sectional slices contaminated by respiratory motion artifacts for both two-dimensional planar photoacoustic imaging and three-dimensional volumetric photoacoustic imaging. SSIM is an index for perceiving image quality, which takes into account the brightness, contrast, and structural information of the image, rather than just pixel-level differences (such as mean square error, MSE). This makes SSIM more effective in detecting image structural changes caused by respiratory motion, such as blurred organ boundaries or distorted textures. Compared with simple threshold methods (such as based on image intensity changes), SSIM reduces the risk of false positives, as it focuses on structural consistency rather than global brightness changes.

[0050] In photoacoustic imaging, respiratory motion often causes non-rigid deformation between image frames. SSIM can sensitively capture these changes, thus accurately identifying contaminated frames.

[0051] The present embodiment processes two-dimensional planar photoacoustic imaging sequences through linear interpolation, specifically: once the cross-sectional slices contaminated by respiratory motion artifacts are identified, linear interpolation of adjacent artifact-free images is used for replacement, which is a simple and fast method. Linear interpolation is computationally lightweight, suitable for real-time or high-frame-rate processing, especially when motion artifacts are intermittent, it can better restore the overall trend of the section. For two-dimensional imaging, linear interpolation is usually sufficient, as the changes between adjacent frames are usually gradual and do not introduce obvious unnatural traces or flaws. In contrast, more complex interpolation methods (such as spline interpolation) may overfit noise or increase computational burden in two dimensions, while linear interpolation performs well in balancing speed and accuracy.

[0052] The present embodiment optimizes noise suppression and feature preservation in two-dimensional photoacoustic imaging sequences through Savitzky-Golay filters, which are a type of smoothing filter based on local polynomial fitting, which can effectively remove high-frequency noise while preserving signal features (such as peaks and edges). In photoacoustic imaging, this helps to maintain important physiological kinetic information (such as blood perfusion delay) during denoising. Filter parameters are carefully selected (such as window size and polynomial order) to balance noise suppression effect and signal feature preservation. This is different from traditional low-pass filters (such as Gaussian filters), which may over-smooth the signal, resulting in loss of detailed features.

[0053] In three-dimensional volumetric photoacoustic imaging, the present embodiment uses three-dimensional spline interpolation of adjacent artifact-free images to replace the removed image frames, which has advantages 1) to 3):

[0054] 1) can maintain the continuity and smoothness of three-dimensional spatial data;

[0055] Three-dimensional spline interpolation (usually cubic spline) is a high-order interpolation method. It estimates the value of unknown points by fitting a smooth polynomial curve (in three dimensions, a surface) between adjacent data points. This not only guarantees the continuity of the interpolated point values, but also guarantees the continuity of the first and second derivatives (i.e. slope and curvature).

[0056] Importance in three-dimensional photoacoustic imaging of biological organs and tissues: biological organs and tissues have continuous and smooth anatomical structures in three-dimensional space (such as curved blood vessels, organ surfaces). Three-dimensional spline interpolation can most realistically restore this natural spatial continuity, making the reconstructed three-dimensional volume data smooth and natural, without appearing harsh "steps" or "slicing" artifacts.

[0057] 2) High-precision fitting, effectively preserving anatomical structure details;

[0058] Because it considers a wider neighborhood of information (not just the nearest two points) and approximates the true physical spatial distribution through polynomial fitting, its interpolation results are more mathematically accurate.

[0059] This is crucial for subsequent three-dimensional visualization (such as volume rendering), quantitative analysis (such as calculating organ volume, blood vessel morphology), and dynamic monitoring (such as tracking blood oxygen changes). High-precision interpolation can ensure the accuracy of these analyses and avoid introducing false trends due to interpolation errors;

[0060] 3) Particularly suitable for handling non-uniform or complex structures;

[0061] In three-dimensional space, the changes in tissue structure may not be linear. For example, an organ may bend or change shape between two adjacent slices. Linear interpolation will simplify it to a straight line connection, while spline interpolation can capture and simulate this curved transition, thus more faithfully reconstructing complex three-dimensional anatomical morphology.

[0062] In addition, the present embodiment selects three-dimensional spline interpolation for processing three-dimensional imaging, which is selected in consideration of the obvious short board of other interpolation methods in the background of three-dimensional imaging, and cannot meet the demand of high-quality three-dimensional reconstruction, for example, (1) nearest neighbor interpolation, which will make the updated photoacoustic image present "mosaic" or "ladder" shape, and seriously damage the spatial continuity and smoothness of the data; at the same time, it completely ignores any gradual change information between adjacent slices, which is unacceptable for photoacoustic imaging which requires accurate geometric shape and signal intensity, that is, although the nearest neighbor interpolation has fast calculation speed, it will seriously reduce the image quality, and is not suitable for scientific analysis and medical imaging. (2) Linear interpolation can only ensure the continuity of the value, but cannot ensure the continuity of the derivative. The result is continuous transition in the slice direction (Z axis), but the transition is not smooth enough, and "faceting" or "edge" phenomenon may occur in three-dimensional isosurface rendering or surface reconstruction; for the nonlinear shape change of the organ contour, linear interpolation will produce deviation, making the reconstructed structure appear stiff and unnatural. That is, linear interpolation is a good way in two-dimensional scene, but in three-dimensional imaging which requires higher spatial continuity, its precision and smoothness are not enough to provide the best reconstruction effect.

[0063] Therefore, in the three-dimensional photoacoustic imaging of the present embodiment, three-dimensional spline interpolation is selected to sacrifice a certain calculation cost, in order to obtain the highest reconstruction precision and spatial smoothness. This is determined by the nature of three-dimensional data and the high requirements of subsequent advanced analysis (such as kinetic monitoring and three-dimensional morphological analysis). It ensures that the final photoacoustic image can truly and reliably reflect the internal anatomical structure and functional information of small animals, and provides a high-quality data basis for scientific research and medical applications.

[0064] By comparing the motion artifact removal method set by the present embodiment with other image processing processes, the following conclusions can be drawn:

[0065] Existing image registration or motion correction algorithms: These methods (such as rigid or non-rigid registration) try to align frames by distorting images, but respiratory motion is often non-rigid, involving complex deformation of multiple organs. Registration may introduce secondary artifacts or errors, especially when the motion amplitude is large, the registration accuracy decreases. In addition, registration has large calculation amount, which is not conducive to high frame rate real-time imaging. The motion artifact removal method set by the present embodiment avoids directly deforming the image, but detects and replaces it more reliably.

[0066] Existing direct filtering or frequency domain processing: If a low-pass filter (such as a Butterworth filter) or frequency domain filtering (such as Fourier transform) is directly applied to remove motion-related frequency components, useful physiological signals may be removed because the respiratory motion frequency may overlap with the signal of interest (such as hemodynamics). The Savitzky-Golay filter set in this embodiment is a time domain smoothing, which pays more attention to the preservation of local features.

[0067] Existing machine learning-based methods: Although deep learning can be used for motion artifact detection and repair, such methods require a large amount of labeled data to train the model, and the model generalization ability may be limited. In photoacoustic imaging, the data variation is large (such as different animals, different physiological states), and it is difficult to train a reliable model. The motion artifact removal method set in this embodiment is based on traditional image processing, without the need for training, and is easier to implement and debug.

[0068] In summary, the motion artifact removal method set in this embodiment achieves a good balance between accuracy, computational efficiency, and practicality, and is particularly suitable for suppressing respiratory motion artifacts in photoacoustic imaging.

[0069] It should be noted that the monitoring of oxygenation kinetics can be based on two-dimensional planar photoacoustic imaging or three-dimensional volumetric photoacoustic imaging. The two-dimensional planar photoacoustic imaging and the three-dimensional volumetric photoacoustic imaging are used to monitor oxygenation kinetics, pharmacokinetics, perfusion kinetics, or nanoparticle kinetics in the body. When each of the body kinetics is monitored, the detection is based on the updated photoacoustic image.

[0070] In one embodiment, when monitoring oxygenation kinetics, the hypoxic state and the hyperoxic state of the small animal are established, and the hypoxic state and the hyperoxic state are alternately performed. In this process, the small animal is photoacoustically imaged by the body multi-parameter kinetics monitoring method to obtain an updated photoacoustic image, based on which the oxygenation kinetics of the small animal in different oxygenation states is monitored. Specifically, the updated photoacoustic image is obtained, the region of interest (such as organs, tissues, etc.) of the photoacoustic image is selected, the average photoacoustic signal intensity or the relative change of the photoacoustic signal intensity of the region of interest is extracted, these extracted quantities are taken as the characteristics of blood oxygen saturation, the time variation curves of the characteristic quantities of different regions are drawn, the variation law of the characteristic quantities is analyzed, such as relative delay, period, sequence, etc., and the region of interest is quantitatively analyzed and compared.

[0071] It should be noted that the small animal can establish the hypoxic state and the oxygen-rich state by breathing in the natural conditions, and the oxygenation kinetics monitoring is performed based on the hypoxic state and the oxygen-rich state, or the small animal can also inhale the preset gas combination through the gas mixer to establish the hypoxic state and the oxygen-rich state. In addition, the small animal can also perform the oxygenation kinetics monitoring in the drug kinetics detection, the perfusion kinetics monitoring or the monitoring of the nanoparticles in the body kinetics. The oxygenation kinetics monitoring results of the small animal under different conditions are obtained.

[0072] The monitoring of the pharmacokinetics: the drug is metabolized in the small animal body, and if the drug changes the photoacoustic signal intensity, the monitoring of the pharmacokinetics can be performed. First, the drug is injected into the small animal, and then the photoacoustic imaging of the small animal is performed by using the in-vivo multi-parameter kinetics monitoring method, so as to quantitatively analyze and compare the metabolism process of the region of interest in the small animal body. Specifically, first, the solution including the drug is injected into the small animal body, the in-vivo multi-parameter kinetics monitoring method is performed, the region of interest (such as an organ, a tissue, etc.) of the photoacoustic image is selected, the metabolism process of the drug is analyzed according to the region of the image representing the strong photoacoustic signal intensity of the drug, the average photoacoustic signal intensity or the change amount of the region of interest at different times is extracted, these quantities are taken as the characteristic indexes of the drug concentration, the change curves of the characteristic quantities of different regions at different times and the fitting curves are drawn, and the change law of the characteristic quantities, such as the sequence, is analyzed. The metabolism process of the region of interest is quantitatively analyzed.

[0073] The monitoring of the perfusion kinetics: the drug is injected into the small animal body, and the perfusion phenomenon exists. If the drug changes the photoacoustic signal intensity, the monitoring of the perfusion kinetics can be performed. Specifically, first, the solution including the drug is injected into the small animal body until the perfusion phenomenon exists, and then the in-vivo multi-parameter kinetics monitoring method is performed. The region of interest (such as an organ, a tissue, etc.) of the photoacoustic image is selected, the average photoacoustic signal intensity or the change amount of the region of interest is extracted, the time change curves of the characteristic quantities of different regions are drawn, the change law of the characteristic quantities, such as the relative delay, the period, the sequence, etc., is analyzed, and the quantitative analysis and comparison of the perfusion dynamics of the region of interest are realized.

[0074] The monitoring of the nanoparticles in the body kinetics: if the nano-drug particles change the photoacoustic signal intensity, the monitoring of the nanoparticles in the body kinetics can be performed. First, the solution including the suspended nano-particles is injected into the small animal body, and then the in-vivo multi-parameter kinetics monitoring method is performed. The region of interest (such as an organ, a tissue, etc.) of the photoacoustic image is selected, the image of the region of interest is extracted, the centroid of the bright spot in each frame image (i.e. the position of the nano-particle) is extracted, the positions of the nano-particles in the frame sequence are connected to form a curve, the trajectory of the in-vivo movement of the nano-drug particles is obtained, and the flow rate and other parameters of the nano-drug particles are quantified.

[0075] It should be noted that by performing photoacoustic imaging on small animals after injecting two different ICG preparations, the excellent effect of the multi-parameter kinetic monitoring method of the embodiment on small animals on pharmacokinetics, perfusion kinetics, and nanoparticle kinetics in vivo is demonstrated.

[0076] The multi-parameter kinetic monitoring method of the embodiment is implemented by a monitoring device, which includes an optical excitation module 1, a small animal 2, an acoustic detection module 3, a motion scanning module 4, a reconstruction processing module 5, a water tank 6, an optical fiber bundle 7, a gas mixer 8, a syringe 9, ICG preparation A, ICG preparation B, etc.

[0077] The optical excitation module 1 can be a pulsed laser for exciting the small animal 2 to generate photoacoustic signals. The acoustic detection module 3 can receive and digitize the photoacoustic signals. The reconstruction processing module 5 can be a computer for reconstructing the photoacoustic signals into photoacoustic images and performing data analysis. The motion scanning module 4 can be an electrically driven translation stage that can move in the vertical direction, so that the method can obtain photoacoustic images of the small animal 2 in a fixed section and different sections, realizing two-dimensional planar photoacoustic imaging and three-dimensional volumetric photoacoustic imaging.

[0078] The water tank 6 is used to hold water to ensure acoustic coupling between the small animal 2 and the acoustic detection module 3. The water tank 6 has a heating device and a water pump to ensure uniform water temperature and maintain it at about 35 degrees Celsius. The optical fiber bundle 7 can be wrapped around the imaging plane of the small animal 2 to achieve uniform laser illumination. The gas mixer 8 can mix oxygen and nitrogen in any proportion for the small animal to inhale, to establish anoxic and hyperoxic states of the small animal. The syringe 9 is used to inject ICG preparation A or ICG preparation B.

[0079] The small animal can be a female BALB / c-nu / nu mouse, 8-12 weeks old, weighing 20-25 grams. ICG preparation A is prepared by dissolving 20 mg of ICG powder in 0.4 mL of DMSO and oscillating it with an ultrasonic oscillator until no visible particulate matter is present. Then, it is diluted with physiological saline to 1 mg / mL to obtain a clear green ICG solution. ICG preparation A is used for pharmacokinetic and perfusion kinetic monitoring of the small animal.

[0080] ICG preparation B is prepared by directly dissolving 20 mg of ICG powder in 20 mL of physiological saline to obtain a turbid ICG solution containing suspended nanoparticles with a concentration of 1 mg / mL. ICG preparation B is used for monitoring the kinetics of nanoparticles in vivo in the small animal.

[0081] Example 1

[0082] This embodiment uses Figure 3The method shown is for in vivo multi-parameter dynamics monitoring of small animals, mainly oxygenation dynamics.

[0083] Taking a mouse as an example, the specific process of the in vivo multi-parameter dynamics monitoring method is as follows: (a1) to (a6):

[0084] (a1) After the mouse is anesthetized, it is fixed in the water tank 6 filled with water, and the head exposed to the water surface is connected to the output end of the gas mixer 8 through the breathing mask.

[0085] (a2) The gas mixer is used to control the small animal to inhale pure oxygen through the breathing mask, and the whole process lasts for 150 seconds, so as to establish an oxygen-rich environment in the mouse body.

[0086] (a3) The gas mixer is used to control the small animal to inhale a mixed gas of 10% oxygen and 90% nitrogen through the breathing mask, and the whole process lasts for 90 seconds, so as to establish an anoxic environment in the mouse body.

[0087] (a4) When performing three-dimensional volumetric photoacoustic imaging, the optical excitation module 1 and the acoustic detection module 3 are moved vertically by the motion scanning module 4, and photoacoustic imaging of two-dimensional sections is performed every 200-250 microns. The reconstruction processing module 5 adopts a layer-by-layer reconstruction technology to generate a three-dimensional data set by splicing photoacoustic images of different sections. The total scanning time for completing the overall volumetric imaging is usually 10 to 20 seconds, and the vertical imaging resolution is about 1.1 millimeters.

[0088] Wherein, when performing three-dimensional volumetric photoacoustic imaging, first, the acoustic detection module 3 is controlled to reach a set position by the motion scanning module 4, the photoacoustic signal of the current section is collected by the acoustic detection module 3, then the acoustic detection module 3 is adjusted to the next set position by the motion scanning module 4, the photoacoustic signal of the corresponding section is collected by the acoustic detection module 3 again, and finally the photoacoustic signals of different sections are spliced and reconstructed into a three-dimensional photoacoustic image by using the layer-by-layer reconstruction technology.

[0089] (a5) In order to remove the respiratory artifacts of the three-dimensional photoacoustic image, the SSIM value of each adjacent two-dimensional image section is calculated, and the section below the threshold value is identified as a cross-sectional section contaminated by respiratory motion artifacts. The contaminated section is removed and replaced by three-dimensional spline interpolation of adjacent non-artifact images. The effect of removing the respiratory artifacts can be seen in Figure 3 .

[0090] (a6) The high-oxygen environment and the anoxic environment are alternately performed, and during this process, the three-dimensional volumetric photoacoustic imaging of the whole mouse is continuously performed, the oxygenation dynamics of the whole mouse under different oxygenation states is monitored, and the monitoring result of the whole oxygenation dynamics can be seen in Figure 4 .

[0091] To observe the oxygenation kinetics at a higher frame rate, two-dimensional planar photoacoustic imaging was performed with the imaging plane controlled to be located at the kidney-spleen junction. Specifically, (b1) to (b4) were performed.

[0092] (b1) First, the oxygen-rich and oxygen-poor environments were established in the mouse in vivo by (a1) to (a3), and then two-dimensional planar photoacoustic imaging was performed. The pulse laser wavelength of the optical excitation module 1 was set to 1064 nm, and the pulse laser was transmitted to the surface of the mouse through the optical fiber bundle 7. At the same time, the photoacoustic signal was collected using the acoustic detection module 3. The GPU-accelerated filtered back-projection algorithm was used in the reconstruction processing module 5 to reconstruct the photoacoustic signal data into a two-dimensional photoacoustic image. The imaging frame rate was 10 Hz, and the resolution in the imaging plane was about 122 microns.

[0093] (b2) To remove the respiratory motion artifacts in the two-dimensional planar photoacoustic image, the SSIM value between each image frame was calculated, and the image frames below the threshold value were identified as being contaminated by respiratory motion artifacts. Subsequently, two-dimensional linear interpolation of adjacent non-artifact images was used for replacement. The result of interpolation and replacement was further denoised by a Savitzky-Golay filter. The filter parameters were carefully selected to balance noise suppression and signal feature preservation.

[0094] (b3) The oxygenation kinetics of the kidney-spleen junction of the mouse under different oxygenation states was monitored by alternately performing the hyperoxic and hypoxic environments. The monitoring results of the oxygenation kinetics of the kidney-spleen junction are shown in Figure 5 .

[0095] (b4) Different wavelengths of laser (700 nm, 797 nm, 1064 nm) were selected, and the oxygenation kinetics of the mouse under different wavelengths of laser was monitored under the condition of alternately performing the hyperoxic and hypoxic environments. The monitoring results of the multi-wavelength oxygenation kinetics are shown in Figure 6 .

[0096] The results of the present embodiment are shown in Figures 3 to 6 .

[0097] Figure 3 The correction process of respiratory motion artifacts in three-dimensional interpolation-based volumetric photoacoustic imaging is shown. During the volumetric scanning of the mouse torso, periodic respiratory motion will produce artifacts in the reconstructed photoacoustic image along the height direction (indicated by the arrow). By calculating the SSIM value of each adjacent two-dimensional image slice, and identifying the slices below the threshold value as being contaminated by respiratory motion artifacts, the contaminated slices were removed and replaced by three-dimensional spline interpolation of adjacent non-artifact images. This method effectively suppresses the periodic respiratory motion artifacts and significantly improves the overall consistency of the image.

[0098] Figure 4The results of monitoring the whole-body oxygenation kinetics of small animals were presented. Figure 4 (a) presents three-dimensional photoacoustic images of mice under continuous alternation between hypoxic and oxygen-rich environments, showing five typical oxygenation states (baseline state, moderate hypoxia, severe hypoxia, partial reoxygenation, and complete reoxygenation). Figure 4 In (a), depth information is color-coded. It can be seen that, due to the lower optical absorptivity of deoxygenated hemoglobin at 1064 nm wavelength compared to oxyhemoglobin, the photoacoustic signal in the mouse trunk decreases with increasing hypoxia and recovers after reoxygenation. During hypoxia, intestinal signals are enhanced because reduced hemoglobin absorption increases the ability of photons to penetrate intestinal tissue, resulting in more photons being absorbed by the intestinal contents. During partial reoxygenation, the abdominal aorta shows discontinuities caused by pulsation (marked by dashed boxes), revealing regional oxygenation heterogeneity. Figure 4 (b) shows the time curves of relative changes in photoacoustic signals in the mouse trunk and intestines. It can be seen that it takes about 2 minutes for the small animal to transition from the baseline state to a state of severe hypoxia, while it only takes about 30 seconds to transition from severe hypoxia to a state of complete reoxygenation. This reflects the gradual nature of oxygen consumption under hypoxic conditions and the rapid response of oxygen supply under hyperoxia conditions.

[0099] Figure 5 The study demonstrated the monitoring of oxygenation kinetics at the kidney-spleen junction in small animals. Figure 5 (a) presents two-dimensional photoacoustic images of five typical oxygenation states (baseline state T1, moderate hypoxia T2, severe hypoxia T3, partial reoxygenation T4, and complete reoxygenation T5) in cross-sections of the kidney-spleen junction of mice under continuous alternation of hypoxia and hyperoxygenation, as well as magnified views within the green box. Figure 5 (b) shows the time curves of the relative changes in photoacoustic signal intensity in different blood vessel regions of this section, where the photoacoustic signal intensity exhibits obvious respiratory motion artifacts. (See enlarged image below.) Figure 5 (c1).

[0100] Figure 5 (c1) identifies artifact-contaminated frames through inter-frame SSIM and performs linear interpolation replacement, thus obtaining the result after motion artifact correction, i.e. Figure 5 (c2). After being smoothed by the Savitzky-Golay filter, (c2) forms... Figure 5 (c3) effectively reduces noise while preserving the basic oxygenation dynamics. Observation Figure 5 As shown in (c3), under hypoxic conditions, respiration periodically produces oxygenated blood clots, resulting in periodic peaks in the photoacoustic signal intensity. Based on the phase difference of the periodic peaks in different blood vessels and organs, the perfusion time of blood from the abdominal aorta to the renal artery is calculated to be approximately 0.3 seconds, while the time required to pass through the capillaries and enter the renal vein is approximately 1.3 seconds longer.

[0101] Will Figure 5 The data in (b) were processed by respiratory artifact linear interpolation removal and smoothing filtering to obtain Figure 5 (d). In Figure 6 In (d), the green arrow indicates a transient overshoot in the abdominal aortic oxygen saturation during reoxygenation, suggesting that oxygen-rich blood is preferentially delivered to the proximal vascular system before being redistributed to distal tissues. The gray arrow indicates a slower response of the spleen to changes in oxygenation, which is related to its lower metabolic demand, blood reserve capacity, and slower perfusion rate. Under hypoxic conditions, the photoacoustic signal exhibits a decreasing gradient from the abdominal aorta through the renal artery to the renal vein, reflecting a gradual change along the circulatory pathway. During reoxygenation, slight overshoot occurs in both renal vessels, due to the preferential redistribution of oxygen to the kidneys, temporarily increasing local blood volume.

[0102] Figure 7 This study presents the three-dimensional volumetric photoacoustic imaging results of small animals under baseline (oxygen-rich) and severely hypoxic conditions using photoacoustic imaging at different laser wavelengths (700 nm, 797 nm, 1064 nm). At 700 nm, the photoacoustic signal intensity is higher in the severely hypoxic state than in the oxygen-rich state because oxyhemoglobin has lower light absorption than deoxyhemoglobin. At 797 nm, the light absorption of oxyhemoglobin and deoxyhemoglobin is comparable, resulting in similar photoacoustic signal intensities in the severely hypoxic and oxygen-rich states. However, at 1064 nm, the light absorption of oxyhemoglobin is higher than that of deoxyhemoglobin, leading to a lower photoacoustic signal intensity in the severely hypoxic state compared to the oxygen-rich state. This result further validates the close correlation between photoacoustic signal intensity and the hemoglobin absorption spectrum.

[0103] In summary, this method fully records the dynamic changes in oxygenation kinetics at the systemic and vascular scales under hypoxic experimental conditions, reveals organ-specific differences in hypoxic responses, clarifies the differences in recovery priorities among different tissues, and captures the wavelength dependence of oxygenation kinetics.

[0104] Example 2

[0105] This embodiment adopts Figure 8 The detection equipment shown performs in vivo multi-parameter kinetics on the small animals to be imaged, mainly referring to the monitoring of pharmacokinetics, perfusion kinetics, and in vivo kinetics of nanoparticles.

[0106] The following explanation uses a mouse as an example. The specific process of this embodiment is as follows (c1) to (c6):

[0107] (c1) Anesthetize and fix a small animal in the water tank 6 filled with water. Inject 100 microliters of ICG preparation A through the tail vein indwelling needle of the mouse under water using the syringe 9.

[0108] (c2) Set the pulsed laser wavelength output from the optical excitation module 1 to 785 nm. Perform three-dimensional volumetric photoacoustic imaging before injection (baseline), 1 minute after injection, and every 5 minutes from 5 minutes after injection, for a total monitoring duration of 300 minutes. To remove respiratory motion artifacts from the three-dimensional photoacoustic images, calculate the SSIM value for each adjacent two-dimensional image slice, and identify slices with SSIM values below a threshold as being contaminated by respiratory motion artifacts. Discard the contaminated slices and replace them with three-dimensional spline interpolation of adjacent artifact-free images, resulting in Figure 9 the monitoring results of ICG pharmacokinetics shown in FIG. 6.

[0109] (c3) Anesthetize and fix another small animal in the water tank 6 filled with water. Inject 100 microliters of ICG preparation A through the tail vein indwelling needle of the mouse under water using the syringe 9.

[0110] (c4) Set the pulsed laser wavelength output from the optical excitation module 1 to 785 nm. To observe the changes in perfusion kinetics at a higher frame rate, perform two-dimensional planar photoacoustic imaging continuously from about 1 minute before injection until 3 minutes after injection, with the imaging plane fixed at the liver section. Use linear interpolation and smoothing filtering to suppress respiratory motion artifacts, resulting in Figure 10 the monitoring results of ICG perfusion kinetics shown in FIG. 7.

[0111] (c5) Anesthetize and fix another small animal in the water tank 6 filled with water. Inject 100 microliters of ICG preparation B through the tail vein indwelling needle of the mouse under water using the syringe 9.

[0112] (c6) Set the pulsed laser wavelength output from the optical excitation module 1 to 785 nm. To observe the changes in nanoparticle kinetics in vivo at a higher frame rate, perform two-dimensional planar photoacoustic imaging continuously before and after injection, with the imaging plane fixed at the liver section. Use linear interpolation and smoothing filtering to suppress respiratory motion artifacts, resulting in Figures 8 to 10 the monitoring results of ICG nanoparticle kinetics in vivo shown in FIG. 8.

[0113] The results of Example 2 are shown in Figure 8 :

[0114] Figure 8 The monitoring of ICG pharmacokinetics in vivo in a small animal is shown. Figure 8(a) presents three-dimensional volumetric photoacoustic images of small animals at five key moments (baseline state, i.e., before injection, 10 minutes, 60 minutes, 150 minutes, and 280 minutes after injection), with the images color-coded using depth information. Figure 8 (b) shows Figure 8 Four different sections (S1, S2, S3, S4) of the three-dimensional photoacoustic images at five time points in (a). Figure 8 (c) details the changes in normalized photoacoustic intensity over time in four typical metabolic organ regions (liver, duodenum, small intestine, and large intestine). The scatter plots represent actual measurements, and the solid lines represent polynomial fitted values. Through analysis... Figure 9 It was observed that after injection of ICG preparation A, the photoacoustic signal in the liver of the small animals gradually increased, indicating that the liver began to take up ICG, and the peak of ICG uptake in the liver occurred within 5-10 minutes. Subsequently, ICG signals appeared in the duodenum between approximately 5 and 140 minutes; ICG signals appeared in the small intestine at approximately 70 minutes; and ICG signals appeared in the large intestine at approximately 200 minutes. This series of changes fully reflects the four typical drug metabolism stages in small animals: hepatic uptake, duodenal drainage, small intestinal clearance, and large intestinal clearance.

[0115] Figure 9 The monitoring of ICG perfusion kinetics in small animals was demonstrated. ICG injection began at 57 seconds and was completed at 63 seconds. Figure 9 (a) presents two-dimensional photoacoustic images of the liver cross section at 0 seconds (baseline), 64 seconds, 80 seconds and 300 seconds. Figure 9 (b) records the time curves of relative changes in photoacoustic intensity in different blood vessels or tissue regions of the liver. Artifacts caused by respiratory movements can be seen in the figure. Figure 9 (c1) is Figure 9 (b) is an enlarged view. Figure 9 (c1) was processed by linear interpolation, which effectively suppressed respiratory motion artifacts, resulting in Figure 9 (c2); (c2) is further smoothed using a filter to obtain Figure 9 (c3). From Figure 9 Analysis of (a) shows that after ICG injection was completed at 63 seconds, the photoacoustic image at 64 seconds showed that ICG was first perfused into the inferior vena cava; the image at 80 seconds showed that the liver began to take up ICG; by 300 seconds, the image showed that the liver’s uptake of ICG increased significantly, and the gradual accumulation of ICG in the peripheral liver enhanced light absorption, limiting photons from penetrating to the central region of the liver, thereby weakening the photoacoustic signal in that region. Figure 10Further, (c) of FIG. 6 shows that after tail vein injection, ICG was first perfused in inferior vena cava, about 1.7 seconds later perfused in abdominal aorta, about 2.1 seconds later entered hepatic artery, about 5 seconds later reached the central liver, and then rapidly spread to the whole liver parenchyma.

[0116] Figure 10 The monitoring of ICG nanoparticle kinetics in small animals in vivo is shown. Figure 10 (a) of FIG. 6 is a photoacoustic image of a liver cross-section before injection, Figure 10 (b) of FIG. 6 is a photoacoustic image of a liver cross-section at the 14th second after injection of ICG preparation B (nanoparticles), and ICG particles are detected in inferior vena cava, liver and subcutaneous tissue (green arrows). Figure 10 (c) of FIG. 6 shows that Figure 10 (b) of FIG. 6 is a series of continuous frames of the region of interest (brown dashed frame), with a frame interval of 0.1 seconds. In order to show the tracking effect of molecular particles, the position of the nanoparticle in each frame is marked with a red circle, the initial position is marked with a brown cross, and the particle path formed by connecting each frame is marked with a brown line. ​ (c) of FIG. 6 shows the transport of an ICG particle in the subcutaneous microcirculation, and the trajectory shows the dynamic movement of the particle from the subcutaneous tissue to the dermis and then back to the subcutaneous tissue, and the estimated flow rate of the particle is 1.1 mm / s.

[0117] In summary, from the perspectives of pharmacokinetics, perfusion kinetics and nanoparticle kinetics in vivo, this method can monitor the clearance of ICG in the circulation system, liver uptake and gastrointestinal excretion in real time, and provides a non-invasive observation window for evaluating liver function, drug metabolism and whole-body pharmacokinetics. In addition, with the help of ICG particle tracking technology in subcutaneous capillaries, local hemodynamic flux and probe transport can be quantified, thereby opening up new research avenues for evaluating microvascular permeability, hemodynamic changes and targeted drug delivery dynamics.

[0118] The embodiment effectively makes up for the shortcomings of CT, magnetic resonance imaging (MRI), positron emission computed tomography (PET), ultrasonic imaging (US) and the like, and realizes accurate capture of multi-parameter information such as functional activity, metabolic process and molecular dynamics. By linear interpolation and smoothing filtering technology, the respiratory motion artifact in two-dimensional plane photoacoustic imaging is effectively suppressed, and only nonlinear processing is applied to three-dimensional respiratory motion artifact suppression, thereby significantly reducing the dependence on image post-processing. The method can in-depth analyze the oxygenation kinetic response of small animal organs, blood vessels and spectral dimensions, accurately depict the metabolic kinetics and perfusion kinetics characteristics of drugs in small animals, and can monitor the flowing nanoparticles in the body at the molecular level. The method successfully solves the short board of the prior art in monitoring rapid, spatially heterogeneous and hierarchical physiological and pathological dynamic changes, fully demonstrates the unique advantages of the method in capturing multi-parameter dynamic physiological changes in the body, and enables the method to be further applied to long-term monitoring of cross-scale physiology and pathology and comprehensive evaluation of treatment and diagnosis.

[0119] The above merely describes a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change according to the technical solution and inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for small animal in vivo multi-parametric kinetic monitoring based on photoacoustic imaging, characterized in that, The method comprises the following steps: acquiring photoacoustic signals generated by a small animal under the excitation of pulsed laser output by an optical excitation module through an acoustic detection module; reconstructing photoacoustic images from the photoacoustic signals through a filtered back-projection algorithm; calculating the similarity between adjacent images in the photoacoustic images, identifying images with a similarity lower than a set frame threshold as cross-sectional slices contaminated by respiratory motion artifacts and capable of being removed, replacing the removed image frames with interpolation of adjacent non-artifact images to obtain updated photoacoustic images; selecting a region of interest in the photoacoustic images to monitor oxygenation kinetics, pharmacokinetics, perfusion kinetics or nanoparticle kinetics in vivo.

2. The small animal in vivo multi-parameter kinetic monitoring method according to claim 1, characterized in that, The acoustic detection module is fixed at a set height by a motion scanning module to acquire photoacoustic signals of selected cross sections of the small animal and reconstruct a two-dimensional photoacoustic image frame sequence from the photoacoustic signals through the filtered back-projection algorithm.

3. The small animal in vivo multi-parametric kinetic monitoring method of claim 1, wherein, The acoustic detection module is controlled by the motion scanning module to move up and down on the outside of the small animal to acquire photoacoustic signals of different cross sections of the small animal and reconstruct a three-dimensional photoacoustic image from the photoacoustic signals through the filtered back-projection algorithm.

4. The small animal in vivo multi-parametric monitoring method of claim 2, wherein, After the two-dimensional photoacoustic image frame sequence is reconstructed from the photoacoustic signals through the filtered back-projection algorithm, the similarity between adjacent frames in the two-dimensional photoacoustic images is calculated, image frames with a similarity lower than a set frame threshold are identified as cross-sectional slices contaminated by respiratory motion artifacts and capable of being removed, the removed image frames are replaced with interpolation of adjacent non-artifact images to obtain updated two-dimensional photoacoustic images.

5. The small animal in vivo multi-parameter kinetic monitoring method according to claim 3, characterized in that, After the three-dimensional photoacoustic image is reconstructed from the photoacoustic signals through the filtered back-projection algorithm, the similarity of each adjacent two-dimensional image slice in the three-dimensional photoacoustic image is calculated, two-dimensional image slices with a similarity lower than a set frame threshold are identified as cross-sectional slices contaminated by respiratory motion artifacts and removed, the removed image frames are replaced with interpolation of adjacent non-artifact images to obtain updated three-dimensional photoacoustic images.

6. The small animal in vivo multi-parametric kinetic monitoring method of claim 4, wherein, The removed image frames are replaced with interpolation of adjacent non-artifact images, wherein linear interpolation is used.

7. The small animal in vivo multi-parametric monitoring method of claim 5, wherein, The removed image frames are replaced with interpolation of adjacent non-artifact images, wherein three-dimensional spline interpolation is used.

8. The small animal in vivo multi-parametric monitoring method of claim 1, wherein, When oxygenation kinetics is monitored, hypoxic and hyperoxic states of the small animal are established, the hypoxic and hyperoxic states are alternated, photoacoustic imaging of the small animal is performed through the in-vivo multi-parameter kinetics monitoring method of the small animal to obtain updated photoacoustic images, and the oxygenation kinetics of the small animal in different oxygenation states is monitored.

9. The small animal in vivo multi-parametric monitoring method of claim 1, wherein, When pharmacokinetics is monitored, after a drug is injected into the small animal, photoacoustic imaging of the small animal is performed through the in-vivo multi-parameter kinetics monitoring method of the small animal to quantitatively analyze and compare the metabolic processes in the region of interest in the small animal; When perfusion kinetics is monitored, after a drug is injected into the small animal and perfusion occurs, photoacoustic imaging of the small animal is performed through the in-vivo multi-parameter kinetics monitoring method of the small animal to obtain updated photoacoustic images, and the perfusion dynamics of the region of interest in the small animal is quantitatively analyzed and compared.

10. The small animal in vivo multi-parametric monitoring method of claim 1, wherein, When the nanoparticles in vivo dynamics is carried out, after the small animal is injected with the solution containing the suspended nanoparticles, the small animal is subjected to photoacoustic imaging by the small animal in-vivo multi-parameter dynamics monitoring method, a renewed photoacoustic image is obtained, the nanoparticles in the image time sequence are detected, positioned and tracked, and thus the trajectory of the nanoparticles in the blood vessel is obtained, and the flow rate of the nanoparticles in the blood vessel is quantified.

Images