JP7828578B2 - Myocardial energy calculation method, myocardial energy calculation system, myocardial energy calculation device, and myocardial energy calculation program - Google Patents

Description

This invention relates to a method for obtaining information on myocardial blood flow and myocardial dynamics (myocardial movement) from image data representing myocardial blood flow obtained by PET (Positron Emission Tomography) devices, SPECT (Single Photon Emission Computed Tomography) devices, etc., which utilize nuclear medicine imaging technology, and calculating an index based on this information. Specifically, it relates to a method, system, apparatus, and program (hereinafter referred to as "myocardial energy calculation method, etc.") for calculating myocardial energy by obtaining quantitative values for myocardial blood flow and myocardial dynamics during the same examination, calculating a new index representing the kinetic energy of the myocardium (hereinafter referred to as "myocardial energy") from these quantitative values, and displaying the myocardial energy based on a segment model that divides the myocardium into multiple regions.

Nuclear medicine imaging devices such as PET and SPECT scanners can image (image) the physiological functions of living organisms, such as blood flow and metabolism in organs. For example, PET and SPECT scanners can administer a radioactive drug (radioactive agent) into the body, and then use special cameras to image the distribution of radiation emitted from the administered radioactive agent from various angles outside the body. This image is then reconstructed by a computer and converted into a tomographic image (for example, Non-Patent Document 1, Non-Patent Document 2).

Yasuhiro Wada, Fundamentals of Nuclear Medicine Technology: "Imaging Principles using PET Scanners," Clinical Nuclear Medicine Vol. 48 No. 1, 2015. URL: http://www.rinshokaku.com/contents/pdf/sec7/6.pdf (Retrieved December 3, 2021) Masaya Suda, Fundamentals of Nuclear Medicine Technology: "From SPECT Imaging to Image Processing," Clinical Nuclear Medicine Vol. 47 No. 3, 2014. URL: http://www.rinshokaku.com/contents/pdf/sec7/2.pdf (Retrieved December 3, 2021)

Previous nuclear medicine imaging devices lacked sufficient sharpness and temporal resolution, limiting their ability to analyze tissue morphology and dynamics. Therefore, it was difficult to simultaneously measure myocardial blood flow and myocardial dynamics from the same tomographic image of the heart. Consequently, despite their close relationship, myocardial blood flow and myocardial dynamics were measured using separate imaging tests. These measurements, obtained from separate imaging tests, did not strictly reflect the patient's condition at a specific point in time, and were used as independent indicators. Furthermore, the image data obtained from these separate tests differed in size and pixel count, requiring adjustments to match, making it difficult to precisely measure the patient's condition at a specific point in time.

While conventional imaging techniques have presented these challenges, recent advancements in nuclear medicine imaging technology have improved the quality of tomographic images, making it possible to obtain clearer outlines of organs such as the heart. Therefore, the present invention provides a method for calculating myocardial energy, which not only calculates blood flow values such as myocardial flow reserve (MFR) based on tomographic images of the heart (e.g., PET images) imaged by a nuclear medicine imaging device (e.g., a PET device), but also applies dynamic analysis techniques such as feature tracking to the tomographic images to calculate dynamic values such as the ratio of myocardial strain under vasodilation load (referred to here as "load strain") to myocardial strain at rest (referred to here as "resting strain") (referred to here as "myocardial strain ratio (MSR)"). This makes it possible to measure myocardial blood flow and myocardial dynamics simultaneously from the same tomographic image of the heart (images of the same size and number of pixels). Furthermore, by calculating myocardial energy, which is calculated by taking the square root of the sum of the squares of these values, it is possible to represent the kinetic energy related to the myocardium with a single indicator, myocardial energy, rather than evaluating myocardial blood flow and myocardial dynamics with separate indicators, thereby facilitating the evaluation of the state of the myocardium.

As one embodiment of the myocardial energy calculation method according to the present invention, the myocardial energy calculation method is:
The stage of acquiring tomographic images of the heart from a nuclear medicine imaging device,
The steps include: calculating the myocardial flow reserve (MFR), which represents the ratio of myocardial blood flow under load to resting blood flow, from the aforementioned tomographic image;
The steps include calculating the peak strain of the myocardium from the aforementioned tomographic image,
The steps include: calculating the myocardial strain ratio (MSR), which represents the ratio of load strain to rest strain of the myocardium, from the aforementioned peak strain;
The step includes calculating myocardial energy (ME), which represents the kinetic energy of the myocardium, based on the MFR and MSR.
The ME is characterized by being calculated as the square root of the sum of the squares of the MFR and the MSR.

In a preferred embodiment of the myocardial energy calculation method according to the present invention, the myocardial energy (ME) is calculated by setting A and B to arbitrary coefficients.
It is characterized by being expressed by the following formula.

In a preferred embodiment of the myocardial energy calculation method according to the present invention, the nuclear medicine imaging device is a PET device or a SPECT device.
The tomographic image is characterized by being a PET image acquired by the PET device or a SPECT image acquired by the SPECT device.

In a preferred embodiment of the myocardial energy calculation method according to the present invention, the tomographic image includes a plurality of cross-sectional images of the entire myocardium.
Each of the aforementioned plurality of cross-sectional images is characterized by including a predetermined number of consecutive images corresponding to the cardiac cycle.

A preferred embodiment of the myocardial energy calculation method according to the present invention is to perform the steps of calculating the MFR and the MSR before the steps of calculating the MFR.
The step of receiving input for determining multiple feature points along the endocardium recorded in a predetermined frame image from a predetermined number of consecutive images,
A step of tracking the plurality of feature points over a predetermined number of consecutive images using feature tracking technology,
The process involves generating a time strain curve with the cardiac cycle and the distance between two adjacent points as axes, for each of the distances between two adjacent points represented by two-dimensional coordinates among the aforementioned plurality of feature points,
The method is characterized by including the step of determining the peak strain in the time strain curve using the peak value as a representative value.

In a preferred embodiment of the myocardial energy calculation method according to the present invention, the load strain and the resting strain are characterized by calculating the endocardial length from the sum of the distances between the plurality of feature points, normalizing the endocardial length at end-diastole during the cardiac cycle, and using its peak value.

A preferred embodiment of the myocardial energy calculation method according to the present invention is characterized in that the ME includes values corresponding to each of the multiple regions of the myocardium, and is displayed by applying it to a segment model in which the myocardium is divided into the multiple regions.

A preferred embodiment of the myocardial energy calculation method according to the present invention is characterized in that the segment model includes segments corresponding to 16 regions of the myocardium.

As one embodiment of the myocardial energy calculation system according to the present invention, the myocardial energy calculation system is
Nuclear medicine imaging device,
Includes an information processing device,
The aforementioned nuclear medicine imaging device acquires and stores cross-sectional images of the heart.
The aforementioned information processing device is
The tomographic image is acquired from the nuclear medicine imaging device.
From the aforementioned tomographic images, the myocardial flow reserve (MFR), which represents the ratio of myocardial blood flow under load to resting blood flow, is calculated.
The peak strain of the myocardium is calculated from the aforementioned tomographic image.
From the aforementioned peak strain, the myocardial strain ratio (MSR), which represents the ratio of load strain to rest strain of the myocardium, is calculated.
Based on the MFR and MSR, the myocardial energy (ME), which represents the kinetic energy of the myocardium, is calculated.
The ME is characterized by being calculated as the square root of the sum of the squares of the MFR and the MSR.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the myocardial energy (ME) is calculated by setting A and B to arbitrary coefficients.
It is characterized by being expressed by the following formula.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the nuclear medicine imaging device is a PET device or a SPECT device.
The tomographic image is characterized by being a PET image acquired by the PET device or a SPECT image acquired by the SPECT device.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the tomographic image includes a plurality of cross-sectional images of the entire myocardium.
Each of the aforementioned plurality of cross-sectional images is characterized by including a predetermined number of consecutive images corresponding to the cardiac cycle.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the information processing device calculates the MFR and the MSR,
The system accepts input for determining multiple feature points along the endocardium recorded in a predetermined frame of a predetermined number of consecutive images.
Using feature tracking technology, the plurality of feature points are tracked across a predetermined number of consecutive images.
For each of the distances between two adjacent points, which are normalized and represented by their peak values, a time-strain curve is generated with the cardiac cycle and the distance between the two points as the axes.
The method is characterized by including determining the peak strain in the time strain curve using the peak value as a representative value.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the load strain and the resting strain are characterized by calculating the endocardial length from the sum of the distances between the plurality of feature points, normalizing the endocardial length at end-diastole during the cardiac cycle, and using its peak value.

In a preferred embodiment of the myocardial energy calculation system according to the present invention, the ME includes values corresponding to each of a plurality of regions of the myocardium,
The information processing device is characterized by displaying the ME by applying it to a segment model in which the myocardium is divided into the plurality of regions.

A preferred embodiment of the myocardial energy calculation system according to the present invention is characterized in that the segment model includes segments corresponding to 16 regions of the myocardium.

As one embodiment of the myocardial energy calculation device according to the present invention, the myocardial energy calculation device is
This method is characterized by performing each step of the myocardial energy calculation method described in any embodiment of the myocardial energy calculation method.

One embodiment of the myocardial energy calculation program according to the present invention is characterized in that the myocardial energy calculation program is executed by a computer, thereby causing the computer to function as the myocardial energy calculation device.

The myocardial energy calculation method according to the present invention calculates Myocardial Flow Rate (MFR) based on tomographic images, such as PET images of the heart, acquired using a nuclear medicine imaging device such as a PET scanner. Simultaneously, it calculates Myocardial Surgery Rate (MSR) using dynamic analysis techniques such as feature tracking on the same tomographic image. This allows for simultaneous measurement of myocardial blood flow and myocardial dynamics from a single tomographic image of the heart, enabling examinations that accurately reflect the patient's condition at a specific point in time. As a result, myocardial blood flow and myocardial dynamics can be measured from image data of the same size and pixel count, eliminating the need for adjustment operations to unify the size and pixel count of image data acquired from separate imaging examinations.

Furthermore, information regarding myocardial blood flow and myocardial dynamics (MFR, MSR) includes numerical data across the entire myocardial region and can be generalized by applying it to segment models that divide the myocardium into multiple regions, such as the conventional myocardial segment model in cardiac nuclear medicine (16 segments excluding the apical segment of the 17-segment model).

Furthermore, by calculating myocardial energy using the square root of the sum of the squares of MFR and MSR, it becomes possible to represent the kinetic energy related to the myocardium with a single indicator—myocardial energy—rather than evaluating myocardial blood flow and myocardial dynamics with separate indicators, making it easier to assess the state of the myocardium.

This figure shows an overview of a myocardial energy calculation system according to one embodiment of the present invention. This flowchart shows the flow of the myocardial energy calculation method related to one embodiment of the present invention. This figure shows an example of a myocardial blood flow image generated by a nuclear medicine imaging device. This figure shows an example of a model for segmenting the myocardium in cardiac nuclear medicine (myocardial segment model). This figure shows an example of how myocardial flow reserve (MFR), calculated from cross-sectional images of the heart (PET images) acquired by a PET scanner, is displayed on the myocardial segment model shown in Figure 4. This figure shows feature points determined from the endocardial contour specified by a pointer in a PET image of the heart at a specific time (any single frame). This diagram shows an overview of the mechanism for tracking feature points using feature tracking technology. This figure shows an example of tracking multiple feature points in a PET image. Figure 8 shows an example of a graph illustrating the results of tracking feature points in a PET image (time strain curve). This figure shows an example of the peak value of a time strain curve. This figure shows examples of displaying a myocardial segment model based on MFR calculated from a patient's PET images and a myocardial segment model based on myocardial energy (ME). This figure shows the results of evaluating the usefulness of myocardial energy using the Receiver Operating Characteristic (ROC) curve. This figure shows the results of evaluating the usefulness of myocardial energy using Kaplan-Meier curves.

An embodiment of the present invention will be described below with reference to the drawings. In all drawings illustrating the embodiments, the same reference numerals are generally used, and repeated descriptions of them will be omitted. The individual embodiments of the present invention are not independent but can be combined and implemented as appropriate.

Figure 1 shows an overview of a functional ischemia detection system according to one embodiment of the present invention. The functional ischemia detection system according to the present invention includes an image analysis device 100, a nuclear medicine imaging device 200 such as a PET device or SPECT device, and a data processing device 210. The nuclear medicine imaging device 200 may be housed in the same casing as the data processing device 210 and configured as an integral unit. The image analysis device 100 and the data processing device 210 are connected via a network N.

The image analysis device 100 and data processing device 210 have a hardware configuration similar to that of a typical computer (information processing device), and exemplified by hardware resources such as a CPU (Central Processing Unit), memory consisting of ROM (Read Only Memory) and RAM (Random Access Memory), a bus, an input/output interface, an input unit, an output unit, a storage unit, and a communication unit.

The CPU executes various processes according to the program stored in memory or the program loaded into memory from the storage unit. For example, the CPU can execute a program to make the computer function as the image analysis device of the present invention. Furthermore, at least some of the functions of the image analysis device can be implemented in hardware using application-specific integrated circuits (ASICs), etc. The same applies to the other data processing devices 210 of the present invention.

The memory also stores data necessary for the CPU to perform various processes. The CPU and memory are interconnected via a bus. An input/output interface is also connected to this bus. The input/output interface is connected to an input unit, an output unit, a storage unit, and a communication unit. The input unit consists of various buttons, a touch panel, or a microphone, and inputs various information according to instructions and operations from users of the image analysis device 100 and the data processing device 210. The output unit consists of a display or speaker, and outputs image data and audio data. The storage unit consists of semiconductor memory such as DRAM (Dynamic Random Access Memory) or a hard disk, and stores various data. The communication unit enables communication with other devices.

The image analysis device 100 can acquire and store cardiac tomographic image data (PET images or SPECT images) from, for example, a nuclear medicine imaging device 200 or a data processing device 210. In the embodiment shown in Figure 1, the image analysis device 100 can acquire tomographic image data transferred from the data processing device 210 connected to the nuclear medicine imaging device 200 via a network N such as a dedicated line or a public line.

Figure 2 shows a flowchart illustrating the flow of a myocardial energy calculation method according to one embodiment of the present invention. Each step of the myocardial energy calculation method shown in Figure 2 is performed, for example, by the image analysis device 100, which is an information processing device in the myocardial energy calculation system shown in Figure 1. First, the image analysis device 100 acquires a tomographic image of the heart from the nuclear medicine imaging device 200 (step S1). For example, the nuclear medicine imaging device 200 may store tomographic images of the chest (heart) of a subject, such as a patient, taken in chronological order in a data processing device 210, and transmit the tomographic images from the data processing device 210 to the image analysis device 100.

The nuclear medicine imaging device 200 can use, for example, a PET device, a SPECT device, etc., and the tomographic images can be PET images or SPECT images acquired from those devices. The tomographic image includes multiple cross-sectional images of the entire myocardium, i.e., images of each cross-section corresponding to each of the entire regions of the myocardium. Each of the multiple cross-sectional images includes a predetermined number of consecutive images corresponding to the cardiac cycle. For example, a PET device includes blood flow images (cross-sectional images) of approximately 15 slices of the entire myocardium, and the blood flow images of these cross-sections include a predetermined number of consecutive images (for example, generally 16 frames) corresponding to the cardiac cycle (whole cardiac phase). In this embodiment, if the size of the cross-sectional images included in the tomographic image is 128 pixels × 128 pixels, then blood flow image information of 128 pixels × 128 pixels × 15 slices × 16 frames can be obtained for both resting and stressed conditions, respectively (see Figure 3).

Figure 3 shows an example of myocardial blood flow images generated by a nuclear medicine imaging device. For comparison, Figure 3 shows stress and rest images of myocardial blood flow obtained from ammonia PET scans, categorized into normal, angina (ischemia-positive), and infarction. Figure 3 also shows normal stress and rest images of myocardial blood flow obtained from SPECT scans.

In ammonia PET scans, myocardial blood flow is depicted as a circular pattern in normal stress and rest images. In contrast, in stress and rest images of angina (ischemia-positive) and infarction, myocardial blood flow is not depicted as a circle, but rather as a partially missing circle (semicircular), allowing the area of reduced myocardial blood flow (indicated by the arrow in Figure 3) to be identified. Similarly, the state of myocardial blood flow can be identified in SPECT scans.

In the ammonia PET embodiment shown in Figure 3, the image analysis device 100 can collect blood flow image data in two sets: one at rest and one under vasodilator loading. That is, the image analysis device 100 can acquire blood flow image data of the entire myocardial region at rest (loaded blood flow image) and blood flow image data of the entire myocardial region under load (resting blood flow image) from the PET device. By dividing the blood flow obtained from the loaded blood flow image by the blood flow obtained from the resting blood flow image (calculating the blood flow obtained from the loaded blood flow image / the blood flow obtained from the resting blood flow image), the myocardial flow reserve (MFR) can be calculated.

Furthermore, the image analysis device 100 can, if necessary, extract MFR data of only the myocardial region from the MFR data obtained from the image data (for example, 128 pixels × 128 pixels × 15 slices × 16 frames), apply it to a conventional myocardial segment model (see Figure 4), and display it in color (in grayscale in Figure 5).

Here, Figure 4 shows an example of a myocardial segmentation model (myocardial segment model) in cardiac nuclear medicine. The example shown in Figure 4 is a typical 17-segment model (segment #17 corresponding to the apex, and segments #1 to #16 corresponding to the myocardial region (base, basal lateral wall, etc.)). The 17-segment model divides the myocardium into 17 segments: 6 segments for the base, 6 segments for the central region, and 4 segments for the apex in the short-axis tomographic view, and one additional segment from the central slice to the apex in the long-axis vertical view. Figure 4 shows the positional relationships of the anterior wall, lateral wall, posterior wall, and septum of the myocardium in the 17-segment model, as well as the positional relationships of the left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary artery (RCA).

Figure 5 shows an example of applying myocardial flow reserve (MFR), calculated from cross-sectional images of the heart (PET images) acquired by a PET scanner, to the myocardial segment model shown in Figure 4. In this way, by applying the MFR calculated from stress blood flow images/resting blood flow images to the myocardial segment model and displaying it in color (shown as grayscale in Figure 5), the state of myocardial blood flow can be easily understood.

Referring to the flowchart in Figure 2, the process for calculating the MFR described above corresponds to the processing after acquiring tomographic image data of the heart (after step S1) (steps S2 to S4). To calculate the MFR and the MSR (described later), the image analysis device 100 first receives input to specify multiple feature points in the cross-sectional image of the heart at a certain time point included in the tomographic image data, and determines the target of analysis (step S2).

Figure 6 shows feature points determined from the endocardial contour specified by a pointer in a PET image of the heart at a specific time (any single frame). For example, a user of the image analysis device 100 can use a pointing device such as a mouse to specify the endocardial contour from a cross-sectional image (PET image) at a certain time using a pointer. The image analysis device 100 then uses dynamic analysis techniques such as existing feature tracking techniques to determine multiple feature points (11 or 12 feature points in the example shown in Figure 6) from that contour. In this way, the object of analysis is determined. Alternatively, feature points can be determined by setting multiple points (for example, 5 to 7 points), including the start and end points and curved portions. The object of analysis can be set arbitrarily, and the object of analysis may be a region, etc. The object of analysis may also be automatically determined using image processing or image recognition using artificial intelligence. Alternatively, a user of the image analysis device 100 can use a pointing device such as a mouse to manually set multiple points on the endocardial contour in the PET image, and then use a density increase/gradient of about 10% to automatically connect the multiple points in a linear fashion and extract the endocardial contour.

After determining the object to be analyzed (stage S2), the image analysis device 100 tracks the object to be analyzed using feature-tracking technology (stage S3). Figure 7 shows an overview of the mechanism for tracking feature points using feature-tracking technology. The image analysis device 100 can apply feature-tracking technology to track feature points using template matching technology. As shown in Figure 7, the image analysis device 100 tracks the changes (movements) of feature points during the cardiac cycle from end-diastole through end-systole to mid-diastole. In template matching, feature points from the end-diastole cross-sectional image (PET image) are used as a template image. In each of the predetermined number of consecutive images corresponding to the cardiac cycle, feature points that substantially match the template image are searched for and tracked (Search Track), thereby enabling the tracking of feature points during the cardiac cycle.

The size of the template image and the search area in template matching can be set arbitrarily. For example, if the PET image size is 128 pixels x 128 pixels, the template image size can be set to 24 pixels x 24 pixels, and the search area to 32 pixels x 32 pixels. Note that Figure 7 shows template matching, search, and tracking for a single feature point for simplicity; however, the mechanism is similar for tracking multiple feature points.

For example, Figure 8 shows an example of tracking multiple feature points in a PET image. Template matching allows tracking of multiple feature points in each of the consecutive images from myocardial expansion to contraction, as shown in Figures 8(a) to (d). Referring to Figure 2, after tracking the analysis target (feature points, etc.) (after step S3), the image analysis device 100 can calculate myocardial flow reserve (MFR), an indicator representing myocardial blood flow, and myocardial strain ratio (MSR), which represents the ratio of myocardial load strain to resting strain, an indicator representing the state of myocardial dynamics (step S4).

MFR is a conventionally known index representing the ratio of myocardial blood flow under load to resting blood flow. As mentioned above, it is calculated by dividing the blood flow obtained from load blood flow images by the blood flow obtained from resting blood flow images (calculating the blood flow obtained from load blood flow images / blood flow obtained from resting blood flow images). An example of MFR for the entire myocardial region applied to a myocardial segment model (see Figure 4) is shown in Figure 5.

Myocardial strain is measured from blood flow image (cross-sectional image) data (e.g., 128 pixels × 128 pixels × 1 slice × 16 frames). This measurement is performed on all slices (e.g., 16 slices) to calculate the strain across the entire myocardial region. Using the example of feature point tracking in PET images shown in Figure 8, to measure myocardial strain, a time-strain curve can be generated by plotting the distance between two adjacent points (represented by two-dimensional coordinates) from among multiple feature points, with the cardiac cycle (time or number of frames) and the distance between the two points as the axes.

Figure 8 shows the tracking of 12 feature points. For example, if we define points 1 through 12 counterclockwise from a given feature point, we can calculate the distances of 12 line segments (distances 1, 2, 3...12) from the 2D coordinates of adjacent feature points. Figure 9 is a curve plotted with the sum of the distances of two adjacent points on the vertical axis and time (number of frames) on the horizontal axis for six regions. It quantifies the movement of each region of the myocardium from diastole to systole and back to diastole during the cardiac cycle. The upper part of Figure 9 reflects the circumferential and longitudinal movement of the myocardium. Similarly, the lower part of Figure 9 reflects the radial movement of the myocardium perpendicular to the circumferential and longitudinal directions.

Based on the time-strain curves generated in this manner, the image analysis device 100 calculates the peak value of myocardial strain (referred to here as "peak strain") from tomographic images such as PET images of the heart during vasodilation and rest. From this peak strain, it calculates the myocardial strain ratio (MSR), which represents the ratio of load strain to rest strain. In the example time-strain curve shown in Figure 9, the peak strain is determined using the peak value as a representative value in the time-strain curve. From this peak strain, the load strain and rest strain of the myocardium can be calculated.

Figure 9 is a graph showing an example of the results (time-strain curve) obtained by tracking feature points in the PET image shown in Figure 8. The horizontal axis of the graph represents the number of frames (corresponding to cardiac phase, or time), and the vertical axis represents the endocardial length calculated from the sum of the distances between feature points. The length is the length at end-diastole, normalized to a percentage of the length in each cardiac phase. As shown in Figure 9, changes in myocardial strain can be measured for each of the multiple feature points determined along the endocardium of the myocardium.

Figure 10 shows an example of the peak value of a time-strain curve. The myocardial strain value can be represented, for example, as the peak value in the time-strain curve. In other words, the image analysis device 100 draws a time-strain curve from the change in endocardial length throughout the cardiac cycle and uses its peak value as a representative value. In the example time-strain curve shown in Figure 10, the horizontal axis represents the cardiac cycle (R-R duration) (corresponding to time and frame count), and the vertical axis represents longitudinal strain (Longitudinal Strain).

Thus, from the calculated myocardial strain across the entire myocardial region, the MSR can be calculated by calculating load strain/resting strain, similar to how the MFR is calculated.

After calculating MFR and MSR (after step S4), as shown in Figure 2, the image analysis device 100 can calculate myocardial energy (ME), which represents the kinetic energy of the myocardium, based on MFR and MSR (step S5). ME is expressed by the square root of the sum of the squares of MFR and MSR. For example, myocardial energy (ME) can be expressed by setting A and B as arbitrary coefficients (e.g., 0.0 < A ≤ 1.0, 0.0 < B ≤ 1.0),
It can be expressed by the following formula. ME includes values corresponding to multiple regions of the myocardium.

After calculating the myocardial energy (ME) (after step S5), the image analysis device 100 can display a myocardial energy (ME) map (step S6). The image analysis device 100 can display the ME values in color (shading) as the myocardial energy (ME) map, for example, using a conventional myocardial segment model in cardiac nuclear medicine (16 segments excluding the apical segment of the 17-segment model) (see Figure 11(b)).

Figure 11 shows an example of a myocardial segment model based on MFR and a myocardial segment model based on ME, both calculated from PET images of a particular patient. This patient required percutaneous coronary intervention (PCI) approximately one year after the PET scan.

Figure 11(a) shows a myocardial segment model based on MFR, and Figure 11(b) shows a myocardial segment model based on ME (ME map). In the MFR-based grayscale display (color display), there is little unevenness in the grayscale display, and abnormal myocardial regions are not detected with the conventional PET index MFR < 2.0 (Figure 11(a)). However, in the ME-based grayscale display (color display), areas of decreased ME can be confirmed in segments #3, #9, #13, and #14 (see Figure 4), which correspond to the lighter areas in the grayscale display. Thus, myocardial abnormalities that cannot be detected with the conventional index MFR can be detected by using the new index ME. If the ME map had been available at the time of this PET scan, it is possible that the highly invasive PCI treatment one year later could not have been performed.

The results of evaluating the usefulness of myocardial energy (ME) are shown in Figures 12 and 13. First, Figure 12 shows the evaluation results of myocardial energy usefulness using ROC curves. The study included 95 patients with coronary artery disease (22 with a future risk of cardiac event). ME was calculated from each patient's PET images, and the entire myocardial region was fitted to a 16-segment model (see Figure 4; segment #17 was excluded from the calculation). The average ME of the 16 segments was defined as G-ME, and the average ME of the segments corresponding to the right coronary artery, left coronary artery, and anterior descending artery were defined as RCA-ME, LAD-ME, and LCX-ME, respectively. For comparison, MFR was calculated from each patient's PET images, and G-MFR, RCA-MFR, LAD-MFR, and LCX-MFR were calculated in the same way as ME.

As an evaluation method, the predictive ability of MFR < 2.0 and ME for cardiac events was compared using ROC analysis. In ROC analysis, a higher position in the upper left of the graph indicates higher predictive accuracy. The area under the curve (AUC) is defined as the area under the curve (AUC), and a higher AUC indicates higher predictive accuracy. The solid line in the graph represents ME, and the dashed line represents the predictive accuracy of MFR.

Figure 12(a) is a graph comparing G-ME and G-MER, (b) is a graph comparing RCA-ME and RCA-MFR, (c) is a graph comparing LAD-ME and LAD-MFR, and (d) is a graph comparing LCX-ME and LCX-MFR. The vertical axis of the graphs represents sensitivity, corresponding to the positive rate, and the horizontal axis represents 1-specificity, corresponding to the false positive rate. As is clear from graphs (a) to (d), the AUC of ME is broader and superior in diagnostic ability compared to the AUC of MFR. Myocardial energy shows superior results in predicting cardiac events in the entire left ventricle and the three major regions when MFR < 2.0.

Next, Figure 13 shows the results of evaluating the usefulness of myocardial energy using Kaplan-Meier curves. Similar to the evaluation using ROC curves, 95 patients with coronary artery disease (22 with a future risk of cardiac event) were included. ME was calculated from each patient's PET images, and the entire myocardial region was fitted to a 16-segment model (see Figure 4; segment #17 was excluded from the calculation). The average ME of the 16 segments was defined as G-ME. For comparison, MFR was also calculated from each patient's PET images, and G-MFR was calculated in the same way as ME.

As an evaluation method, Kaplan-Meier analysis was used. The subjects were divided into two groups using ME=62.4 as the cutoff, and the number of months was plotted on the x-axis, with the event-free rate in each group plotted on the y-axis (see Figure 13(a)). Similarly, the subjects were divided into two groups using MFR=2.0 as the cutoff, and the number of months was plotted on the x-axis, with the event-free rate in each group plotted on the y-axis (see Figure 13(b)).

The p-value calculated from the Kaplan-Meier curve based on ME was 0.0034, and the p-value calculated from the Kaplan-Meier curve based on MFR was 0.017. In statistical hypothesis testing, the p-value represents the probability that the test statistic will reach that value under the null hypothesis. A smaller p-value means that the test statistic is less likely to reach that value. The Kaplan-Meier curve indicates that the greater the separation between the two curves, the better the prognostic indicator. Therefore, comparing Figures 13(a) and (b), in long-term follow-up of 60 months or more, using G-ME ≥ 62.4 and G-ME < 62.4 as criteria for diagnosing the risk of cardiac event is more accurate than using G-MFR ≥ 2.0 and G-MFR < 2.0 as criteria (note that n in parentheses indicates the number of patients).

The evaluation results shown in Figures 12 and 13 confirm that myocardial energy (ME) can be a more useful indicator compared to the conventional index (MFR). In the embodiment of one invention, ammonia PET data is shown for myocardial blood flow PET preparations, but similar analysis is possible for myocardial glucose metabolism (FluoroDeoxyGlucose; FDG) in addition to myocardial blood flow PET preparations (ammonia). In other words, instead of calculating myocardial energy based on myocardial flow reserve (MFR) and myocardial strain (MSR), an energy map calculated based on glucose metabolism and strain can also be created.

Thus, the myocardial energy calculation method according to the present invention calculates the myocardial flow rate (MFR) based on tomographic images, such as PET images of the heart, acquired by a nuclear medicine imaging device such as a PET scanner. Furthermore, it calculates the myocardial surge ratio (MSR) using dynamic analysis techniques such as feature tracking on the same tomographic image. This allows for the simultaneous measurement of myocardial blood flow and myocardial dynamics from a single tomographic image of the heart, enabling examinations that accurately reflect the patient's condition at a specific point in time. As a result, myocardial blood flow and myocardial dynamics can be measured from image data of the same size and pixel count, eliminating the need for adjustment operations to unify the size and pixel count of image data acquired from separate imaging examinations.

Furthermore, by calculating myocardial energy using the square root of the sum of the squares of MFR and MSR, it becomes possible to represent the kinetic energy related to the myocardium with a single indicator—myocardial energy—rather than evaluating myocardial blood flow and myocardial dynamics with separate indicators, making it easier to assess the state of the myocardium.

Furthermore, since the display using the myocardial segment model (for example, the map display of the myocardial segment model shown in Figure 11) is a common display method for all imaging modalities such as MRI, CT, ultrasound, and nuclear medicine (PET), it is possible to calculate a new clinical value called ME by processing the information of 16 segments, even with different modalities.

The myocardial energy calculation method and other related technologies according to the present invention can be used to support the diagnosis of heart disease, etc.

100 Image analysis device 200 Nuclear medicine imaging device 210 Data processing device N Network

Claims (18)

A method for calculating myocardial energy,
The stage of acquiring tomographic images of the heart from a nuclear medicine imaging device,
The steps include: calculating the myocardial flow reserve (MFR), which represents the ratio of myocardial blood flow under load to resting blood flow, from the aforementioned tomographic image;
The steps include calculating the peak strain of the myocardium from the aforementioned tomographic image,
The steps include: calculating the myocardial strain ratio (MSR), which represents the ratio of load strain to rest strain of the myocardium, from the aforementioned peak strain;
The step includes calculating myocardial energy (ME), which represents the kinetic energy of the myocardium, based on the MFR and MSR.
A method for calculating myocardial energy, characterized in that the ME is calculated using the square root of the sum of the squares of the MFR and the MSR. The myocardial energy (ME) is calculated by setting A and B as arbitrary coefficients.
The method for calculating myocardial energy according to claim 1, characterized by being expressed by the following formula. The nuclear medicine imaging device is a PET device or a SPECT device.
The method for calculating myocardial energy according to claim 1 or 2, characterized in that the tomographic image is a PET image acquired by the PET device or a SPECT image acquired by the SPECT device. The aforementioned tomographic images include multiple cross-sectional images of the entire myocardium,
The myocardial energy calculation method according to any one of claims 1 to 3, characterized in that each of the plurality of cross-sectional images includes a predetermined number of consecutive images corresponding to the cardiac cycle. Prior to the steps of calculating the MFR and the MSR,
The step of receiving input for determining multiple feature points along the endocardium recorded in a predetermined frame image from a predetermined number of consecutive images,
A step of tracking the plurality of feature points over a predetermined number of consecutive images using feature tracking technology,
The process involves generating a time strain curve with the cardiac cycle and the distance between two adjacent points, each of which represents the distance between two adjacent points expressed in two-dimensional coordinates, from among the aforementioned plurality of feature points;
The method for calculating myocardial energy according to claim 4, characterized in that it includes the step of determining the peak strain using the peak value as a representative value in the time strain curve. The method for calculating myocardial energy according to claim 5, characterized in that the load strain and the resting strain are each calculated from the sum of the distances between the plurality of feature points to obtain the length of the endocardium, normalized to the length of the endocardium at the end of diastole in the cardiac cycle, and the peak value thereof. The myocardial energy calculation method according to any one of claims 1 to 6, characterized in that the ME includes values corresponding to each of multiple regions of the myocardium, and is displayed by applying it to a segment model in which the myocardium is divided into the multiple regions. The method for calculating myocardial energy according to claim 7, characterized in that the segment model is a 16- segment model that includes segments #1 to #16 corresponding to 16 regions of the myocardium, excluding segment #17 corresponding to the apex of the myocardium in a 17-segment model corresponding to 17 regions of the myocardium. A myocardial energy calculation system,
Nuclear medicine imaging device,
Includes an information processing device,
The aforementioned nuclear medicine imaging device acquires and stores cross-sectional images of the heart.
The aforementioned information processing device is
The tomographic image is acquired from the nuclear medicine imaging device.
From the aforementioned tomographic images, the myocardial flow reserve (MFR), which represents the ratio of myocardial blood flow under load to resting blood flow, is calculated.
The peak strain of the myocardium is calculated from the aforementioned tomographic image.
From the aforementioned peak strain, the myocardial strain ratio (MSR), which represents the ratio of load strain to rest strain of the myocardium, is calculated.
Based on the MFR and MSR, the myocardial energy (ME), which represents the kinetic energy of the myocardium, is calculated.
A myocardial energy calculation system characterized in that the ME is calculated using the square root of the sum of the squares of the MFR and the MSR. The myocardial energy (ME) is calculated by setting A and B as arbitrary coefficients.
The myocardial energy calculation system according to claim 9, characterized in that it is expressed by the formula shown above. The nuclear medicine imaging device is a PET device or a SPECT device.
The myocardial energy calculation system according to claim 9 or 10, characterized in that the tomographic image is a PET image acquired by the PET device or a SPECT image acquired by the SPECT device. The aforementioned tomographic images include multiple cross-sectional images of the entire myocardium,
The myocardial energy calculation system according to any one of claims 9 to 11, characterized in that each of the plurality of cross-sectional images includes a predetermined number of consecutive images corresponding to the cardiac cycle. The information processing device calculates the MFR and the MSR,
The system accepts input for determining multiple feature points along the endocardium recorded in a predetermined frame of a predetermined number of consecutive images.
Using feature tracking technology, the plurality of feature points are tracked across a predetermined number of consecutive images.
For each of the distances between two adjacent points represented by two-dimensional coordinates among the aforementioned multiple feature points, a time strain curve is generated with the cardiac cycle and the distance between the two points as the axes.
The myocardial energy calculation system according to claim 12, characterized in that it includes determining the peak strain using the peak value as a representative value in the time strain curve. The myocardial energy calculation system according to claim 13, characterized in that the load strain and the resting strain are each calculated from the sum of the distances between the plurality of feature points to obtain the endocardial length, normalized to the endocardial length at end-diastole in the cardiac cycle, and the peak value thereof. The aforementioned ME includes values corresponding to each of several regions of the myocardium,
The myocardial energy calculation system according to any one of claims 9 to 14, characterized in that the information processing device displays the ME by applying it to a segment model in which the myocardium is divided into a plurality of regions. The myocardial energy calculation system according to claim 15, characterized in that the segment model is a 16-segment model that includes segments #1 to #16 corresponding to 16 regions of the myocardium, excluding segment #17 corresponding to the apex of the myocardium in a 17-segment model corresponding to 17 regions of the myocardium. A myocardial energy calculation device characterized by performing each step of the myocardial energy calculation method described in any one of claims 1 to 8. A myocardial energy calculation program characterized by causing the computer to function as the myocardial energy calculation device described in claim 17, by being executed by the computer.