JPWO2020059135A1 - Data processing methods, programs, data processing equipment and positron emission tomographic imaging equipment - Google Patents

Abstract

The data processing method includes a step (S2) of converting the TOF measurement data into non-TOF measurement data that does not include TOF information, and at least random correction processing, scattering correction processing, and reconstruction processing for the non-TOF measurement data. By performing this, a step (S3) of generating a non-TOF reconstructed image, a step of generating TOF projection data including TOF information by forward-projecting the non-TOF reconstructed image (S4), and TOF measurement data A step (S5) of estimating a bias component included in the TOF projection data based on the TOF projection data is provided.

Description

The present invention relates to a data processing method, program, data processing apparatus and positron emission tomography for TOF measurement data including TOF (Time of Flight) information obtained by positron emission tomography (PET). Regarding the device.

A positron emission tomography device (hereinafter referred to as "PET device") is known as a nuclear medicine diagnostic device that obtains medical data of a subject based on the radiation emitted from the subject to which a radiopharmaceutical is administered. ..

The PET device detects two γ-rays generated by the disappearance of positrons (Positron) with a plurality of detectors. Specifically, a radiopharmaceutical (radiotracer) containing a positron emitting nuclide is administered to a subject, and annihilation γ-rays emitted from the administered subject are detected by a large number of radiation detectors. When γ-rays are detected by two radiation detectors within a certain period of time, the PET device counts them as a pair of annihilation γ-rays and collects measurement data indicating the two detected radiation detection positions and the number of counts. Generate. The TOF type PET apparatus generates measurement data including TOF information indicating the two radiation detection positions where the pair annihilation γ-ray pair is detected and the count number, and the detection time difference between the pair annihilation γ-ray pair. Here, the measurement data that does not include the TOF information is referred to as "non-TOF measurement data", and the measurement data that includes the TOF information is referred to as "TOF measurement data". The TOF information is information for identifying the generation point of the pair annihilation γ-ray. Therefore, it is possible to more accurately identify the radiation concentration at each point in the subject by using the TOF measurement data than by using the non-TOF measurement data.

In order to quantitatively identify the radiation concentration in the subject from the TOF measurement data, various correction processes are performed. Typical correction processing includes sensitivity correction processing, attenuation correction processing, scattering correction processing, random correction processing, attenuation correction processing, and dead time correction processing. Among these correction processes, the random correction process is a process of removing a component of random coincidence counting (hereinafter, referred to as “random component”) included in the TOF measurement data from the TOF measurement data. Similarly, the scattering correction process is a process of removing a component of simultaneous scattering counting (hereinafter, referred to as “scattering component”) included in the TOF measurement data from the TOF measurement data. In the present specification, a component in which a random component and a scattering component are combined is referred to as a "bias component".

As a method of random correction processing for TOF measurement data, the same method as random correction processing for non-TOF measurement data can be used. This is because random coincidence counting does not depend on TOF information. On the other hand, the scattering coincidence counting depends on the TOF information. Therefore, as a method of scattering correction processing for TOF measurement data, a method of estimating a scattering component according to TOF information and correcting using the estimated scattering component has been developed. The scattering correction processing method includes a method of differentiating the estimated scattering component from the TOF measurement data and converting it into data in which the influence of the scattering coincidence counting is excluded, and a calculation formula of the estimated scattering component for image reconstruction. A method of obtaining a reconstructed image in which the influence of the scattering coincidence counting is eliminated by incorporating the method is included.

For example, U.S. Pat. No. 7,129,496 (Patent Document 1) estimates a virtual scattering synogram without measurement error of TOF information, and convolves the scattering synogram with the timing response of a PET device to obtain a scattering component. The method of estimation is disclosed. Further, US Pat. No. 7,397,035 (Patent Document 2) is based on a calculation formula obtained by extending the non-TOF-SSS (Single Scatter Simulation) algorithm for estimating the scattering component of non-TOF measurement data. Therefore, a method of estimating the scattering component of the TOF measurement data is disclosed.

U.S. Pat. No. 8,265,365 (Patent Document 3) discloses a method of indirectly estimating a scattering component based on measured coincidence counting data. Specifically, the true TOF is achieved by converting the TOF projection data into non-TOF projection data that does not include TOF information, and forward-projecting the corrected reconstructed image obtained based on the non-TOF projection data. Obtain distribution data. Then, the scattering component is estimated by subtracting the measured random simultaneous counting data and the true TOF distribution data from the prompt simultaneous counting data.

U.S. Pat. No. 7,129,496 U.S. Pat. No. 7,397,035 U.S. Pat. No. 8,265,365

In the method described in Patent Documents 1 and 2 above, the scattering component is calculated directly from the TOF measurement data. The TOF measurement data has a dimension of TOF information added to the non-TOF measurement data. Therefore, when the scattering component is calculated directly from the TOF measurement data, the calculation time becomes long.

In the method described in Patent Document 3 above, since the random simultaneous counting data and the true TOF distribution data are subtracted from the prompt simultaneous counting data, the statistical error is greatly propagated by the subtraction, and the accuracy of the obtained scattering component is obtained. Will be low.

The present invention has been made to solve such a problem, and an object of the present invention is a data processing method, program, or data capable of easily and accurately estimating a bias component contained in TOF measurement data. It is to provide a processing apparatus and a positron emission tomography apparatus.

The present invention is a data processing method for TOF measurement data including TOF information obtained by positron emission tomographic imaging. The data processing method includes a step of converting the TOF measurement data into non-TOF measurement data that does not include TOF information, and at least performing random correction processing, scattering correction processing, and reconstruction processing on the non-TOF measurement data. Based on the step of generating a non-TOF reconstructed image, the step of generating TOF projection data including TOF information by forward-projecting the non-TOF reconstructed image, and the TOF measurement data and the TOF projection data. , The step of estimating the bias component included in the TOF projection data is provided.

Preferably, the reconstruction process is a process according to an analytical reconstruction method. Alternatively, the reconstruction process is a process according to the successive approximation reconstruction method.

ここで、ｉは同時計数線のインデックスであり、ｔはＴＯＦ情報のビンのインデックスであり、ｊは非ＴＯＦ再構成画像におけるボクセルのインデックスであり、ｐは前記ＴＯＦ投影データであり、ＡＣＦは減弱補正係数であり、λは非ＴＯＦ再構成画像の画素値であり、Ｃは３次元システムマトリクスである。Preferably, the TOF projection data is calculated according to the following equation (1).

Here, i is the index of the coincidence counting line, t is the index of the bin of TOF information, j is the index of voxels in the non-TOF reconstructed image, p is the TOF projection data, and ACF is attenuated. It is a correction coefficient, λ is a pixel value of a non-TOF reconstructed image, and C is a three-dimensional system matrix.

ここで、ｉは同時計数線のインデックスであり、ｔはＴＯＦ情報のビンのインデックスであり、ｔ’はＴＯＦ情報のインデックスであり、ｊは非ＴＯＦ再構成画像におけるボクセルのインデックスであり、ｐは前記ＴＯＦ投影データであり、ＡＣＦは減弱補正係数であり、λは非ＴＯＦ再構成画像の画素値であり、Ｒは３次元回転行列であり、Ｇはガウシアンカーネルである。Alternatively, the TOF projection data is calculated according to the following equation (2).

Here, i is the index of the coincidence counting line, t is the index of the bin of TOF information, t'is the index of TOF information, j is the index of voxels in the non-TOF reconstructed image, and p is the index of the voxel. In the TOF projection data, ACF is an attenuation correction coefficient, λ is a pixel value of a non-TOF reconstructed image, R is a three-dimensional rotation matrix, and G is a voxel kernel.

According to the present invention, the bias component included in the TOF measurement data can be estimated easily and with high accuracy.

It is an overall schematic view of the PET apparatus which concerns on this embodiment. It is a flowchart which shows the flow of the data processing of the data processing apparatus provided in the PET apparatus shown in FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The same or corresponding parts in the drawings are designated by the same reference numerals, and the description thereof will not be repeated.

FIG. 1 is an overall schematic view of a PET apparatus (positron emission tomography imaging apparatus) according to the present embodiment. With reference to FIG. 1, the PET device 100 includes a gantry device 10, a control device 60, an operation unit 70, a data processing device 80, and a display device 90. Note that FIG. 1A is a front view of the gantry device 10, and FIG. 1B is a side view of the gantry device 10.

The gantry device 10 includes a top plate 20 for placing the subject 15, a moving device 22 for moving the top plate 20, a substantially cylindrical gantry 30 having an opening, and a detection arranged in the gantry 30. Includes vessel ring 40 and.

The subject 15 is placed on a cushion 24 provided on the top plate 20. The top plate 20 is provided so as to penetrate the openings of the gantry 30 and the detector ring 40 in the Z direction indicated by the arrow AR in the drawing, and can reciprocate along the Z direction. The moving device 22 is controlled by a drive signal from the control device 60, adjusts the height of the top plate 20, moves the top plate 20 along the Z direction, and is placed on the top plate 20. Specimen 15 is introduced into the opening of the gantry 30.

The detector ring 40 is configured by arranging a plurality of unit rings formed by arranging a plurality of γ-ray detectors 42 radially on a plane perpendicular to the Z direction in the Z direction. The γ-ray detector 42 includes a scintillator block composed of a plurality of array-shaped scintillators and a light receiving sensor. The scintillator converts γ-rays 52 emitted from a radiopharmaceutical (radiotracer) 50 (for example, fluorodeoxyglucose (FDG)) containing positron-emitting nuclides administered to subject 15 into fluorescence having a peak in the ultraviolet region. To do. The light receiving sensor multipliers the photoelectrons converted by the corresponding scintillator and converts them into an electrical signal. The γ-ray detector 42 outputs the generated electric signal to the control device 60.

The control device 60 includes a data collecting unit 61 and a driving unit 62. The data collection unit 61 and the drive unit 62 are composed of, for example, a microprocessor or an FPGA (Field Programmable Gate Array).

The drive unit 62 controls the drive of the mobile device 22 by outputting a drive signal that specifies the height of the top plate 20 and the coordinates (Z coordinates) in the Z direction to the mobile device 22. The drive unit 62 outputs the Z coordinate of the current top plate 20 to the data collection unit 61.

The data collection unit 61 generates TOF measurement data for each subject 15 based on the signal received from the γ-ray detector 42, and outputs the generated TOF measurement data to the data processing device 80. The TOF measurement data includes TOF information indicating the detection time difference between two γ-ray detectors 42 that simultaneously detect γ-rays.

Based on the signals received from each γ-ray detector 42, the data collection unit 61 has a window width in which both signal energies are within a certain energy window width and the γ-ray incident timing (detection time) is a certain time. Search for a combination of two γ-ray detectors 42 within. The data collection unit 61 generates TOF measurement data based on the combination of the two γ-ray detectors 42 extracted by the search and the detection time difference when the two γ-ray detectors 42 detect the γ-rays.

The format of the TOF measurement data is not particularly limited, and is, for example, a list mode format or a synogram format. The TOF measurement data in the list mode format includes LOR information that identifies a coincidence counting line (LOR: Line of Response), which is a line connecting the two γ-ray detectors 42 extracted by the above search, and the two γ-rays. This is data in which TOF information indicating the detection time difference by the detector 42 is arranged in time series. The distance between the γ-ray detector 42 located at one end point of the LOR and the radioactive agent 50 existing on the LOR is TA, and the distance between the γ-ray detector 42 located at the other end point of the LOR and the radioactive agent 50 is TB. , When the speed of light is c, the TOF information is indicated by (TA-TB) / c. When the length of LOR is d, the TOF information can take a value from −d / c to + d / c.

The TOF measurement data in the synogram format is data indicating the number of counts at each point in the five-dimensional space whose coordinates (variables) are the projected position, the projected angle, the slice position, the tilt angle, and the TOF information. When generating TOF measurement data in synogram format, LORs are grouped by their direction. Then, for each group, the LOR belonging to the group is projected on the projection plane orthogonal to the LOR. Here, the projection angle indicates the direction of LOR. The projection position indicates the position of the projection point on the projection surface. The tilt angle indicates the angle formed by the plane perpendicular to the Z direction and the LOR.

In the following, a case where the TOF measurement data is in the synogram format will be described as an example.

The operation unit 70 includes a pointing device such as a keyboard, a touch panel, and a mouse (none of which are shown). The operation unit 70 receives an instruction for operating the moving device 22 of the gantry device 10 and an instruction such as start / stop of imaging from the operator. The operation unit 70 outputs a signal according to the operation of the operator to the control device 60. The drive unit 62 of the control device 60 drives the mobile device 22 in response to a signal from the operation unit 70.

The data processing device 80 performs data processing for estimating a bias component that is a combination of a random component and a scattering component included in the TOF measurement data. The data processing device 80 is composed of a computer including a storage device such as a CPU (central processing unit), a GPU (Graphics Processing Unit), and a memory, for example.

FIG. 2 is a flowchart showing the flow of data processing of the data processing device 80. As shown in FIG. 2, first, in step S1, the data processing device 80 acquires TOF measurement data from the data collecting unit 61.

Next, in step S2, the data processing device 80 converts the TOF measurement data into non-TOF measurement data that does not include the TOF information. Specifically, the data processing device 80 generates non-TOF measurement data by integrating the TOF measurement data with the TOF information.

In step S3, the data processing device 80 performs a correction process and a reconstruction process on the non-TOF measurement data to obtain a distribution image of the radioactive drug 50 in the subject 15 (hereinafter, referred to as a non-TOF reconstruction image). Generate.

The correction process includes at least a random correction process and a scattering correction process. A known technique may be used as a correction process for non-TOF measurement data. As a known random correction process for non-TOF measurement data, there is a method of estimating random coincidence counting by a delayed coincidence counting method, an estimation method based on single counting, or the like, and subtracting the estimated random coincidence counting from the data to be processed. As a known scattering correction process for non-TOF measurement data, the scattering coincidence count is estimated by the non-TOF-SSS algorithm method, the energy window method, the convolution method, etc., and the estimated scattering coincidence count is subtracted from the data to be processed. There is a way. The correction process may include other correction processes (for example, sensitivity correction, attenuation correction process, attenuation correction process, dead time correction process, etc.).

A known technique may be used as the reconstruction process for the non-TOF measurement data. For example, an analytical reconstruction method or a successive approximation reconstruction method can be used. As an analytical reconstruction method, there is an FBP (Filtered Back Projection) method. As the successive approximation reconstruction method, there are MLEM (Maximum Likelihood Expectation Maximization) method, OSEM (Ordered Subset Expectation Maximization) method, and DRAMA (Dynamic Row-Action Maximum-likelihood Algorithm) method.

The data processing device 80 may perform correction processing and reconstruction processing only on the two-dimensional data having an inclination angle of 0 degrees among the non-TOF measurement data to generate a non-TOF reconstruction image. As a result, the calculation time required to generate the non-TOF reconstructed image can be shortened.

The data processing device 80 may also perform a known noise reduction process at the time of the reconstruction process for the non-TOF measurement data. As the noise reduction processing, image processing using a Gaussian filter or a Non-Local Means Filter is known. The noise reduction processing may be any of pre-processing applied to the non-TOF measurement data, processing in parallel with the reconstruction processing, and post-processing after the generation of the non-TOF reconstructed image.

Next, in step S4, the data processing device 80 generates TOF projection data including TOF information by forward-projecting the non-TOF reconstructed image. A known technique may be used as a method for forward-projecting a non-TOF reconstructed image. For example, the data processing apparatus 80 performs TOF forward projection calculation using the system matrix of the PET apparatus 100 according to the following equation (1).

In equation (1), i is the LOR index, t is the index of the bin of TOF information, j is the voxel index in the non-TOF reconstructed image, and p is the TOF projection data. ACF is an attenuation correction coefficient, λ is a pixel value (count number) of a non-TOF reconstructed image, and C is a three-dimensional system matrix of a PET device that associates image data with projected data.

Alternatively, the data processing apparatus 80 may perform TOF forward projection according to the following equation (2) without using the system matrix.

In equation (2), i is the LOR index, t is the index of the TOF information bin, t'is the TOF information index, and j is the voxel index in the non-TOF reconstructed image. Yes, p is the TOF projection data, ACF is the attenuation correction coefficient, and λ is the pixel value (count number) of the non-TOF reconstructed image. R _i, t'is a three-dimensional rotation matrix for rearranging voxels along the direction of the LOR of the target i and the direction orthogonal to the direction. G is a Gaussian kernel which is set according to the time resolution of the PET device 100 is convolved _{R i, t 'λ j.}

Next, in step S5, the data processing device 80 estimates the bias component included in the TOF measurement data based on the TOF measurement data and the TOF projection data. For example, the data processing device 80 estimates the bias component by subtracting the TOF projection data from the TOF measurement data. Alternatively, the data processing device 80 may estimate the bias component by subtracting the TOF projection data multiplied by the weighting coefficient from the TOF measurement data. At this time, the data processing device 80 may perform noise reduction processing on at least one of the TOF measurement data and the TOF projection data. Further, the data processing device 80 may perform noise reduction processing on the estimated bias component. The type and parameters of the noise reduction processing are not particularly limited, and a known method can be used. For example, image processing using a Gaussian filter or a Non-Local Means Filter can be used.

The data processing device 80 may appropriately display the image shown by each data obtained in steps S1 to S5 on the display device 90.

As described above, in the present embodiment, a data processing method for TOF measurement data including TOF information obtained by PET (positron emission tomography) is disclosed. This data processing method reconstructs the non-TOF measurement data with step S2 for converting the TOF measurement data into non-TOF measurement data that does not include TOF information, and correction processing and reconstruction including at least random correction processing and scattering correction processing for the non-TOF measurement data. Step S3 to generate a non-TOF reconstructed image by performing processing, step S4 to generate TOF projection data including TOF information by forward-projecting a non-TOF reconstructed image, and TOF measurement data and TOF. A step S5 for estimating a bias component included in the TOF measurement data based on the projection data is provided.

According to the above method, the scattering component is not directly calculated from the TOF measurement data including the TOF information as disclosed in Patent Documents 1 and 2, and is included in the TOF measurement data easily and in a short time. The bias component can be estimated.

In the technique disclosed in Patent Document 3, the scattering component is estimated by subtracting the measured random simultaneous counting data and the true TOF distribution data from the prompt simultaneous counting data. Here, the prompt simultaneous counting data is P, the measured random simultaneous counting data is D, the true TOF distribution data is T, the estimated scattering component is S, and the scattering component and the random component (random simultaneous counting) are combined. Let B be the bias component, and let δP, δD, δS, and δB be the statistical errors of P, D, S, and B, respectively. However, for the sake of simplicity, the statistical error of T is set to 0. Assuming that the statistical errors of each component are independent, since S = ^{PTD, δS = {(δP) 2} + ( ^{δD) 2} } ^1/2 , and δB = {(δS) ² + (ΔD) ² } ^1/2 = {(δP) ² + 2 (δD) ² } ^1/2 .

On the other hand, the statistical error of the bias component estimated by the present embodiment is as follows. That is, in the same manner as above, the TOF measurement data is P', the TOF projection data is T', the bias component is B', the statistical errors of P'and B'are δP'and δB', respectively, and the statistical error of T'. Is set to 0. Assuming that the statistical errors of each component are independent, B'= P'-T', so δB'= δP'. Since δP and δP'are the same, the statistical error of the bias component estimated by the present embodiment is smaller than the statistical error of the bias component estimated by Patent Document 3.

As described above, according to the present embodiment, the bias component included in the TOF measurement data can be estimated easily and with high accuracy.

The estimated bias component can be used for various data processing. For example, the data processing device 80 generates non-bias measurement data that does not include a bias component based on the TOF measurement data and the bias component, and performs reconstruction processing using the TOF information on the non-bias measurement data. To perform the step of generating the reconstructed image. Then, the data processing device 80 displays the generated reconstructed image on the display device 90. The data processing device 80 generates non-bias measurement data by, for example, subtracting a bias component from the TOF measurement data. At this time, the data processing device 80 may appropriately perform correction processing other than the random correction processing and the scattering correction processing.

^{The reconstructed image I conventional} generated by subtracting the bias component estimated by Patent Document 3 from the prompt coincidence counting data ^{is expressed as I conventional} = PSD using the above P, S, and D. Will be done. Therefore, the statistical error δI ^conventional of the reconstructed image is expressed by the following equation.
δI ^conventional = {(δP) ² + (δS) ² + (δD) ² } ^1/2
= {(ΔP) ² + (δP) ² + (δD) ² + (δD) ² } ^1/2
= 2 ^1/2 {(δP) ² + (δD) ² } ^1/2

^{On the other hand, the reconstructed image I proposal} generated by subtracting the bias component estimated by the present embodiment from the TOF measurement data uses the above ^{P'and B'and I proposal} = P'-. Expressed as B'. Therefore, the statistical error δI ^proposal of the reconstructed image is expressed by the following equation.
^{δI proposal = {(δP ')} 2 + (δB') 2} 1/2
= 2 ^1/2 (δP')

As described above, the statistical error of the reconstructed image generated by using the bias component estimated by the present embodiment is based on the statistical error of the reconstructed image generated by using the bias component estimated by Patent Document 3. Also becomes smaller. As a result, a highly accurate reconstructed image can be obtained.

In the above embodiment, the data processing device 80 is provided in the PET device 100. However, the data processing device 80 is an external device of the PET device 100, and may be connected to the control device 60 by wire or wirelessly to acquire TOF measurement data from the data collection unit 61.

In the above embodiment, a processing example of the data processing device 80 when the TOF measurement data is in the synogram format has been described. When the TOF measurement data is in the list data format, the data processing device 80 may generate non-TOF measurement data by deleting the TOF information from the TOF measurement data in step S2. Further, in step S5, the data processing apparatus 80 may estimate the bias component by, for example, converting the TOF measurement data into data in the synogram format and subtracting the TOF projection data from the converted TOF measurement data.

A program for causing the data processing device 80 to execute the above operations (steps S1 to S5 shown in FIG. 2) may be provided. Such a program is provided as a program product by recording it on a computer-readable recording medium such as a flexible disk attached to a computer, a CD-ROM (Compact Disk-Read Only Memory), a ROM, a RAM, and a memory card. You can also do it. Alternatively, the program can be provided by recording on a recording medium such as a hard disk built in the computer. The program can also be provided by downloading via the network.

The provided program product is installed and executed in a program storage unit such as a hard disk. The program product includes the program itself and a recording medium on which the program is recorded.

It should be considered that the embodiments disclosed this time are exemplary in all respects and not restrictive. The scope of the present invention is shown by the scope of claims rather than the above description, and it is intended that all modifications within the meaning and scope equivalent to the scope of claims are included.

10 gantry, 15 subject, 20 top plate, 22 moving device, 24 cushion, 30 gantry, 40 detector ring, 42 γ-ray detector, 50 radioactive drug, 52 γ-ray, 60 control device, 61 data collector, 62 Drive unit, 70 operation unit, 80 data processing device, 90 display device, 100 PET device.

Claims (8)

A data processing method for TOF measurement data including TOF information obtained by positron emission tomography.
The step of converting the TOF measurement data into non-TOF measurement data that does not include TOF information, and
A step of generating a non-TOF reconstructed image by performing at least random correction processing, scattering correction processing, and reconstruction processing on the non-TOF measurement data.
A step of generating TOF projection data including TOF information by forward-projecting the non-TOF reconstructed image, and
A data processing method comprising a step of estimating a bias component included in the TOF projection data based on the TOF measurement data and the TOF projection data. The data processing method according to claim 1, wherein the reconstruction process is a process according to an analytical reconstruction method. The data processing method according to claim 1, wherein the reconstruction process is a process according to the successive approximation reconstruction method. The TOF projection data is calculated according to the following equation (1).

i is the index of the coincidence counting line, t is the index of the bin of TOF information, j is the index of voxels in the non-TOF reconstructed image, p is the TOF projection data, and ACF is the attenuation correction coefficient. , Λ is the pixel value of the non-TOF reconstructed image, and C is the three-dimensional system matrix.
The data processing method according to any one of claims 1 to 3. The TOF projection data is calculated according to the following equation (2).

i is the index of the coincidence counting line, t is the index of the bin of TOF information, t'is the index of TOF information, j is the index of voxels in the non-TOF reconstructed image, and p is the index of the TOF. It is projection data, ACF is an attenuation correction coefficient, λ is a pixel value of the non-TOF reconstructed image, R is a three-dimensional rotation matrix, and G is a voxel kernel.
The data processing method according to any one of claims 1 to 3. A program that causes a computer to execute the data processing method according to any one of claims 1 to 5. A data processing apparatus that executes the data processing method according to any one of claims 1 to 5. A positron emission tomographic imaging apparatus including the data processing apparatus according to claim 7.

Images