JP2020069110A - Image processing device and image processing program - Google Patents

Abstract

An object of the present invention is to accurately detect a feature amount of a test site such as an artery and blood vessel when the shapes of a plurality of movable parts of a test subject change, as compared with a conventional example. Kind Code: A1 An image processing apparatus according to the present invention includes imaging means for imaging a test site into a three-dimensional image by irradiating a test site of a subject with light. Here, the imaging means images a test site including a plurality of movable parts into a three-dimensional image, and the image processing device bends the plurality of movable parts based on the imaged three-dimensional image. A processing means is provided for calculating feature amounts of the relevant test sites before and after. Further, the processing means corrects the positions when the plurality of movable parts bend after calculating the characteristic amount of the test site before and after the plurality of movable parts bend. Furthermore, the processing means extracts the test site from the three-dimensional image, and then divides the test site into a plurality of regions corresponding to the movable parts, thereby dividing the plurality of regions into Perform the specified preprocessing. [Selection drawing] Fig. 2

Description

  The present invention relates to an image processing apparatus that performs a predetermined analysis after three-dimensionally imaging a plurality of movable parts of a subject using an imaging method such as an optical ultrasonic imaging method (also referred to as a photoacoustic imaging method). The present invention relates to a method and an image processing program using the image processing method.

  Photoacoustic Imaging (hereinafter, also referred to as PA imaging or PA imaging) is a promising new technique for early clinical diagnosis of neovascularization of cancer and many other diseases. PA imaging is capable of non-invasive visualization of blood vessels in a living body with high spatial resolution even without a contrast agent, and is useful for many medical applications. Here, PA imaging utilizes the photoacoustic effect. That is, the imaged tissue (i.e., blood vessel) absorbs the laser energy of the short pulse near infrared radiation and converts it into heat, resulting in thermoelastic expansion and emission of ultrasonic waves. By detecting the emitted photoacoustic waves, the three-dimensional structure of the object can be reconstructed.

  For example, Patent Document 1 discloses a technique of improving the accuracy of the oxygen saturation value calculated by the photoacoustic apparatus and imaging an arterial blood vessel, for example. The image processing device using the photoacoustic apparatus, by using the signal derived from the acoustic wave generated by the irradiation of the light of a plurality of wavelengths to the subject, image data showing the oxygen saturation distribution inside the subject. An image processing apparatus including a processing unit for acquiring, wherein the processing unit acquires oxygen saturation at a specific position of a blood vessel of a subject, which is generated by image reconstruction using a signal, and at a specific position of the blood vessel. A first correction amount for correcting the oxygen saturation is acquired, and a second correction amount for correcting the oxygen saturation of a portion that is not the specific position of the subject is acquired based on the first correction amount. ..

  Further, Patent Document 2 discloses to provide a finger pathological condition evaluation device capable of quantitatively evaluating a motor paralysis state of a finger. The hand pathological condition evaluation device includes a range image capturing unit that captures a range image in which the distance data to the subject is included in the pixel value of each pixel, and each finger from each frame of the range image captured with the subject's moving fingers as the subject. A virtual phalange image generation means for obtaining a phalange position of the virtual phalange and generating an image of the virtual phalange including pixel data of the phalange position, a joint angle between the virtual phalanges, an acceleration representing a motion state of the virtual phalange, and a virtual The virtual phalangeal calculation means for calculating at least one of the movement amounts of the phalangeal movements, the reference data in the pathological condition serving as the reference of the finger, and the value calculated by the virtual phalangeal calculation means are compared, and the pathological condition of the finger of the subject is compared. And a finger pathological condition determining means for determining a level. With this finger pathological condition evaluation device, it is possible to quantitatively evaluate by comparing the pathological condition level of the finger with a reference level.

JP, 2008-102359, A JP, 2017-158808, A

  However, in the above-mentioned conventional example, when the shapes of a plurality of movable parts (for example, finger joints) of the subject change, it is not possible to accurately detect the feature amount of the test site such as the arterial blood vessel. was there.

  The object of the present invention is to solve the above problems, and when the shapes of a plurality of movable parts of the subject are changed, compared to the conventional example, it is possible to accurately detect the feature amount of the test site such as arterial blood vessel. An object is to provide an image processing apparatus and method, and an image processing program that can be performed.

The image processing apparatus according to an aspect of the present invention is an image processing apparatus including an imaging unit that images a test site of a test subject into light by irradiating the test site of the test subject with light.
The imaging means images a test site including a plurality of movable parts into a three-dimensional image,
The image processing device,
It is characterized by further comprising processing means for calculating a feature amount of the test site before and after the plurality of movable parts are bent based on the imaged three-dimensional image.

  Therefore, according to the image processing apparatus and the like according to the present invention, when the shapes of the plurality of movable parts of the subject are changed, the feature amount of the test site such as the arterial blood vessel can be accurately detected as compared with the conventional example. You can

It is a schematic cross section and a block diagram showing an example of composition of an optical ultrasonic imaging system concerning one embodiment of the present invention. 3 is a flowchart showing optical ultrasonic image processing executed by the optical ultrasonic imaging apparatus 10 of FIG. 1. 3 is a flowchart showing a blood vessel curvature calculation process which is a subroutine (S4) of FIG. 2. 3 is a flowchart showing a blood vessel position correction process which is a subroutine (S5) of FIG. 2. 9 is a photographic image showing an example of imaging of a bent finger and segmentation based on a phalangeal joint. FIG. 6 is a plan view showing a point of interest P _i and a plurality of points in the vicinity thereof. It is a figure showing an example of calculation of a plurality of paths using dynamic programming (DP method). It is a graph which shows the curvature plot calculated | required from the imaging example of a bent finger | toe, and three local maximum values. It is a figure which shows an example of the radius of the blood vessel of a spiral shape. It is a graph which shows the example which shows the result of carrying out position correction using DP method to two curvature curves. 2 is a photographic image showing an imaging example of the radial index artery (RIA) imaged by the optical ultrasonic imaging apparatus 10 of FIG. 1. It is a side view when the finger of a 22-year-old man is bent and the RIA is deformed, which is imaged by the optical ultrasonic imaging apparatus 10 of FIG. 1. FIG. 2 is an oblique rear view of a 22-year-old man who has bent a finger and deformed an RIA imaged by the optical ultrasonic imaging apparatus 10 of FIG. 1. It is a graph which shows the average curvature of each area | region of RIA when not bending a finger joint of a young person, and their statistical difference. It is a graph which shows the average curvature of each area | region of RIA when the joint of a finger of a young person is bent 30 degrees, and their statistical difference. It is a graph which shows the average curvature of each area | region of RIA when the joint of a finger of a young person is bent 60 degrees, and those statistical differences. It is a graph which shows the average curvature of each area | region of RIA when the joint of a finger of a young person is bent 90 degrees, and their statistical difference. It is a photograph image which shows the example of imaging of RIA when the joint of the finger of a 59-year-old man is bent. FIG. 14B is an enlarged view of FIG. 14A. It is a graph which shows the curvature of RIA when a finger joint is bent at 30 degrees and 90 degrees, (a) is a graph in case of 25 years old, (b) is a graph in case of 42 years old, (C) is a graph at the age of 56.

  Hereinafter, embodiments and examples according to the present invention will be described with reference to the drawings. In addition, in each of the following embodiments, the same components are denoted by the same reference numerals.

  FIG. 1 is a schematic sectional view and a block diagram showing a configuration example of an optical ultrasonic imaging system according to an embodiment of the present invention.

  In FIG. 1, the optical ultrasonic imaging system according to this embodiment includes an optical ultrasonic light source / detection device 20 and an optical ultrasonic imaging device 10.

  In the optical ultrasonic light source and detection device 20, a plurality of sensors 22 for detecting ultrasonic waves are provided on a sensor support portion 21 having a hemispherical shape, for example, at predetermined intervals. Water 24 is immersed in the inside of the hemisphere to support the mesh plate 23 on which the inspection object 1 is placed. It is assumed that the inspection object 1 is submerged in the water 24. A laser light source 25 for irradiating the inspection object 1 with laser light of a predetermined wavelength for PA imaging is provided at the bottom of the sensor support portion 21. Here, the laser light source 25 is controlled by the controller 11 described later, and the ultrasonic waves emitted from the inspection object 1 in response to the laser light by PA imaging are detected by the plurality of sensors 22. The ultrasonic waves detected by the plurality of sensors 22 are converted into electric signals and then sent to the signal receiving circuit 12 described later.

  The inspection target 1 is an imaging target. Specific examples include a living body such as a finger, its arterial blood vessel, and a breast, and a phantom simulating the acoustic and optical characteristics of the living body when adjusting the device. The acoustic characteristics are specifically the propagation velocity and attenuation rate of an acoustic wave, and the optical characteristics are specifically the absorption coefficient and scattering coefficient of light. In addition, it is preferable that the mesh plate 23 supports the inspection object 1 and has a transmission characteristic with respect to light and ultrasonic waves.

  The optical ultrasonic imaging apparatus 10 includes a controller 11, which is a digital computer such as a computer that controls the operation of the apparatus, a signal receiving circuit 12, a data memory 13, an operation unit 14, a display unit 15, and an image processing unit. And 16 are provided. The signal receiving circuit 12 receives the plurality of electric signals converted by the plurality of sensors 22, A / D-converts the signals at a predetermined sampling frequency, and stores the data in the data memory 13. The operation unit 14 includes a mouse and a keyboard for the operator of the optical ultrasonic imaging apparatus 10 to input data or instructions. The display unit 15 displays the input data or instructions, and visualizes and displays a result image during or after the optical ultrasonic imaging process. The image processing unit 16 is controlled by the controller 11 and executes the program of the optical ultrasonic imaging processing of FIG. 2 described later in detail.

  In addition, in the embodiment of FIG. 1, the hemispherical sensor support portion 21 is used, but the present invention is not limited to this, and the directional axes of the plurality of sensors 22 are gathered in a predetermined area to form a predetermined detection area. Any arrangement is possible.

  In the optical ultrasonic imaging system of FIG. 1, by using the optical ultrasonic imaging method, a photoacoustic effect is utilized to non-invasively, for example, a blood vessel of a living body (where hemoglobin, which is a light absorber passes), in a three-dimensional manner. Imaging in images can be visualized and is useful for early clinical diagnosis of cancer and many other diseases. The tissue to be imaged, which is the examination object 1, absorbs the laser energy, converts it into thermoelastic expansion and emits ultrasonic waves. The three-dimensional structure of the object of the inspection object 1 can be reconstructed by detecting the optical ultrasonic waves.

  In the following embodiment, the optical ultrasound imaging system of FIG. 1 takes advantage of non-contrast imaging of fibrillation veins, and it is possible to perform imaging by dynamic changes of joints and the like, which are difficult to be imaged by a conventional modality. A three-dimensional image of the deformation of an artery is provided and a method of analyzing the deformation is provided. As a result, it is possible to provide the medical staff with various parameters related to the dynamic deformation of blood vessels and their numerical values, as well as aging or pathological changes in the numerical values of the various parameters. By this, the doctor can be useful for diagnosis, treatment, determination of treatment effect, and the like of diseases related to blood vessels.

  The following embodiments are characterized in that the blood vessel of a finger (excluding the thumb) having a plurality of movable parts of joints is used as a test site and the analysis process is performed as follows.

(1) Identification of Deformation Region For convenience, a region is annotated in a test region of a finger blood vessel, which is three-dimensionally imaged using the optical ultrasonic imaging system of FIG. 1, and which region has an increased curvature is determined. Analyze statistically. According to the examples described in detail below, the region of maximum curvature increase is the D-Px region described later, and based on the finding that the amount of change in curvature at the D-Px decreases with age, for example, Age can be estimated. In addition, each area | region of the finger | toe in an Example is as follows.
Base: base of finger;
MP: Metacarpophalangeal interphalangeal joint (hereinafter also referred to as MP joint);
DIP: distal interphalangeal joint (hereinafter referred to as DIP joint);
D-Mx: a region corresponding to the distal half of the middle phalanx;
D-Px: the region corresponding to the distal third of the proximal phalanx;
M-Px: the region corresponding to the middle third of the proximal phalanx;
PIP: Proximal interphalangeal joint (hereinafter referred to as PIP joint);
P-Mx: A region corresponding to the proximal half of the middle phalanx.

(2) Characterization of Deformation of Test Site in Which Deformation Occurs In analysis processing, parameters such as the amount of change in curvature, amplitude, twist ratio, and expansion / contraction ratio are calculated. For example, since blood vessels are markerless, it is difficult to track the same blood vessel segment due to deformation. Therefore, by performing matching (DP matching) by a dynamic programming method (hereinafter referred to as DP method), even if the fingers are bent and the positions of a plurality of joints change, the positions of blood vessels are corrected so as to be associated with each other. The following is calculated as the analysis parameter in this case.

(A) How much deformation is obtained by calculating the average curvature change amount of a certain section.
(B) The principal component vector is obtained for the quasi-helical form in a certain section, the maximum distance from it is obtained, and the amplitude of the helical deformation is calculated. This is to evaluate how small the spiral deformation is in the space.
(C) In order to evaluate the three-dimensionality of the quasi-helical deformation in a certain section, the torsion is calculated, which is estimated to change from two-dimensional to three-dimensional as the blood vessel is more flexible. Is.
(D) By calculating the actual distance (the length of the curve) in a certain section and calculating the rate of change in length according to the movement of the joint, it is possible to know how much compression has been performed as a parameter for evaluating the elasticity of blood vessels.

  According to the results of Examples described later, it is expected that the parameters (A) and (B) will decrease the deformation amount and the expansion / contraction rate as the blood vessel becomes harder due to aging. For example, a curvature change amount of a blood vessel in the D-Px region of a subject when a finger is bent at a predetermined angle is calculated, and a curvature change amount table stored in advance (standard curvature of the D-Px region with respect to age) is calculated. The amount of change is stored in the data memory 13 of FIG. 1), and the subject's blood vessel age can be evaluated based on the calculated curvature. Here, the above-mentioned feature amounts (A) to (D) are change information of blood vessels, and may be spatial expansion of blood vessels, physical property information, and elasticity information.

  FIG. 2 is a flowchart showing optical ultrasonic image processing executed by the optical ultrasonic imaging apparatus 10 of FIG.

  In step S1 of FIG. 2, for example, when a plurality of joints (movable parts) of a finger are bent at a plurality of bending angles using the optical ultrasonic light source and the detection device 20, the plurality of bending angles at each bending angle (k) are increased. An optical ultrasonic image of a part including the joint and its vicinity is acquired.

  Next, in the preprocessing P1 of step S2, the blood vessel to be analyzed is extracted from the acquired optical ultrasonic image. The blood vessels may be automatically extracted or the doctor may manually extract the annotations. Specifically, the main artery of the index finger (index finger) to be analyzed can be annotated from proximal to distal using, for example, the Kurumi system.

Next, in the preprocessing P2 of step S3, each region is identified in the blood vessel having the reference bending angle. Now, analyze the relationship between the phalanx and the blood vessel. For example, the three joints (DIP, PIP, MP joints) of the index finger are annotated, and the connection points of these are called S _dip , S _pipe , and S _base , respectively. Then, the midpoints of consecutive connecting points are defined as _{Sm -mx} and _{Sm -px} . Based on these five points, as shown in FIG. 5, four phalange lines D-Mx (Mx-d in FIG. 5), P-Mx (Mx-p in FIG. 5), and D-Px (in FIG. 5). Px-d) and M-Px (Px-p in FIG. 5) are defined.

  Next, in the blood vessel position correction processing Q1 of step S4, the blood vessel curvature calculation processing is executed as follows, like the processing of FIG.

  First, it is difficult to take a plurality of consecutive annotation points at equal intervals. We resampled the arterial annotation points by spline interpolation because they may affect the calculation of vessel curvature. In this resampling process (S11 in FIG. 3), the equal intervals are set to one pixel, and this interpolation smoothes the curvature.

The curvature of each point in the blood vessel is calculated using the points near the interest point P _i , and is repeated as follows for all the points Pi of the blood vessel coordinates (S12 to S16 in FIG. 3). Vector of each point P _i on the blood vessel curve
The curvature κ _i of the point P _i is calculated by using the following equation (S15 in FIG. 3).

here,
When
Is
Is a neighborhood point of. X represents a cross product. | · | Is the Euclidean norm of the vector. Since the curvature value depends on the sampling interval, instead of using the annotation points directly, for example, 20 neighboring points
Vector using principal component analysis (PCA) for
Is calculated (S13 to S14 in FIG. 3). The normalized first principal component is a vector
Used for.

Similarly, the vector
Is also calculated (S14 in FIG. 3).

FIG. 6 shows an exemplary diagram of the neighbor points and the first principal component vector. This process is repeatedly executed for each interest point P _i (S12 to S16 in FIG. 3). here,
Is
Is calculated as (S13 to S14 in FIG. 3). Here, the number of adjacent points does not have to be 20, and may be large or small. The more points used as the near points, the smoother the curvature curve.

  Next, in the blood vessel position correction process Q2 of step S5, the blood vessel position correction process is executed as shown in FIG. In this process, as shown in FIG. 4, the following process is repeated for each bending angle (S21 to S24 in FIG. 4).

First, in order to measure the motion of arteries as the joint angle changes, we first use dynamic programming matching (DP matching) to align the vessels at different angles. Here, all the blood vessel point sets when the angle is d ₁ or d ₂ are C _d1 and C _d2 , respectively. DP matching produces a list of matching combinations of all points on the vessel point set C _d1 and points on the vessel point set C _d2 by minimizing the matching cost between the two sequential data. The i-th point on the blood vessel point set C _{d1 is} defined as C _d1 ^i, and the k-th point on the blood vessel point set C _d2 is defined as C _d2 ^k . Here, | C _d1 | is the number of points of the blood vessel point set C _d1 . The cost matrix between the two sequences is constructed to compute the matching between the vessel point set C _d1 and the vessel point set C _d2 , denoted by matrix G ⁿ .

The cost matrix G ⁿ is a matrix of | C _d1 | × | C _d2 |, and the element at the i-th row and the j-th column of the matrix is represented by the following equation (S22 in FIG. 4).

(1)

  Here, the function min (·) is a function indicating the minimum value of the three numerical values in the argument. Also,

Is the 1-norm distance (absolute value of difference) between the curvature at the point C _d1 ⁱ and the curvature at the point C _d2 ^j for i, j> 0. As another example, another index indicating the magnitude of the difference in curvature such as the absolute value of the difference may be used. FIG. 8 shows an example of three routes. The DP method can find the minimum path by storing as a memory which value was the minimum.

  The value of g (i, j) for i ≦ 0 or / and j ≦ 0 is defined by the following equation.

  Here, Inf means an infinite number, and, for example, g (i = 0, j = 1) takes an infinite value. This means that the beginning of the two vessel sequences will always match.

Next, the back trace processing (S23 in FIG. 4) will be described below. After calculating the cost matrix, the DP method selects the one with the smallest value in the row g (| C _d1 |, :) corresponding to the last point of the reference blood vessel, and selects the corresponding value on the corresponding blood vessel corresponding to the last point. Determine the point. From there, back trace of the minimum path from the final point to the initial point as to which path is selected (i.e., the minimum) in each of i and j in Expression (1). FIG. 8 shows an example of the alignment result between the curvature curves 30-60 and 30-90. It can be seen that the curvature peaks are aligned and position-corrected between the two data.

  In the following steps S6 to S9, each feature amount is sequentially calculated, but the present invention is not limited to this, and four feature amounts may be calculated in parallel using a plurality of processors, or four feature amounts may be calculated. At least one of the feature quantities may be calculated. Alternatively, a plurality of combinations may be used for calculation.

First, in the curvature change amount calculation process of step S6, the change in the average value of the curvatures of the D-Mx, P-Mx, P-Mx, and M-Px regions of each segment when the finger is bent is calculated. Here, the average change in curvature in the segment between the angles d ₁ and d ₂ can be calculated as:

Here, κ _i ^d2 is the curvature of the i-th point when the angle of S _pip is d ₂ . S is a set of subscripts of blood vessel points. | S | is the number of elements of S. In the example, the following equation was set.

(D ₁ , d ₂ ) = (0,60)
(D ₁ , d ₂ ) = (0, 90)

Next, in step S7, a spiral region amplitude calculation process is executed. This analysis shows the relationship between age and radius of blood vessels in the segment containing three peaks in the P-Mx region. To determine the segment, first find the peak in the P-Mx region. All maxima in the curve of the blood vessel are detected. If the maximum value is larger than the threshold value 0.05 and the adjacent maximum value does not have the maximum value, the maximum point is registered as the maximum point. Next, the three largest peaks in the P-Mx region are selected. The selected peaks are named peaks P1, P2, and P3 in the order of A _pip to Am-mx, as shown in FIG. 7. The vessel curve between peak P1 and peak P3 is defined as the target segment. The first principal component vector of the segments P1 to P3 is calculated. Then, all points of the segment are projected onto the plane whose normal vector is the first principal component vector. The maximum distance from the first component vector to a point on the projection plane is defined as the radius of the circumscribed circle. A scatter plot of radius and age is plotted as in FIG.

  Next, in step S8, the expansion / contraction ratio calculation process is executed as follows. Here, the expansion / contraction ratio in the local region of the blood vessel is defined by the expansion / contraction ratio = the length of the blood vessel after the deformation / the length of the blood vessel before the deformation.

  First, the length of the blood vessel in the region of interest before deformation is obtained. Since the re-sampling is performed at equal intervals, the length of the blood vessel can be obtained by the number of samples on the blood vessel × interval. Of the deformed blood vessels, the segment that matches the blood vessel in the region of interest before the deformation is obtained by DP. The length of the blood vessel can be obtained by the same method as that of the blood vessel before the deformation.

  Deflection of blood vessels is also required as an option. The degree of deflection is defined by the following formula. Stretch rate = Blood vessel length / Shaft length. The length of the shaft can be calculated in two ways:

(Option 1) The axis vector is obtained by obtaining the first principal component of the blood vessel sampling point group in the local region. Project vertically from the end point of the sampling point onto the axis, and use that point as the end point of the axis. Calculate the length of the axis from the end points of the axis.
(Option 2) Simply define the straight line connecting the end points of the blood vessel sampling points as the axis. Therefore, the length of the shaft is the distance connecting the end points.

  Next, in step S9, a twist rate calculation process is executed. Here, the twist ratio τ (t) is defined and calculated by the following known equation.

here,
Is the derivative at the point x,
Is the second derivative at the point x,
Is the third derivative at the point x.

  The discrete data can be calculated as follows.

Similarly to the curvature, the calculation method of can be calculated by taking the first principal component.

  Further, in step S10, statistical analysis processing is executed and the optical ultrasonic image processing is ended. In the embodiment of the statistical analysis process, the “D-Px region of the finger (the region between the PIP joint and the base, which is the distal third of the proximal phalanx portion) obtained in an example described later in detail is obtained. Based on the knowledge that the amount of change in the three peaks of the blood vessel curvature when the flexion changes at a predetermined angle (for example, from 30 degrees to 90 degrees) in the area corresponding to , The age of the subject can be estimated. Specifically, the amount of change in the three peaks of the blood vessel curvature when the flexion changes at the predetermined angle in the region is stored in advance as the data of a large number of subjects of a large number of ages to create average data. , Age can be estimated based on the amount of change of the three peaks of the blood vessel curvature.

  Although the optical ultrasonic imaging system of FIG. 1 is used as the imaging device in the above embodiments, the present invention is not limited to this, and other three-dimensional imaging systems such as MRI, MRA, and CT are used. May be.

  In the following examples, the optical ultrasonic imaging system shown in FIG. 1 is referred to as "PAT system".

  Hereinafter, the “deformation of the finger artery with respect to the movement of the proximal interphalangeal joint (PIP joint)” according to the embodiment performed by the inventors will be described.

Background Finger reconstruction has been developed on the basis of anatomical studies of the finger arteries. In particular, anatomical studies on the microvessels of the distal node contribute to the establishment of re-adhesion of the distal node amputation (Nam et al., 2017), and anatomical studies on the branching of the finger arteries include reverse pedicle digital island flap. ) (Kojima et al., 1990) and dorsal island flap (Endo et al., 1990) and other arterialized flaps. On the other hand, little is known about which region of the finger arteries is deformed in which flexion phase. Among the three phalangeal joints present in the finger, the PIP joint is a hinge joint having a wide bending and stretching range of 100 degrees. Since the finger arteries run on the palm side of the flexion / extension axis of the PIP joint, the arteries are not only bent by the bending of the joint, but are also subjected to axial pressure. This predicts that the artery will deform in three dimensions. In a healthy finger, peripheral blood circulation is maintained even when the PIP joint is in the deepest flexion state. This suggests the existence of a deformation mechanism in which the digital arteries do not kink even when the joint is flexed (kink (excessive bending occurs)). To clarify the native deformation mechanism of the digital arteries independently of aging requires non-invasive and non-contact in vivo imaging that enables three-dimensional observation of subcutaneous blood vessels.

  Photoacoustic tomography (PAT) is a new diagnostic imaging technique that visualizes blood vessels with near-infrared pulsed laser light and ultrasound (Wang and Hu, 2012). The main purpose of this study is to visualize the deformation of human finger arteries under articulation using optical ultrasound imaging and to characterize the deformation three-dimensionally and quantitatively. In addition, we will examine the differences in the deformity mechanism of finger arteries in young and middle-aged individuals.

Methodology studies were conducted under the approval of the institutional ethics committee and the Declaration of Helsinki. Written consent was obtained from all subjects. Subjects incorporated conditions, healthy person 20 years of age or older, BMI is within the normal range (Men 18.5~27.7kg / ^{m 2,} women ^{15.1~24.7kg / m 2) (Yamakado et} al, 2015). He had no history of hypertension, diabetes, or any other vascular-related illness, or a history of hand trauma. Twenty-two fingers of 12 volunteers were included in this study. The breakdown was 7 fingers aged 22 to 33, 7 fingers aged 38 to 43, and 8 fingers aged 50 to 59. Blood pressure and pulse were measured immediately before the PAT examination, and it was confirmed that the blood pressure was within the normal range and that there was no arrhythmia.

In this research, the same equipment and the same image processing system (PAI system) as those used in the existing research by our research group were used. Briefly, our optical ultrasound system has an ultrasonic array (HAD) embedded in the wall of a hemispherical cup, with the near-infrared light emitter incorporated in the center of the cup. HAD is filled with degassed water for transmission of ultrasonic waves, and ultrasonic waves are generated by irradiating a blood vessel with laser light, which is detected by a large number of ultrasonic wave detectors incorporated in the inner wall of the cup. .. The laser used was a Q-switched laser with a wavelength of 795 nm or 797 nm, and the maximum laser power was limited to less than 65% of the maximum allowable exposure recommended by the American National Standards Institute (15 mJ / cm). ² , 20 Hz repetition rate). In order to visualize the digital arteries in the flexed position over the entire length of the finger, it was necessary to position the finger so that the side of the finger faces the laser emitting device. Due to the limitation of the method, the radialis indicis artery (RIA) on the radial side of the index finger (index finger) was selected as a test object. There are no anatomical landmarks because PAT does not show bone or soft tissue. Therefore, purple ink (detectable by PAT) is marked on the dorsal end of the three proximal, middle, and distal physiodal creases on the palm side of the finger (proximal, middle, and distal digital crease). I went. These markings are landmarks at the base (a level of about one third of the metacarpophalangeal joint (MP) to PIP), the PIP level of each finger and the level of the distal interphalangeal (DIP) joint. Was used as. The hand of the subject was placed in the cup with the pronation of the forearm so that the side surface on the radial side of the index finger faced the laser beam emitting device. The MP joint was maintained in the extended position during the scan. The first scan was performed in the finger extension position, followed by automatic bending and holding at 30, 60, and 90 degrees. The position of the PIP joint was confirmed using a goniometer by a surgeon with over 20 years of hand surgery experience. The position of the DIP joint changed with the automatic movement of the PIP joint (Hahn et al., 1995). The scan took approximately 2 minutes for each joint angle. The finger was confirmed by a camera installed in the center of the HDA. The two index fingers of different individuals were photographed twice on different dates to examine the reproducibility of the observed deformation.

The 3D image is 14 × 14 cm ² area with image analysis software (Kurumi, http://www.kuhp.kyoto-u.ac.jp/~diag_rad/intro/tech/kurumi.html) developed in the laboratory. Was created with a resolution of 1120 × 1120 pixels. In order to extract the coordinates of RIA, image processing was performed as follows. First, the subcutaneous vein networks of the fingers and their deep RIA were visualized in 8-bit images. Next, a program called cross simulation (Sekiguchi, 2017) was used to identify the skin surface of the finger, and the blood vessels were color-coded according to the depth from the surface (FIG. 11 (a)). By deleting the signal of the superficial subcutaneous layer, the network of subcutaneous veins was deleted, and as a result, the RIA was independently visualized (FIG. 11 (b)). Finally, the trajectory of the three-dimensional coordinates of the artery was semi-automatically identified based on the brightness information of the blood vessel. The blood vessel locus was browsed by color according to the depth from the identified skin surface (FIG. 11 (c)). In order to grasp the anatomical position of the RIA in each flexion position of the PIP joint, the RIA is used as a base phalangeal region (Px) and a middle phalanx region by using a plane including three anatomical landmarks. It was divided into (Mx). In the flexed position, the arteries near the PIP were divided in a plane that bisects the angle between the base-PIP axis (Base-PIP axis) and the PIP-DIP axis (PIP-DIP axis). Furthermore, each area was divided into two areas by a plane including the midpoint of each axis, and a total of four areas were divided. The four subregions were defined as follows from the proximal to the distal direction. For each subregion, the central third of the proximal phalanx (M-Px), the distal third of the proximal phalanx (D-Px), the proximal half of the middle phalanx (P-Mx). , Each region of the distal half of the medial phalanx (D-Mx) was defined.

  The curvature was measured based on the coordinate data of the artery. Since the curvature calculation is affected by the interval of the sampled points, we calculated the curvature by using the principal component vector by performing the principal component analysis on 20 points at the intervals of 0.125 mm. Correlation of blood vessel segments was performed using dynamic programming (DP). To observe the three-dimensional quasi-helical structure, the amplitude was quantified using the first principal component vector from the coordinate values of the quasi-helical structure.

Data analysis Data with normal distribution are shown as the mean (standard deviation). One-factor repeated measures ANOVA and Tukey's test were used to determine the effect of one factor on the variation of relevant data. A P value of less than 0.05 was judged to have a significant difference.

Results and Discussion The morphological changes of RIA in young people aged 22 to 33 years were evaluated. 12A and 12B are the youngest (22 years old) PAT images. At 0 degrees and 30 degrees, the average curvature of the D-Mx region was significantly higher than that of the other regions, but no significant difference was observed between the other three regions (FIGS. 13A and 13B). .. At 60 degrees, the average curvature was significantly higher than the adjacent areas in the D-Mx area and the D-Px area, and at 90 degrees, only the D-Px area showed a significantly high value. Three-dimensional observation was performed in order to confirm the deformation that caused the increase in curvature in the D-Mx region and the D-Px region. At 0 to 30 degrees, the RIA was running in a straight or gentle curve as a whole (Fig. 12A), but at 60 degrees, the reverse W-shaped wavy deformation occurred in the D-Mx region and the D-Px region. Appeared (Fig. 12B). At 90 degrees, the inverse W-shaped deformation of the D-Px region became a three-dimensional shape due to the increase of the curve curvature and the twist. This deformation consisted of the following three consecutive curves in proximal to distal order: A curve that first faces the back scale side (hereinafter referred to as curve DU1), a second curve that faces the palm radial direction (referred to as curve PR), and a curve that faces the back scale side again (hereinafter referred to as curve DU2). .),Met. This configuration was reproducible and was observed in all subjects. The curve DU1 was located at an average of 57% from the finger base to the PIP joint, and the curve DU2 was located at 83% (Table 1). The average amplitude was 1.7 mm (Table 2).

  Next, RIA variants were assessed using middle-aged individual data. 14A and 14B show optical ultrasound images of the oldest participant, a 59 year old man. At 90 degrees of flexion, the expression of three consecutive twisting deformations in the D-Mx region and the D-Px region was observed similarly to the findings of the young person. The positions of these three consecutive twists were similar to those in young people (Table 1). However, the three consecutive twist deformations were preserved even in the finger extension position (FIG. 14B). FIG. 15 shows curvature graphs of joint flexion of 30 degrees and 90 degrees obtained by matching arteries with DP. Three-peak morphology was observed in the D-Px region, which was consistent with the three consecutive twist distortions of the PAT image. Table 2 shows data on the morphology and deformation of RIA. Morphological parameters in the D-Px region at 90 degrees of flexion were similar in the three age groups. However, the average curvature change in the D-Px region decreased with increasing age.

Discussion We used PAT to acquire non-invasive in vivo 3D data on morphological changes of the finger arteries in individuals of various ages. Reportedly, the performance of PAT differs depending on the site in the body (Toi et al., 2017; Tsuge et al., 2018). The maximum imaging depth of subcutaneous blood vessels was 13 mm in a femoral perforator vessel study (Tsuge et al., 2018). This depth is sufficient to image the RIA. We believe that PAT will be a useful tool in anatomical and biomechanical studies of the finger arteries.

  The RIA and the proper digital artery have similar anatomical features, although they differ in diameter and origin (Coleman and Anson, 1961). RIA diameters in the proximal phalanx are 0.9 and 1.4 mm (Braga-Silva et al., 2002; Leslie et al., 1987). Both the RIA and the proper digital artery have multiple branches. On the palmar side, there are three large arches of the finger arteries: called the proximal, central and distal lateral palmar arches (Endo et al., 1992; Strauch and de Moura, 1990). The proximal lateral arch passes through a tunnel formed from the branches of the proximal (C1) cruciate ligament (Yousif et al., 1985). Similarly, the mid-lateral palmar arch passes through the leg formed by the distal (C3) cruciate ligament. The digital arteries have multiple branches on the dorsal side. In the index finger (index finger), the dorsal branch is located 12-16 mm proximal to the PIP joint and 9-12 mm distal to the PIP joint (Braga-Silva et al., 2002; Kostopoulos et al., 2016). .. These dorsal branches pass through the Cleland ligament, run near the phalanx, and dorsally to the fingers (Endo et al., 1992; Strauch and de Moura, 1990). In particular, one of the dorsal branches in the distal third of the proximal phalanx originates at the same level as the proximal lateral arch (Strauch and de Moura, 1990; Yousif et al., 1985). Therefore, the finger arteries at this site are connected by two branches in different directions. We found that deformation of the RIA occurs at 57% of the distance from the base to the PIP joint. This corresponds to the level of the proximal lateral palm arch. This suggests that the longitudinal motion of the artery is restricted at the site of the arch and that the RIA deformation is localized in the near-joint region.

  We have found that the RIA deformation in the D-Px region is characterized by three continuous twist curves. Moreover, the amplitude of those curves was small. A similar deformation pattern and amplitude suggests that the deformation has an anatomical basis. Smith et al., (1991) reported that 50% of 134 digital arteries had multiple wavy deformations near the DIP joint. They described the deformation formed by three continuous curves 90 degrees to each other as a "270 degree bend". They believed that this curved section was constructed to protect the vessel against the limited elasticity that occurs when the DIP joint flexes and extends, like a protective spring. They did not describe arterial deformation proximal to the PIP joint. Although the operating ranges of the DIP joint and the PIP joint are different, it is rational to presume that this type of spring-like deformation mechanism is common to the hinge joint. Anatomically, the finger arteries near the DIP and PIP joints are wrapped in adipose tissue and are associated with Grayson's ligament (palm) (Grayson, 1941) and Cleland's ligament (dorsal) ( Chrysopoulo et al., 2002; Zwanenburg et al., 2014). These ligaments are thought to play a role in stabilizing the neurovascular bundle during flexion of the knuckles (Zwanenburg et al., 2014). The presence of Cleland ligaments in the vicinity of the interphalangeal joints suggests a proximal, mechanical association between these ligaments and three continuous twist curve deformations.

  Since immediate exercise may cause damage to the vascular anastomosis, it has been practiced to delay PIP joint exercise training with a wide range of motion after revascularization (Silverman and Gordon, 1996). We found that the deformations associated with PIP articulation were localized to the distal third of the proximal phalanx. This indicates that early ankle motion therapy can be tolerated if an anastomosis is performed in the midregion of the proximal phalanx and the branch located at the site of the proximal palmar arch is intact. Flexion position of MP joints affects rehabilitation after flexor tendon repair (Tang, 2007). We did not study the effect of flexion angle of MP joints on deformation associated with PIP joint movement. Future studies of this relationship will contribute to a better understanding of digital arterial deformity in the clinic.

  Previous PAT studies have found that proper finger arterial tortuosity increases with age (Matsumoto et al., 2018). Morphological changes from 30 ° to 90 ° flexion decreased with age, and trimodal morphology was observed in middle-aged and elderly people even at 30 ° flexion. It is important to recognize that the elasticity of the finger arteries may decline with age when performing procedures such as advancing flaps.

Chart description

(Explanation of FIG. 11)
FIG. 11 is an image of a RIA using PAT and related image processing.
(A) The side surfaces of the index finger and the middle finger are depicted. This image is color-coded according to the depth from the skin surface (however, in the application drawings, it is a black and white image).
(B) This figure shows only the deep layer.
(C) Four regions were identified along the arterial pathway.
DIP: distal interphalangeal joint;
D-Mx: a region corresponding to the distal half of the middle phalanx;
D-Px: a region corresponding to the distal third of the proximal phalanx;
M-Px: a region corresponding to the middle third of the phalanx.
PIP: proximal interphalangeal joint;
P-Mx: A region corresponding to the proximal half of the middle phalanx.

(Explanation of FIGS. 12A and 12B)
12 and 12B show a modified (FIG. 12A) side view and (FIG. 12B) oblique rear view of the RIA in a 22 year old man. Along the path of the arteries, color corresponds to depth from the surface of the skin. At 60 and 90 degrees of flexion, arterial deformation in the form of three curves can be observed in the D-Px and D-Mx regions (indicated by bars). However, in other regions, the linear shape is maintained (arrow). DU1, PR and DU2 represent the three curves formed by the deformation.
Base: The base of the finger.
DIP: distal interphalangeal joint;
D-Mx: a region corresponding to the distal half of the middle phalanx;
D-Px: a region corresponding to the distal third of the proximal phalanx;
M-Px: a region corresponding to the middle third of the phalanx.
PIP: proximal interphalangeal joint;
P-Mx: A region corresponding to the proximal half of the middle phalanx.

(Description of FIGS. 13A to 13D)
13A to 13D show the average curvature of each region at each joint flexion angle of a young person. 13A is 0 degrees, FIG. 13B is 30 degrees, FIG. 13C is 60 degrees, and FIG. 13D is 90 degrees.
D-Mx: a region corresponding to the distal half of the middle phalanx;
D-Px: a region corresponding to the distal third of the proximal phalanx;
M-Px: a region corresponding to the middle third of the proximal phalanx;
P-Mx: A region corresponding to the proximal half of the middle phalanx.
Bars show standard deviation.
* P <0.05, ** p <0.01.

(Explanation of Table 1)
Table 1 shows the location of the RIA deformation in the region corresponding to the distal third of the proximal phalanx. The position of the structure is defined as the percentage of distance from the base to the proximal interphalangeal joint. Data are presented as the mean (standard deviation).

(Explanation of Table 2)
Table 2 shows the parameters related to RIA morphology and deformation. Data are presented as the mean (standard deviation). K indicates the curvature.
D-Mx: a region corresponding to the distal half of the middle phalanx;
D-Px: a region corresponding to the distal third of the proximal phalanx;
M-Px: a region corresponding to the middle third of the phalanx.
PIP: proximal interphalangeal joint;
P-Mx: A region corresponding to the proximal half of the middle phalanx.

Reference list:
(1) American National Standards Institute. American national standard for safe use of lasers: American National Standards Institute, 2014.https: //www.ansi.org/ (accessed 10 June 2015).
(2) Bise R, Kang L, Eom S, Kanade T. Reliably tracking partially overlapping neural stem cells in DIC microscopy image sequences. MICCAI Workshop on OPTIMHisE. 2009, 5: 67-77.
(3) Braga-Silva J, Kuyven CR, Fallopa F, Albertoni W. An anatomical study of the dorsal cutaneous branches of the digital arteries, J Hand Surg Br. 2002, 27: 577-9.
(4) Chrysopoulo MT, McGrouther DA, Jeschke MG, Kaufman BR. Cleland's ligaments: an anatomic study. Plast Reconstr Surg. 2002, 109: 566-72.
(5) Coleman SS, Anson BJ. Arterial patterns in the hand based upon a study of 650 specimens. Surg Gynecol Obstet. 1961, 113: 409-24.
(6) Endo T, Kojima T, Hirase Y. Vascular anatomy of the finger dorsum and a new idea for coverage of the finger pulp defect that restores sensation.J Hand Surg Am. 1992, 17: 927-32.
(7) Geiger D, Gupta A, Costa LA, Vlontzos J. Dynamic programming for detecting, tracking, and matching deformable contours.IEEE Trans Pattern Analysis and Machine Intell. 1995, 17: 294-302.
(8) Gokgol C, Diehm N, Kara L, Buchler P. Quantification of popliteal artery deformation during leg flexion in subjects with peripheral artery disease: a pilot study. J Endovasc Ther. 2013, 20: 828-35.
(9) Grayson J. The cutaneous ligaments of the digits. J Anat. 1941, 75: 164.
(10) Hahn P, Krimmer H, Hradetzky A, Lanz U. Quantitative analysis of the linkage between the interphalangeal joints of the index finger.An in vivo study.J Hand Surg Br. 1995, 20: 696-9.
(11) Kojima T, Tsuchida Y, Hirase Y, Endo T. Reverse vascular pedicle digital island flap. Br J Plast Surg. 1990, 43: 290-5.
(12) Kostopoulos E, Agiannidis C, Konofaos P et al. Predictable pattern digital artery perforator flap: an alternative concept in digital reconstruction.Hand (NY) .2016, 11: 88-96.
(13) Leslie BM, Ruby LK, Madell SJ, Wittenstein F. Digital artery diameters: an anatomic and clinical study. J Hand Surg Am. 1987, 12: 740-3.
(14) Matsumoto Y, Asao Y, Yoshikawa A et al. Label-free photoacoustic imaging of human palmar vessels: a structural morphological analysis.Sci Rep. 2018, 8: 786.
(15) Nam YS, Jun YJ, Kim IB, Cho SH, Han HH. Anatomical study of the fingertip artery in Tamai zone I: clinical significance in fingertip replantation. J Reconstr Microsurg. 2017, 33: 45-8.
(16) Sekiguchi H. Development of the rapid MIP viewer for PAT data-KURUMI: Kyoto University Rapid and Universal MIP Imager. IEICE Tech Report, Med Imaging. 2017, 116: 163-7.
(17) Silverman PM, Gordon L. Early motion after replantation. Hand Clin. 1996, 12: 97-107.
(18) Smith DO, Oura C, Kimura C, Toshimori K. Artery anatomy and tortuosity in the distal finger. J Hand Surg Am. 1991, 16: 297-302.
(19) Strauch B, de Moura W. Arterial system of the fingers. J Hand Surg Am. 1990, 15: 148-54.
(20) Tang JB. Indications, methods, postoperative motion and outcome evaluation of primary flexor tendon repairs in Zone 2. J Hand Surg Eur. 2007, 32: 118-29.
(21) Toi M, Asao Y, Matsumoto Y et al. Visualization of tumor-related blood vessels in human breast by photoacoustic imaging system with a hemispherical detector array.Sci Rep. 2017, 7: 41970.
(22) Tsuge I, Saito S, Sekiguchi H et al. Photoacoustic tomography shows the branching pattern of anterolateral thigh perforators in vivo.Plast Reconstr Surg. 2018, 141: 1288-92.
(23) Wang LV, Hu S. Photoacoustic tomography: in vivo imaging from organelles to organs. Science. 2012, 335: 1458-62.
(24) Yamakado M, Ichihara K, Matsumoto Y et al. Derivation of gender and age-specific reference intervals from fully normal Japanese individuals and the implications for health screening. Clin Chim Acta. 2015, 447: 105-14.
(25) Yousif NJ, Cunningham MW, Sanger JR, Gingrass RP, Matloub HS. The vascular supply to the proximal interphalangeal joint. J Hand Surg Am. 1985, 10: 852-61.
(26) Zwanenburg RL, Werker PM, McGrouther DA. The anatomy and function of Cleland's ligaments. J Hand Surg Eur. 2014, 39: 482-90.

  As described above in detail, according to the image processing apparatus and the like according to the present invention, when the shapes of the plurality of movable parts of the subject are changed, the feature amount of the test site such as arterial blood vessel is compared with the conventional example. Can be accurately detected. In particular, the movable part is the joint of the finger, and the feature amount such as the amount of change in curvature, the amplitude of the spiral region, and the twist ratio in the test site such as the arterial blood vessel is calculated, and based on the feature amount, for example, age estimation, etc. Predetermined analysis can be performed.

1 inspection object,
10 optical ultrasound imaging device,
11 controller,
12 signal receiving circuit,
13 data memory,
14 Operation part,
15 Display,
16 Image processing unit,
20 optical ultrasonic light source and detection device,
21 sensor support,
22 sensors,
23 mesh board,
24 water,
25 laser light source,
P1, P2 pretreatment,
Q1, Q2 Blood vessel position correction processing.

Claims (8)

By irradiating the test site of the subject with light, in an image processing apparatus equipped with an imaging means for imaging the test site into a three-dimensional image,
The imaging means images a test site including a plurality of movable parts into a three-dimensional image,
The image processing device,
An image processing apparatus comprising: a processing unit that calculates a feature amount of the test site before and after the plurality of movable portions are bent based on the imaged three-dimensional image.   The processing means calculates the characteristic amount of the test site before and after the plurality of movable parts are bent, and then corrects the position when the plurality of movable parts is bent in correspondence with the plurality of movable parts. The image processing apparatus according to claim 1, wherein   The processing means, after extracting the test site from the imaged three-dimensional image, divides the test site into a plurality of regions corresponding to the movable parts, thereby specifying the plurality of regions. The image processing apparatus according to claim 1, wherein the image processing apparatus executes preprocessing.   The feature amount is at least one of a change amount of curvature of the test site, an amplitude of a spiral region of the test site, an expansion / contraction rate of the test site, and a twist rate of the test site. The image processing apparatus according to any one of claims 1 to 3, which is characterized.   The image processing device according to claim 1, wherein the imaging unit is an optical ultrasonic imaging device. The test site is a finger,
The processing means bends at a predetermined angle in a region between the proximal interphalangeal joint of the finger and the base, which corresponds substantially to the distal third of the base. The image according to any one of claims 1 to 5, wherein the age of the subject is estimated based on the amount of change in the three peaks of the blood vessel curvature of the finger when changed. Processing equipment. Imaging means, by irradiating the test site of the subject with light, the step of imaging the test site including a plurality of movable parts into a three-dimensional image,
The processing means calculates the characteristic amount of the test site before and after the plurality of movable parts are bent based on the imaged three-dimensional image, and the image processing method.   An image processing program executed by a computer, comprising each step according to claim 7.