Magnetic particle imaging system, magnetic particle imaging method, and magnetic particle imaging program

Abstract

A processor (21) calculates a system function by a first deconvolution operation based on a set of first detection signals obtained when a calibration sample is disposed in an inspection region and a numerical model of a spatial distribution of magnetic particles included in the calibration sample. The processor (21) obtains a spatial distribution of the magnetic particles included in the test sample by a second deconvolution operation based on the system function and a group of second detection signals obtained when the test sample is disposed in the test region.

Description

Technical Field

[0001] The present disclosure relates to a magnetic particle imaging system, a magnetic particle imaging method and a magnetic particle imaging procedure. Background Art

[0002] A magnetic particle imaging system is known in the past. The magnetic particle imaging system comprises: a selection unit that generates a selection magnetic field having a low magnetic field region and a high magnetic field region in an inspection region; a driving unit that changes the relative positional relationship between the two regions in the inspection region with respect to the magnetic particles; an excitation unit that provides an excitation magnetic field in a manner that changes the magnetization of the magnetic particles; and a receiving unit that receives the magnetization change of the magnetic particles as a detection signal.

[0003] The value of the detection signal of the magnetic particle imaging system is a value obtained by convolving the spatial distribution of the magnetic particles with the system function. Therefore, in order to image the spatial distribution of the magnetic particles based on the detection signal, image reconstruction is performed with deconvolution of the system function.

[0004] The system function includes the magnetization curve of the magnetic particles and the information of the system (selection unit, drive unit, excitation unit, and receiving unit), so generally, each time a particle for inspection is selected and for each setting condition of the system, the data of the system function is obtained in advance before the inspection. As a method for obtaining the system function, a method based on actual measurement is known.

[0005] In the method based on actual measurement, a tiny point-shaped calibration sample is moved little by little in the inspection area, and the measured value is obtained as a system function. The magnetic particles contained in the calibration sample are very small, so it takes time to measure little by little to obtain a sufficient SN ratio. As a result, a lot of time is required to obtain the system function.

[0006] To solve this problem, Patent Document 1 uses a method based on numerical calculation. Patent Document 1 makes use of the spatial symmetry of the theoretical system function. For example, when the system function distribution is symmetrical relative to two axes, a point-shaped calibration sample is set in only one-fourth of the area and moved to obtain the signal intensity at each position. After that, the system function is created by replicating the measured signal in a manner that conforms to the symmetry of the system function.

[0007] Prior art literature

[0008] Patent Literature

[0009] Patent Document 1: Japanese Patent Application No. 2012-510847 Summary of the invention

[0010] However, the method described in Patent Document 1 cannot cope with the asymmetry of the system function caused by manufacturing errors. As a result, the quality of the reconstructed image deteriorates due to artifacts, reduced spatial resolution, reduced quantification, and the like.

[0011] Therefore, an object of the present disclosure is to provide a magnetic particle imaging system, a magnetic particle imaging method, and a magnetic particle imaging program that can obtain a system function in a shorter time than a method based on actual measurement and can generate a high-quality reconstructed image.

[0012] The magnetic particle imaging system for imaging the spatial distribution of magnetic particles in the inspection area disclosed in the present invention comprises: a selector for generating a selection magnetic field having a spatial pattern of magnetic field strength in such a manner as to form a first partial area with a low magnetic field strength and a second partial area with a higher magnetic field strength in the inspection area; an exciter for providing an excitation magnetic field in such a manner as to change the magnetization of magnetic particles present in the selection magnetic field; a receiver for receiving the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal; and a processor. The detection signal is represented by the convolution of the spatial distribution of the magnetic particles and the system function. The processor calculates the system function by performing a first deconvolution operation based on a data set of the first detection signal obtained when the calibration sample is arranged in the inspection area and a numerical model of the spatial distribution of the magnetic particles contained in the calibration sample, and obtains the spatial distribution of the magnetic particles contained in the inspection sample by performing a second deconvolution operation based on a data set of the second detection signal obtained when the inspection sample is arranged in the inspection area and the system function.

[0013] The magnetic particle imaging method for imaging the spatial distribution of magnetic particles in the inspection region disclosed in the present invention comprises: a step in which a selector generates a selection magnetic field having a spatial pattern of magnetic field strength in such a manner as to form a first partial region with a low magnetic field strength and a second partial region with a higher magnetic field strength in the inspection region; a step in which an exciter provides an excitation magnetic field in such a manner as to change the magnetization of magnetic particles present in the selection magnetic field; and a step in which a receiver receives the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal. The detection signal is represented by the convolution of the spatial distribution of the magnetic particles and the system function. The magnetic particle imaging method comprises: a step in which a processor calculates the system function by performing a first deconvolution operation based on a data set of a first detection signal obtained when the calibration sample is arranged in the inspection region and a numerical model of the spatial distribution of magnetic particles contained in the calibration sample; and a step in which the processor obtains the spatial distribution of magnetic particles contained in the inspection sample by performing a second deconvolution operation based on a data set of a second detection signal obtained when the inspection sample is arranged in the inspection region and the system function.

[0014] In the magnetic particle imaging program used in the magnetic particle imaging system for imaging the spatial distribution of magnetic particles in the inspection region of the present disclosure, the magnetic particle imaging system generates a selection magnetic field having a spatial pattern of magnetic field strength in such a manner that a first partial region having a low magnetic field strength and a second partial region having a higher magnetic field strength are formed in the inspection region, provides an excitation magnetic field in such a manner that the magnetization of magnetic particles existing in the selection magnetic field changes, and receives the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal. The magnetic particle imaging system includes a processor. The detection signal is represented by the convolution of the spatial distribution of magnetic particles and a system function. The magnetic particle imaging program causes the processor to execute: a step of calculating the system function by a first deconvolution operation based on a data set of a first detection signal obtained when the calibration sample is arranged in the inspection region and a numerical model of the spatial distribution of magnetic particles contained in the calibration sample; and a step of obtaining the spatial distribution of magnetic particles contained in the inspection sample by a second deconvolution operation based on a data set of a second detection signal obtained when the inspection sample is arranged in the inspection region and the system function.

[0015] According to the present disclosure, a system function can be acquired in a short time and a high-quality reconstructed image can be generated. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 : is a diagram showing an example of the overall structure of a magnetic particle imaging system.

[0017] Figure 2 1 is a diagram showing an example of the hardware configuration of the information processing device 11 .

[0018] Figure 3 : is a flowchart showing the process of the magnetic particle imaging method of the embodiment.

[0019] Figure 4 It is shown Figure 3 Flow chart of the calibration sample measurement process of step S104.

[0020] Figure 5 It is a diagram showing an example of arrangement of calibration samples.

[0021] Figure 6 It is shown Figure 4 Flowchart of the process of the subroutine of the system function generation (first deconvolution operation) of step S206.

[0022] Figure 7 It is shown Figure 3 This is a flowchart of the process of the diagnostic measurement (test sample measurement) of step S107.

[0023] Figure 8 It is shown Figure 7Flowchart of the process of spatial distribution imaging processing (second deconvolution operation) of step S406.

[0024] Fig. 9 This is a flowchart showing the process of measuring the calibration sample in Patent Document 1.

[0025] Fig.10 This is a diagram showing an example of arrangement of calibration samples in Patent Document 1.

[0026] Fig.11 : is a flowchart showing the process of the system function generation subroutine of Patent Document 1. DETAILED DESCRIPTION

[0027] Hereinafter, embodiments will be described with reference to the drawings.

[0028] Implementation method 1.

[0029] Figure 1 : is a diagram showing an example of the overall structure of a magnetic particle imaging system.

[0030] The magnetic particle imaging system includes an exciter 2, a first selector 3a, a second selector 3b, a receiver 4, an excitation magnetic field application power supply 5, a first selection magnetic field power supply 7, a second selection magnetic field power supply 8, a filter 9, a signal amplifier 10, and an information processing device 11.

[0031] The first selector 3a and the second selector 3b generate a selective magnetic field having a spatial pattern of magnetic field strength so as to form a first partial region with a low magnetic field strength and a second partial region with a higher magnetic field strength in an inspection region where the subject is placed.

[0032] Thus, the region of magnetic particles that can contribute to the measurement signal among the magnetic particles present in the inspection region is limited to the vicinity of the first partial region. The region in the first partial region, in particular, where the magnetic field strength takes a value close to zero, is called a zero magnetic field region (Field Free Region (FFR)). For example, depending on the shape, the zero magnetic field region FFR is also called a point zero magnetic field region (Field Free Point (FFP)), a linear zero magnetic field region (Field Free Line (FFL)), or a planar zero magnetic field region. Regarding FFL, the zero magnetic field region can extend in a certain direction, for example, it can also be a rectangular shape (in this case, the length direction of the long side is the extension direction) or an elliptical shape. For example, the planar zero magnetic field region and FFL have a larger area that can contribute to the signal than FFP, so they have the advantage of obtaining a sufficient SN ratio for measurement and image reconstruction in a short time.

[0033] The first selector 3a includes a first electromagnet. The second selector 3b includes a second electromagnet. The first electromagnet and the second electromagnet are arranged opposite to each other in such a manner that a reverse magnetic field is generated in the inspection area. The first electromagnet is connected to a first selective magnetic field power supply 7. The second electromagnet is connected to a second selective magnetic field power supply 8. A current flows from the first selective magnetic field power supply 7 to the first electromagnet, and a current flows from the second selective magnetic field power supply 8 to the second electromagnet, thereby generating a selective magnetic field. By changing the magnitude of the current from the first selective magnetic field power supply 7 and the second selective magnetic field power supply 8, the relative positional relationship of the selective magnetic field (more specifically, the first partial area and the second partial area) with respect to the magnetic particles can be changed (electrical movement).

[0034] For example, by changing the current balance of the first electromagnet and the second electromagnet, the position of the zero magnetic field region FFR in the inspection area can be moved in the parallel direction or the rotational direction. In this way, the first partial area (especially the zero magnetic field region FFR) in the scanning area is driven to scan. The method of selecting the generation of the magnetic field is not limited to the electromagnet. Instead of the first electromagnet and the second electromagnet, two permanent magnets arranged opposite to each other, or a combination of a permanent magnet and an electromagnet can also be used. In addition, the scanning method of the magnetic field (zero magnetic field region FFR in this example) is not limited to the above method. For example, it can also be set to scan and drive the magnetic field by physical movement of the first selector 3a and the second selector 3b, or a combination of electrical movement and physical movement. Alternatively, the magnetic field can be scanned relatively with respect to the subject by fixing the position of the magnetic field and moving the subject.

[0035] The actuator 2 applies an excitation magnetic field in such a way that the magnetization of magnetic particles present in the selected magnetic field of the inspection area of ​​the inspection object such as the calibration sample or the inspection sample is changed. The calibration sample is a sample used to determine the system function and the magnetic particle distribution is known. The inspection sample is a sample whose magnetic particle distribution is to be inspected.

[0036] For example, the exciter 2 includes a coil connected to an excitation magnetic field applying power source 5. When an alternating current flows from the excitation magnetic field applying power source 5 to the coil, an alternating magnetic field is applied as an excitation magnetic field to an inspection region where the subject is placed.

[0037] By applying an excitation magnetic field to the subject, the magnetic particles contained in the subject generate a magnetic signal of a fundamental wave with the same frequency as the excitation magnetic field and a magnetic signal of a higher harmonic than the fundamental wave. The magnetic particles are modified with substances such as proteins that are coupled to target substances contained in the subject through antigen-antibody reactions.

[0038] The receiver 4 receives the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal. The receiver 4 includes, for example, a coil. The receiver 4 can be an element capable of detecting a changing magnetic field, such as a Hall element, a magnetoresistive element (AMR (Anisotropic Magneto Resistive) element, SMR (Semiconductor Magneto Resistive) element, TMR (Tunnel Magneto Resistive) element, etc.), an MI (Magneto Impedance) sensor, etc. The detection signal is input to the information processing device 11, for example, via a filter 9 for noise removal and a signal amplifier 10.

[0039] Figure 2 1 is a diagram showing an example of the hardware configuration of the information processing device 11 .

[0040] The information processing device 11 includes a processor 21 , a RAM (Random Access Memory) 25 , a reader 26 , an internal storage 27 , a display 22 , an operation unit 23 , and a communication interface 24 .

[0041] The processor 21 is, for example, a CPU (Central Processing Unit), and performs calculation processing.

[0042] The RAM 25 stores temporary information generated in conjunction with the calculation processing of the processor 21 .

[0043] The processor 21 reads out the programs (including the system function generation program and the spatial distribution imaging program) stored in the internal storage unit 27 , expands the programs in the RAM 25 , and executes them.

[0044] The reading unit 26 reads information recorded on an optical storage medium such as a CD-ROM (Compact Disk Read Only Memory), for example.

[0045] The internal storage unit 27 is, for example, a hard disk drive and stores various programs such as a system function generation program and a spatial distribution imaging program, and various data such as a numerical model of a calibration sample.

[0046] The display unit 22 is, for example, a liquid crystal display, and displays a screen generated by the calculation processing of the processor 21 .

[0047] The operation unit 23 includes, for example, a keyboard, a mouse, etc. The operation unit 23 receives operation input by an operator.

[0048] The communication interface 24 communicates with an external device (eg, a server device) via a network.

[0049] The system function generation program includes a command group for processing to generate a system function based on the measurement signal of the calibration sample. The spatial distribution imaging program includes a command group for processing related to imaging of the spatial distribution of magnetic particles present in the patient's biological body and other inspection samples. These programs are recorded on an optical recording medium, read by the readout unit 26 and stored in the internal storage unit 27. Alternatively, these programs can also be downloaded from the server device using the communication interface 24 and stored in the internal storage unit 27.

[0050] The processor 21 stores the detection signal from the receiver 4. When the detection signal is measured as a voltage, the zero magnetic field region exists at the parallel position r. i and the angle θ j The kth harmonic component U of the detection signal k (r i ,θ j ) Use system function S k (p) and the magnetic particle distribution c(p). p is a vector representing a three-dimensional position (x, y, z). The convolution operation is represented by the following equation, for example. System function S k (p) is determined by the effect of the magnetization curve of the magnetic particles and the characteristics of the device used to calibrate the magnetic particle imaging system. 3 p represents dx×dy×dz.

[0051] [Formula 1]

[0052] U k (r i ,θ j )=∫S k (p)·c(p)d 3 p …(1)

[0053] System function S k (p) is the kth harmonic component U of the detection signal, with the input being the magnetic particle distribution c(p) and the output being the detection signal k (r i ,θ j ) is the transfer function when .

[0054] The kth harmonic component U of the detection signal will be used k (r i ,θ j ) and magnetic particle distribution c(p) to calculate the system function S k (p) is called the first deconvolution operation. Through the first deconvolution operation, even if the system function S is unknown, kThe characteristics of (p) can also be determined based on U k (r i ,θ j ) and magnetic particle distribution c(p) to obtain S k (p).

[0055] An example of the first deconvolution operation is: k (r i ,θ j ) divided by the Fourier transform value of c(p) and perform inverse Fourier transform to obtain S k (p) method.

[0056] In the present embodiment, the processor 21 uses the data group of the first detection signal, which is the detection signal collected as the relative position of the calibration sample and the selection magnetic field changes, to perform a deconvolution operation. Hereinafter, the deconvolution operation performed using the data group of the first detection signal is referred to as the first deconvolution operation. Here, the "data group of the first detection signal" refers to a collection of data of the first detection signal collected at multiple positions. Therefore, each data of each element of the data group of the first detection signal corresponds to the collection position of the first detection signal. For example, whenever the position is changed, the data of the first detection signal corresponding to the position is collected. The collection of the data of the first detection signal of each position collected becomes the data group of the first detection signal. The first deconvolution operation can be performed using the data group of the first detection signal corresponding to all positions at the end of the scan, or can be performed successively using the data group of the first detection signal corresponding to each position before and after the change whenever the scanning position changes.

[0057] At this time, as a specific process, the processor 21 may also update the expected system function in such a way that the data set of the first expected detection signal obtained by the convolution operation of the magnetic particle distribution of the calibration sample and the expected system function and the data set of the first detection signal are as close as possible to each other, thereby performing the first deconvolution operation. As an example, the processor 21 may also update the expected system function in such a way that the sum of squares of the errors between each element of the data set of the first expected detection signal and each element of the data set of the first detection signal is reduced.

[0058] The kth harmonic component U of the detection signal will be used k (r i ,θ j ) and system function S k (p) to calculate the magnetic particle distribution c(p) is called the second deconvolution operation.

[0059] An example of the second deconvolution operation is: k (r i ,θj ) divided by S k A method of obtaining c(p) by inverse Fourier transforming the value obtained by Fourier transforming the value of (p).

[0060] In the present embodiment, the processor 21 uses the data group of the second detection signal as the detection signal collected with the relative position change of the inspection sample and the selected magnetic field to perform the deconvolution operation. Hereinafter, the deconvolution operation performed using the data group of the second detection signal is referred to as the second deconvolution operation. Here, the "data group of the second detection signal" refers to a collection of data of multiple collected second detection signals. Therefore, each data of each element of the data group of the second detection signal corresponds to the collection position of the second detection signal. For example, whenever the position is changed, the data of the second detection signal corresponding to the position is collected. The collection of the data of the second detection signal of each position collected becomes the data group of the second detection signal. The second deconvolution operation can be performed using the data group of the second detection signal corresponding to all positions at the end of the scan, or can be performed successively using the data group of the second detection signal corresponding to each position before and after the change whenever the scanning position changes. At this time, the inspection sample can be stationary at one place in the inspection area, can be stationary at multiple places, and can also be moved relative to the selected magnetic field.

[0061] The processor 21 may also update the expected magnetic particle distribution in a manner that the data set of the second expected detection signal obtained by the convolution operation of the system function and the expected magnetic particle distribution and the data set of the second detection signal are as close as possible to each other, thereby performing the second deconvolution operation. As an example, the processor 21 may also update the expected magnetic particle distribution in a manner that the sum of the squares of the errors between the elements of the data set of the second detection signal and the elements of the data set of the second expected detection signal are reduced.

[0062] Figure 3 : is a flowchart showing the process of the magnetic particle imaging method of the embodiment.

[0063] In step S101, a test agent corresponding to a target substance to be imaged is selected.

[0064] In step S102, system conditions suitable for the subject and the inspection agent are set. The inspection agent is a superparamagnetic magnetic particle modified with a protein coupled to a target substance through an antigen-antibody reaction. Depending on the type of the inspection agent, the magnetization characteristics of the inspection agent are different. Furthermore, when the magnetization characteristics are different, the system function is also different. Therefore, it is necessary to use a system function corresponding to the selected inspection agent. The magnetization characteristics are not only affected by the characteristics of the magnetic particles themselves, such as the size of the core particle diameter of the magnetic particles and the core particle diameter distribution, but also by the changes in the particle diameter of the fluid mechanics caused by the differences in the types of antibody molecules modified on the magnetic particles and the surrounding environment such as the viscosity near the lesion.

[0065] The system conditions typically include: excitation conditions such as excitation intensity distribution and excitation frequency, selection conditions such as intensity distribution of the selected magnetic field, driving conditions for driving the selected magnetic field, reception conditions such as sensitivity distribution of the receiving coil, filter characteristics, and characteristics of the signal amplifier. The system conditions are system setting values ​​that contribute to the measurement of signal strength.

[0066] In order to ensure the reliability of the reconstructed image obtained by imaging the spatial distribution of magnetic particles, it is preferable to use a system function obtained under the same conditions as those for the diagnostic measurement (test sample measurement).

[0067] In step S103, it is determined whether a system function obtained under the same conditions as those of the diagnostic measurement (test sample measurement) is stored in the information processing device 11. If such a system function is not stored, the process proceeds to step S104.

[0068] In step S104, the calibration sample is measured. Thus, the calibration sample is measured before the diagnostic measurement (test sample).

[0069] In step S105, the acquisition period of the system function is determined. In the case where the system function is not a function acquired within a certain period, the process returns to step S104. The reason is that even if the required system function has been stored in the information processing device 11, since the state of the magnetic particle imaging system changes over time, it is best to implement regular inspections of the calibration sample at each certain period. In the case of performing diagnostic measurements using a system function that does not meet the type of inspection agent or system conditions, the reconstructed image obtained by spatial distribution imaging becomes an image containing artifacts, and the quantitative nature of the magnetic particles will be impaired. As a result, the diagnostic measurements are adversely affected.

[0070] In step S106, a system function is selected. Here, for example, the latest system function is selected.

[0071] After the system function is selected, in step S107, a diagnostic measurement (test sample measurement) is performed. Figure 3 In the magnetic particle imaging method shown, first, it is determined whether there is a system function that meets various conditions (such as the type of inspection agent, system conditions or period conditions). If there is no matching system function, a calibration sample measurement is performed before the inspection sample measurement to obtain a system function that meets the conditions.

[0072] Figure 4 It is shown Figure 3 Flow chart of the calibration sample measurement process of step S104.

[0073] Figure 4 The processing of steps S201 to S206 in the flowchart shown is realized by the processor 21 executing the program expanded in the RAM 25 .

[0074] In step S200, a calibration sample is placed in an inspection area.

[0075] Figure 5 : is a diagram showing an example of the configuration of the calibration sample. In this example, Figure 5 One grid in corresponds to one pixel of the reconstructed image of the spatial distribution of magnetic particles. The size of the calibration sample may be larger than the pixel size of the reconstructed image. Figure 5 The example shown is an example of using a linear zero magnetic field region (FFL) as the zero magnetic field region. The calibration sample is placed, for example, approximately in the center of the inspection region.

[0076] In step S201, the processor 21 generates an instruction for controlling the power supply to the first electromagnet and the second electromagnet, and outputs the generated instruction to the first selective magnetic field power supply 7 and the second selective magnetic field power supply 8. The first selective magnetic field power supply 7 and the second selective magnetic field power supply 8 start supplying power to the first electromagnet and the second electromagnet according to the instruction. As a result, a selective magnetic field is generated in the inspection area.

[0077] In step S202, the processor 21 generates a command for controlling the power supply to the exciter 2, and outputs the generated command to the excitation magnetic field applying power supply 5. The excitation magnetic field applying power supply 5 starts supplying power to the exciter 2 according to the command. As a result, an AC excitation magnetic field is applied to the subject.

[0078] In step S203, the processor 21 scans the selective magnetic field in the inspection area by adjusting the current balance from the first selective magnetic field power supply 7 and the second selective magnetic field power supply 8 to the first electromagnet and the second electromagnet. For example, the relative position of the selective magnetic field is changed for a calibration sample placed at one location in the inspection area. Furthermore, in the case of FFL, rotation scanning may also be performed.

[0079] In step S204 , the receiver 4 receives the change in magnetizing torque of the magnetic particles excited by the excitation magnetic field as a detection signal, and inputs the detection signal to the information processing device 11 via the filter 9 and the signal amplifier 10 for noise removal.

[0080] In step S205, the processor 21 determines whether the scanning of the selected magnetic field in the inspection area is completed according to the pre-set end condition. If the scanning is not completed, the processing returns to step S203. If the scanning is completed, the processing enters step S206. For example, as an example when the zero magnetic field region FFR is a linear zero magnetic field region (FFL), the scanning is completed by rotating the FFL at an angle scale specified in the range of 0 degrees to 180 degrees and moving the FFL to the position of the entire range in the inspection area at each angle. When the zero magnetic field region FFR is other shapes, the scanning end condition is different.

[0081] In step S206 , the processor 21 executes a process (first deconvolution operation) of generating a system function using the group of detection signals stored in step S204 .

[0082] In addition, the order of the selection magnetic field generation in step S201 and the excitation magnetic field generation in step S202 may be reversed. The order of the selection magnetic field scanning drive in step S203 and the signal detection in step S204 may also be reversed.

[0083] Figure 6 It is shown Figure 4 Flowchart of the process of the subroutine of the system function generation (first deconvolution operation) of step S206.

[0084] In step S301, the processor 21 performs the following operations according to the Figure 4 The correction measurement sinusoidal diagram is generated by the detection signal group stored in step S204 and the information of the scanning position of the zero magnetic field area. The correction measurement sinusoidal diagram is a k-order higher harmonic component U of the detection signal at the order k of the higher harmonic component, the translation position r and the angle θ. k The mapping of (r, θ).

[0085] U k (r i ,θ j ) Use system function S k (p) and the convolution of the magnetic particle distribution c(p) of the calibration sample.

[0086] In step S302, the processor 21 sets the expected system function S2 k (p). In the first time of step S302, the expected system function S2 k(p) is set to a predetermined initial value.

[0087] In step S303, the processor 21 executes the expected system function S202 set in step S202. k (p) and the magnetic particle distribution c(p) of the calibration sample to calculate the k-th harmonic component U2 of the expected detection signal k (r, θ).

[0088] The magnetic particle distribution c(p) of the calibration sample is represented by a numerical model representing the shape of the calibration sample and the concentration of the magnetic particles. When magnetic particles uniformly filled in a columnar space with a concentration ct are used as the calibration sample, the numerical model is represented as follows using a step function H(p). For example, when the calibration sample is a cylindrical shape (a circle with a diameter R in the YZ plane and a length L in the X direction), the numerical model is represented by the following formula.

[0089] c(r)=ct×[H((R / 2) 2 -(z-z0) 2 -(y-y0) 2 )×H(L / 2-(x-x0))]…(A1)

[0090] When the calibration sample is in the shape of a quadrangular prism (length Lx in the X direction, length Ly in the Y direction, and length Lz in the Z direction), the numerical model is represented by the following equation.

[0091] c(r)=ct×[H(Lx / 2-(x-x0))×H(Ly / 2-(y-y0))×H(Lz / 2-(z-z0))]…(A2)

[0092] In equations (A1) and (A2), x0, y0, and z0 are the center coordinates of the calibration sample.

[0093] The processor 21 generates a corrected predicted sinusoidal diagram. The corrected predicted sinusoidal diagram is a diagram showing the kth harmonic component U2 of the predicted detection signal at the order k of the harmonic component, the translation position r, and the angle θ. k The mapping of (r, θ).

[0094] In step S304 , the processor 21 calculates the sum of squares E1 of the errors between each element of the corrected measured sinusoidal graph and each element of the corrected expected sinusoidal graph.

[0095] For example, when the expected system function S2 is represented k When the tensor of (p) is denoted as Sass and the tensor of the numerical model representing the magnetic particle distribution of the calibration sample is denoted as Cmodel, the tensor Uass representing the calibration prediction sinusoidal graph is expressed by the following equation. Sass has k, r i,θ j , x, y, z dimensions. Cmodel has x, y, z dimensions. Uass has k, r i ,θ j Dimension.

[0096] Uass=Sass·Cmodel…(B1)

[0097] When the tensor representing the corrected measured sinogram is Ucal, E1 is expressed by the following equation. Ucal has k, r i ,θ j The following formula is the sum of squares of the errors of the elements of the two tensors.

[0098] E1=|Ucal-Uass| 2 …(B2)

[0099] In step S305, the processor 21 determines whether E1 is equal to or less than a predetermined convergence condition. If E1 is equal to or less than a predetermined reference value, the process returns to step S302. If E1 exceeds a predetermined reference value, the process proceeds to step S306.

[0100] In step S302, the processor 21 updates the expected system function S2 k (p). The processor updates the expected system function S2 by the gradient descent method as shown in the following formula: k (p). η is the acceleration factor that determines the update speed.

[0101]

[0102] In step S306, the processor 21 converts the expected system function S2 corresponding to the corrected expected sinusoidal diagram that meets the convergence condition into k (p) is determined as the system function S k (p).

[0103] In general, the system function has a smooth distribution, but in the assumed system function S2 k (p) is a non-smooth distribution such as one containing noise, resulting in a spatial distribution image obtained by imaging the spatial distribution, which also has noise. k (p) Smoothing in post-processing or in determining the system function S k It suffices to impose constraints on the convergence calculation of (p).

[0104] Figure 7 It is shown Figure 3 This is a flowchart of the process of the diagnostic measurement (test sample measurement) of step S107.

[0105] Figure 7 The processing of steps S401 to S406 in the flowchart shown is realized by the processor 21 executing the program expanded in the RAM 25 .

[0106] In step S400, an inspection sample is placed in an inspection area.

[0107] The processing of steps S401 to S405 is Figure 4 The processing of steps S201 to S205 is the same, so the description will not be repeated.

[0108] In step S406 , the processor 21 performs spatial distribution imaging processing (second deconvolution operation) using the group of detection signals stored in step S404 .

[0109] Figure 8 It is shown Figure 7 Flowchart of the process of spatial distribution imaging processing (second deconvolution operation) of step S406.

[0110] In step S501, the processor 21 performs the following operations according to the Figure 7 The inspection and measurement sinusoidal diagram is generated by combining the detection signal group stored in step S404 and the information of the scanning position of the zero magnetic field region. The inspection and measurement sinusoidal diagram is a diagram showing the kth harmonic component U of the detection signal at the order k of the harmonic component, the translation position r, and the angle θ. k The mapping of (r, θ).

[0111] U k (r i ,θ j ) Use system function S k (p) and the convolution representation of the magnetic particle distribution c(p) of the inspected sample.

[0112] In step S502, the processor 21 sets the expected magnetic particle distribution c2(p). In the first time of step S502, the expected magnetic particle distribution c2(p) is set to a predetermined initial value.

[0113] In step S503, the processor 21 calculates the expected magnetic particle distribution c2(p) and the system function S k (p) is convolution operation, and the k-order harmonic component U2 of the expected detection signal is calculated. k (r, θ).

[0114] The processor 21 generates an expected inspection sinusoidal diagram. The expected inspection sinusoidal diagram is a diagram showing the kth order harmonic component U2 of the expected detection signal at the order k of the harmonic component, the translation position r, and the angle θ. kThe mapping of (r, θ).

[0115] In step S504 , the processor 21 calculates the sum of squares E2 of the errors between each element of the measured inspection sinusoidal graph and each element of the expected inspection sinusoidal graph.

[0116] For example, in the representation of the system function S k When the tensor of (p) is S and the tensor representing the expected magnetic particle distribution is Cexp, the tensor Uexp representing the expected inspection sinusoidal graph is expressed by the following equation. S has k, r i ,θ j , x, y, z dimensions. Cexp has dimensions x, y, z. Uexp has dimensions k, r i ,θ j Dimension.

[0117] Uexp=S·Cexp…(C1)

[0118] When the tensor representing the inspection and measurement sinogram is Uins, E2 is expressed by the following equation. Uins has k, r i ,θ j The following formula is the sum of squares of the errors of the elements of the two tensors.

[0119] E2=|Uins-Uexp| 2 …(C2)

[0120] In step S505, the processor 21 determines whether E2 is less than a predetermined convergence condition. If E2 is less than a predetermined reference value, the process returns to step S502. If E2 exceeds a predetermined reference value, the process proceeds to step S506.

[0121] In step S502, the processor 21 updates the expected magnetic particle distribution c2(p). The processor 21 updates the expected magnetic particle distribution c2(p) by the gradient descent method as shown in the following formula.

[0122]

[0123] In step S506, the processor 21 determines the expected magnetic particle distribution c2(p) satisfying the convergence condition as the magnetic particle distribution c(p). The image representing the magnetic particle distribution c(p) is spatial distribution imaging data, that is, a reconstructed image.

[0124] Next, Patent Document 1 and the differences between Patent Document 1 and the embodiment will be described.

[0125] Fig. 9 This is a flowchart showing the process of measuring the calibration sample in Patent Document 1.

[0126] Fig. 9 Flowchart and Figure 4 The difference between the flowcharts is Fig. 9 The flowchart of FIG. 1 includes steps S800 and S806 instead of steps S200 and S206 .

[0127] Fig.10 This is a diagram showing an example of arrangement of calibration samples in Patent Document 1.

[0128] In Patent Document 1, in step S801, a dot-shaped calibration sample is arranged at a spatial position corresponding to one pixel in the inspection region, and the calibration sample is sequentially moved in increments corresponding to one pixel.

[0129] Fig.11 : is a flowchart showing the process of the system function generation subroutine of Patent Document 1.

[0130] In step S901 , the symmetry of the system function is read out.

[0131] In step S902 , the system function is created by replicating the data of the measurement signal (detection signal) so as to conform to the symmetry of the system function.

[0132] In step S903, the system function data is output.

[0133] In the present embodiment, the distribution of the calibration sample is deconvolved, so unlike the method of patent document 1, it is not necessary to make the size of the calibration sample conform to the size corresponding to the pixel of the reconstructed image, so the size of the calibration sample can be made larger than the size of the pixel of the reconstructed image. Since the calibration sample is large, the measurement signal when measuring the calibration sample becomes larger, so even if the measurement time is short, a sufficient SN ratio can be obtained. As a result, the measurement of the calibration sample can be performed in a shorter time. Therefore, in the present embodiment, the regular inspection of the magnetic particle imaging system can be implemented at a higher frequency than before, so the image quality of the spatial distribution imaging is improved.

[0134] In addition, in the present embodiment, the measurement of the calibration sample can be performed through the same process as the diagnostic measurement (measurement of the inspection sample). That is, it is not necessary to move the calibration sample to the corresponding position of the pixel of the reconstructed image and perform a drive scan that changes the relative position relationship between the calibration sample and the selected magnetic field. For example, when the FFL is used as the zero magnetic field area, the measurement of the calibration sample can also be performed for a certain calibration sample in the inspection area by measuring the parallel scanning and rotation scanning accompanied by the FFL. In such a case, mechanical scanning of the calibration sample is not required, so the time for the calibration measurement is shortened, and a drive mechanism for performing mechanical scanning is not required.

[0135] The reconstructed image is generally represented by pixels divided into a grid, so in the numerical modeling of the calibration sample, the discretization error at the end of the calibration sample becomes smaller when a quadrangular prism shape is used compared to a cylindrical shape. Since the error of the numerical modeling is reduced, the error of the generated system function and the error of the spatial distribution image can be reduced.

[0136] The embodiments disclosed herein should be considered to be illustrative and non-restrictive in all respects. The scope of the present disclosure is not indicated by the above description but by the claims, and is intended to include all changes within the meaning and scope equivalent to the claims. It is obvious to those skilled in the art that the contents described in the embodiments other than the matters described in claim 1 of the claims are not essential.

[0137] (Explanation of symbols)

[0138] 2: exciter; 3a: first selector; 3b: second selector; 4: receiver; 5: power supply for applying excitation magnetic field; 7: power supply for generating and driving first selected magnetic field; 8: power supply for generating and driving second selected magnetic field; 9: filter; 10: signal amplifier; 11: information processing device; 21: processor; 22: display unit; 23: operation unit; 24: communication interface; 25: RAM; 26: read unit; 27: internal storage unit.

Claims

1. A magnetic particle imaging system for imaging the spatial distribution of magnetic particles in an inspection area, the magnetic particle imaging system comprising: a selector for generating a selection magnetic field having a spatial pattern of magnetic field strength in such a manner as to form a first partial region having a low magnetic field strength and a second partial region having a higher magnetic field strength in the inspection region; an actuator providing an excitation magnetic field in such a manner that the magnetization of the magnetic particles present in the selection magnetic field changes; a receiver, receiving the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal; as well as processor, The detection signal is represented by the convolution of the spatial distribution of the magnetic particles and the system function, The processor calculates the system function by performing a first deconvolution operation based on a data group of a first detection signal obtained when the calibration sample is configured in the inspection area and a numerical model of the spatial distribution of magnetic particles contained in the calibration sample, and obtains the spatial distribution of magnetic particles contained in the inspection sample by performing a second deconvolution operation based on a data group of a second detection signal obtained when the inspection sample is configured in the inspection area and the system function.

2. The magnetic particle imaging system according to claim 1, wherein: The numerical model represents the shape of the calibration sample and the concentration of the magnetic particles contained in the calibration sample.

3. The magnetic particle imaging system according to claim 1 or 2, wherein: The size of the calibration sample is larger than the pixel size of the obtained image representing the spatial distribution of the magnetic particles.

4. The magnetic particle imaging system according to any one of claims 1 to 3, wherein: The calibration sample is formed by filling a columnar space with magnetic particles. The processor generates a numerical model of the calibration specimen using a step function.

5. The magnetic particle imaging system according to any one of claims 1 to 3, wherein: The calibration sample is formed by filling the space of a quadrangular prism with magnetic particles. The processor generates a numerical model of the calibration specimen using a step function.

6. The magnetic particle imaging system according to any one of claims 1 to 5, wherein: The processor performs the first deconvolution operation on the calibration sample placed still at one location in the inspection region using a data group of first detection signals collected along with a relative position change including rotational scanning of the selective magnetic field.

7. The magnetic particle imaging system according to any one of claims 1 to 6, wherein: The processor updates the expected system function in such a manner that the sum of squares of errors between each element of the data group of the first detection signal and each element of the data group of the first expected detection signal obtained by the convolution operation of the expected system function and the numerical model of the calibration sample is reduced, thereby performing the first deconvolution operation.

8. The magnetic particle imaging system according to any one of claims 1 to 7, wherein: The processor updates the expected magnetic particle distribution in such a way that the sum of squares of errors between each element of the data group of the second detection signal and each element of the data group of the second expected detection signal obtained by convolution operation of the system function calculated by utilizing the first deconvolution operation and the expected magnetic particle distribution is reduced, thereby performing the second deconvolution operation.

9. A magnetic particle imaging method for imaging the spatial distribution of magnetic particles in an inspection area, the magnetic particle imaging method comprising: A step of generating a selection magnetic field having a spatial pattern of magnetic field strength by a selector so as to form a first partial region having a low magnetic field strength and a second partial region having a higher magnetic field strength in the inspection region; a step of providing an excitation magnetic field by a magnetic exciter in such a way that the magnetization of said magnetic particles present in said selected magnetic field changes; as well as a step of receiving a magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal by a receiver, The detection signal is represented by the convolution of the spatial distribution of the magnetic particles and the system function, The magnetic particle imaging method comprises: The processor calculates the system function by performing a first deconvolution operation based on a data set of a first detection signal obtained when the calibration sample is arranged in the inspection area and a numerical model of the spatial distribution of magnetic particles contained in the calibration sample; as well as The processor obtains the spatial distribution of magnetic particles included in the test sample by performing a second deconvolution operation based on a data group of a second detection signal obtained when the test sample is arranged in the test area and the system function.

10. A magnetic particle imaging procedure for use in a magnetic particle imaging system for imaging the spatial distribution of magnetic particles in an examination region, wherein: The magnetic particle imaging system generates a selection magnetic field having a spatial pattern of magnetic field strength in such a manner as to form a first partial region having a low magnetic field strength and a second partial region having a higher magnetic field strength in the inspection region, provides an excitation magnetic field in such a manner as to cause a magnetization change in the magnetic particles existing in the selection magnetic field, receives the magnetization change of the magnetic particles excited by the excitation magnetic field as a detection signal, and has a processor, wherein the detection signal is represented by a convolution of the spatial distribution of the magnetic particles and a system function. The magnetic particle imaging program causes the processor to execute: A step of calculating the system function by a first deconvolution operation based on a data set of a first detection signal obtained when the calibration sample is arranged in the inspection area and a numerical model of the spatial distribution of magnetic particles contained in the calibration sample; and A step of obtaining a spatial distribution of magnetic particles included in the test sample by performing a second deconvolution operation based on a data group of a second detection signal obtained when the test sample is arranged in the test area and the system function.

Images