Magnetic resonance imaging apparatus, image processing apparatus, and image processing method

Abstract

To provide a technology to composite a blood vessel image stably while suppressing deterioration in brightness of a blood vessel when compositing a blood vessel image.SOLUTION: Optimization of an imaging parameter set for quantitative value map calculation is executed by adding at least one condition that there is no decrease in brightness of a blood vessel in a region in which a blood vessel is not depicted either in an original image or in a quantitative value image. An MRI device of the present invention includes: a measuring unit for measuring an echo signal generated from a subject; an image reconfiguration unit for generating a reconfiguration image from the measured echo signal; an imaging condition determination unit for determining a combination of a plurality of imaging conditions with different combinations; a quantitative value image generation unit for generating a quantitative value image with a quantitative value on the subject as a pixel value; and a composite image generation unit for compositing a blood vessel image. The imaging condition determination unit includes at least one imaging condition that a decrease in local brightness of a blood vessel is small in the combination of the plurality of imaging conditions.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to magnetic resonance imaging technology, and particularly to a technology for improving the ability to depict blood vessels in angiography using a calculated image.

Background Art

[0002] In the images obtained by a magnetic resonance imaging apparatus (hereinafter referred to as an MRI apparatus), in addition to the image reconstructed from the nuclear magnetic resonance signal (echo signal), based on a plurality of images (original images) taken with different imaging parameters, the subject parameters and apparatus parameters are calculated pixel by pixel, and there is a quantitative value image (also referred to as a calculated image) having these parameters as pixel values.

[0003] The quantitative value image is obtained by finding the least squares fit of the signal function with respect to the imaging parameters of a plurality of original images and their pixel values using a signal function that represents the relationship between the signal of the original images, the subject parameters, the imaging parameters, and the apparatus parameters. Here, the imaging parameters are the repetition time, the set intensity of the high-frequency magnetic field, the phase of the high-frequency magnetic field, etc., the subject parameters are the longitudinal relaxation time T1, the transverse relaxation time T2, the spin density ρ, the resonance frequency f0, the diffusion coefficient D, the irradiation intensity B1 distribution of the high-frequency magnetic field, etc., and the apparatus parameters are the magnetic field intensity B0, the reception sensitivity distribution s, etc.

[0004] [[ID=二十]]There are several methods for creating a quantitative value image. For example, as one method that enables simultaneous acquisition of calculated images of a plurality of subject parameters and apparatus parameters, a method of estimating a quantitative value image (map) by constructing a signal function by numerical simulation has been proposed (Patent Document 1).

[0005] On the other hand, various magnetic resonance angiography (MRA) techniques have been developed for imaging blood vessels. One such technique involves combining two or more MR images with different emphasized physical properties using a composite function optimized to enhance the brightness of blood vessels, thereby obtaining a vascular image (Patent Document 2). Patent Document 2 also proposes using T1-weighted images and PD images, as well as computational images where the aforementioned quantitative values ​​such as T1, T2, and PD are used as pixel values, as the images to be combined. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2011-024926 [Patent Document 2] Patent No. 6979151 [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] The method described in Patent Document 2 allows for the acquisition of MRA images by setting synthesis parameters for each region of the subject's tissue and then synthesizing them. However, if the brightness of a certain region of a blood vessel is very low in all the source images used for synthesis, even if a synthesis function optimized to ensure uniform blood vessel brightness is used, the brightness of that region may still be low in the synthesized blood vessel image.

[0008] This invention has been made in view of the above circumstances, and provides a technique for stably synthesizing vascular images while suppressing a decrease in the brightness of blood vessels when synthesizing vascular images. [Means for solving the problem]

[0009] This invention optimizes the imaging parameter set by adding a condition that prevents a decrease in brightness even in certain areas of blood vessels with low brightness when generating computational images used for synthesizing blood vessel images.

[0010] Specifically, the MRI apparatus of the present invention comprises: a measurement unit that applies a high-frequency magnetic field and a gradient magnetic field to a subject placed in a static magnetic field according to predetermined imaging conditions and a predetermined imaging sequence, and measures the echo signal generated from the subject; an image reconstruction unit that creates a reconstructed image from the measured echo signal; an imaging condition determination unit that determines a combination of multiple imaging conditions with different combinations; a quantitative value image generation unit that generates a quantitative value image in which the pixel value is a quantitative value related to the subject, using multiple reconstructed images created by the image reconstruction unit from the echo signals acquired under each of the multiple imaging conditions and a signal function determined by the imaging sequence; and a composite image generation unit that synthesizes a vascular image by combining part or all of the original image including the reconstructed image and the quantitative value image. The imaging condition determination unit includes at least one imaging condition in the combination of the multiple imaging conditions in which the local brightness reduction of blood vessels is small. [Effects of the Invention]

[0011] According to the present invention, when determining imaging parameters, it is possible to ensure that there is always brightness in at least one of the images used for synthesis in all regions of the blood vessels, thus preventing a decrease in brightness in some blood vessels in the synthesized vascular image. This allows for the stable acquisition of synthesized vascular images. [Brief explanation of the drawing]

[0012] [Figure 1] This is a block diagram showing the schematic configuration of an MRI apparatus according to an embodiment of the present invention. [Figure 2] This is a functional block diagram of a computer according to an embodiment of the present invention. [Figure 3] This is a processing flow in an embodiment of the present invention. [Figure 4] Figure 3 is a flowchart detailing the process. [Figure 5] The diagram shows sequence diagrams in embodiments of the present invention, where (a) shows a general gradient echo pulse sequence and (b) shows a gradient echo pulse sequence with added flow compensation pulses. [Figure 6] This is a diagram showing a part of the signal function in an embodiment of the present invention. [Figure 7] These are the tissues to be optimized for the imaging parameter set in an embodiment of the present invention, along with their T1 and T2 values. [Figure 8] This is a flowchart showing details of a part of the process in FIG. 5. [Figure 9] This is a diagram showing the original image for synthesizing the blood vessel image in an embodiment of the present invention. [Figure 10] (a) to (c) are diagrams for explaining details of the synthesis process in an embodiment of the present invention. [Figure 11] This is a diagram for explaining the effect of improving the image quality of the blood vessel image in an embodiment of the present invention. [Figure 12] This is a diagram for explaining the effect of improving the image quality of the blood vessel image in an embodiment of the present invention.

Embodiments for Carrying Out the Invention

[0013] Hereinafter, embodiments to which the present invention is applied will be described. In the following description, the imaging conditions include parameters (also referred to as imaging parameters) that can be arbitrarily set by the user at the time of executing the imaging sequence. The imaging pulse sequence itself may also be included in the imaging conditions. In addition, parameters depending on the subject are called subject parameters, and parameters specific to the MRI apparatus are called apparatus parameters.

[0014] The imaging conditions (imaging parameters) include, for example, repetition time (TR), echo time (TE), set intensity of the high-frequency magnetic field (flip angle (FA)), phase (θ) of the high-frequency magnetic field, etc. Subject parameters include longitudinal relaxation time (T1), transverse relaxation time (T2), spin density (ρ), resonance frequency difference (Δf0), diffusion coefficient (b), irradiation intensity distribution (B1) of the high-frequency magnetic field, etc. Apparatus parameters include static magnetic field intensity (B0), sensitivity distribution (Sc) of the receiving coil, etc. Note that the resonance frequency difference Δf0 is the difference between the resonance frequency of each pixel and the reference frequency f0.

[0015] First, referring to FIG. 1, an embodiment of the MRI apparatus of the present invention will be described. In all the figures for explaining the embodiment of the present invention, those having the same function are denoted by the same reference numerals, and the repeated description thereof will be omitted.

[0016] FIG. 1 is a block diagram showing a schematic configuration of an MRI apparatus 10 according to the present embodiment. The MRI apparatus 10 includes, as a measurement unit 100 for measuring echo signals, a magnet 101 that generates a static magnetic field, a gradient magnetic field coil 102 that generates a gradient magnetic field, a sequencer 104, a gradient magnetic field power supply 105, a high-frequency magnetic field generator 106, a transmission / reception coil 107 that irradiates a high-frequency magnetic field and detects a nuclear magnetic resonance signal, and a receiver 108. The MRI apparatus 10 also includes a computer 200 that controls the measurement unit 100 and performs various operations such as image reconstruction using the measured echo signals, a display 111, and a storage medium 112. The transmission / reception coil 107 is shown as a single coil in the figure, but it may separately include a transmission coil and a reception coil.

[0017] A subject (e.g., a living body) 103 is placed on a bed (table) within the static magnetic field space generated by the magnet 101.

[0018] The sequencer 104 sends commands to the gradient magnetic field power supply 105 and the high-frequency magnetic field generator 106 to generate a gradient magnetic field and a high-frequency magnetic field, respectively. The high-frequency magnetic field is applied to the subject 103 through the transmission / reception coil 107. The nuclear magnetic resonance signal generated from the subject 103 is received by the transmission / reception coil 107 and detected by the receiver 108. The nuclear magnetic resonance frequency (detection reference frequency f0) used as a reference for detection is set by the sequencer 104. The sequencer 104 usually controls the operation of each device at a preset timing and intensity. Among the programs, in particular, those describing the timing and intensity of the high-frequency magnetic field, gradient magnetic field, and signal reception are called pulse sequences (imaging sequences).

[0019] The computer 200 controls the measurement unit via the sequencer 104 and receives the signal detected by the receiver 108, performing processing such as image reconstruction. The results are displayed on the display 111. If necessary, the detected signal and shooting conditions can also be stored in the storage medium 112.

[0020] The processing performed by the computer 200 in this embodiment includes, as processing specialized for angiography, the generation of quantitative images and the processing of combining quantitative images and reconstructed images to generate a composite vascular image.

[0021] To realize these processes, the computer 200 of this embodiment, as shown in Figure 2, includes a measurement control unit 210 that controls the measurement unit 100, an image reconstruction unit 220 that obtains a reconstructed image from the measured echo signal, a quantitative value image generation unit 230 that generates a quantitative value image using reconstructed images acquired with a combination of multiple different shooting conditions, a shooting condition determination unit 240 that determines multiple shooting conditions for generating a quantitative value image, and a composite image generation unit 250 that synthesizes at least two or more original images including the reconstructed image and the quantitative value image to generate a composite image.

[0022] Furthermore, the quantitative value image generation unit 230 may include a signal function generation unit 231, a parameter estimation unit 232, etc., which generate a signal function for estimating the quantitative value parameter. The quantitative value is at least one parameter, which is a parameter dependent on the subject and a parameter specific to the device, and a quantitative value is obtained for each pixel. The distribution (map) of the quantitative values ​​obtained for each pixel is the quantitative value image. The composite image generation unit 250 may also include an image selection unit 251 for selecting multiple source images and a composite parameter setting unit 252 for setting the parameters of the composite function.

[0023] In the above configuration, as shown in Figure 3, the measurement control unit 210 controls the measurement unit 100 and the image reconstruction unit 220 based on a combination of multiple shooting conditions determined by the shooting condition determination unit 240 (S301), and obtains multiple reconstructed images with different combinations of shooting conditions (S302, S303). The quantitative value image generation unit 230 generates a quantitative value image using the multiple reconstructed images. For example, the signal function generation unit 231 generates a signal function for each shooting sequence by numerical simulation (S304), the parameter estimation unit 232 estimates the subject parameters for each pixel using the signal function for each shooting sequence, obtains a subject parameter distribution, and generates a quantitative value image of the subject from this subject parameter distribution (S305).

[0024] The composite image generation unit 250 combines two or more original images, consisting of quantitative value images and multiple reconstructed images, using a predetermined synthesis function (S306). The synthesis function and its parameters (synthesis parameters) used for synthesis may be predetermined, or the synthesis parameter setting unit 252 may set optimized ones. The composite vascular image generated by the composite image generation unit 250 is displayed on the display 111 (S307).

[0025] The functions of the computer 200 are realized by the CPU of the computer 200 loading a program stored in the storage medium 112 into memory and executing it. Alternatively, these functions may be realized by hardware such as a PLC (programmable logic device). Furthermore, all or part of the functions of the quantitative image generation unit 230 and the composite image generation unit 250 may be realized by a computer (image processing device) provided independently of the MRI device 10, which is capable of sending and receiving data with the computer 200 of the MRI device 10.

[0026] The specific details of the processing within the computer 200 during angiography will be explained below with reference to Figure 4. Figure 4 is a detailed diagram of the processing shown in Figure 3, and steps with the same processing content as in Figure 3 are indicated by the same symbols.

[0027] Here, as an example, we will explain the case where the GE-type RF-spoiled GE sequence (hereinafter abbreviated as RF-spoiled GE) shown in Figure 5(a) is used as the imaging sequence. In Figure 5, RF, Gs, Gp, and Gr represent the high-frequency magnetic field, slice gradient magnetic field, phase-encoded gradient magnetic field, and readout gradient magnetic field, respectively. The "-1" and "-2" in "501-1" and "501-2" are codes used to distinguish the same pulse from each repetition.

[0028] In this pulse sequence, first, a slice gradient magnetic field pulse 501 is applied along with a radio frequency (RF) pulse 502 to excite the magnetization of a slice within the target object. Next, a slice rephase gradient magnetic field pulse 503, a phase encoding gradient magnetic field pulse 504 to add positional information in the phase encoding direction to the magnetization phase, and a dephase readout gradient magnetic field 505 are applied. Then, while applying a readout gradient magnetic field pulse 506 to add positional information in the readout direction, a magnetic resonance signal (echo) 507 is measured. Finally, a dephase phase encoding gradient magnetic field pulse 509 is applied.

[0029] The above procedure is repeated for repetition time TR while varying the intensity (phase encoding amount kp) of phase-encoded gradient magnetic field pulses 504 and 509, and changing the phase increment of the RF pulse by 117 degrees (the phase of the nth RF pulse is θ(n) = θ(n-1) + 117n), and the echoes required to obtain one image are measured.

[0030] RF-spoiled GE yields images with enhanced T1 (longitudinal relaxation time).

[0031] [Generating signal functions: S304] The signal function generation unit 321 generates a signal function to be used for determining the imaging conditions (S301). In this embodiment, the RF-spoiled GE signal function 412 is created by numerical simulation. The RF-spoiled GE signal function fs is expressed as follows.

number

[0032] Here, T1, T2, ρ, and B1 are the longitudinal relaxation time, transverse relaxation time, spin density, and RF irradiation intensity of the subject parameters, respectively. Also, Sc is the receiving coil sensitivity of the instrument parameter. B1 is a coefficient of FA (flip angle) during imaging, so it is expressed as a product of FA. ρ and Sc act as proportionality constants with respect to signal intensity, so they are taken outside the function.

[0033] The adjustable imaging parameters in this RF-spoiled GE are FA (flip angle), TR (repetition time), TE (echo time), and θ (RF phase increment). Of these, the RF phase increment is generally fixed at 117 degrees to obtain an image contrast with low T2 dependence, similar to FLASH (Fast Low-angle Shot), one of the high-speed imaging methods. Changing θ significantly alters the T2 dependence of the image contrast.

[0034] In numerical simulations, signals are generated by comprehensively varying the imaging parameters FA, TR, and θ for arbitrary values ​​of the subject parameters T1 and T2, and then a signal function is created by interpolation. The spin density ρ of the object being scanned, as well as B1 and Sc, are assumed to be constant (e.g., 1).

[0035] The imaging parameters and subject parameters are changed as follows, for example. Each parameter should include the range of imaging parameters used in the actual imaging and the T1 and T2 ranges of the subject, respectively. 4 TRs [ms]: 10, 20, 30, 40 FA 10 points [degrees]: 5, 10, 15, 20, 25, 30, 35, 40, 50, 60 θ 10 [degrees]: 170, 171, 172, 173, 174, 175, 176, 177, 178, 179 T2 17 pieces [s]: 0.01, 0.02, 0.03, 0.04, 0.05, 0.07, 0.1, 0.14, 0.19, 0.27, 0.38, 0.53, 0.74, 1.0, 1.4, 2.0, 2.8 T1 15 pieces [s]: 0.05, 0.07, 0.1, 0.14, 0.19, 0.27, 0.38, 0.53, 0.74, 1.0, 1.5, 2.0, 2.8, 4.0, 5.6

[0036] A set of 102,000 imaging parameters 411 is constructed, consisting of all combinations of the above imaging parameters and subject parameters, and the signal value for each is calculated by computer simulation.

[0037] Numerical simulation takes a subject model with spins arranged on a grid, the imaging sequence, imaging parameters, and instrument parameters as input, and solves Bloch's equation, the fundamental equation of the magnetic resonance phenomenon, to output a magnetic resonance signal. The subject model is given as the spatial distribution of spins (γ, M0, T1, T2, Sc), where γ is the gyromagnetic ratio and M0 is the thermal equilibrium magnetization (spin density). By reconstructing the image from the magnetic resonance signal, an image can be obtained under the given conditions.

[0038] Bloch's equation is a first-order linear ordinary differential equation and can be expressed as follows:

number

[0039] The signal function (fs) 412 is constructed from the signal values ​​obtained by computer simulation through interpolation. Linear interpolation of order 1 to 3 or spline interpolation can be used for this purpose.

[0040] Figure 6 shows a portion of the signal function intensity created as described above. Figure 6 displays the case where T1=900 ms, T2=100 ms, and θ=5 degrees, with the horizontal and vertical axes representing FA and TR, respectively. Once the signal function is created and saved, it does not need to be created again each time a computational image is taken, and can be used repeatedly.

[0041] [Determination of shooting conditions: S301] The shooting condition determination unit 240 uses the signal function 412 created as described above to determine the shooting parameter set 413 for acquiring the calculated image. Specifically, the shooting condition determination unit 240 uses the law of error propagation to search for multiple shooting conditions (imaging parameter sets) 413 to be used for capturing the original image.

[0042] For example, the law of error propagation is used to optimize the estimation error of each map (T1, T2, B1, and a) for the major tissues of the human head (gray matter GM, white matter WM, fat, and cerebrospinal fluid CSF) to the minimum. The subject parameters to be targeted are shown in Figure 7. Since there are 4 unknowns (T1, T2, B1, and a), the imaging parameter set will be increased to 6 sets.

[0043] Examples of values ​​for each imaging parameter (FA, TR, θ) are shown below. FA 4 pieces [degrees]: 10, 20, 30, 40 4 TRs [ms]: 10, 20, 30, 40 θ (20 values): From 160 degrees to 179 degrees (in 1-degree increments)

[0044] Furthermore, the TE of the first echo and the ΔTE between echoes are usually kept as short as possible to increase the signal-to-noise ratio (SNR) for quantitative estimation, thereby increasing the number of multi-echo images. Under typical imaging conditions, the shortest TE and ΔTE are set to, for example, 4.6 ms, and as many multi-echo images as possible that fit within the TR are acquired. The number of these multi-echo images is typically only 1 echo when the TR is 10 ms, and a maximum of 8 echoes (with the last echo's TE being 36.8 ms) when the TR is 40 ms.

[0045] One set is formed by combining any of these, resulting in a total of 320 sets (4 x 4 x 20). The search range is defined as the combination where the sum of the TRs of the 6 sets falls within a predetermined range. The restriction on the sum of TRs is in place to prevent the search range from becoming too broad and to avoid excessively long exposure times. Specifically, the sum of TRs is set to approximately 100 ms to 120 ms.

[0046] Furthermore, at least one of the six sets of imaging parameters should be predetermined to ensure that the vascular brightness does not decrease locally. This condition can be determined by considering the causes of the decrease in vascular brightness as follows.

[0047] One of the main factors contributing to the decrease in vascular brightness in gradient echocardiography is the saturation of the blood signal. Specifically, during imaging, blood flowing in from the lower part of the head is continuously irradiated with RF pulses as it flows to the top of the head. As a result, the magnetization gradually saturates, the transverse magnetization decreases, and the signal intensity decreases. Consequently, the brightness of the blood vessels decreases towards the top of the head.

[0048] The way to suppress this signal degradation is to make FA small and TR long. In other words, under imaging conditions where FA is small and TR is long, for example, FA=10, TR=40, and θ is, for example, 170. This condition is always included in the six combinations of imaging conditions and optimized.

[0049] As mentioned above, the determination of the imaging conditions is based on the law of error propagation. In this process, optimization may be performed using the law of error propagation, including the condition that suppresses the decrease in blood signal intensity, or six sets of conditions may be determined by optimization without including the condition that suppresses the decrease in blood signal intensity, and then the condition that suppresses the decrease in blood signal intensity may be added as the seventh set of conditions. In the former case, stable optimization processing can be performed. In the latter case, the weight of the condition that suppresses the decrease in blood signal intensity increases, improving the blood signal suppression effect in quantitative images.

[0050] Alternatively, instead of changing imaging parameters such as TE and FA, or in addition to changing them, an imaging sequence that suppresses the decrease in blood signal may be used. Specifically, in RF-spoiled GE shown in Figure 5(a), a flow compensation gradient pulse is added between the RF pulse and the echo signal acquisition (A / D). Adding a flow compensation gradient pulse is effective in suppressing the decrease in vascular brightness caused by turbulence. Regions where turbulence is likely to occur are regions where the direction of blood vessels changes rapidly, and in gradient echo imaging, the signal tends to decrease in regions where blood flow is turbulent. By adding a flow compensation gradient pulse, the vascular signal can be made brighter, and the decrease in brightness can be suppressed. In other words, it becomes an imaging condition that suppresses the decrease in blood signal.

[0051] Figure 5(b) shows an example of a pulse sequence in which flow compensation gradient pulses are added to RF-spoiled GE.

[0052] In this sequence, a pair of phase-encoded gradient pulses with inverted signs are added to the phase-encoded gradient pulse 504 in Figure 5(a). As a result, for the phase-encoded gradient Gp, pulse 512 has the same area as the phase-encoded gradient pulse 504 but a different sign, and pulse 513 has twice the area of ​​the phase-encoded gradient pulse 504 but the same sign. For the slice gradient Gs, a rephase gradient pulse 510 has twice the area of ​​the rephase gradient pulse 503 in the slicing direction, and a gradient pulse 511 has the same area as the rephase gradient pulse 503 but with the sign inverted. For the readout gradient Gr, a gradient pulse 514 has the same area as the dephase gradient pulse 505 but a different sign, and a gradient pulse 515 has twice the area of ​​the dephase gradient pulse 505.

[0053] In the pulse sequence shown in Figure 5(b), a flow compensation pulse is added, so the TE of the first echo needs to be greater than the shortest TE mentioned above, but an image with a high vascular signal is obtained.

[0054] [Measurement of echo signal: S302] Once multiple imaging conditions (combinations of imaging parameters) including imaging conditions that suppress the decrease in blood signal have been determined as described above, the measurement unit 100 instructs the sequencer 104 to measure the echo signal according to the predetermined imaging sequence and the determined set of imaging parameters. The measured echo signal is then placed in k-space to obtain measurement data 414.

[0055] [Image reconstruction: S303] The image reconstruction unit 220 performs an inverse Fourier transform on the echo signals (measurement data 414) placed in k-space by the imaging unit and reconstructs the image. Part (or all) of this reconstructed image can be used as the source image 415 for the vascular image synthesis described later.

[0056] [Target parameter estimation: S3051] The parameter estimation unit 232 uses the six reconstructed images generated by the image reconstruction unit 220 and the signal function generated by the signal function generation unit 231 to estimate a (=ρSc), which is the product of the subject parameters T1, T2, B1, and ρ and the device parameter Sc. Specifically, the subject parameters and device parameter 416 are estimated by fitting the signal value I for each pixel of the reconstructed image to the function f of equation (3), which is a modified version of equation (1).

[0057]

number

[0058] Function fitting is performed using the least squares method, as shown in equation (4).

number

[0059] Based on the above, an image (quantitative value image) is obtained in which the parameter values ​​obtained from parameter estimation are used as pixel values.

[0060] [Vascular image synthesis: S3052] The composite image generation unit 250 synthesizes the image reconstructed by the image reconstruction unit 220 and the quantitative value image to generate a vascular image. While it is possible to synthesize the vascular image using simple addition or weight addition on a pixel-by-pixel basis, in this embodiment, a method (described in Patent Document 2) is adopted in which the region is divided and each region is synthesized using a predetermined synthesis function. The procedure of this method will be explained with reference to Figure 8.

[0061] As shown in Figure 8, the synthesis of multiple images includes a step S810 to determine the source images, a step S820 to set the synthesis function and its parameters (synthesis parameters) to be used for synthesis, and a step S830 to synthesize the images.

[0062] [Original image determined by: S810] The image selection unit 251 selects two or more images to be used as source images for synthesis from among multiple (in this case, six sets) reconstructed images generated by the image reconstruction unit 220, composite images such as T1-weighted images and diffusion-weighted images generated from the reconstructed images, and quantitative value images for each parameter generated by the quantitative value image generation unit 230, such as T1 images, T2 images, and PD (proton density ρ) images. In this case, the source images should include at least one quantitative value image. Preferably, the two or more source images are images with different emphasized physical properties (quantitative values), in which case only the quantitative value image may be used as the source image. Furthermore, as the reconstructed image, an image acquired under conditions where the blood flow signal is high intensity (for example, a T1-weighted image or an image acquired with an imaging sequence that adds a flow compensation pulse) may be used. By selecting appropriate images and reducing the number of source images, the amount of data can be reduced and the processing time can be shortened.

[0063] Here, for example, as shown in Figure 9, we assume that one type of reconstructed image (e.g., PD-weighted image) 901 and one type of quantitative image (T1 image) 902 were determined as the original images.

[0064] [Determination of synthesis parameters: S820] The synthesis parameter setting unit 252 sets the parameters (synthesis parameters) of the synthesis function used when synthesizing the determined original images.

[0065] If the pixel values ​​of the source images 901 and 902 to be combined are I1 and I2, respectively, and the pixel value after combination is Ic, then the composite function can generally be written as Ic = f(I1, I2). Here, we will explain the case where a linear polynomial given by equation (5) is used as an example of a composite function.

number

[0066] The synthesis parameters are set for each region by dividing the image into multiple regions. For this reason, the synthesis parameter setting unit 252 first identifies the region of interest to be used for calculating the synthesis parameters and divides the region of interest into multiple regions (S821). Then, the synthesis parameters are set for each of the multiple divided regions. The division method is preferably along the direction in which there is a difference in how the brightness values ​​of blood vessels (specific tissues) change between the two images 901 and 902 to be synthesized, as shown in Figures 9(c) and (d). For example, in the case of blood vessels, it is preferable to divide along the main direction of the blood vessel's course (in a plane perpendicular to the direction of course).

[0067] Figure 10(a) shows how the region of interest is divided. In this example, the region of interest 910 within the imaging area set for the human head 900 is divided into N division regions 911-1 to 911-N parallel to the XY plane. The region of interest 910 is all or part of the imaging area and can be set by setting a predetermined location within the imaging area or by accepting user instructions using the user interface. If the imaging areas of multiple images are different, the region of interest will be the entire or a part of the imaging area common to all images. The direction and number of divisions may also be specified by the user.

[0068] In this process, based on standard data of pixel values ​​for blood vessels and non-vascular tissues, the synthesis parameters are set for each divided region such that the combined pixel values ​​of blood vessels are greater than the pixel values ​​of non-vascular tissues, and the representative value of the pixel values ​​for each divided region is a common value across all divided regions.

[0069] As standard data, for example, data from multiple types of images obtained in advance by the measurement unit 100 capturing a predetermined imaging area on a healthy volunteer under the same imaging conditions as those applied to the subject can be used, and the pixel value data for blood vessels and non-vascular areas (pre-image pixel value data) are used as standard data. Information (labels) indicating whether a pixel is a blood vessel or not is assigned to this pre-image pixel value data. Label information is set, for example, by manually setting it while viewing the acquired MRA image obtained by imaging the imaging area of ​​a healthy volunteer using conventional MRA methods such as TOF, or by automatically setting it using a conventional MRA image with binarization processing. This label information is added to the pre-image pixel value data.

[0070] The synthesis parameter setting unit 252 sets synthesis parameters (a, b, c) for each of the thus divided regions (step S822).

[0071] When plotting the range of possible values ​​for pixel value pairs I1 and I2 for pixels labeled as blood vessels and pixels labeled as non-blood vessels, using the standard data (prior pixel value data) described above, a graph like the one shown in Figure 10(b) is obtained. The range 921 for pixel value pairs of blood vessels and the range 922 for pixel value pairs of non-blood vessels are located in different places on the plot. Furthermore, the range 922 for pixel value pairs of non-blood vessels includes the distribution of pixels of multiple biological tissues such as gray matter and white matter, so the range takes on an elliptical shape.

[0072] The synthesis using the composite function in equation (1) above means projecting the pair I1 and I2 onto the axis indicated by Ic in Figure 10(b). Therefore, the synthesis parameter setting unit 252 determines the slope of the axis (Ic) such that the range of possible pixel values ​​Ic for blood vessels and non-vascular tissues does not overlap as much as possible when the composite function is applied to the prior pixel value data I1 and I2 (step S822-1). The slope of the axis is determined by the ratio of the synthesis parameters a and b.

[0073] Specifically, first, Fisher's linear discriminant analysis is used to determine the vector (a',b') that represents the ratio of a and b. (a,b) is determined to be proportional to (a',b').

[0074]

number

[0075]

number

[0076] Next, the signs of a and b are determined such that the average pixel value of blood vessels in the composite image (pixel value Ic) is greater than the average pixel value of non-vascular tissues (step S822-2). This makes the pixel values ​​of blood vessels larger than the pixel values ​​of non-vascular tissues in the composite function. Although the description above explains the case where the conditions for determining the signs of the composite parameters a and b are set so that the pixel values ​​of blood vessels are greater than the pixel values ​​of non-vascular tissues in each divided region, it is also possible to set them so that the pixel values ​​of blood vessels are smaller than the pixel values ​​of non-vascular tissues.

[0077] Next, a, b, and c are determined such that the above conditions (i.e., the ratio and sign conditions of a and b) are satisfied, and the mean and variance of the pixel values ​​Ic for non-vascular tissues are constant (step S822-3). The following describes the case where the mean is 0 and the variance is 1. First, when a' and b' are substituted as a and b as they are into equation (5), the variance Sco of the pixel values ​​Ic in non-vascular tissues is expressed by the following equation (8).

number

[0078] Therefore, the (a,b) that makes the variance of the pixel value Ic of non-vascular tissues equal to 1 can be determined by the following equation (9).

number

[0079] Furthermore, c is determined by the following equation (10) so that the average of the pixel values ​​Ic other than blood vessels becomes 0.

number

[0080] By setting the mean to 0 and the variance to 1 in this way, all non-vascular tissues are low in intensity (close to 0), and the range of intensity becomes approximately constant across all divided regions, thus reducing the number of non-vascular tissues that mistakenly appear high in intensity, similar to blood vessels.

[0081] In step S822-3, the mean and variance were kept constant across the entire divided region as conditions for determining a, b, and c. However, instead of the mean and variance, similar representative values ​​that indicate the range of pixel value Ic, such as the median and interquartile range, may be kept constant (hereinafter, these will be collectively referred to as representative values).

[0082] In step S822-3, the synthesis parameters were determined based on prior pixel value data to keep the representative values ​​constant. However, the synthesis parameters may also be determined based on the image data used for synthesis to keep the representative values ​​(e.g., mean and variance) constant. Since blood vessels are significantly fewer in number than other tissues, the mean and variance of non-vascular tissues will be approximately constant. Compared to determining the parameters using prior pixel value data, this method has the advantage of producing images that are not affected by individual differences such as the proportion of various tissues included in non-vascular tissues.

[0083] The composite parameter setting unit 252 repeats the above composite parameter setting process (S822-1 to S822-3) for divided regions 911-1 to 911-N (S823), and calculates the composite parameters for each divided region. As shown in Figures 9(c) and (d), the pixel values ​​I1 and I2 of blood vessels change significantly in the Z direction. In addition, the range of pixel values ​​I1 and I2 of tissues other than blood vessels also changes to some extent due to changes in the proportion of tissues included. Therefore, the composite parameters suitable for creating a composite image also differ for each divided region depending on the position in the Z direction. By finding composite parameters that make blood vessels much brighter than other tissues in each divided region and setting them in each divided region, the accuracy of the composite image can be improved.

[0084] [Step S830] The composite image generation unit 250 generates a composite image from the subject images 901 and 902 selected in S810 and the composite parameters set by the composite parameter setting unit 252 for each divided region. Specifically, for each pixel in the imaging region, the pixel values ​​I1 and I2 of the same pixel in multiple (in this case, two) images 901 and 902 selected from the image reconstructed by the image reconstruction unit 220 and the image generated by the quantitative value image generation unit 230, and the composite parameters a, b, and c set by the composite parameter setting unit 252 for the divided region 911-n in which that pixel is included, are substituted into equation (1) to obtain the pixel value of the composite image for that pixel.

[0085] This results in a composite image 930 that shows the entire blood vessel, as shown in Figure 10(c). In reality, the composite image 930 is a three-dimensional image, but for illustrative purposes, the projection is shown here using the Maximum Intensity Projection (MIP) method.

[0086] Figure 11 shows an example of a synthesized vascular image of a healthy person's head. For comparison, Figure 12 shows a vascular image taken using a set of imaging parameters that were explored without excluding imaging conditions with poor image quality. In Figure 12, some blood vessels (indicated by arrows) and blood vessels near the top of the head are missing, whereas in Figure 11, such missing vessels are suppressed.

[0087] According to this embodiment, the quantitative value image used when generating the composite image is composed of parameter values ​​estimated using multiple images, including a reconstructed image acquired under imaging conditions that suppress the decrease in blood flow signal. Therefore, the brightness of some blood vessels does not decrease in the composite vascular image, and a stable composite vascular image can be obtained. [Explanation of Symbols]

[0088] 10: MRI device, 100: Measurement unit, 101: Magnet for generating static magnetic field, 102: Gradient magnetic field coil, 103: Subject, 104: Sequencer, 105: Gradient magnetic field power supply, 106: High-frequency magnetic field generator, 107: Probe, 108: Receiver, 111: Display, 112: Storage medium, 200: Computer, 210: Measurement control unit, 220: Image reconstruction unit, 230: Quantitative value image generation unit, 240: Shooting condition setting unit, 250: Composite image generation unit.

Images