JP7153261B2 - IMAGE PROCESSING DEVICE, OPERATING METHOD OF IMAGE PROCESSING DEVICE, AND IMAGE PROCESSING PROGRAM - Google Patents

Description

TECHNICAL FIELD The present invention relates to image processing of medical images, and more particularly to an image processing apparatus, an operating method of the image processing apparatus, and an image processing apparatus program for analyzing an image obtained by photographing the liver.

Fibrotic diseases are diseases that occur in all organs throughout the body, and cause organ failure in various organs. Fibrosis is a state in which a protein called "fiber (collagen)", which is produced when repairing damage caused by inflammation, etc., increases and spreads. ceases to function normally. Cirrhosis is one of the fibrotic diseases.

Conventionally, a liver biopsy has been performed to confirm the state of liver cirrhosis. Alternatively, as liver cirrhosis progresses, the liver surface becomes uneven, so using images obtained by abdominal CT (Computed Tomography), the degree of unevenness of the liver surface and / or the balance of the size of the right and left lobes of the liver. Diagnosis of liver cirrhosis is made by examining the Further, in Patent Document 1, an MRI (Magnetic Resonance Imaging) apparatus is used to capture images before and after the liver undergoes a morphological change (such as inhalation and exhalation), and two cross-sectional images are obtained from each image. Then, the edges of the liver appearing in the two cross-sectional images are extracted, and information about the hardness of the liver is obtained from the index representing the change in the edges.

JP-A-6-311978

Liver biopsy is the most accurate way to actually know the degree of liver cirrhosis, but it is not suitable for observation of changes over time because it is difficult to reproduce sampling at the same position in liver biopsy. On the other hand, the technique of Patent Document 1 only analyzes a two-dimensional image of a cross-sectional image, and it is considered that the result changes depending on how the cross-section is selected. In addition, since only peripheral information is used, there is a possibility that accurate results cannot be obtained because the information inside the organ is not utilized. Thus, no effective means has been established for simply and objectively confirming changes in the state of liver stiffness or liver fibrosis.

Therefore, in order to solve the above-described problems, the present invention provides an image processing apparatus, an image processing apparatus operation method, and an image processing apparatus program that can easily and objectively confirm the state of liver fibrosis. intended to provide

An image processing apparatus according to the present invention includes an acquisition unit that acquires first volume data and second volume data obtained by imaging a subject's liver in two states in which morphological changes have occurred; By performing a non-rigid registration process between the liver region of the data and the liver region of the second volume data, the deformation vector representing the direction and distance that the corresponding positions in the two liver regions were deformed was analyzed and obtained. and an index output unit that outputs an index representing liver stiffness.

A method of operating an image processing apparatus according to the present invention is a method of operating an image processing apparatus including an acquisition unit and an index output unit, wherein the acquisition unit detects two states in which a morphological change has occurred in the liver of a subject. acquires the first volume data and the second volume data captured in , and the index output unit performs non-rigid registration processing between the liver region of the first volume data and the liver region of the second volume data. By doing so, an index representing the hardness of the liver obtained by analyzing the deformation vector representing the direction and distance in which the corresponding positions in the two liver regions are deformed is output.

The image processing program of the present invention comprises a computer, an acquisition unit that acquires first volume data and second volume data obtained by imaging the liver of a subject in two states in which morphological changes have occurred; By performing a non-rigid registration process between the liver region of one volume data and the liver region of the second volume data, a deformation vector representing the direction and distance that the corresponding positions in the two liver regions are deformed is analyzed. function as an index output unit that outputs an index representing the hardness of the liver obtained by

Moreover, it is preferable that the index output unit outputs the deformation amount obtained based on the magnitude of the deformation vector as the index value representing the index.

Further, the deformation amount may be an average value of magnitudes of deformation vectors.

Further, the first volume data may be data captured with the subject in the supine position, and the second volume data may be data captured with the subject in the prone position. .

The first volume data is data captured with the subject's lung field expanded, and the second volume data is the subject's lung field being larger than the lung field of the first volume data. Data captured in a reduced state may also be used.

Further, the index output unit uses the difference value in the body axis direction between the upper end position of the liver region of the first volume data and the upper end position of the liver region of the second volume data to calculate the hardness level represented by the index. It is desirable to perform adjustment such that the hardness level represented by the index increases as the value increases, or the hardness level represented by the index decreases as the difference value decreases.

A "high level of firmness" refers to a condition in which the liver is stiffer than a "low level of firmness". “Adjustment to raise the level of hardness represented by the index” adjusts the index so that it indicates a more rigid state of the liver than the current index, and “Adjustment to lower the level of hardness represented by the index” is adjusted to indicate that the liver is softer than the current index.

Alternatively, the index output unit may output a value obtained by dividing the deformation amount obtained based on the magnitude of the deformation vector by the difference value as the index value representing the index.

Alternatively, the index output unit may adjust using a table in which indices are prepared in advance.

Alternatively, the index output unit may acquire an index for each of a plurality of liver segments forming the liver region of the first volume data and the liver region of the second volume data.

Alternatively, the index output unit may acquire the index by analyzing deformation vectors of the upper right lobe of the liver in the liver region of the first volume data and the liver region of the second volume data.

In addition, the index is the liver fibrosis level, and the index output unit provides first volume data and second volume data obtained by imaging in two states in which the liver fibrosis level and the patient's liver undergo morphological changes. The first volume data and the second volume data of the subject are input to a discriminator that has learned the relationship between the liver fibrosis level and the deformation vector using a teacher data set having a liver fibrosis level may be obtained.

Also, the index output unit acquires the magnitude of the deformation vector at each position of the liver region of the first volume data and the liver region of the second volume data as indices, Each position on the outline shape representing the outline of the liver region corresponding to each position of the liver region may be displayed in a color according to the index of each position of the liver region of the volume data.

Further, the index output unit acquires deformation vectors at respective positions of the liver region of the first volume data and the liver region of the second volume data as indices, and outputs the liver region of the first volume data and the second volume data. A deformation vector at each position of the liver region may be displayed at each position on the contour shape representing the contour of the liver region corresponding to each position of the liver region.

Also, the index output unit preferably performs rigid registration before non-rigid registration.

Moreover, it is desirable that the rigid body alignment should match the position of the center of gravity of the liver region of the first volume data with the position of the center of gravity of the liver region of the second volume data.

Further, the rigid registration may be performed using landmarks of the liver region of the first volume data and landmarks of the liver region of the second volume data.

Also, the first volume data and the second volume data are preferably data obtained by imaging with a CT device, an MRI device, or an ultrasonic device.

Another image processing apparatus of the present invention comprises a memory storing instructions for execution by a computer, and a processor configured to execute the stored instructions, the processor morphologically transforming the liver of a subject. First volume data and second volume data obtained by imaging in two changed states are acquired, and non-rigid registration is performed between the liver region of the first volume data and the liver region of the second volume data. By performing the processing, a deformation vector representing the direction and distance in which the corresponding positions in the two liver regions are deformed is obtained, the deformation vector is analyzed to obtain an index representing the stiffness of the liver, and the index is output. Execute the process.

According to the present invention, non-rigid registration processing is performed between the liver regions of the first volume data and the second volume data obtained by photographing the liver of the subject in two states in which morphological changes have occurred. acquires a deformation vector representing the direction and distance in which the corresponding positions in the two liver regions are deformed, and outputs an index representing the hardness of the liver. And it becomes possible to confirm objectively.

Diagram showing the schematic configuration of the medical information system 1 is a diagram showing a schematic configuration of an image processing apparatus according to a first embodiment of the present invention; FIG. Images of normal cases of the liver taken with the subject in the supine and prone positions Images of abnormal cases of the liver taken with the subject in the supine and prone positions Normal images of the liver taken by the subject in the exhalation and inhalation states Images of abnormal cases of the liver taken in the exhalation and inhalation states of the subject Diagram to explain non-rigid alignment An example of the distribution of deformation vector magnitudes Diagram for explaining a plurality of segments that make up the liver region An example of displaying deformation vectors at each position inside the liver region with arrows Example of color-coded display of the magnitudes of deformation vectors at multiple locations in the liver region Flowchart showing the flow of processing of the image processing apparatus according to the first embodiment FIG. 2 is a diagram showing a schematic configuration of an image processing apparatus according to a second embodiment of the present invention; Flowchart showing the flow of processing of the image processing apparatus according to the second embodiment

A medical information system equipped with an image processing apparatus according to a first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a medical information system according to this embodiment.

In the medical information system of the present embodiment, as shown in FIG. 1, an image processing apparatus 1, a medical image storage server 2, and an imaging apparatus 3 (hereinafter referred to as a modality) can communicate with each other via a network 4. It is configured by connecting with

The modality 3 is, for example, a CT device, an MRI device, an ultrasound device, or the like, and captured three-dimensional volume data is transmitted over the network 4 according to a storage format and communication standard conforming to the DICOM (Digital Imaging and Communication in Medicine) standard. It is transmitted to the medical image storage server 2 via and stored therein.

The image processing apparatus 1 is a general-purpose computer, and includes a CPU (Central Processing Unit), a memory (main storage device), a storage (auxiliary storage device), an input/output interface, a communication interface, an input device, a display device, and a data bus. It has a well-known hardware configuration such as, and a well-known operating system is installed. It also has a liquid crystal display or the like as a display device, and a keyboard and/or a pointing device such as a mouse as an input device. The storage is composed of a hard disk, SSD (Solid State Drive), or the like. Note that the computer may be provided with a GPU (Graphics Processing Unit) as necessary. This computer functions as the image processing apparatus 1 by installing the image processing program of the present embodiment. The image processing apparatus 1 also has a function of requesting the medical image storage server 2 to send an image and receiving an image from the medical image storage server 2. Each function is performed by executing a software program. .

The image processing program is recorded on a recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) for distribution, and is installed in a computer from the recording medium. Alternatively, the image processing program is stored in a storage device of a server computer connected to a network or network storage in an externally accessible state, downloaded to the computer in response to an external request, and then installed. You may do so.

As shown in FIG. 2, the image processing apparatus 1 of the first embodiment includes an acquisition unit 10, a liver region extraction unit 11, a global alignment unit 12, a local alignment unit 13, an index acquisition unit 14, and an index output unit. A part 15 is provided.

The acquiring unit 10 acquires from the medical image storage server 2 first volume data V1 and second volume data V2 obtained by imaging the subject's liver in two states in which morphological changes have occurred. The volume data V1 and V2 are three-dimensional image data captured by a CT device, an MRI device, an ultrasonic imaging device, or the like in this embodiment. Volume data consists of data representing physical quantities at voxel positions that finely divide a three-dimensional space. , is represented by a value corresponding to the ultrasonic wave reflected by an organ or tissue.

The first volume data V1 and the second volume data V2 are images captured in two states before and after morphological changes occur in the liver. Specifically, morphological changes in the liver are caused by changes in posture, influence of respiration, or the like.

For example, a morphological change occurs in the liver depending on whether the subject is lying on his or her back (supine position) or lying down (prone position). Therefore, an image captured with the subject in the supine position is defined as first volume data V1, and an image captured with the subject in the prone position is defined as second volume data V2. FIG. 3A shows images of a normal liver taken by the subject in the supine and prone positions. The images under a on the left side are an axial image, a sagittal image, and a coronal image taken in the supine position. The images under b on the right side are the axial, sagittal, and coronal images taken in the prone position. FIG. 3B shows images of a liver with advanced cirrhosis taken by the subject in the supine position (a on the left) and in the prone position (b on the right). The shape of the upper part and the lower part of the liver (black line) in the image of the normal example in FIG. When the two are compared, the non-normal cases show less shape change than the normal cases.

Alternatively, respiration causes morphological changes in the liver due to changes in the size of the lung fields. Therefore, an image captured with the subject's lung field expanded is used as the first volume data V1, and an image captured with the subject's lung field reduced from the lung field of the first volume data V1. may be used as the second volume data V2. It is preferable that the first volume data V1 and the second volume data V2 have a large change in the size of the lung field. It is preferable to take images in the state of minimum exhalation. FIG. 4A is an image of a normal liver taken by a subject in an inspiratory state and an expiratory state. The images under a on the left side are an axial image, a sagittal image, and a coronal image taken in the inspiratory state. Images under b on the right side are an axial image, a sagittal image, and a coronal image taken in the exhaled state. FIG. 4B is an image of a liver with advanced cirrhosis taken by a subject in an inspiratory state (a on the left) and in an expiratory state (b on the right). Comparing the shape of the upper part of the liver in the image of the normal example in FIG. 4A (white arrow) and the shape of the upper part of the liver in the image of the abnormal example in FIG. Little change in shape.

The first volume data V1 and the second volume data V2 are preferably captured in the state of maximum inspiration and minimum expiration in which the size of the lung field changes greatly, but it is sufficient if the size of the lung field is different. For example, it is possible to use images obtained by administering a contrast agent and performing imaging in a plurality of phases. Normally, when performing CT imaging using a contrast agent, imaging is performed in four phases according to the state of diffusion of the contrast agent. First, breathing is stopped while inhaling, and 2 phases (early arterial phase, late arterial phase) are imaged. Volume data can be captured while the size of the field is changed. It is also possible to select volume data in which the size of the lung field is greatly changed from the four-phase volume data, and use them as the first volume data V1 and the second volume data V2.

The volume data V1 and V2 are stored in the medical image storage server 2 in advance together with patient identification information. The acquisition unit 10 acquires the first volume data V1 and the second volume data V2 of the patient to be diagnosed from the medical image storage server 2 based on the patient identification information input by the user using an input device such as a keyboard. , and stored in a storage (not shown).

The liver region extraction unit 11 extracts liver regions from each of the first volume data V1 and the second volume data V2. The liver region is extracted according to pixel values and the like according to the imaging modality, and various known techniques can be used.

The global registration unit 12 performs global registration of the liver by rigid registration. In rigid registration, registration is performed using only translation and rotation movements so as not to change the shape of the liver. Specifically, in order to match the position of the center of gravity of the liver region of the first volume data V1 with the position of the center of gravity of the liver region of the second volume data V2, Perform parallel movement. Furthermore, the six parameters of translation and rotation in the X, Y, and Z axial directions of the rigid body transformation model are optimized so that the overlapping volume of the two liver regions is maximized. to Alternatively, parallel movement and/or rotational movement is performed so that the landmarks of the liver region of the second volume data V2 corresponding to the landmarks of the liver region of the first volume data V1 come closer. The landmark may be a position representative of the liver, such as the upper edge of the liver region, the boundary between the right and left lobes, the portal vein, or the hepatic artery, and one of these landmarks may be Using the above landmarks, alignment is performed so that the sum of the distances between the two liver landmarks is the smallest.

Next, the local registration unit 13 performs detailed registration within the liver region by non-rigid registration with respect to the liver region for which rigid registration has been performed by the global registration unit 12 . In non-rigid registration, registration is performed by changing the shape of at least one of the two livers to be registered. Linear registration such as affine transformation and non-linear registration using B-splines are known as non-rigid registration for registering between two different volume data. Medical images are registered using such linear registration and non-linear registration.

The registration of medical images is performed by moving each point _xM in the image I _M of the second volume data V2 while keeping each point _xF in the image I _F of the first volume data V1 fixed. , to find the geometric transformation function T that matches the position of each point x _M in the image I _M with the position of each point x _F in the anatomically corresponding image I _F . As shown in FIG. 5, this problem can be solved as an optimization problem for maximizing the degree of matching between one image IF and the other image _IM transformed by _T as an evaluation function. can. Since there is no way to obtain a global optimum solution, the solution is found by setting some initial values and then iteratively searching for a better solution using the gradient method or the like. Mutual information can be used as the degree of matching. Mutual information is used for registering images taken with different modalities such as CT images and MRI images, but in the case of images taken with the same modality, a cross-correlation function may be used.

Specifically, first, a B-spline transformation model is used as a geometric transformation model to provide grid points in the volume data, and rough alignment is performed by moving the grid points to search for matching points. conduct. Then fine pixel-by-pixel alignment is performed with a Diffecomorphic transformation model. That is, after obtaining the optimum solution of the B-spline transformation model, the optimum solution of the Diffecomorphic transformation model is obtained using the result as an initial value. Thereby, corresponding positions in the two liver regions can be found between the two liver regions of the first volume data and the second volume data after rigid registration.

The index acquisition unit 14 acquires a deformation vector representing the direction and distance in which the two liver regions are deformed from the corresponding positions in the two liver regions obtained by the non-rigid registration of the local registration unit 13 . That is, a vector connecting each point x _M in the liver region of image I _M corresponding to each point x _F in the liver region of image I _F found by non-rigid registration is obtained as a deformation vector. I _F and image I _M are the images after the liver region has been rigid registration). Furthermore, an index representing liver stiffness is obtained from the deformation vector field. For example, the deformation amount of this deformation vector is calculated as an index value.

FIG. 6 shows an example of the distribution of magnitudes of deformation vectors. In FIG. 6, the horizontal axis indicates the magnitude of the deformation vector (deformation vector length), and the vertical axis indicates the relative frequency. In addition, the solid line shows an example of the distribution of liver deformation vectors in a state where fibrosis is advanced enough to require transplantation, and the dashed line shows the distribution of liver deformation vectors in a state where fibrosis is present but not serious. , and the dashed-dotted line shows an example of the distribution of deformation vectors of a liver in a healthy state. An average value of magnitudes of such deformation vectors can be obtained and used as an index value. As for the mean value of the magnitude of the deformation vector, the smaller the value, the more advanced the fibrosis. Alternatively, the average value and standard deviation of the magnitudes of deformation vectors may be calculated as index values representing liver stiffness. Alternatively, the relationship between the distribution of deformation vector magnitudes and the liver fibrosis level may be statistically analyzed from a large number of data, and the liver fibrosis level representing the fibrosis stage may be used as an index. Preferably, liver fibrosis levels correspond to fibrosis stages.

In addition, the index may not be a numerical value, and may represent the level of liver stiffness in letters. For example, "A" for healthy liver without evidence of cirrhosis, "B" for mild cirrhosis, moderate cirrhosis, depending on the mean magnitude of the deformation vector or the level of liver fibrosis. A letter index representing the level of liver stiffness may be used, such as "C" for severe cirrhosis and "D" for severe cirrhosis. Furthermore, the index is not limited to letters, and may be numbers or a combination of letters and numbers. Furthermore, the index may be a sentence such as "moderate liver cirrhosis".

Alternatively, the index acquisition unit 14 may acquire an index for each of a plurality of segments forming the liver region. The liver is divided into left and right lobes by the falciform ligament of the liver as a boundary, and is further divided into four segments: the lateral segment, the medial segment, the anterior segment, and the posterior segment. These segments are further divided into two segments: S1, caudate segment; S2, posterolateral segment (lateral superior segment); S3, anterolateral segment (lateral inferior segment); S4, medial inferior segment (quadrata) , the lower front section of S5, the lower rear section of S6, the upper rear section of S7, and the upper front section of S8 (see FIG. 7). An index may be obtained for each of these areas.

In addition, since the upper right lobe of the liver has a thin and flat shape among other livers, its morphology changes greatly depending on the patient's body position and breathing condition. Therefore, it is preferable to acquire the index by analyzing the deformation vector of the upper right lobe of the liver in the liver region.

Furthermore, the index acquisition unit 14 may acquire the liver fibrosis level as an index value using a discriminator. For example, data in which the liver fibrosis level is associated with the first volume data V1 and the second volume data V2 obtained by imaging the patient's liver in two states in which morphological changes have occurred is set as a teacher data set. A number of discriminators are prepared in advance, and the relationship between the liver fibrosis level and the deformation vector is machine-learned. The liver fibrosis level may be acquired by inputting the volume data V2.

The index output unit 15 outputs and displays an index value or characters representing the index on a display device such as a liquid crystal display. Specifically, numerical values and characters representing the hardness level are displayed on the screen displaying the volume data as shown in FIGS. 3A, 3B, 4A, and 4B. Alternatively, the index may be output to a printer and printed on paper or the like.

Alternatively, the index acquisition unit 14 may acquire the deformation vectors themselves at a plurality of positions within the liver region as indices. In this case, the index output unit 15 may display the deformation vector of each position of the liver region at each position on the outer shape representing the outer shape of the liver region corresponding to each position of the liver region. FIG. 8 is an example of displaying deformation vectors at respective positions inside the liver region in an axial section of the liver. In FIG. 8, the black arrow indicates the + direction and the white arrow indicates the - direction of displacement in the Z direction (body axis direction). In FIG. 8, the displacement in the Z direction is indicated by a black arrow and a white arrow for the sake of convenience, but it may be displayed in different colors such as red and blue on the display.

Furthermore, the index acquisition unit 14 may acquire, as an index, magnitudes obtained from deformation vectors at a plurality of positions within the liver region. In this case, the index output unit 15 may color-code and display each position on the outer shape representing the outer shape of the liver region in a color corresponding to the index of each position of the liver region. FIG. 9 is an example in which color coding is indicated by hatching for convenience. Different hatching in FIG. 9 corresponds to the red and blue color coding, respectively, where red indicates that the magnitude of the displacement vector is greater than or equal to a predetermined value, and blue indicates that the magnitude of the displacement vector is less than or equal to the predetermined value. may be shown.

Next, the processing flow of the image processing apparatus 1 of this embodiment will be described with reference to the flowchart shown in FIG.

First, when the patient's identification information is input by the user, the acquiring unit 10 acquires two volume data V1 and V2 from the patient's contrast-enhanced CT image (ST10).

Next, a liver region is extracted from the two volume data V1 and V2 acquired by the acquisition unit 10 using the liver region extraction unit 11 (ST11). In the global registration unit 12, parallel translation is performed in each of the X, Y, and Z axial directions of rigid body transformation so that the center of gravity of the liver coincides. Furthermore, with the parameters for translation that match the center of gravity of the liver as initial values, six parameters for translation and rotation in the X, Y, and Z axial directions are optimized to perform rigid body alignment (ST12). .

Furthermore, the local registration unit 13 performs detailed registration within the liver region by performing non-rigid registration on the liver region for which the rigid registration has been performed by the global registration unit 12, and the first volume The corresponding positions within the liver region of the data V1 and the second volume data V2 are found (ST13).

The index acquisition unit 14 acquires a deformation vector representing the direction and distance of deformation of the two liver regions of the first volume data V1 and the second volume data V2 based on the result of the non-rigid registration (ST14). . Furthermore, an index is obtained from the deformation vector (ST15). The index output unit 15 displays the index on the display (ST16).

As described in detail above, the deformation vector of the liver is obtained by non-rigid registration, and the hardness level of the liver is obtained. Therefore, the state of liver fibrosis can be easily and objectively confirmed becomes possible.

In the above-described first embodiment, the case where rigid body registration is performed before non-rigid body registration has been described. If two pieces of volume data are captured without a large change in axial position, only non-rigid alignment may be performed without performing rigid alignment.

Next, a second embodiment will be described. In the present embodiment, a method for obtaining the hardness level more accurately when the morphological change of the liver occurs depending on the size of the lung field will be described. The same reference numerals as in the first embodiment are assigned to the same configurations as in the first embodiment, detailed descriptions thereof are omitted, and only different configurations are described in detail.

The medical information system of the second embodiment is similar to that of the first embodiment, so detailed description is omitted. As shown in FIG. 11, the image processing apparatus 1a of the second embodiment includes an acquisition unit 10, a liver region extraction unit 11, a global alignment unit 12, a local alignment unit 13, an index acquisition unit 14a, and an index output unit. A part 15 is provided. Further, the index acquisition unit 14a of the present embodiment differs from that of the first embodiment in that it includes an adjustment unit 16 that adjusts the index.

As in the first embodiment, the index acquisition unit 14a obtains the first volume data V1 and the second volume data V2 based on the result of the non-rigid registration performed by the local registration unit 13. A deformation vector representing the direction and distance of deformation of one liver region is obtained, and an index is obtained based on the deformation vector.

However, even with the same subject, the results will be different if the exhalation and inhalation conditions are different. For example, if the breathing conditions when two volume data V1 and V2 are captured are maximum inhalation and minimum exhalation, the size of the lung field changes greatly when liver cirrhosis does not progress, so the amount of deformation of the liver region is Although the value is large, when the change in the size of the lung field when the two volume data V1 and V2 are captured is small, the deformation amount of the liver region is a relatively small value even if liver cirrhosis has not progressed. On the other hand, when liver cirrhosis is progressing, the deformation amount becomes a small value even if the size of the lung field changes greatly or the size of the lung field does not change much.

During exhalation and inhalation, the position of the diaphragm changes greatly, and the position of the liver also moves up and down accordingly. When imaging the two volume data V1 and V2 with different breathing conditions, the body position of the subject does not change with respect to the imaging device. can be estimated from changes in

Therefore, the adjustment unit 16 calculates the difference value in the body axis direction between the upper end position of the liver region of the first volume data V1 and the liver region of the second volume data V2 before the global registration unit 12 performs the registration. Then, the hardness level represented by the index is adjusted so that the hardness level represented by the index increases as the difference value increases, and the hardness level represented by the index decreases as the difference value decreases. . In other words, when the size of the lung field changes greatly, the size of the deformation vector is larger than when the size of the lung field does not change much, and the index is that the liver is soft (low hardness level). , so adjust to increase the hardness level. On the other hand, when the size of the lung field does not change much, the size of the deformation vector becomes smaller than when the size of the lung field changes greatly, and the index is that the liver is harder than it actually is (the hardness level is high), so adjust to lower the hardness level. This makes it possible to add a correction that absorbs the difference in lung field size that varies depending on the state of expiration or inspiration.

Specifically, the index acquisition unit 14a divides the deformation amount acquired based on the magnitude of the deformation vector obtained as the index value representing the index by the difference value to obtain the index value. Alternatively, the hardness level may be represented by an index represented by letters "A" to "D" according to the value obtained by dividing the deformation amount by the difference value.

The adjustment unit 16 of the index acquisition unit 14a prepares a table representing an index (an index value or an index representing a hardness level in characters) according to the deformation amount and the difference value of the deformation vector in advance, You may make it adjust an index. Alternatively, a table representing index values corresponding to indices obtained from deformation vectors and difference values may be prepared to adjust the indices.

Next, the processing flow of the image processing apparatus 1 of this embodiment will be described with reference to the flowchart shown in FIG.

First, when the user inputs patient identification information, the acquiring unit 10 acquires two volume data V1 and V2 from a contrast-enhanced CT image of the patient (ST20).

Next, a liver region is extracted from the two volume data V1 and V2 acquired by the acquisition unit 10 using the liver region extraction unit 11 (ST21). Here, the adjustment unit 16 acquires the difference value in the body axis direction between the upper end positions of the liver regions of the two volume data V1 and V2 (ST22).

Next, the global registration unit 12 translates the livers in parallel along the X, Y, and Z axes of rigid body transformation so that the centers of gravity of the livers match. Further, the parameters for parallel movement that match the center of gravity of the liver are used as initial values, and the six parameters for parallel movement and rotational movement in the respective axial directions of X, Y, and Z are optimized to perform rigid body alignment (ST23). .

Furthermore, the local registration unit 13 performs detailed registration within the liver region by performing non-rigid registration on the liver region for which the rigid registration has been performed by the global registration unit 12, and the first volume The corresponding positions within the liver region of the data V1 and the second volume data V2 are found (ST24).

The index acquisition unit 14a acquires a deformation vector representing the direction and distance in which the two liver regions of the first volume data V1 and the second volume data V2 are deformed based on the result of the non-rigid registration (ST25). . Furthermore, an index is obtained from the deformation vector (ST26). Here, the adjustment unit 16 adjusts the hardness level represented by the index so that the greater the difference value, the higher the hardness level represented by the index, and the smaller the difference value, the lower the hardness level represented by the index. Adjustment is performed (ST27). The index output unit 15 displays the adjusted index on the display (ST28).

As described in detail above, the deformation vector of the liver is obtained by non-rigid registration to obtain the stiffness level of the liver. By doing so, it becomes possible to more accurately acquire the state of liver fibrosis.

In the above description, the case where one computer functions as an image processing apparatus has been described, but functions may be distributed to a plurality of computers.

Moreover, although the case where a general-purpose computer functions as an image processing apparatus has been described above, it may be implemented by a dedicated computer. The dedicated computer may be firmware that executes a program recorded in a non-volatile memory such as a built-in ROM (Read Only Memory) or flash memory. Furthermore, dedicated circuits such as ASICs (Application Specific Integrated Circuits) and FPGAs (field programmable gate arrays) that permanently store programs for executing at least part of the functions of the image processing apparatus are provided. You may make it provide. Alternatively, program instructions stored in dedicated circuitry may be combined with program instructions executed by a general-purpose CPU programmed to utilize the dedicated circuitry program. As described above, any combination of computer hardware configurations may be used to execute program instructions.

1, 1a Image processing device 2 Medical image storage server 3 Imaging device 4 Network 10 Acquisition unit 11 Liver area extraction unit 12 Global registration unit 13 Local registration unit 14, 14a Index acquisition unit 15 Index output unit 16 Adjustment unit V1 First volume data V2 of second volume data

Claims (16)

Volume data obtained by imaging the subject's liver in two states in which morphological changes have occurred, wherein the first volume data includes an image of the liver in a state in which the subject's lung field is expanded; an acquisition unit that acquires second volume data including an image of the liver in a state in which the lung field of the subject is smaller than the lung field of the first volume data;
For the liver region of the first volume data and the liver region of the second volume data in which rigid body registration processing is performed between the liver region of the first volume data and the liver region of the second volume data performing non-rigid registration processing to analyze deformation vectors representing directions and distances in which corresponding positions in the two liver regions are deformed, and outputting an index representing liver stiffness according to the distances; an index output unit;
with
The index output unit
upper edge position of the liver region of the first volume data before the rigid registration processing and the non-rigid registration processing between the liver region of the first volume data and the liver region of the second volume data are performed; and the upper end position of the liver region of the second volume data in the body axis direction, the hardness level represented by the index is increased as the difference value increases. or to lower the level of hardness represented by the index as the difference value becomes smaller. 2. The image processing apparatus according to claim 1, wherein said index output unit outputs a deformation amount obtained based on the magnitude of said deformation vector as an index value representing said index. 3. The image processing apparatus according to claim 2, wherein the deformation amount is an average value of magnitudes of the deformation vectors . 2. The image processing apparatus according to claim 1, wherein the index output unit outputs, as the index value representing the index, a value obtained by dividing the deformation amount obtained based on the magnitude of the deformation vector by the difference value. 4. The image processing apparatus according to any one of claims 1 to 3 , wherein the index output unit adjusts the index using a table prepared in advance. 6. The index output unit according to any one of claims 1 to 5 , wherein the index output unit acquires the index for each of a plurality of liver segments forming a liver region of the first volume data and a liver region of the second volume data. The image processing device according to . 6. The index output unit obtains the index by analyzing deformation vectors of the upper right lobe of the liver of the liver region of the first volume data and the liver region of the second volume data. 2. The image processing device according to item 1. the indicator is liver fibrosis level,
The index output unit uses a teacher data set having first volume data and second volume data obtained by imaging the liver fibrosis level and two states in which the patient's liver undergoes morphological changes. The liver fibrosis level is obtained by inputting the first volume data and the second volume data of the subject to a discriminator that has learned the relationship between the liver fibrosis level and the deformation vector. The image processing apparatus according to claim 1. The index output unit acquires the magnitude of the deformation vector at each position of the liver region of the first volume data and the liver region of the second volume data as the index, each position on the outline shape representing the outline of the liver region corresponding to each position of the liver region corresponding to each position of the liver region of the second volume data is displayed in a color according to the index of each position of the liver region of the second volume data. 1. The image processing apparatus according to 1. The index output unit acquires the deformation vector at each position of the liver region of the first volume data and the liver region of the second volume data as the index, 2. The image processing apparatus according to claim 1, wherein the deformation vector of each position of the liver region of the second volume data is displayed at each position on an outer shape representing the outer shape of the liver region corresponding to each position of the liver region. The image processing apparatus according to any one of claims 1 to 10 , wherein the index output unit performs rigid registration before non-rigid registration. 12. The image processing apparatus according to claim 11 , wherein the rigid body registration matches the position of the center of gravity of the liver region of the first volume data and the position of the center of gravity of the liver region of the second volume data. 12. The image processing apparatus according to claim 11 , wherein the rigid registration is performed using landmarks of the liver region of the first volume data and landmarks of the liver region of the second volume data. The first volume data and the second volume data according to any one of claims 1 to 13 , wherein the data are obtained by imaging with a CT device, an MRI device, or an ultrasonic device. Image processing device. A method of operating an image processing device comprising an acquisition unit and an index output unit, comprising:
The acquisition unit obtains volume data obtained by imaging the subject's liver in two states in which morphological changes have occurred, the first volume data including an image of the liver in a state in which the lung field of the subject is expanded. and second volume data including an image of the liver in a state in which the lung field of the subject is smaller than the lung field of the first volume data,
The index output unit performs rigid body alignment processing between the liver region of the first volume data and the liver region of the second volume data, and the liver region of the first volume data and the second volume data. By performing non-rigid body registration processing on the liver regions of the two liver regions, a deformation vector representing the direction and distance in which the corresponding positions in the two liver regions are deformed is analyzed, and the stiffness of the liver according to the distance outputs an index representing
upper edge position of the liver region of the first volume data before the rigid registration processing and the non-rigid registration processing between the liver region of the first volume data and the liver region of the second volume data are performed; and the upper end position of the liver region of the second volume data in the body axis direction, the hardness level represented by the index is increased as the difference value increases. or adjusting the hardness level indicated by the index to decrease as the difference value decreases. the computer,
Volume data obtained by imaging the subject's liver in two states in which morphological changes have occurred, wherein the first volume data includes an image of the liver in a state in which the subject's lung field is expanded; an acquisition unit that acquires second volume data including an image of the liver in a state in which the lung field of the subject is smaller than the lung field of the first volume data;
For the liver region of the first volume data and the liver region of the second volume data in which rigid body registration processing is performed between the liver region of the first volume data and the liver region of the second volume data performing non-rigid registration processing to analyze deformation vectors representing directions and distances in which corresponding positions in the two liver regions are deformed, and outputting an index representing liver stiffness according to the distances; An image processing program that functions as an index output unit,
The first volume data before the rigid registration processing and the non-rigid registration processing between the liver region of the first volume data and the liver region of the second volume data are performed by the index output unit. using the difference value in the body axis direction between the upper end position of the liver region of the second volume data and the upper end position of the liver region of the second volume data. An image processing program for performing adjustment to increase the level of hardness or to decrease the level of hardness represented by the index as the difference value decreases.

Images