Nuclear medicine diagnosis apparatus and method

Abstract

To improve the workflow related to the respiratory gating.SOLUTION: A nuclear medicine diagnosis apparatus according to an embodiment includes: a control unit configured to generate a feature vector for each of the mini frames from list mode data acquired by performing a nuclear medicine scan and to calculate a Euclidean distance between each of the feature vectors corresponding to frames and a reference phase vector; a determination unit configured to calculate, on the basis of the Euclidean distances, a mean Euclidean distance for a plurality of conditions obtained by varying a percentage of counts to be used in a reconstructing process, and to determine the percentage on the basis of a relationship between the percentage and the mean Euclidean distance; and a reconstruction unit configured to generate a gated image by carrying out an image reconstruction while using the determined percentage as a gating condition.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The embodiments disclosed in this specification and the drawings relate to nuclear medicine diagnostic apparatuses and methods.

Background Art

[0002] Data-driven Deviceless Gating is a technique for reconstructing respiratory-gated images and reducing motion blur without requiring an external motion sensor. Conventionally, in order to perform data-driven deviceless gating, it has been necessary for the user to select gating conditions.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to improve the workflow related to respiratory gating. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. It is also possible to position the problems corresponding to the respective effects of each configuration shown in the embodiments described later as other problems.

Means for Solving the Problems

[0005] The nuclear medicine diagnostic apparatus according to the embodiment includes a control unit that generates a feature vector for each mini-frame from list mode data collected by performing a nuclear medicine scan and calculates the Euclidean distance between the feature vector of each frame and a reference phase vector, and calculates an average Euclidean distance for a plurality of conditions in which the ratio of counts used for reconstruction processing is changed based on the Euclidean distance, and a determination unit that determines the ratio based on the relationship between the ratio and the average Euclidean distance, and a reconstruction unit that executes image reconstruction using the determined ratio as a gating condition to generate a gated image.

Brief Description of Drawings

[0006]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 3C

Figure 3D

Figure 3E

Figure 3F

Figure 3G

Figure 4A

Figure 4B

Figure 4C

Figure 4D

Figure 4E

Figure 4F

Figure 5

Figure 6

Figure 7

[0007] Hereinafter, embodiments of a nuclear medicine diagnostic apparatus and method will be described with reference to the drawings.

[0008] (Embodiment) The nuclear medicine diagnostic apparatus 10 according to the embodiment is a medical imaging diagnostic apparatus (modality) capable of performing a nuclear medicine scan. A nuclear medicine scan is a scan performed by administering a drug labeled with a radionuclide to a subject, and typical examples include a PET (Positron Emission Computed Tomography) scan and a SPECT (Single Photon Emission Computed Tomography) scan.

[0009] In this embodiment, taking the case where the nuclear medicine diagnostic apparatus 10 is a PET apparatus as an example, it will be described with reference to FIG. 1. FIG. 1 is a diagram showing an example of the configuration of the nuclear medicine diagnostic apparatus 10 according to the embodiment. The nuclear medicine diagnostic apparatus 10 shown in FIG. 1 includes a gantry apparatus 110 and a console apparatus 120. The gantry apparatus 110 includes a detector 130, a front-end circuit 112, a top plate 113, a couch 114, and a couch driving unit 116.

[0010] The detector 130 is a detector that detects radiation. For example, the detector 130 is a detector that detects radiation by detecting scintillation light (fluorescence), which is light re-emitted when a substance in an excited state returns to the ground state when gamma rays generated by the annihilation of positrons emitted from a drug administered to and accumulated in the subject P interact with the surrounding tissue electrons interact with the phosphor. Also, in the embodiment, the detector 130 can also detect Cherenkov light. The detector 130 detects the energy information of the radiation of gamma rays generated by the annihilation of positrons emitted from a drug administered to and accumulated in the subject P with the electrons of the surrounding tissue. A plurality of detectors 130 are arranged so as to surround the periphery of the subject P in a ring shape, and are composed of, for example, a plurality of detector blocks.

[0011] Detector 130 typically consists of a scintillator crystal and a light detection surface composed of a light detection element. As the material of the scintillator crystal, for example, materials suitable for generating Cherenkov light, such as bismuth germanate (BGO: Bismuth Germanium Oxide), lead glass (SiO2 + PbO), lead fluoride (PbF2), lead tungstate (PWO: PbWO4) and other lead compounds can be used. As another example, for instance, scintillator crystals such as LYSO (Lutetium Yttrium Oxyorthosilicate), LSO (Lutetium Oxyorthosilicate), LGSO (Lutetium Gadolinium Oxyorthosilicate) and BGO can also be used. The light detection element constituting the light detection surface typically consists of a plurality of pixels, and each of these pixels is, for example, composed of a SPAD (Single Photon Avalanche Diode). Also, the configuration of detector 130 is not limited to the above examples. As an example, the light detection element may use, for example, a SiPM (Silicon photomultiplier) or a photomultiplier tube. Further, the scintillator crystal may be a monolithic crystal, and the light detection surface composed of the light detection element may be arranged, for example, on the six faces of the scintillator crystal.

[0012] In addition, the gantry device 110 generates count (count number) information from the output signal of the detector 130 by the front-end circuit 112, and stores the generated count information in the memory 124 of the console device 120. Note that the detector 130 may be divided into a plurality of blocks and may include the front-end circuit 112.

[0013] The front-end circuit 112 converts the output signal from the detector 130 into digital data and generates count information. This count information includes, for example, the detection position, energy value, and detection time of the annihilation gamma rays. For example, the front-end circuit 112 identifies a plurality of photodetector elements that have converted scintillation light into electrical signals at the same timing. Then, the front-end circuit 112 identifies the scintillator number (P) indicating the position of the scintillator where the annihilation gamma rays have entered. The means for identifying the position of the scintillator where the annihilation gamma rays have entered may be identified by performing a centroid calculation based on the position of each photodetector element and the intensity of the electrical signal. Also, when the element sizes of each of the scintillator and the photodetector element correspond, for example, the scintillator corresponding to the photodetector element that has obtained the maximum output is assumed to be the position of the scintillator where the annihilation gamma rays have entered, and finally, it is identified considering inter-scintillator scattering and the like.

[0014] Also, the front-end circuit 112 performs an integral calculation of the intensity of the electrical signal output from each photodetector element or measures the time (Time over Threshold) when the electrical signal intensity exceeds a threshold value to identify the energy value (E) of the annihilation gamma rays incident on the detector 130. Also, the front-end circuit 112 identifies the detection time (T) when scintillation light due to the annihilation gamma rays is detected by the detector 130. Note that the detection time (T) may be an absolute time or an elapsed time from the start time of imaging. In this way, the front-end circuit 112 generates count information including the scintillator number (P), energy value (E), and detection time (T).

[0015] Note that the front-end circuit 112 is realized by a circuit such as a CPU (Central Processing Unit), a GPU (Graphical Processing Unit), an application specific integrated circuit (ASIC), or a programmable logic device (for example, a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)).

[0016] The top plate 113 is a bed on which the subject P is placed and is arranged above the bed 114. The bed drive unit 116 moves the top plate 113 under the control of the control function 125a of the processing circuit 125. For example, the bed drive unit 116 moves the subject P into the imaging port of the gantry device 110 by moving the top plate 113.

[0017] The console device 120 receives the operation of the nuclear medicine diagnostic apparatus 10 by the operator, controls the execution of the nuclear medicine scan, and reconstructs a nuclear medicine image from the collected list mode data. When the nuclear medicine diagnostic apparatus 10 is a PET apparatus, the console device 120 controls the execution of the PET scan and reconstructs a PET image from the list mode data. As shown in FIG. 1, the console device 120 includes a communication interface 121, an input interface 122, a display 123, a memory 124, and a processing circuit 125.

[0018] The communication interface 121 controls the transmission and communication of various data transmitted and received between the console device 120 and other devices and systems connected via a network. Specifically, the communication interface 121 is connected to the processing circuit 125, outputs data received from other devices and systems to the processing circuit 125, or transmits data output from the processing circuit 125 to other devices and systems. For example, the communication interface 121 is realized by a network card, a network adapter, a NIC (Network Interface Controller), etc.

[0019] The input interface 122 receives various input operations from the operator, converts the received input operations into electrical signals, and outputs them to the processing circuit 125. For example, the input interface 122 is realized by a mouse, a keyboard, a trackball, a switch, a button, a joystick, a touch pad that performs an input operation by touching an operation surface, a touch screen in which a display screen and a touch pad are integrated, a non-contact input circuit using an optical sensor, a voice input circuit, etc. Note that the input interface 122 may be composed of a tablet terminal or the like that can communicate wirelessly with the console device 120 main body. Also, the input interface 122 may be a circuit that receives an input operation from the operator by motion capture. For example, the input interface 122 can receive the body movement, line of sight, etc. of the operator as input operations by processing signals acquired via a tracker or images collected about the operator. Also, the input interface 122 is not limited to those having physical operation components such as a mouse and a keyboard. For example, an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the console device 120 and outputs this electrical signal to the processing circuit 125 is also included in the examples of the input interface 122.

[0020] The display 123 displays various types of information. For example, the display 123 displays various types of medical information such as nuclear medicine images collected from the subject P, respiratory waveforms, and heartbeat waveforms. Also, for example, the display 123 displays a GUI (Graphical User Interface) for receiving various instructions, settings, etc. from the operator via the input interface 122. For example, the display 123 is a liquid crystal display or a CRT (Cathode Ray Tube) display. The display 123 may be a desktop type, or may be configured as a tablet terminal or the like that can communicate wirelessly with the console device 120 main body.

[0021] The memory 124 is realized by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, hard disks, optical disks, etc. For example, the memory 124 stores various types of medical information and programs for the circuits included in the console device 120 to realize their functions. The memory 124 may also be realized by a server group (cloud) connected to the console device 120 via the network NW.

[0022] The processing circuit 125 controls the operation of the entire console device 120 by functioning as a control function 125a, a determination function 125b, and a reconstruction function 125c. For example, the processing circuit 125 functions as the control function 125a by reading and executing a program corresponding to the control function 125a from the memory 124.

[0023] For example, the control function 125a controls the operations of various components included in the gantry device 110 to perform a nuclear medicine scan on the subject P and acquire list mode data indicating the radiation detection results. Further, the control function 125a performs display control on the display 123. Also, the control function 125a controls the transmission and reception of each data via the network. For example, the control function 125a transmits and registers the list mode data collected by the nuclear medicine scan and the nuclear medicine image reconstructed based on the list mode data to the server of the PACS (Picture Archiving and Communication System).

[0024] Similarly, the processing circuit 125 functions as a determination function 125b and a reconstruction function 125c. The control function 125a is an example of a control unit. The determination function 125b is an example of a determination unit. The reconstruction function 125c is an example of a reconstruction unit.

[0025] In the nuclear medicine diagnostic apparatus 10 shown in FIG. 1, each processing function is stored in the memory 124 in the form of a program executable by a computer. The processing circuit 125 is a processor that realizes the functions corresponding to each program by reading and executing the program from the memory 124. In other words, the processing circuit 125 in the state of having read the program has the functions corresponding to the read program.

[0026] Note that in FIG. 1, it has been described that the control function 125a, the determination function 125b, and the reconstruction function 125c are realized by a single processing circuit 125, but it is also possible to configure the processing circuit 125 by combining a plurality of independent processors and have each processor execute a program to realize the functions. Also, each processing function of the processing circuit 125 may be appropriately distributed or integrated into a single or a plurality of processing circuits and realized.

[0027] Further, the processing circuit 125 may also implement functions by using the processor of an external device connected via the network NW. For example, the processing circuit 125 reads out and executes a program corresponding to each function from the memory 124, and also uses a group of servers (cloud) connected to the nuclear medicine diagnostic apparatus 10 via the network NW as computing resources, thereby implementing each function shown in FIG. 1.

[0028] The configuration example of the nuclear medicine diagnostic apparatus 10 has been described above. Here, the list mode data collected by nuclear medicine scanning may be affected by respiratory motion. Also, respiratory gating is known as a method for reducing blurring due to respiratory motion. Methods for implementing respiratory gating include a device-based method using an external device and a device-less method not using an external device.

[0029] In device-based respiratory gating, an external device such as a respiratory synchronization monitor is attached to the subject during the execution of nuclear medicine scanning. Thereby, while collecting list mode data, the respiratory waveform of the subject can be measured by the external device, and the list mode data can be associated with the respiratory waveform. However, device-based respiratory gating has various demerits such as deterioration of the workflow, increase in cost, and influence on the skin of the subject.

[0030] Conventionally, device-less respiratory gating has been executed by a user selecting various gating conditions. The inventors proposed a method for automatically selecting a static respiration phase based on the calculation of characteristic motion vectors in past research, but in this proposal, the user's selection was required for counting.

[0031] Specifically, when performing respiratory gating, among all the collected data, the more counts are used for the reconstruction process, the more the noise decreases and the SD value improves. On the other hand, the more counts are used, the more susceptible it is to the influence of movements such as respiratory motion, and the greater the blurring of the image becomes. When the blurring of the image increases, the contrast of the site that the user wants to observe, such as a tumor, may decrease in some cases.

[0032] Therefore, considering the balance between noise and blurring, it is necessary to appropriately set, as a gating condition, how many counts are to be used for the reconstruction process. Hereinafter, the gating condition indicating the ratio of the counts used for the reconstruction process is also referred to as % count. As described above, regarding % count, conventionally, it has been subject to the user's selection.

[0033] As one method of automating the control of % count, it is also conceivable to use a recommended fixed value. However, the appropriate % count varies depending on the position of the bed, the respiratory motion (fast / slow / irregular, etc.) of the subject, the dose, the BMI (Body Mass Index) of the subject, and the like. Therefore, it is difficult to appropriately cover a plurality of subjects with a single recommended value of % count. The factor that particularly affects the selection of the optimal % count is the respiratory motion (fast breathing / slow breathing / asymmetric respiratory cycle, etc.) that varies from subject to subject. For example, when the stationary period (less movement) is short (for example, in the case of fast breathing / irregular breathing), the optimal % count becomes small, and when the stationary period is long (for example, in the case of slow and regular breathing), the optimal % count becomes large.

[0034] The nuclear medicine diagnostic apparatus 10 according to the embodiment appropriately performs automatic count control in device - less respiratory gating, and improves the clinical workflow while ensuring an image quality with reduced blurring due to respiratory motion. The count optimization algorithm according to the embodiment generally optimizes the count by detecting a sharp transition in the motion characteristics from the exhalation phase to the inhalation phase.

[0035] In this algorithm, as also explained in the inventors' past research, the movement characteristics of the subject are utilized, which are represented by feature vectors obtained from short-term (mini-frame) PET images reconstructed in each phase of the respiratory cycle. The period of the mini-frame images is small enough to capture the fluctuations of respiratory motion, but large enough to exclude the motion of the heart with a period of less than one second (by averaging due to period overlap).

[0036] First, the reference phase vector is calculated by averaging the vectors corresponding to the stationary phases over all respiratory cycles. Note that the reference phase vector is also simply referred to as the reference vector. Next, for the vectors of the respiratory phases, the Euclidean distance from the reference vector is calculated.

[0037] Next, they are sorted from the ones with low Euclidean distances to the ones with high Euclidean distances, and for each % count (for example, from "5%" to "100%"), the Mean Euclidean Distance (MED) is calculated. The optimal % count is selected from the plot of "MED vs % count" by detecting a sharp transition (change in motion) of the MED from the exhalation phase to the inhalation phase. More generally, the optimal % count can be selected by modeling and analyzing the MED as a function of the count. Such a plot can be represented as an MED curve as shown in the right figure of Figure 2. When collecting data for multiple beds, the optimal % count for all beds is selected from the MED curves of each bed and applied to all beds.

[0038] Alternatively, the following additional criteria can be applied to select the optimal count for each bed: (a) Detect the maximum value of the respiratory motion (the MED curve showing the most rapid change) from among the MED curves of all beds. (b) The user can specify multiple beds, and the algorithm selects the optimal count for each bed based on the specified criteria. (c) Use automatic anatomical identification techniques (such as AI-based segmentation) to identify the optimal count for each bed (such as identifying the lung region, liver region, etc.), and select the optimal count for each region based on the anatomical motion of one or more organs.

[0039] A specific example of the above process is shown in FIG. 2. As described above, the count optimization method shown in FIG. 2 reduces blurring due to movement by detecting a sharp transition of the MED from the exhalation phase to the inhalation phase. The % count is optimized by modeling the MED as a function of weights (Wt, A, B) derived from the MED curve (with upper and lower limits of the count).

[0040] The "preprocessing module for device-free gating reconstruction" shown in FIG. 2 corresponds to the control function 125a and executes steps S1 to S4. Also, the "automatic % count module" shown in FIG. 2 corresponds to the decision function 125b and executes step S5. Also, the "device-free gating reconstruction module" shown in FIG. 2 corresponds to the reconstruction function 125c and executes step S6.

[0041] First, the processing circuit 125 performs a nuclear medicine scan and collects list mode data (step S1).

[0042] Next, the processing circuit 125 performs short-term image reconstruction on the list mode data (step S2). For example, the processing circuit 125 can capture the variation of respiratory motion and set a period of such length that it can exclude the motion of the heart, and reconstruct mini-frame images for each mini-frame corresponding to the period.

[0043] Next, the processing circuit 125 encodes the movement by respiratory gate movement analysis using a latent vector (step S3).

[0044] That is, the processing circuit 125 quantifies the respiratory movement in each respiratory phase by generating a latent vector from the mini-frame image reconstructed in step S2. Although not particularly limited to the type of encoder, as an example, an autoencoder based on principal component analysis (PCA) and a neural network that performs waveform analysis can be used. The latent vector is an example of a feature vector.

[0045] Next, the processing circuit 125 performs Euclidean distance analysis using the latent vectors over a plurality of respiratory cycles (step S4).

[0046] For example, the processing circuit 125 extracts the latent vectors corresponding to the rest phases of each of the plurality of respiratory cycles and calculates a reference phase vector by averaging them. Further, the processing circuit 125 calculates the Euclidean distance between the latent vector of each frame (i.e., each point of the respiratory waveform) and the reference phase vector.

[0047] The Euclidean distance is calculated based on, for example, the following formula (1). In formula (1), "lv" n " is the n-dimensional point of each latent vector (lv). "lv_ref" n " is the n-dimensional point of the reference phase vector (lv_ref). "k" is the number of dimensions of the latent vector.

[0048]

Equation

[0049] Next, the processing circuit 125 automatically determines the % count using the plot of "MEDvs% count" (step S5).

[0050] For example, the processing circuit 125 first sorts the Euclidean distances calculated in step S4 in ascending order. That is, the processing circuit 125 aligns the data based on the similarity of the respiratory motion.

[0051] Also, the processing circuit 125 calculates the MED for each % count. That is, the processing circuit 125 calculates the MED for a plurality of conditions in which the ratio of the count used for the reconstruction process is changed based on the Euclidean distance calculated in step S4. As a result, a plot of "MED vs % count" is obtained. More generally speaking, the MED as a function of the count is obtained. The plot of "MED vs % count", or the MED as a function of the count, is an example of the relationship between the ratio of the count used for the reconstruction process and the MED.

[0052] Here, the processing circuit 125 can determine the % center so that the MED is minimized for each % count. The % center is the phase center corresponding to the phase with the least movement (rest phase), and is the gating condition used in the reconstruction process of step S6.

[0053] Next, the processing circuit 125 performs fitting on the plot of "MED vs % count" to generate an MED curve. Thereby, the influence of local non-uniformity is reduced. Further, as also shown in FIG. 2, the processing circuit 125 determines "MED" based on the following formula (2). Auto_%Count ".

[0054]

Equation

[0055] Specifically, the processing circuit 125 first sets an upper limit value "High_Thr" and a lower limit value "Low_Thr". Next, the processing circuit 125 sets the value of the MED at the point corresponding to the upper limit value "High_Thr" on the MED curve to "MED" High_Thris specified as. Further, the processing circuit 125 determines the MED value of the point corresponding to the lower limit value "Low_Thr" on the MED curve as "MED Low_Thr ". Then, as shown in Equation (2), the processing circuit 125 can automatically determine the % count by calculating the weighted average of "MED High_Thr " and "MED Low_Thr ".

[0056] That is, according to Equation (2), the MED point "MED Auto_%Count " is determined, and the % count corresponding to "MED Auto_%Count " is automatically determined. The selection of the MED point can be optimized by adjusting the weights "Wt", "A", and "B".

[0057] "Auto %Count" is a condition determined for each bed. When there are multiple beds, the processing circuit 125 determines "Auto %Count" for each bed. For example, after the processing circuit 125 determines "Auto %Count" for an arbitrary bed, it determines whether there are still beds for which "Auto %Count" has not been determined. If there are, the above processing is repeated to determine "Auto %Count". On the other hand, when "Auto %Count" has been determined for all beds, the processing circuit 125 determines the % count (Final Auto %Count) applied to all beds.

[0058] For example, the processing circuit 125 selects the bed with the minimum "Auto %Count" as the "Final Auto %Count". Although the magnitude of the respiratory movement varies from bed to bed, the "Final Auto %Count" can also handle the bed with the most intense respiratory movement.

[0059] As described above, the processing circuit 125 can automatically set various respiratory gating conditions including % count and % center. Further, the processing circuit 125 executes data-driven device-less image reconstruction based on the set respiratory gating conditions (step S6) to generate a device-less gated image.

[0060] The series of processes shown in FIG. 2 are executed in a device-less manner, and the workflow is improved compared to device-based respiratory gating. Further, the series of processes shown in FIG. 2 can be automatically executed by the processing circuit 125 without requiring a user to select gating conditions. Thus, the nuclear medicine diagnostic apparatus 10 according to the embodiment can improve the workflow related to respiratory gating.

[0061] The inventors evaluated the image quality of five clinical data sets using the count optimization method described above. The conditions for the evaluation will be described with reference to FIGS. 3A to 3G. FIGS. 3A, 3B, 3C, 3D, and 3E show the "MED vs % count" curves for five clinical data sets collected on multiple beds. Further, FIG. 3F is a legend of the % count shown in the curves of FIGS. 3A to 3E. That is, in the curves of FIGS. 3A to 3E, the % count (Auto %Count) automatically determined for each bed is indicated by a star symbol. In the curves of FIGS. 3A to 3E, in FIG. 3, the % count (Final Auto %Count) applied to all beds is indicated by a dashed line. FIG. 3G shows an overview of the statistical information regarding the subject to be evaluated.

[0062] In addition, FIGS. 4A to 4F and FIG. 5 show the evaluation results. The plots in FIGS. 4A, 4B, 4C, 4D, and 4E show that for the liver, as the count increases, the SD (noise) decreases and the SNR increases. By count optimization, as shown in the table of FIG. 4F, a stable ratio of SD and SNR in the liver and the contrast (SUV) of the tumor with respect to the ungated image are achieved.

[0063] As shown in FIG. 5, for dataset 4, as the % count increases, the blurring due to movement increases and the SD of the liver decreases. The method according to the embodiment selects a % count (e.g., 60%) that achieves a balance between the IQ of the liver and the blurring due to movement.

[0064] By optimizing the count, compared with the ungated image, the blurring due to respiratory motion decreases, a stable ratio between the IQ of the liver and the IQ of the tumor is obtained, and the image quality is improved. Specifically, as shown in FIG. 4F, the IQ of the liver is "SD Auto / SD ungated = 0.19±0.02", "SNR Auto / SNR ungated =11.56±1.32", and the IQ of the tumor is "SUV max(Auto) / SUV max(ungated) =1.27±0.23", "SUV peak(Auto) / SUV peak(ungated) =1.17±0.1", "SUV Avg(Auto) / SUV Avg(ungated) =1.28±0.22", showing a stable ratio.

[0065] So far, the method for determining the % count using Equation (2) has been described, but the embodiment is not limited thereto.

[0066] As shown in FIG. 6, the processing circuit 125 may determine the % count by identifying a point on the MED curve where the distance D from the straight line L to the MED curve is maximized.

[0067] Specifically, the processing circuit 125 performs fitting on the plot of "MED vs % count" to generate the MED curve shown in FIG. 6. Further, the processing circuit 125 sets an upper limit value and a lower limit value for the % count. In the example of FIG. 6, "25%" is set as the lower limit value and "70%" is set as the upper limit value. Also, the processing circuit 125 sets a straight line L passing through the point corresponding to the upper limit value on the MED curve and the point corresponding to the lower limit value on the MED curve. Further, the processing circuit 125 calculates the distance between each point on the MED curve and the straight line L, and identifies the point on the MED curve where the distance is maximized. Then, the processing circuit 125 determines the % count indicated by the identified point as the gating condition.

[0068] As another example, as shown in FIG. 7, the processing circuit 125 may determine the % count by obtaining the second derivative of the MED curve.

[0069] Specifically, the processing circuit 125 performs fitting on the plot of "MED vs % count" to generate the MED curve shown in the left diagram of FIG. 7. Next, the processing circuit 125 obtains the first derivative curve shown in the middle diagram of FIG. 7 by performing the first differentiation on the fitted MED curve, and further applies curve fitting to the first derivative curve. Next, the processing circuit 125 obtains the second derivative curve shown in the right diagram of FIG. 7 by performing the second differentiation on the curve-fitted first derivative curve. Then, the processing circuit 125 determines the % count based on the position where the second derivative curve becomes "0". For example, the processing circuit 125 determines, as the gating condition, the value among the % count values with data that is closest to the position where the second derivative curve becomes "0". The % count determined in this way indicates the first inflection point of the MED curve, that is, the point where the MED curve begins to increase exponentially.

[0070] As described above, the processing circuit 125 can determine the % count in a plurality of ways. Here, the processing circuit 125 may select any one of the plurality of methods described above, or may combine the plurality of methods described above.

[0071] For example, the processing circuit 125 determines the % count in the manner shown in FIG. 6. Hereinafter, the % count determined by the method shown in FIG. 6 will also be referred to as "Count_bed_dist". "Count_bed_dist" is an example of a first ratio.

[0072] Also, the processing circuit 125 determines the % count in the manner shown in FIG. 7. Hereinafter, the % count determined by the method shown in FIG. 7 will also be referred to as "Count_bed_der". "Count_bed_der" is an example of a second ratio.

[0073] Next, the processing circuit 125 substitutes "Count_bed_dist" and "Count_bed_der" into the following formula (3) to obtain "MED" Auto_%Count ". Then, by determining the MED point called "MED" Auto_%Count ", the % count corresponding to "MED" Auto_%Count is automatically determined. That is, the processing circuit 125 can automatically determine the % count by obtaining the weighted average of the first ratio and the second ratio.

[0074]

Equation

[0075] The selection of the MED point can be optimized by adjusting the weights "Wt", "A", and "B". For example, when it is desired to select a midpoint between the method shown in FIG. 6 and the method shown in FIG. 7, weights such as "Wt = 0.5", "A = 1", and "B = 1" are set.

[0076] Although the method for automatically setting the % count and % center has been described, the processing circuit 125 can also perform settings for other conditions.

[0077] Examples of other conditions include the type and intensity of the post-filter. As an example, the processing circuit 125 can automatically set a three-dimensional Gaussian filter with a full-width at half maximum (FWHM) of "5 mm" and a CaLM (Clear adaptive Low-noise Method) of "strong".

[0078] Also, examples of other conditions include the reconstruction parameters when performing reconstruction processing by ordered subsets expectations maximization (OSEM). The reconstruction parameters of OSEM are, for example, the number of iterations and the number of subsets.

[0079] Also, examples of other conditions include the reconstruction parameters when performing AI-based reconstruction processing. The reconstruction parameters in AI-based reconstruction processing are, for example, network parameters and filter intensity.

[0080] The above various conditions are adjusted according to the subject's BMI, scan dose (injection dose and scan delay time), scan period, etc., so as to obtain optimal image quality.

[0081] For example, the processing circuit 125 cumulatively quantifies the scan dose, acquisition time, BMI, etc. For example, the processing circuit 125 can use the Cumulative Dose Product (CDP), which is the product of concentration and time period, for quantifying the scan dose and acquisition period. Also, the processing circuit 125 can use a function of CDP and BMI (for example, "f(CDP,BMI)=CDP / BMI") as an index. For example, when "f(CDP,BMI)" is small (i.e., there is a lot of noise), the gating and reconstruction parameters are optimized to perform stronger noise suppression.

[0082] The term "processor" used in the above description means, for example, circuits such as a CPU, GPU, ASIC, programmable logic device (for example, simple programmable logic device (SPLD), complex programmable logic device (CPLD), and field programmable gate array (FPGA)). When the processor is, for example, a CPU, the processor realizes its function by reading and executing a program stored in the storage circuit. On the other hand, when the processor is, for example, an ASIC, instead of storing a program in the storage circuit, the function is directly incorporated as a logic circuit in the circuit of the processor. Note that each processor in the embodiment is not limited to being configured as a single circuit for each processor, and a plurality of independent circuits may be combined to be configured as one processor to realize its function. Further, a plurality of components in each figure may be integrated into one processor to realize its function.

[0083] Each component of each device according to the above-described embodiments is a functional concept, and does not necessarily have to be physically configured as shown in the drawings. That is, the specific forms of distribution and integration of each device are not limited to those shown in the drawings, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads, usage situations, etc. Further, each processing function performed by each device can be realized by all or any part of it being realized by a CPU and a program analyzed and executed by the CPU, or can be realized as hardware by wired logic.

[0084] Also, the method described in the above-described embodiments can be realized by executing a pre-prepared program on a computer such as a personal computer or a workstation. This program can be distributed via a network such as the Internet. Further, this program can be recorded on a non-transitory recording medium readable by a computer such as a hard disk, a flexible disk (FD), a CD-ROM, an MO, a DVD, etc., and can also be executed by being read from the recording medium by a computer.

[0085] According to at least one of the embodiments described above, the workflow related to respiratory gating can be improved.

[0086] Although several embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, changes, and combinations of embodiments can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0087] 10: Nuclear medicine diagnostic device 110: Gantry device 112: Front-end circuit 113: Ceiling 114: Bed 116: Bed drive unit 120: Console device 121: Communication interface 122: Input interface 123: Display 124: Memory 125: Processing circuit 125a: Control function 125b: Decision function 125c: Reconfiguration function 130: Detector

Claims

1. A control unit that generates a feature vector for each mini-frame from list mode data collected by performing a nuclear medicine scan, and calculates the Euclidean distance between the feature vector of each frame and a reference phase vector; A determination unit that calculates an average Euclidean distance for a plurality of conditions in which the ratio of the counts used for the reconstruction process is changed based on the Euclidean distance, and determines the ratio based on the relationship between the ratio and the average Euclidean distance; A reconstruction unit that generates a gated image by performing image reconstruction using the determined ratio as a gating condition A nuclear medicine diagnostic apparatus comprising:

2. The nuclear medicine diagnostic apparatus according to claim 1, wherein the relationship is an MED plot in which the value of the average Euclidean distance is plotted against the ratio.

3. The nuclear medicine diagnostic apparatus according to claim 2, wherein the determination unit determines the ratio based on an MED curve obtained by curve fitting the MED plot.

4. The nuclear medicine diagnostic apparatus according to claim 3, wherein the determination unit sets an upper limit value and a lower limit value for the ratio, and obtains a weighted average of the value of the average Euclidean distance of the point corresponding to the upper limit value on the MED curve and the value of the average Euclidean distance of the point corresponding to the lower limit value on the MED curve, thereby determining the ratio.

5. The nuclear medicine diagnostic apparatus according to claim 3, wherein the determination unit sets an upper limit value and a lower limit value for the ratio, and identifies a point on the MED curve at which the distance from the straight line passing through the point corresponding to the upper limit value on the MED curve and the point corresponding to the lower limit value on the MED curve is maximized, thereby determining the ratio.

6. The nuclear medicine diagnostic apparatus according to claim 3, wherein the determination unit determines the ratio by obtaining a second derivative of the MED curve.

7. The nuclear medicine diagnostic apparatus according to claim 3, wherein the determination unit sets an upper limit value and a lower limit value for the ratio, and identifies a point on the MED curve at which the distance from the straight line passing through the point corresponding to the upper limit value on the MED curve and the point corresponding to the lower limit value on the MED curve is maximized, thereby determining a first ratio, determines a second ratio by obtaining a second derivative of the MED curve, and obtains a weighted average of the first ratio and the second ratio, thereby determining the ratio.

8. The nuclear medicine diagnostic apparatus according to claim 1, wherein the control unit generates a mini-frame image obtained by reconstructing the list mode data for each mini-frame, and generates the feature vector from the mini-frame image.

9. The nuclear medicine diagnostic apparatus according to claim 8, wherein the feature vector is a latent vector obtained by encoding the mini-frame image.

10. The nuclear medicine diagnostic apparatus according to claim 1, wherein the control unit generates the reference phase vector by averaging the feature vectors corresponding to the stationary phases.

11. Generating a feature vector for each mini-frame from list mode data collected by performing a nuclear medicine scan, calculating the Euclidean distance between the feature vector of each frame and the reference phase vector, Calculating an average Euclidean distance for a plurality of conditions in which the ratio of the count used for the reconstruction process is changed based on the Euclidean distance, determining the ratio based on the relationship between the ratio and the average Euclidean distance, Performing image reconstruction using the determined ratio as a gating condition to generate a gated image A method including this.

Images