PET data-driven cardiac rest period extraction

Abstract

A data-driven approach is provided for identifying periods of no motion (or rest) for imaging the heart. This timing information is not readily available from a conventional electrocardiogram (ECG) trace. The approach is based on beat-by-beat identification of periods of no motion.

Description

Technical Field

[0001] This disclosure generally relates to the field of nuclear imaging systems, and more particularly to a novel method for imaging the heart. Background Technology

[0002] refer to Figure 1A-1B , Figure 1A An external electrocardiogram (ECG) and left ventricular volume plot are shown during one cardiac cycle at rest; and Figure 1B The diagram shows an ECG and left ventricular volume plot during one cardiac cycle under stress. When the subject is at rest, the diastolic phase is typically longer than the systolic phase, while the opposite can be observed when the subject is under stress, as the diastolic phase is significantly shorter than the systolic phase. The duration of systole is relatively constant during both rest and stress conditions, while the diastolic phase is significantly shortened during cardiac stress. Figure 1A and 1B In the ECG, LV stands for left ventricle. The P wave, Q wave, R wave, S wave, T wave, and U wave components of the ECG cycle are... Figure 1A and 1B The resting period of the heart is indicated by a double-headed arrow.

[0003] Imaging of the systolic phase of the heart is typically performed by gating the acquisition or reconstruction with a reference ECG device. Conventional ECG gating simply divides each heartbeat into a predetermined number of equal phases. In some cases, it may be desirable to image the heart during the resting phase when the myocardium is relatively at rest (mid- to late diastole), making the resulting images less susceptible to motion artifacts. However, existing PET-based cardiac imaging methods do not provide a useful method for obtaining PET images of the heart during its resting phase. Some recent efforts in the industry have demonstrated the ability to discern heartbeats directly from PET data using the centroid method. However, such methods result in relatively coarse signals because the centroid method primarily captures the apex-base motion of the heart. Much of the signal from this alternative trace of cardiac motion is masked by the symmetry of the heart's shape and systolic contractility. This is because the short axis of the heart includes the bulge and the contractility is relatively homogeneous, so the centroid does not change much during the cardiac cycle.

[0004] Therefore, there is a desire for an improved means of generating PET scan images of the heart during its resting period. Summary of the Invention

[0005] This disclosure provides a data-driven method for identifying non-motor (resting) periods of the heart for cardiac imaging.

[0006] The method disclosed herein can include: (a) performing a PET scan on a patient and generating list-pattern PET scan data including the patient's heart; (b) extracting tissue image data at high temporal resolution; (c) cropping the tissue image data to isolate the tissue image data of the ventricles of the patient's heart using bounded boxes extracted from reconstructed PET scan data segmented for the ventricles of the patient's heart; (d) identifying cardiac anatomical coordinates; (e) extracting the temporal process of the projection along a cardiac axis (which may be a radial and / or longitudinal cardiac axis); (f) extracting the normalized temporal process of the projection along the cardiac axis (radial and / or longitudinal); and (g) defining a threshold to determine the resting period of the heart.

[0007] According to some other embodiments, the method of this disclosure applied to obtain resting images of heart valves can include: (a) performing a PET scan on a patient and generating list-pattern PET scan data including the patient's heart; (b) extracting tissue image data at high temporal resolution; (c) cropping the tissue image data to isolate the tissue image data of the heart valves of the patient's heart using bounding boxes extracted from reconstructed PET scan data segmented for the heart valves of the patient's heart; (d) identifying cardiac anatomical coordinates; (e) extracting the temporal process of the projection along the longitudinal cardiac axis; (f) extracting the normalized temporal process of the projection along the longitudinal cardiac axis; and (g) defining a threshold to determine the resting period of the heart.

[0008] This disclosure also provides an imaging system utilizing the methods disclosed herein. Attached Figure Description

[0009] The features of the embodiments described herein will be more fully disclosed in the following detailed description, which should be considered together with the accompanying drawings, in which similar numbers refer to similar parts.

[0010] Figure 1A The ECG and left ventricular volume plots are shown for one cardiac cycle under resting conditions, showing the systolic and diastolic phases.

[0011] Figure 1B The ECG and left ventricular volume plots are shown for one cardiac cycle under stress conditions, showing the systolic and diastolic phases.

[0012] Figure 2 This is a flowchart depicting the method of this disclosure.

[0013] Figure 3A This is a diagram showing the radial projection through the left ventricular wall during end-diastole.

[0014] Figure 3B This is a diagram showing the radial projection through the left ventricular wall at the end of systole.

[0015] Figure 4A Indicates the distance along the polar angle Figure 3A The image intensity of each radial projection is shown in the figure.

[0016] Figure 4B Indicates the distance along the polar angle Figure 3B The image intensity of each radial projection is shown in the figure.

[0017] Figure 5 It is a plot based on the normalized image intensity of the cardiac cycle.

[0018] Figure 6 It is a diagram showing the longitudinal projection of the valves (such as the aortic valve) in the ventricles of the heart.

[0019] Figure 7 This is an illustration of an imaging system according to an embodiment of the present disclosure. Detailed Implementation

[0020] The description of exemplary embodiments is intended to be read in conjunction with the accompanying drawings, which are considered an integral part of the entire written description. The drawings are not necessarily drawn to scale, and for clarity and brevity, certain features may be enlarged to scale or shown in some schematic form. In this description, related terms such as “horizontal,” “vertical,” “up,” “down,” “top,” and “bottom,” and their derivatives (e.g., “horizontally,” “downward,” “upward,” etc.) should be interpreted as referring to orientations as shown in the drawings described thereafter or discussed. These relative terms are for convenience of description and are not normally intended to require a specific orientation. Where appropriate, terms including “inward” relative to “outward,” “longitudinal” relative to “lateral,” and the like, should be interpreted relative to each other or relative to an axis of extension or rotation or a center. Terms relating to attachment, coupling, and the like, such as “connected” and “interconnected,” refer to a relationship in which structures are directly or indirectly secured or attached to each other by intervening structures, and both movable and rigid attachments or relationships, unless otherwise expressly described. The term “operably connected” refers to an attachment, coupling, or connection that allows related structures to operate as intended by means of that relationship.

[0021] A data-driven method is provided for identifying resting periods of the heart based on PET scan data, allowing PET scan data corresponding to these resting periods to be selectively used to reconstruct images of the heart. The resulting PET scan images of the heart are minimally affected by the heart's contractile motion.

[0022] Data-driven measurements of cardiac contractility are similar to ventricular volume curves and provide detail about the resting phase, which is not readily apparent in conventional ECG traces. For ECGs, for example, the action potentials during left ventricular systole comprise the primary signal for gating, and there is no suitable ECG signature to pinpoint the resting phase. Data-driven signals, on the other hand, will be based solely on cardiac motion that should be more readily apparent during the resting phase. All counts obtained during the resting phase of each individual heartbeat can be included in the reconstruction of cardiac images for a single motion freeze. The inventive techniques disclosed herein can also be employed in PET scans where ECG gating is not typically performed.

[0023] Figure 2 A flowchart 100 is provided, which illustrates an imaging method according to some embodiments of the present disclosure.

[0024] Using imaging of the left ventricle of the heart as an example, the concept of the disclosed invention will be described; however, those skilled in the art will understand that the disclosed method can be applied to imaging different parts of the heart. For example, the method of the present invention can be applied to define the resting periods of the right ventricle, left atrium, and right atrium.

[0025] Furthermore, the disclosed method can be applied to define the resting period of valves in the heart. Details on how to apply the disclosed method to valves will be discussed below in the discussion of the example of the left ventricle.

[0026] Referring back to flowchart 100, the disclosed method includes the procedures identified in portions A and B of flowchart 100. The procedures identified in portion A are performed using reconstructed image data from PET, CT, or MR scans. The procedures identified in portion B are performed based on raw PET scan data of the heart (i.e., list-pattern data). The output from the procedures in portion A is utilized by some procedures in portion B, which ultimately result in identifying the resting period of the heart and selecting only the tissue image data of the heart corresponding to the resting period.

[0027] In part A of flowchart 100, starting with a reconstructed PET, CT, or MR scan image 110 including the subject's heart, the reconstructed image of the region of interest (in this example, the left ventricle) is segmented, see step 111. Then, a bounded box defining the region surrounding the region of interest (e.g., the left ventricle) is extracted from the segmented reconstructed image, see step 112. Next, an appropriate cardiac anatomical coordinate system is identified for the region of interest. For this example of the left ventricle, radial (and / or longitudinal) cardiac vectors are extracted for the left ventricle. r nThe array is shown in step 113. For the radial cardiac vector, polar coordinates will be defined for the region of interest (left ventricle in this example). As discussed below, for imaging heart valves, a more appropriate cardiac anatomical coordinate system could be primarily the longitudinal cardiac vector. See [link to relevant documentation]. Figure 3A and 3B The stylized diagram shows the radial cardiac vectors superimposed on cross-sectional views of the left ventricle at end-diastole and end-systole, respectively. r n An array. In other words, Figure 3A and 3B The view shown is a slice taken through the long cardiac axis (i.e., the longitudinal axis) of the left ventricle. At end-diastole, the myocardium 10 is relaxed and therefore thinner than at end-systole.

[0028] In part B of flowchart 100, starting with the raw PET scan data (list-mode data) of heart 120, tissue image data is extracted at high temporal resolution, see step 130. The high time-of-flight (TOF) performance of modern digital PET scanners enables the direct generation of tissue images by accumulating event locations in the center of the TOF window. These tissue images are non-quantitative but can be generated very quickly without performing conventional reconstruction.

[0029] Next, since cardiac motion can include the patient's breathing and / or bulk motion, optional preprocessing of the tissue image data can be performed to remove any respiratory and / or bulk motion of the patient, see step 135. This ensures that only the motion component associated with cardiac contractility remains.

[0030] Next, using the bounding box extracted in step 112 of part A, the tissue image data is cropped to isolate the tissue image data of the left ventricle of the scanned heart, see step 140.

[0031] Next, the radial cardiac vector identified and extracted in step 113 of part A is used. r n For the array, extract the time process of the radial projection, see step 150. The term "time process" here refers to the time sequence of events. See also... Figure 4A and 4B Plotting the polar angle relative projection image intensity at the end of diastole and the end of systole. Figure 4A and 4B Each represents the image intensity at a given time, projected along each radial direction according to the polar angle ϕ. Therefore, at end-diastole, for each slice taken at a given time along the long cardiac axis of the left ventricle, similar to... Figure 4AA plot of the diagram shown is generated. Similarly, at end-systole, for each slice taken along the long cardiac axis of the left ventricle at a given time, a similar process is performed. Figure 4B One plot of the diagram shown is generated. The number of slices taken along the long cardiac axis can be optimized to provide meaningful data with minimal computational time burden.

[0032] Next, the normalized time process of the projection along the cardiac axis is extracted, see step 160. This produces Figure 5 The intensity position of the resulting projected image is plotted relative to the cardiac cycle. In other words, Figure 5 The drawing in the figure is a global plot of the time process of radial projections onto all slices taken along the long cardiac axis of the left ventricle. This drawing is a substitute for ventricular volume.

[0033] Here, I will take a moment to consider how the disclosed method can be applied to image acquisition of heart valves rather than ventricles. When applied to valves such as the aortic valve, the array of cardiac vectors representing valve motion used to extract the time-series projection will be longitudinal vectors rather than radial vectors, because the direction of valve motion is substantially orthogonal to the direction of myocardial motion in the ventricles. This is in Figure 6 It is shown in the figure. Figure 6 The heart valve 15 in ventricle V is shown. Vector L n The array represents the motion of valve 15, and the vector r n The array represents the contractile motion of the myocardium in the ventricle. Vector L n Basically in the longitudinal direction along the longitudinal axis of the ventricle V, while the vector r n Basically in the radial direction relative to the longitudinal axis of the ventricle V.

[0034] Next, a threshold T is defined to determine the boundaries of the resting period. The threshold T is a free parameter that can be optimized to maximize the amount of signal that will be retained as part of the resting cycle of the heartbeat, while not overdefining the length of the resting period to minimize the amount of signal from non-resting periods that is excluded from being defined as part of the resting period.

[0035] like Figure 5 As shown, the threshold T defines time points t1 and t2, and time points t1 and t2 define the resting period. Therefore, in Figure 5The time interval between t1 and t2 in the cardiac cycle shown is the non-resting period, and tissue image signals from this part of the cardiac cycle will be discarded. Only tissue image signals from the resting period will be used to generate the final reconstructed image of the heart. For example, the threshold T can be set as a certain percentage (%) of the tissue image signal from the heartbeat.

[0036] exist Figure 2 As summarized in flowchart 100, processes 150, 160, 170, and 180 in part B will be repeated for each heartbeat. In other words, processes 150, 160, 170, and 180 are executed for the tissue image dataset corresponding to each heartbeat.

[0037] Figure 5 The resulting plot shown is an alternative trace of the cardiac cycle derived directly from PET tissue images.

[0038] Still referencing Figure 2 In some embodiments, the method of this disclosure may include: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart, see box 120; (b) Extract tissue image data at high temporal resolution, see box 130; (c) Using bounded boxes extracted from reconstructed PET scan data segmented from the ventricles of the patient’s heart, the tissue image data is cropped to isolate the tissue image data of the ventricles of the patient’s heart, see box 140; (d) Identify cardiac anatomical coordinates, see box 113; (e) Extract the time process of the projection along the cardiac axis (which can be the radial and / or longitudinal cardiac axis), see box 150; (f) Extract the normalized time process of the projection (radial and / or longitudinal) along the cardiac axis, see box 160; and (g) Define thresholds to determine the resting period of the heart, see box 170.

[0039] In some embodiments, the method further includes selecting only tissue image data corresponding to the resting period, see box 180.

[0040] As discussed above, in some embodiments, step (d) includes defining an array of radial and / or longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the left ventricle of the patient's heart.

[0041] Typically, tissue image data can be very noisy. Therefore, in some embodiments, step (b) can include applying a known denoising process. In some embodiments, such a denoising process can be a neural network that denoises and improves the quality of the tissue image data. In some embodiments, the neural network can be a neural network trained using pairs of tissue images and fully reconstructed images.

[0042] According to some other embodiments, the inventive method described above, which uses the left ventricle as an example, is equally applicable to the right ventricle.

[0043] According to some other embodiments, the inventive method described above using the left ventricle as an example is equally applicable to heart valves. The method of this disclosure applied to obtain resting images of heart valves can include: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart; (b) Extract tissue image data with high temporal resolution; (c) Using bounding boxes extracted from reconstructed PET scan data of heart valves segmented from the patient’s heart, the tissue image data is cropped to isolate the tissue image data of the heart valves of the patient’s heart. (d) Identify cardiac anatomical coordinates; (e) Extracting the time process of the projection along the longitudinal cardiac axis; (f) Extracting the normalized time process of the projection along the longitudinal cardiac axis; and (g) Define a threshold to determine the resting period of the heart.

[0044] In some embodiments, the method for obtaining resting images of heart valves may further include (h) selecting only tissue image data corresponding to the resting period for image reconstruction.

[0045] In some embodiments, the heart valve can be an aortic valve, a pulmonary valve, a tricuspid valve, or a mitral valve.

[0046] In some embodiments, step (d) includes defining an array of longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the ventricles of a patient's heart.

[0047] In some embodiments, step (b) includes using a neural network to improve the quality of tissue image data.

[0048] In some embodiments, the neural network is trained using pairs of tissue images and fully reconstructed images.

[0049] This disclosure also provides an imaging system 200 utilizing the methods summarized in flowchart 100. Such an imaging system 200 can include: PET scanner 230, which scans the subject's heart to generate image data of the heart; Memory 252, provided in controller 250, stores instructions thereon; and Processor 253, provided in controller 250, is configured to read instructions to: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart, see box 120 of flowchart 100; (b) Extract tissue image data at high temporal resolution, box 130 of flowchart 100; (c) Cropping tissue image data to isolate tissue image data of the ventricles of the patient’s heart, see box 140 of flowchart 100; (d) Identify cardiac anatomical coordinates; (e) Extract the time process projection along the cardiac axis (which can be the radial and / or longitudinal cardiac axis), see box 150 of flowchart 100; (f) Extract the normalized time process for all projections (radial and / or longitudinal) along the cardiac axis, see box 160 of flowchart 100; and (g) Define the threshold for the rest period, see box 170 of flowchart 100.

[0050] In some embodiments of the imaging system 200, the instructions also include instructions for causing the processor 253 to select only tissue image data corresponding to the resting period, see block 180 of flowchart 100. Such an imaging system 200 will also include some basic components such as a patient response unit 210 and a patient tunnel 220.

[0051] The following is a list of non-limiting illustrative embodiments disclosed herein: Illustrative Example 1. A method comprising: (a) performing a PET scan on a patient and generating list-pattern PET scan data including the patient's heart; (b) extracting tissue image data at high temporal resolution; (c) cropping the tissue image data to isolate the tissue image data of the ventricles of the patient's heart using bounded boxes extracted from reconstructed PET scan data segmented for the ventricles of the patient's heart; (d) identifying cardiac anatomical coordinates; (e) extracting the temporal process of the projection along the cardiac axis; (f) extracting the normalized temporal process of the projection along the cardiac axis; and (g) defining a threshold to determine the resting period of the heart.

[0052] Illustrative Example 2. The method according to Illustrative Example 1 further includes (h) selecting only tissue image data corresponding to the resting period for image reconstruction.

[0053] Illustrative Example 3. In the method according to any one of the foregoing illustrative examples, the central ventricle is the left ventricle.

[0054] Illustrative Example 4. The method according to any one of the foregoing illustrative Examples 1 and 2, wherein the central ventricle is the right ventricle.

[0055] Illustrative Example 5. The method according to one of the foregoing illustrative examples, wherein step (d) includes defining an array of radial and / or longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the ventricles of a patient's heart.

[0056] Illustrative Example 6. The method according to one of the foregoing illustrative examples, wherein step (b) includes using a neural network to improve the quality of tissue image data.

[0057] Illustrative Example 7. According to the method of Illustrative Example 6, the neural network is trained using pairs of tissue images and fully reconstructed images.

[0058] Illustrative Example 8. According to one of the foregoing illustrative examples, tissue image data is preprocessed to remove respiratory, bodily, and other motion components whose temporal frequency is lower than the time frequency defining cardiac contractility.

[0059] Illustrative Example 9. An imaging system includes a PET scanner that scans the heart of a subject to generate image data of the heart; a memory storing instructions thereon; and a processor configured to read instructions to: (a) perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart; (b) extract tissue image data at high temporal resolution; (c) crop the tissue image data to isolate tissue image data of the ventricles of the patient's heart using bounded boxes extracted from reconstructed PET scan data segmented for the ventricles of the patient's heart; (d) identify cardiac anatomical coordinates; (e) extract the temporal process of the projection along the cardiac axis; (f) extract the normalized temporal process of the projection along the cardiac axis; and (g) define a threshold to determine the resting period of the heart.

[0060] Illustrative Example 10. The imaging system according to Illustrative Example 9, wherein the instructions further include instructions for causing the processor to select only tissue image data corresponding to the resting period.

[0061] Illustrative Example 11. In the imaging system according to any one of the foregoing illustrative Examples 9 to 10, the central ventricle is the left ventricle.

[0062] Illustrative Example 12. In the imaging system according to any one of the foregoing illustrative Examples 9 to 10, the central ventricle is the right ventricle.

[0063] Illustrative Example 13. An imaging system according to any one of the foregoing illustrative Examples 9 to 12, wherein step (d) includes defining an array of radial cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the ventricles of a patient's heart.

[0064] Illustrative Example 14. An imaging system according to any one of the foregoing illustrative Examples 9 to 13, wherein step (b) includes using a neural network to improve the quality of tissue image data.

[0065] Illustrative Example 15. An imaging system according to Illustrative Example 14, wherein the neural network is trained using pairs of tissue images and fully reconstructed images.

[0066] Illustrative Example 16. The imaging system according to Illustrative Example 9, wherein tissue image data is preprocessed to remove respiratory, bodily, and other motion components whose temporal frequency is lower than the time frequency defining cardiac contractility.

[0067] Illustrative Example 17. A method comprising: (a) performing a PET scan on a patient and generating list-pattern PET scan data including the patient's heart; (b) extracting tissue image data at high temporal resolution; (c) cropping the tissue image data to isolate the tissue image data of the patient's heart valves using bounding boxes extracted from reconstructed PET scan data segmented for the heart valves of the patient's heart; (d) identifying cardiac anatomical coordinates; (e) extracting the temporal process of the projection along the longitudinal axis; (f) extracting the normalized temporal process of the projection along the longitudinal axis; and (g) defining a threshold to determine the resting period of the heart.

[0068] Illustrative Example 18. The method according to Illustrative Example 17 further includes (h) selecting only tissue image data corresponding to the resting period for image reconstruction.

[0069] Illustrative Example 19. The method according to any one of the foregoing illustrative Examples 17 and 18, wherein the heart valve is an aortic valve, a pulmonary valve, a tricuspid valve, or a mitral valve.

[0070] Illustrative Example 20. The method according to any one of the foregoing illustrative Examples 17 to 19, wherein step (d) includes defining an array of longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the heart valves of a patient's heart.

[0071] Illustrative Example 21. The method according to any one of the foregoing illustrative Examples 17 to 20, wherein step (b) includes using a neural network to improve the quality of tissue image data.

[0072] Illustrative Example 22. According to the method described in the aforementioned illustrative Example 21, the neural network is trained using pairs of tissue images and fully reconstructed images.

[0073] It will be understood that the foregoing description is an exemplary embodiment of the invention, and the invention is not limited to the specific forms shown. Modifications may be made to the design and arrangement of the elements without departing from the scope of the invention.

Claims

1. A method comprising: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart; (b) Extract tissue image data with high temporal resolution; (c) Using bounding boxes extracted from reconstructed PET scan data segmented from the ventricles of the patient’s heart, the tissue image data is cropped to isolate the tissue image data of the ventricles of the patient’s heart; (d) Identify cardiac anatomical coordinates; (e) Extracting the time course of the projection along the cardiac axis; (f) Extract the normalized time process of the projection along the cardiac axis; as well as (g) Define a threshold to determine the resting period of the heart.

2. The method according to claim 1, further comprising (h) selecting only the tissue image data corresponding to the resting period for image reconstruction.

3. The method according to claim 1, wherein the ventricle is the left ventricle.

4. The method according to claim 1, wherein the ventricle is the right ventricle.

5. The method of claim 4, wherein step (d) comprises defining an array of radial and / or longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the ventricles of the patient's heart.

6. The method of claim 1, wherein step (b) comprises using a neural network to improve the quality of the tissue image data.

7. The method of claim 6, wherein the neural network is trained using paired tissue images and fully reconstructed images.

8. The method of claim 1, wherein the tissue image data is preprocessed to remove respiratory, bodily, and other motion components whose temporal frequency is lower than the time frequency defining cardiac contractility.

9. An imaging system, comprising: A PET scanner, which scans the subject's heart to generate image data of the heart; It contains a memory that stores instructions; as well as A processor configured to read the instructions to: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart; (b) Extract tissue image data with high temporal resolution;    (c) Using bounding boxes extracted from reconstructed PET scan data segmented from the ventricles of the patient’s heart, the tissue image data is cropped to isolate the tissue image data of the ventricles of the patient’s heart;    (d) Identify cardiac anatomical coordinates;    (e) Extracting the time course of the projection along the cardiac axis;    (f) Extract the normalized time process of the projection along the cardiac axis; as well as (g) Define a threshold to determine the resting period of the heart.

10. The imaging system of claim 9, wherein the instructions further include instructions for causing the processor to select only the tissue image data corresponding to the resting period.

11. The imaging system of claim 9, wherein the ventricle is the left ventricle.

12. The imaging system of claim 9, wherein the ventricle is the right ventricle.

13. The imaging system of claim 9, wherein step (d) includes defining an array of radial cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the ventricles of the patient's heart.

14. The imaging system of claim 9, wherein step (b) includes using a neural network to improve the quality of the tissue image data.

15. The imaging system of claim 14, wherein the neural network is trained using paired tissue images and fully reconstructed images.

16. The imaging system of claim 9, wherein the tissue image data is preprocessed to remove respiratory, bodily, and other motion components whose temporal frequency is lower than the time frequency defining cardiac contractility.

17. A method comprising: (a) Perform a PET scan on the patient and generate list-pattern PET scan data including the patient's heart; (b) Extract tissue image data with high temporal resolution; (c) Using bounding boxes extracted from reconstructed PET scan data segmented for the heart valves of the patient’s heart, the tissue image data is cropped to isolate the tissue image data of the heart valves of the patient’s heart; (d) Identify cardiac anatomical coordinates; (e) Extract the time process of the projection along the longitudinal axis; (f) Extract the normalized time process of the projection along the longitudinal axis; and (g) Define a threshold to determine the resting period of the heart.

18. The method of claim 17, further comprising (h) selecting only the tissue image data corresponding to the resting period for image reconstruction.

19. The method of claim 17, wherein the heart valve is an aortic valve, a pulmonary valve, a tricuspid valve, or a mitral valve.

20. The method of claim 17, wherein step (d) includes defining an array of longitudinal cardiac vectors extracted from reconstructed PET, CT, or MR scan data segmented for the heart valves of the patient's heart.

21. The method of claim 17, wherein step (b) includes using a neural network to improve the quality of the tissue image data.

22. The method of claim 21, wherein the neural network is trained using paired tissue images and fully reconstructed images.

Images