List mode image reconstruction method and nuclear medicine diagnostic device

Abstract

The list-mode image reconstruction method includes the following steps: dividing list-mode data into multiple subsets; and obtaining a subset balance coefficient based on the number of events in the multiple subsets.

Description

Technical Field

[0001] The present invention relates to a list mode image reconstruction method and a nuclear medicine diagnostic apparatus. Background Art

[0002] Conventionally, a list mode image reconstruction method is known which reconstructs a radioactive distribution of a subject by iterative calculation based on list mode data collected by a nuclear medicine diagnostic apparatus. For example, such a list mode image reconstruction method is disclosed in Wang, W., et al., "Systematic and distributed time-of-flight list mode PET reconstruction." 2006 IEEE Nuclear Science Symposium Conference Record. Vol. 3. IEEE, 2006. (hereinafter simply referred to as "Non-Patent Document 1").

[0003] In the above Non-Patent Document 1, a list mode image reconstruction method is disclosed which reconstructs a radioactive distribution of a subject by iterative calculation based on list mode data collected by a PET apparatus (nuclear medicine diagnostic apparatus).

[0004] Prior Art Documents

[0005] Patent Documents

[0006] Non-Patent Document 1: Wang, W., et al., "Systematic and distributed time-of-flight list mode PET reconstruction." 2006 IEEE Nuclear Science Symposium Conference Record. Vol. 3. IEEE, 2006. Summary of the Invention

[0007] Problems to be Solved by the Invention

[0008] However, in the conventional list mode image reconstruction method as described in the above Non-Patent Document 1, when the number of events between subsets obtained by dividing list mode data is uneven, the calculated value of the iterative calculation sometimes does not converge to a value representing the radioactive concentration of the subject due to the uneven number of events between subsets. In this case, there is an adverse situation where a quantitative radioactive distribution image cannot be generated. In addition, when the number of subsets is changed, there is also an adverse situation where the pixel value changes depending on the subset. Therefore, in the conventional list mode image reconstruction method, there is a problem that it is difficult to generate a quantitative radioactive distribution image without depending on the number of subsets.

[0009] The present invention is completed to solve the above problems, and an object of the present invention is to provide a list mode image reconstruction method and a nuclear medicine diagnostic apparatus capable of generating a quantitative radioactive distribution image without depending on the number of subsets.

[0010] Solution to the problem

[0011] In order to achieve the above object, a list mode image reconstruction method according to a first aspect of the present invention is a list mode image reconstruction method for reconstructing a radioactive distribution of a subject by iterative calculation based on list mode data collected by a nuclear medicine diagnostic apparatus. The list mode image reconstruction method includes the following steps: dividing the list mode data into a plurality of subsets; obtaining a subset balance coefficient based on the number of events of the plurality of subsets; obtaining a backprojection value based on the list mode data; obtaining a backprojection image based on the backprojection value; multiplying the backprojection value or the backprojection image by the subset balance coefficient; and updating the radioactive distribution image based on the backprojection image. In addition, the list mode data refers to data obtained by storing radiation detection event information (detector number, detection time, radiation energy, etc.) in time series.

[0012] In addition, a nuclear medicine diagnostic apparatus according to a second aspect of the present invention includes: a detection unit that detects radiation generated from a radioactive agent in a subject; and an arithmetic unit that reconstructs the radioactive distribution of the subject by iterative calculation based on list mode data that is a detection result of the radiation by the detection unit. The arithmetic unit is configured to divide the list mode data into a plurality of subsets, obtain a subset balance coefficient based on the number of events of the plurality of subsets, obtain a backprojection value based on the list mode data, obtain a backprojection image based on the backprojection value, multiply the backprojection value or the backprojection image by the subset balance coefficient, and update the radioactive distribution image based on the backprojection image.

[0013] Effect of the invention

[0014] According to the present invention, as described above, list mode data is divided into a plurality of subsets, a subset balance coefficient is obtained based on the number of events in the plurality of subsets, a backprojection value is obtained based on the list mode data, a backprojection image is obtained based on the backprojection value, the backprojection value or the backprojection image is multiplied by the subset balance coefficient, and a radioactive distribution image is updated based on the backprojection image. Thus, by introducing the subset balance coefficient, it is possible to adjust the non-uniformity of the number of events between subsets, and therefore it is possible to suppress the state where the calculated values (pixel values) of the iterative calculation do not converge to the value representing the radioactive concentration of the subject due to the non-uniformity of the number of events between subsets. In other words, it is possible to make the calculated values of the iterative calculation converge to the value representing the radioactive concentration of the subject, and therefore it is possible to generate a quantitative radioactive distribution image. In addition, by introducing the subset balance coefficient, it is possible to adjust the non-uniformity of the number of events between subsets even when the number of subsets is changed, and therefore it is possible to suppress the change in pixel values depending on the number of subsets. That is, it is possible to obtain the same pixel values regardless of the number of subsets. As a result, it is possible to generate a quantitative radioactive distribution image regardless of the number of subsets. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a schematic diagram showing the structure of the PET apparatus according to the first embodiment.

[0016] Figure 2 is a schematic perspective view showing the structure of the radiation (gamma ray) detector according to the first embodiment.

[0017] Figure 3 is a flowchart showing the reconstruction process according to the first embodiment.

[0018] Figure 4 is a diagram for explaining the effect of the subset balance coefficient according to the first embodiment and is a schematic curve showing the change in the counting rate with respect to the measurement time.

[0019] Figure 5 is a diagram for explaining the effect of the subset balance coefficient according to the first embodiment and is a schematic diagram for explaining the concept of the subset balance coefficient.

[0020] Figure 6 is a diagram for explaining the effect of the subset balance coefficient according to the first embodiment and is a schematic diagram showing the state of the area of each subset before the subset balance coefficient correction.

[0021] Figure 7 is a diagram for explaining the effect of the subset balance coefficient according to the first embodiment and is a schematic diagram showing the state of the area of each subset after the subset balance coefficient correction.

[0022] Figure 8This is a diagram showing an example of an actual reconstructed image reconstructed without using a subset balance coefficient, an example of a ground truth image, and a difference image between the reconstructed image and the ground truth image.

[0023] Figure 9 This is a diagram showing an example of an actual reconstructed image reconstructed using a subset balance coefficient, an example of a ground truth image, and a difference image between the reconstructed image and the ground truth image.

[0024] Figure 10 This is a schematic diagram showing the structure of the PET device according to the second embodiment. Detailed Embodiment

[0025] Hereinafter, embodiments in which the present invention is embodied will be described based on the drawings.

[0026] (Structure of PET Device)

[0027] Refer to Figure 1 and Figure 2 to describe the structure of the PET (Positron Emission Tomography) device 1 according to the first embodiment.

[0028] As Figure 1 shown, the PET device 1 is a device that photographs the subject 100 by detecting radiation (gamma rays) generated from within the subject 100 due to a radioactive agent previously injected into the subject 100. The radiation (gamma rays) is annihilation radiation generated within the subject 100 due to the pair annihilation of a positron generated from the radioactive agent and an electron possessed by an atom in the vicinity of the positron. The PET device 1 is configured to generate a radioactive distribution image of the subject 100 based on the photographing result of the subject 100. In addition, the PET device 1 may be configured to be able to photograph the whole body of the subject 100, or may be configured to be able to photograph a part (such as the breast, head, etc.) of the subject 100. Further, the PET device 1 is an example of the "nuclear medicine diagnostic device" of the present disclosure.

[0029] The PET device 1 includes a detector ring 2 that surrounds the subject 100. The detector ring 2 is provided so as to be stacked in multiple layers in the body axis direction of the subject 100. A plurality of radiation (gamma ray) detectors 3 are provided inside the detector ring 2 (refer to Figure 2 ). Thus, the detector ring 2 is configured to detect radiation (gamma rays) generated from the radioactive agent within the subject 100. In addition, the detector ring 2 is an example of the "detection unit" of the present disclosure.

[0030] In addition, the PET device 1 includes a control unit 4. The control unit 4 includes a coincidence counting circuit 40 and an arithmetic circuit 41. Further, in Figure 1only shows two wirings from the radiation detector 3 (see Figure 2 ), but actually, the same number of wirings as the total number of channels of the photomultiplier tube (PMT: Photo Multiplier Tube) 33 (see Figure 2 ) described later of the radiation detector 3 are connected to the control unit 4 (coincidence counting circuit 40). In addition, the arithmetic circuit 41 is an example of the "arithmetic unit" of the present disclosure. Additionally, sensors other than PMT, such as SiPM (Silicon Photomultiplier), may sometimes be used.

[0031] As Figure 2 shown, the radiation detector 3 includes a scintillator block 31, a light guide 32, and a photomultiplier tube 33. Additionally, the light guide 32 may sometimes not be used.

[0032] The scintillator block 31 converts radiation (gamma rays) generated from the subject 100 (see Figure 1 ) injected with a radioactive agent into light. When a radioactive agent is injected into the subject 100, the positrons of the positron-emitting RI annihilate, thereby generating two radiations (gamma rays). Each scintillator element constituting the scintillator block 31 emits light as the radiation (gamma rays) is incident, thereby converting the radiation (gamma rays) into light.

[0033] The light guide 32 is optically coupled to the scintillator block 31 and the photomultiplier tube 33 respectively. The light emitted from the scintillator elements of the scintillator block 31 diffuses in the scintillator block 31 and is input to the photomultiplier tube 33 via the light guide 32.

[0034] The photomultiplier tube 33 multiplies the light input via the light guide 32 and converts it into an electrical signal. This electrical signal is sent to the coincidence counting circuit 40 (see Figure 1 ).

[0035] The coincidence counting circuit 40 (see Figure 1 ) generates detection signal data (count value) based on the electrical signal sent from the photomultiplier tube 33.

[0036] Specifically, the coincidence counting circuit 40 (see Figure 1)Check the position of the scintillator block 31 and the incidence timing of the radiation (gamma ray). Only when the radiation (gamma ray) simultaneously enters two scintillator blocks 31 located on both sides of the subject 100 (on the diagonal line centered on the subject 100), the transmitted electrical signal is determined as appropriate data. That is, the coincidence counting circuit 40 detects the situation where radiation (gamma ray) is simultaneously observed (i.e., coincidence counting) in two radiation detectors 3 located on both sides of the subject 100 (on the diagonal line centered on the subject 100) based on the above electrical signal.

[0037] The detection signal data (count value) composed of appropriate data determined as coincidence counting by the coincidence counting circuit 40 is sent to the arithmetic circuit 41 (refer to Figure 1 ). The arithmetic circuit 41 acquires list mode data as the detection result of the radiation (gamma ray) detected by the detector ring 2. The list mode data refers to the data obtained by saving the detection event information (detector number, detection time, energy of the radiation (gamma ray), etc.) of the radiation (gamma ray) in time series. The arithmetic circuit 41 reconstructs the radioactive distribution of the subject 100 through iterative calculation based on the list mode data.

[0038] (Structure related to the reconstruction of the radioactive distribution)

[0039] Next, with reference to Figure 3 's flowchart, the reconstruction process of the radioactive distribution of the subject 100 using the list mode data by the PET device 1 of the first embodiment will be described. In addition, each process of the flowchart is performed by the arithmetic circuit 41 of the control unit 4.

[0040] First, as Figure 3 shown, in step 101, the subject 100 is photographed (measured) by the PET device 1 for a specified measurement time (e.g., 30 minutes, etc.), thereby obtaining the list mode data within the measurement time range. The list mode data within the measurement time range contains multiple events (measured events).

[0041] Next, in step 102, the list mode data within the measurement time range is divided into multiple subsets. Specifically, in step 102, the list mode data is divided into multiple subsets by using any one of the equal event number division method, the equal ideal event number division method, the equal event interval division method, and the equal time interval division method. In addition, the subset division method and the number of subsets can be the unique subset division method and the number of subsets predetermined in the PET device 1, or the subset division method and the number of subsets input and specified by the user to the PET device 1.

[0042] The equal number of events segmentation method is a subset segmentation method that divides the measured events in the reconstruction time range (i.e., the measurement time range) into subsets with equal (roughly equal) numbers of events. For example, consider the case of dividing list mode data with a total of 100,000 measured events into 10 subsets. In this case, in the equal number of events segmentation method, the list mode data is equally segmented in a way that the first subset contains the measured events with event numbers {1, 2, …, 10000}, the second subset contains the measured events with event numbers {10001, 10002, …, 20000}, and so on, starting from the first event in groups of 10000 each. In addition, in the case of a remainder, the remainder is included in the last subset (the tenth subset).

[0043] The equal ideal number of events segmentation method is a subset segmentation method that divides the measured events in the reconstruction time range into subsets with equal (roughly equal) ideal numbers of events.

[0044] The ideal number of events is a number obtained by correcting the measured number of events using at least one factor selected from four physical factors: the physical decay of the radionuclide, the counting loss of detector 3, the deviation of the detection efficiency of detector 3, and the photon absorption by the subject 100. That is, the ideal number of events is the ideal number of events in the absence of the influence of physical factors. When the factor coefficient based on at least one factor selected from the four physical factors of the physical decay of the radionuclide, the counting loss of detector 3, the deviation of the detection efficiency of detector 3, and the photon absorption by the subject 100 is set as η t the ideal number of events for one measured event can be represented by the reciprocal of η t (i.e., 1 / η t ). Therefore, the total ideal number of events can be represented by the following formula (1).

[0045] [Mathematical formula 1]

[0046]

[0047] Here,[[]]END]]

[0048] X: Total ideal number of events

[0049] N: Total measured number of events

[0050] t: Event number

[0051] η t : Factor coefficient.

[0052] In addition, the factor used for the factor coefficient among the four physical factors can be a unique factor predetermined in the PET device 1 or a factor specified by the user by inputting to the PET device 1.

[0053] The equal ideal event number segmentation method is a subset segmentation method for segmenting the measured events in the reconstructed time range in such a way that the ideal event number K of each subset is X / M when the number of subsets is set to M. Specifically, in the equal ideal event number segmentation method, the ideal event numbers (i.e., 1 / η j ) of each measured event are accumulated, and if the cumulative value of the ideal event numbers of the measured events up to the event number j 1 exceeds K, the measured events with event numbers {1, 2,..., j 1} are set as the first subset, and if the cumulative value of the ideal event numbers of the measured events up to the event number j 2 exceeds K, the measured events with event numbers {j 1 + 1, j 1 + 2,..., j 2} are set as the second subset, and so on. In this way, the list mode data is equally segmented starting from the first event so that the ideal event numbers are each K. In addition, in the case of a remainder, the remainder is included in the last subset.

[0054] The equal event interval segmentation method is a subset segmentation method for segmenting the measured events in the reconstructed time range at equal event intervals. For example, consider the case of segmenting list mode data with a total of 100,000 measured events into 10 subsets. In this case, in the equal event interval segmentation method, the measured events with event numbers {1, 11, 21,..., 99991} are included in the first subset, the measured events with event numbers {2, 12, 22,..., 99992} are included in the second subset, and so on. In this way, starting from the first event, the list mode data is segmented by skipping 10 events (the number of events corresponding to the number of subsets).

[0055] The equal time interval segmentation method is a subset segmentation method for segmenting the measured events in the reconstructed time range at equal time intervals. For example, consider the case of segmenting list mode data with a reconstructed time range of 30 minutes into 10 subsets. In this case, in the equal time interval segmentation method, the measured events in the reconstructed time range from 0 minutes to 3 minutes are included in the first subset, the measured events in the reconstructed time range from 3 minutes to 6 minutes are included in the second subset, and so on. In this way, starting from the first event, the list mode data is segmented at equal time intervals.

[0056] Next, in step 103, an absorption coefficient image for correcting the absorption of radiation (gamma rays) in the subject 100 is estimated. The method for estimating the absorption coefficient image is not particularly limited, and for example, known methods such as a CT (Computed Tomography) image conversion method and a simultaneous estimation method can be used. In addition, as the simultaneous estimation method, for example, an MLAA (Maximum Likelihood Estimation of Attenuation and Activity) method, an MLACF (Maximum Likelihood Attenuation Correction Factors) method, etc. can be used.

[0057] Next, in step 104, scattered ray distribution projection data for correcting the scattering of radiation (gamma rays) in the subject 100 is estimated. The method for estimating the scattered ray distribution projection data is not particularly limited, and for example, known methods such as a single scattering simulation method and a convolution method can be used.

[0058] Here, in the first embodiment, in step 105, a subset balance coefficient for adjusting the non-uniformity of the number of events between subsets is obtained based on the number of events in a plurality of subsets. In step 105, the subset balance coefficient is obtained based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset. Specifically, the subset balance coefficient is obtained based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset. More specifically, the subset balance coefficient represented by the following formulas (2) to (4) is obtained.

[0059] [Mathematical formula 2]

[0060] c l =W / W l …(2)

[0061] [Mathematical formula 3]

[0062]

[0063] [Mathematical formula 4]

[0064]

[0065] Here,

[0066] W: The average ideal number of events in the reconstruction time range

[0067] W l : The average ideal number of events in the l-th subset

[0068] l: Subset number

[0069] C l : Subset balance coefficient

[0070] S l : Set of measured events belonging to the l-th subset

[0071] j: Pixel number

[0072] η t : Factor coefficient.

[0073] For example, when the list mode data is divided into 10 subsets, according to Equations (2) to (4), obtain c corresponding to the first subset to the tenth subset respectively 1 -c 10 These 10 subset coefficients. That is, in step 105, obtain the subset balance coefficient of each subset. The subset balance coefficient is a positive coefficient that depends on the subset number. In addition, the details of the effect of the subset balance coefficient are described later.

[0074] Next, in step 106, perform a reconstruction calculation to reconstruct the radioactive distribution of the subject 100 by iterative calculation based on the list mode data. Specifically, in step 106, perform a reconstruction calculation including the following steps: obtain a backprojection value based on the list mode data; obtain a corrected backprojection value by multiplying the backprojection value by the subset balance coefficient; obtain a backprojection image based on the corrected backprojection value; and update the radioactive distribution image based on the backprojection image. More specifically, perform the reconstruction calculation through the following Equations (5) to (8). In addition, introduce the subset balance coefficient into Equation (5). In addition, Equation (5) is an equation using the DRAMA (Dynamic Row-Action Maximum Likelihood Algorithm) method including backprojection calculation as the list mode reconstruction algorithm.

[0075] [Mathematical formula 5]

[0076]

[0077] [Mathematical formula 6]

[0078]

[0079] [Mathematical formula 7]

[0080]

[0081] [Mathematical formula 8]

[0082] λ (k，l) =β 0 / (β 0(1 + γkL)…(8)

[0083] Here,

[0084] k: number of iterations

[0085] l: subset number

[0086] L: number of subsets

[0087] S l : set of measured events belonging to the l-th subset

[0088] t: event number

[0089] i: detector number

[0090] j: pixel number

[0091] x j : j-th pixel value of the radioactive distribution image of the object to be estimated

[0092] a ij : probability that the radiation generated in the j-th pixel is detected by the i-th detector (value independent of time)

[0093] r i : average coefficient rate of background events

[0094] T acq : measurement time (sec)

[0095] h t : product of the physical attenuation coefficient at the detection time of the t-th measured event and the count loss coefficient for the t-th measured event

[0096] β 0 , γ: relaxation parameters.

[0097] In addition, in Equation (5), the backprojection value is the part represented by the following Equation (9), the corrected backprojection value is the part represented by the following Equation (10), and the backprojection image is the part represented by the following Equation (11).

[0098] [Equation 9]

[0099]

[0100] [Equation 10]

[0101]

[0102] [Equation 11]

[0103]

[0104] As shown in Equation (5), the reconstruction calculation includes: a first step of updating the radioactive distribution image for each subset; and a second step of iterating (repeating) the first step for an iteration number (i.e., k times). In the first step, the following steps are performed for each subset: obtaining a back-projection value based on the above list-mode data; obtaining a corrected back-projection value by multiplying the back-projection value by a subset balance coefficient; obtaining a back-projection image based on the corrected back-projection value; and updating the radioactive distribution image based on the back-projection image.

[0105] For example, in the case where the list-mode data is divided into 10 subsets, in the first step, using the subset balance coefficient c corresponding to the first subset 1 and updating the radioactive distribution image using the first subset, using the subset balance coefficient c corresponding to the second subset 2 and updating the radioactive distribution image using the second subset (the radioactive distribution image updated using the first subset),... In this way, the radioactive distribution image is updated by sequentially performing calculations from the first subset to the tenth subset. In the second step, the calculation of the first step is repeated for the iteration number.

[0106] When the reconstruction calculation including the first step and the second step is completed, a quantitative radioactive distribution image in which the pixel values of each pixel converge to the value representing the radiation concentration of the subject 100 is obtained.

[0107] (Explanation of the effect of the subset balance coefficient)

[0108] Next, with reference to Figures 4 - 7 the effect of the subset balance coefficient will be explained.

[0109] Figure 4 is a schematic curve showing the change in the count rate with respect to the measurement time when photographing the subject 100. In Figure 4 the curve graph, the vertical axis represents the count rate (cps: count / sec), and the horizontal axis represents the time (sec). In addition, in Figure 4 the curve graph, the solid line represents the time change of the measured count rate, and the dashed-dotted line represents the time change of the ideal count rate. In addition, as described above, the time change of the ideal count rate is obtained by multiplying the time change of the measured count rate by the reciprocal of the factor coefficient η t (i.e., 1 / η t ). That is, the time change of the ideal count rate is the time change of the ideal count rate without the influence of physical factors.

[0110] In addition, Figure 4 the curve graph shows the case where the list-mode data is divided into 6 subsets using the equal-event-number division method. In Figure 4In the curve graph, the six subsets are respectively illustrated as sub1, sub2, sub3, sub4, sub5, and sub6. In Figure 4 In the example of

[0111] Figure 5 Since the equal-event-number segmentation method is used for segmentation, the measured event numbers in the six subsets of sub1 to sub6 are equal (roughly equal). On the other hand, in the six subsets of sub1 to sub6, the time widths of each subset are not equal, and the ideal event numbers of each subset are not uniform.

[0112] As Figure 5 shown, regarding the average ideal event number in the reconstruction time range, which is the numerator of the subset balance coefficient represented by (average ideal event number in the reconstruction time range) / (average ideal event number in the subset), it can be expressed as the value obtained by dividing the area of the time variation of the ideal count rate for the entire reconstruction time range by the area of the time variation of the measured count rate for the entire reconstruction time range. Additionally, regarding the average ideal event number in the subset (sub2), which is the denominator of the subset balance coefficient, it can be expressed as the value obtained by dividing the area of the time variation of the ideal count rate for the subset (sub2) in the reconstruction time range by the area of the time variation of the measured count rate for the subset (sub2) in the reconstruction time range. Furthermore, since the area is obtained by cps × sec, the unit of each area is count (count).

[0113] In addition, in Figure 5 an example of obtaining the subset balance coefficient of sub2 is shown, but for sub1 and sub3 to sub6 other than sub2, the subset balance coefficient can also be obtained through the same calculation.

[0114] Figure 6 is a schematic diagram showing the state of the areas of each subset (sub1 to sub6) before the subset balance coefficient correction (before the multiplication operation). In Figure 6 the area of the time variation of the ideal count rate for each subset is represented by a rectangle using the average ideal count rate of each subset. Additionally, the average ideal count rate of each subset refers to the average of the ideal count rates in the time range of each subset. In Figure 6 the area of the time variation of the ideal count rate for each subset is represented by a rectangle with the average ideal count rate of each subset as the upper end value. In addition, in Figure 6 only the average ideal count rate of sub5 is shown, but for sub1 to sub4 and sub6 other than sub5, the upper end value of the rectangle is also the average ideal count rate. In Figure 6Among sub1 to sub6, the areas of the respective subsets are uneven. That is, among sub1 to sub6, the ideal event numbers of the respective subsets are uneven. In this case, since the ideal event numbers of the respective subsets are uneven and inconsistent, a quantitative radioactive concentration cannot be obtained.

[0115] Figure 7 is a schematic diagram showing the state of the areas of the respective subsets after subset balance coefficient correction (after multiplication). In Figure 7 for easy understanding, the positions of the average ideal count rates of the respective subsets before subset balance coefficient correction are indicated by dashed lines (i.e., Figure 6 the upper end positions of the rectangles representing the areas of the respective subsets). In Figure 7 by multiplying the area of each subset by the subset balance coefficient, the areas of sub1 to sub3 with relatively small areas become larger, and the areas of sub4 to sub6 with relatively large areas become smaller.

[0116] Multiplying by the subset balance coefficient means, as described above, dividing the area of the subset by the "average ideal event number of the subset" and then multiplying by the "average ideal event number of the reconstruction time range" that does not depend on the subset. Therefore, by multiplying the area of each subset by the subset balance coefficient, the areas of the respective subsets become consistent (roughly equal). That is, by multiplying by the subset balance coefficient, the ideal event numbers of the respective subsets become consistent (roughly equal), and thus a quantitative radioactive concentration can be obtained.

[0117] (Actual image)

[0118] Next, referring to Figure 8 and Figure 9 the actual reconstructed image (radioactive distribution image) will be described. Figure 8 is a diagram showing an example of an actual reconstructed image reconstructed without using the subset balance coefficient, an example of a correct answer image, and a difference image between the reconstructed image and the correct answer image. In addition, Figure 9 is a diagram showing an example of an actual reconstructed image reconstructed using the subset balance coefficient, an example of a correct answer image, and a difference image between the reconstructed image and the correct answer image.

[0119] In Figure 8 and Figure 9 in any of the cases, the subset division method is set to the equal event number division method, and the number of subsets is set to 100. In addition, in the reconstruction calculation of the reconstructed image in Figure 8 the above formula (5) without the subset balance coefficient is used, and in the reconstruction calculation of the reconstructed image in Figure 9 the above formula (5) including the subset balance coefficient is used. In addition, in Figure 8 and Figure 9In the case of any of them, as the correct image, use the reconstructed image when the number of subsets is set to 1.

[0120] As Figure 8 shown, when the subset balance coefficient is not used, the difference image between the reconstructed image and the correct image is not zero. That is, the calculated value (pixel value) calculated iteratively in the reconstructed image does not converge to the value (the value of the correct image) representing the radioactive concentration of the subject 100. Therefore, when the subset balance coefficient is not used, a quantitative reconstructed image (radioactive distribution image) cannot be obtained.

[0121] On the other hand, as Figure 9 shown, when the subset balance coefficient is used, the difference image between the reconstructed image and the correct image is zero. That is, the calculated value (pixel value) calculated iteratively in the reconstructed image converges to the value (the value of the correct image) representing the radioactive concentration of the subject 100. Therefore, by using the subset balance coefficient, a quantitative reconstructed image (radioactive distribution image) can be obtained. In addition, the same result can be obtained when the number of subsets is changed.

[0122] (Effect of the first embodiment)

[0123] In the first embodiment, the following effects can be obtained.

[0124] In the first embodiment, as described above, the list mode data is divided into a plurality of subsets, the subset balance coefficient is obtained based on the number of events of the plurality of subsets, the backprojection value is obtained based on the list mode data, the corrected backprojection value is obtained by multiplying the backprojection value by the subset balance coefficient, the backprojection image is obtained based on the corrected backprojection value, and the radioactive distribution image is updated based on the backprojection image. Thus, by introducing the subset balance coefficient, the unevenness of the number of events between subsets can be adjusted, so that the state in which the calculated value (pixel value) of the iterative calculation does not converge to the value representing the radioactive concentration of the subject due to the unevenness of the number of events between subsets can be suppressed. In other words, the calculated value of the iterative calculation can be made to converge to the value representing the radioactive concentration of the subject, so that a quantitative radioactive distribution image can be generated. In addition, by introducing the subset balance coefficient, the unevenness of the number of events between subsets can also be adjusted when the number of subsets is changed, so that the change of the pixel value depending on the number of subsets can be suppressed. That is, the same pixel value can be obtained regardless of the number of subsets. As a result, a quantitative radioactive distribution image can be generated regardless of the number of subsets.

[0125] In addition, in the first embodiment, as described above, the subset balance coefficient is obtained based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset. Thus, it is possible to easily adjust the non-uniformity of the ideal number of events between subsets based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset, and therefore it is possible to easily converge the calculated value of the iterative calculation to the value representing the radioactivity concentration of the subject.

[0126] In addition, in the first embodiment, as described above, the subset balance coefficient is obtained based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset. Thus, it is possible to more easily adjust the non-uniformity of the ideal number of events between subsets based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset, and therefore it is possible to more easily converge the calculated value of the iterative calculation to the value representing the radioactivity concentration of the subject. In addition, this structure is particularly effective when the subsets are divided using the equal number of events division method.

[0127] In addition, in the first embodiment, as described above, the ideal number of events is a number obtained by correcting the measured number of events using at least one factor selected from the four physical factors of the physical decay of the radionuclide, the counting loss of the detector, the deviation of the detection efficiency of the detector, and the photon absorption by the subject. Thus, it is possible to correct the measured number of events using at least one factor selected from the four physical factors of the physical decay of the radionuclide, the counting loss of the detector, the deviation of the detection efficiency of the detector, and the photon absorption by the subject, and therefore it is possible to accurately obtain the ideal number of events.

[0128] In addition, in the first embodiment, as described above, the subset balance coefficient for each subset is obtained. Thus, it is possible to adjust the non-uniformity of the ideal number of events between subsets using the subset balance coefficient suitable for each subset, and therefore it is possible to reliably adjust the non-uniformity of the ideal number of events between subsets.

[0129] In addition, in the first embodiment, as described above, the list mode data is divided into a plurality of subsets using any one of the equal number of events division method, the equal ideal number of events division method, the equal event interval division method, and the equal time interval division method. Thus, it is possible to divide the list mode data into a plurality of subsets using any one of the equal number of events division method, the equal ideal number of events division method, the equal event interval division method, and the equal time interval division method, and therefore it is possible to easily divide the list mode data into a plurality of subsets.

[0130] [Second Embodiment]

[0131] Next, referring to Figure 10A description will be given of the second embodiment of the present invention. In the second embodiment, an example will be described in which a subset balance coefficient different from that of the above-described first embodiment is used. In addition, for the same structures as those of the above-described first embodiment, the same reference numerals are used in the drawings and their descriptions are omitted.

[0132] (Structure of PET device)

[0133] In the second embodiment, as Figure 10 shown, the PET device 201 includes an arithmetic circuit 241 instead of the arithmetic circuit 41 of the above-described first embodiment. In addition, the PET device 201 is an example of the "nuclear medicine diagnostic device" of the present disclosure. Further, the arithmetic circuit 241 is an example of the "arithmetic unit" of the present disclosure.

[0134] The arithmetic circuit 241 is configured to obtain a subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the value obtained by multiplying the ideal number of events in each subset by the number of subsets. Specifically, the arithmetic circuit 241 is configured to obtain the subset balance coefficient represented by the following equations (12) to (14).

[0135] [Mathematical formula 12]

[0136] c l = W / W l …(12)

[0137] [Mathematical formula 13]

[0138]

[0139] [Mathematical formula 14]

[0140]

[0141] Here,

[0142] W: Ideal number of events in the reconstruction time range (total ideal number of events)

[0143] W l : Value obtained by multiplying the ideal number of events in the l-th subset by the number of subsets

[0144] l: Subset number

[0145] C l : Subset balance coefficient

[0146] S l : Set of measured events belonging to the l-th subset

[0147] j: Pixel number

[0148] η t : Factor coefficient.

[0149] In the second embodiment, the arithmetic circuit 241 is configured to perform reconstruction calculation on an expression obtained by introducing the subset balance coefficient of Expression (12) into Expression (5) of the first embodiment by using a subset balance coefficient that replaces the expression (2) of the first embodiment. In addition, the details of the reconstruction calculation are the same as those of the first embodiment, and thus detailed description thereof is omitted.

[0150] In addition, other configurations of the second embodiment are the same as those of the first embodiment.

[0151] (Effect of the second embodiment)

[0152] In the second embodiment, the following effects can be obtained.

[0153] In the second embodiment, as described above, the subset balance coefficient is obtained based on the ratio between the ideal number of events within the reconstruction time range and the value obtained by multiplying the ideal number of events of each subset by the number of subsets. Thus, compared with the case where the subset balance coefficient is obtained based on the ratio between the ideal number of events within the reconstruction time range and the ideal number of events of a subset (the case of the first embodiment), a subset balance coefficient with higher versatility can be obtained, and thus it is possible to more easily adjust the unevenness of the ideal number of events between subsets.

[0154] In addition, other effects of the second embodiment are the same as those of the first embodiment.

[0155] [Modification example]

[0156] In addition, it should be considered that the embodiments disclosed this time are illustrative in all aspects and not restrictive. The scope of the present invention is represented not by the description of the above embodiments but by the claims, and also includes all modifications (modification examples) within the meaning and scope equivalent to the claims.

[0157] For example, in the above first and second embodiments, an example where the nuclear medicine diagnostic device is a PET device is shown, but the present invention is not limited thereto. For example, the nuclear medicine diagnostic device may also be an SPECT (Single photon emission computed tomography) device other than a PET device.

[0158] In addition, in the above-described first and second embodiments, examples of obtaining the subset balance coefficients represented by formulas (2) to (4) or formulas (12) to (14) are shown, but the present invention is not limited thereto. In the present invention, as long as the non-uniformity of the number of events between subsets can be adjusted, subset balance coefficients other than those represented by formulas (2) to (4) or formulas (12) to (14) can also be obtained.

[0159] In addition, in the above-described first and second embodiments, an example of introducing the subset balance coefficient into formula (5) in which the DRAMA method including backprojection calculation is used as the list mode reconstruction algorithm is shown, but the present invention is not limited thereto. In the present invention, the subset balance coefficient can also be introduced into formulas in which methods other than the DRAMA method including backprojection processing, such as the OSEM (Ordered Subsets Expectation Maximization) method, are used as the list mode reconstruction algorithm. That is, in the present invention, the introduction of the subset balance coefficient is not limited to being applied to a specific list mode reconstruction algorithm.

[0160] In addition, in the above-described first and second embodiments, an example of reconstructing the radioactive distribution of the subject based on the list mode data after the shooting of the subject is completed is shown, but the present invention is not limited thereto. In the present invention, the radioactive distribution of the subject can also be reconstructed in real time based on the list mode data during the shooting of the subject. For example, the list mode data at the completion of the shooting of the subject can be estimated based on the list mode data (intermediate list mode data) during the shooting of the subject, and the radioactive distribution of the subject can be reconstructed based on the estimated list mode data. In this case, the estimated list mode data can be divided into a plurality of subsets in the same manner as in the above-described first or second embodiment, and the subset balance coefficient can be obtained in the same manner as in the above-described first or second embodiment. In addition, for example, the radioactive distribution of the subject can be directly reconstructed based on the list mode data (intermediate list mode data) during the shooting of the subject. In this case, the list mode data (intermediate list mode data) during the shooting of the subject can be divided into a plurality of subsets in the same manner as in the above-described first or second embodiment, and the subset balance coefficient can be obtained in the same manner as in the above-described first or second embodiment.

[0161] In addition, in the above-described first and second embodiments, for the sake of convenience of explanation, the processes of the arithmetic circuit 41 (241) are described using a "flow-driven" flowchart, but the present invention is not limited thereto. In the present invention, the above-described processes may also be performed in an "event-driven" manner executed in units of events. In this case, it may be performed in a completely event-driven manner, or may be performed by combining event-driven and flow-driven.

[0162] In addition, in the above-described first and second embodiments, an example in which the backprojection value is multiplied by the subset balance coefficient is shown, but the present invention is not limited thereto. In the present invention, the backprojection image may also be multiplied by the subset balance coefficient. That is, since the subset balance coefficient does not depend on the event number within the subset, the formula (11) representing the backprojection image can be expressed by the following formula (15).

[0163] [Mathematical formula 15]

[0164]

[0165] [Mode]

[0166] Those skilled in the art understand that the above exemplary embodiments are specific examples of the following modes.

[0167] (Item 1)

[0168] A list mode image reconstruction method for reconstructing the radioactive distribution of a subject by iterative calculation based on list mode data collected by a nuclear medicine diagnostic device, the list mode image reconstruction method comprising the following steps:

[0169] Dividing the list mode data into a plurality of subsets;

[0170] Obtaining a subset balance coefficient based on the number of events of the plurality of subsets;

[0171] Obtaining a backprojection value based on the list mode data;

[0172] Obtaining a backprojection image based on the backprojection value;

[0173] Multiplying the backprojection value or the backprojection image by the subset balance coefficient; and

[0174] Updating the radioactive distribution image based on the backprojection image.

[0175] (Item 2)

[0176] The list mode image reconstruction method according to item 1, wherein the step of obtaining the subset balance coefficient includes the following steps: obtaining the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset.

[0177] (Item 3)

[0178] The list mode image reconstruction method according to item 2, wherein the step of obtaining the subset balance coefficient includes the following steps: obtaining the subset balance coefficient based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset.

[0179] (Item 4)

[0180] The list mode image reconstruction method according to item 2, wherein

[0181] the step of obtaining the subset balance coefficient includes the following steps: obtaining the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the value obtained by multiplying the ideal number of events in each subset by the number of subsets.

[0182] (Item 5)

[0183] The list mode image reconstruction method according to any one of items 2 to 4, wherein the ideal number of events is a number obtained by correcting the measured number of events using at least one factor selected from the four physical factors of physical decay of the radionuclide, counting loss of the detector, deviation of the detection efficiency of the detector, and photon absorption by the subject.

[0184] (Item 6)

[0185] The list mode image reconstruction method according to any one of items 1 to 5, wherein the step of obtaining the subset balance coefficient includes the following steps: obtaining the subset balance coefficient for each subset.

[0186] (Item 7)

[0187] The list mode image reconstruction method according to any one of items 1 to 6, wherein the step of dividing the list mode data into the plurality of subsets includes the following steps: dividing the list mode data into the plurality of subsets using any one of the equal number of events division method, equal ideal number of events division method, equal event interval division method, and equal time interval division method.

[0188] (Item 8)

[0189] A nuclear medicine diagnostic device, comprising:

[0190] A detection unit that detects radiation generated from a radioactive agent within a subject; and

[0191] An arithmetic unit that reconstructs the radioactive distribution of the subject through iterative calculation based on list mode data that is the detection result of radiation by the detection unit,

[0192] wherein the arithmetic unit is configured to: divide the list mode data into a plurality of subsets, obtain a subset balance coefficient based on the number of events of the plurality of subsets, obtain a backprojection value based on the list mode data, obtain a backprojection image based on the corrected backprojection value, multiply the backprojection value or the backprojection image by the subset balance coefficient, and update a radioactive distribution image based on the backprojection image.

[0193] (Item 9)

[0194] The nuclear medicine diagnostic apparatus according to Item 8, wherein the arithmetic unit is configured to obtain the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset.

[0195] (Item 10)

[0196] The nuclear medicine diagnostic apparatus according to Item 9, wherein the arithmetic unit is configured to obtain the subset balance coefficient based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset.

[0197] (Item 11)

[0198] The nuclear medicine diagnostic apparatus according to Item 9, wherein the arithmetic unit is configured to obtain the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the value obtained by multiplying the ideal number of events in each subset by the number of subsets.

[0199] Explanation of reference numerals

[0200] 1, 201: PET apparatus (nuclear medicine diagnostic apparatus); 2: detector ring (detection unit); 41, 241: arithmetic circuit (arithmetic unit); 100: subject.

Claims

1. A list mode image reconstruction method, which reconstructs the radioactive distribution of a subject through iterative calculation based on list mode data collected by a nuclear medicine diagnostic device. The list mode image reconstruction method comprises the following steps: Dividing the list mode data into a plurality of subsets; Obtaining a subset balance coefficient based on the number of events in the plurality of subsets; Obtaining a backprojection value based on the list mode data; Obtaining a first backprojection image based on the backprojection value, multiplying the first backprojection image by the subset balance coefficient to obtain a second backprojection image, and updating the radioactive distribution image based on the second backprojection image; or Multiplying the backprojection value by the subset balance coefficient to obtain a corrected backprojection value, obtaining a third backprojection image based on the corrected backprojection value, and updating the radioactive distribution image based on the third backprojection image.

2. The list mode image reconstruction method according to claim 1, wherein, The step of obtaining the subset balance coefficient comprises the following steps: obtaining the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset.

3. The list mode image reconstruction method according to claim 2, wherein, The step of obtaining the subset balance coefficient comprises the following steps: obtaining the subset balance coefficient based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset.

4. The list mode image reconstruction method according to claim 2, wherein, The step of obtaining the subset balance coefficient comprises the following steps: obtaining the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the value obtained by multiplying the ideal number of events in each subset by the number of subsets.

5. The list mode image reconstruction method according to claim 2, wherein, The ideal number of events is a number obtained by correcting the measured number of events using at least one factor selected from the four physical factors of physical decay of the radionuclide, counting loss of the detector, deviation of the detection efficiency of the detector, and photon absorption by the subject.

6. The list mode image reconstruction method according to claim 1, wherein, The step of obtaining the subset balance coefficient comprises the following steps: obtaining the subset balance coefficient for each subset.

7. The list mode image reconstruction method according to claim 1, wherein, The step of dividing the list mode data into the plurality of subsets comprises the following steps: dividing the list mode data into the plurality of subsets using any one of the equal number of events division method, equal ideal number of events division method, equal event interval division method, and equal time interval division method.

8. A nuclear medicine diagnostic device, comprising: A detection unit that detects radiation generated from a radioactive agent in a subject; and An arithmetic unit that reconstructs the radioactive distribution of the subject through iterative calculation based on list mode data that is the detection result of the radiation by the detection unit, wherein, The arithmetic unit is configured to: Divide the list mode data into a plurality of subsets, Obtain a subset balance coefficient based on the number of events in the plurality of subsets, Obtain back-projection values based on the list-mode data, obtain a first back-projection image based on the back-projection values, multiply the first back-projection image by the subset balance coefficient to obtain a second back-projection image, and update the radioactive distribution image based on the second back-projection image; or multiply the back-projection values by the subset balance coefficient to obtain corrected back-projection values, obtain a third back-projection image based on the corrected back-projection values, and update the radioactive distribution image based on the third back-projection image.

9. The nuclear medicine diagnostic apparatus according to claim 8, wherein, the operation unit is configured to obtain the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the ideal number of events in each subset.

10. The nuclear medicine diagnostic apparatus according to claim 9, wherein, the operation unit is configured to obtain the subset balance coefficient based on the ratio of the average ideal number of events in the reconstruction time range to the average ideal number of events in each subset.

11. The nuclear medicine diagnostic apparatus according to claim 9, wherein, the operation unit is configured to obtain the subset balance coefficient based on the ratio of the ideal number of events in the reconstruction time range to the value obtained by multiplying the ideal number of events in each subset by the number of subsets.

Images