Magnetic resonance signal acquisition method and system for cardiac magnetic resonance imaging

Abstract

The invention relates to a magnetic resonance signal acquisition method for cardiac imaging, the method comprising an acquisition sequence (SE) comprising: acquiring signals comprising, during each of a succession of pairs consisting of two successive cardiac cycles: - acquiring black-blood signals from a region of interest of a heart of a subject to be imaged by black-blood-delayed gadolinium-enhanced magnetic resonance; and - acquiring bright-blood signals from the region of interest by bright-blood-delayed gadolinium-enhanced magnetic resonance; the bright-blood signals acquired during the succession of pairs being acquired by implementing a relaxation measurement sequence.

Description

Technical Field

[0001] This invention relates to the field of cardiac magnetic resonance imaging (MRI) via delayed gadolinium enhancement (LGE). Background Technology

[0002] A reference technique for characterizing regional lesions, including myocardial fibrosis, is bright blood-delayed gadolinium enhancement (BR-LGE) imaging using inversion recovery (e.g., PSIR (phase-sensitive inversion recovery) sequences). In this type of imaging, the inversion recovery pulses cancel out the signal from healthy myocardium, enabling high-contrast visualization of lesions (especially intramyocardial lesions, i.e., gadolinium retention areas where gadolinium is cleared more slowly than from healthy myocardium), with high contrast between healthy and lesions. However, for lesions adjacent to the cardiac chambers (right and left ventricles), the high intensity of the blood signal and therefore the low contrast between the lesion and the blood hinder automated, accurate, reliable, and robust detection.

[0003] To address this issue, black-blood LGE (BL-LGE) imaging technology has been developed. This technology enables the simultaneous elimination of signals from healthy myocardium where gadolinium is rapidly cleared, as well as signals from the blood, thereby providing high contrast between lesions and healthy myocardium with rapidly cleared gadolinium, and the blood. These techniques are particularly useful for identifying focal (i.e., localized) lesions.

[0004] However, black-blood imaging technology does not allow for accurate localization of the position of myocardial lesions relative to healthy myocardium because the contrast between blood and healthy myocardium is insufficient to clearly distinguish the boundaries of the myocardium.

[0005] The inventors of this application provided a method for acquiring and fusing images of black blood and bright blood in a previous patent application FR2203782, which enables the detection and better localization of these lesion areas while maintaining a shorter acquisition time.

[0006] However, this technique does not allow for the quantification of diffuse changes in the tissue of the area being imaged, particularly the myocardium. This type of change is especially likely to occur during the acute phase of a cardiac event. Summary of the Invention

[0007] One object of the present invention is to provide a solution for acquiring signals, according to which lesion areas can be detected and located and diffuse changes in tissue can be quantified.

[0008] One objective of one embodiment is to provide a method for limiting at least one of the aforementioned disadvantages.

[0009] For this purpose, one aspect of the present invention is a method for acquiring magnetic resonance signals for cardiac imaging, the method comprising an acquisition sequence including: The signals are acquired, including during each of multiple consecutive pairs (each pair consisting of two consecutive cardiac cycles): - Black-blood signals are acquired from the imaging region of an individual's heart using black-blood-delayed gadolinium-enhanced magnetic resonance imaging (MRI), which enables the generation of a basic black-blood image of the imaging region. - Bright blood signals are acquired from the region to be imaged using bright blood-delayed gadolinium-enhanced magnetic resonance imaging (MRI), and these dark blood signals enable the generation of a basic dark blood image of the region to be imaged. The bright blood signal acquired during this consecutive multi-pair period was obtained by performing a relaxation measurement sequence.

[0010] According to one embodiment, the black blood signal enables the generation of a basic black blood image of the region to be imaged, and the black blood signal enables the generation of a basic black blood image of the region to be imaged.

[0011] The present invention also relates to a magnetic resonance cardiac imaging method including the acquisition method.

[0012] In other words, the present invention relates to a magnetic resonance imaging method for the heart, the method comprising an acquisition sequence including: The signals are acquired, including during multiple consecutive pairs (consisting of two consecutive cardiac cycles): - Acquire black blood signals from the imaging area of ​​an individual's heart using black blood-delayed gadolinium-enhanced magnetic resonance imaging. - Bright blood signals were acquired from the region to be imaged using bright blood-delayed gadolinium-enhanced magnetic resonance imaging. The bright blood signal acquired during this consecutive multi-pair period was obtained by performing a relaxation measurement sequence.

[0013] According to one embodiment, the black blood signal enables the generation of a basic black blood image of the region to be imaged, and the black blood signal enables the generation of a basic black blood image of the region to be imaged.

[0014] According to one embodiment, the black blood signal is acquired by performing the same black blood acquisition sequence over a series of consecutive pairs.

[0015] According to one embodiment, the method includes: - Signals acquired during continuous cardiac cycles are processed by computer, including: - Generate a tissue relaxation time map of the region to be imaged from the bright blood signal acquired during the consecutive cardiac cycles of this region.

[0016] According to one embodiment, processing the signals acquired during the multiple pairs of pairs by a computer includes: - Generate a black blood image from the black blood signal acquired during the period of these consecutive pairs. - Generate a bright blood image from the bright blood signal acquired during multiple consecutive pairs of periods.

[0017] According to one embodiment, processing signals acquired during multiple pairs of events via a computer includes: - Determine at least one piece of data from at least one of the dark blood images and at least one of the bright blood images that relates to the region of the undetermined cardiac lesion in the reference dark blood image.

[0018] According to one embodiment, the reference black blood image is a combination of the black blood images.

[0019] According to one embodiment, at least one piece of data is determined from bright blood signals acquired only during one of the multiple pairs.

[0020] According to one embodiment, at least one piece of data is determined from bright blood signals collected during multiple pairs.

[0021] According to one embodiment, at least one piece of data is determined from the bright blood signals collected during the multiple pairs.

[0022] According to one embodiment, processing signals acquired during multiple pairs of events via a computer includes: - The reference black blood image was segmented by computer to determine the area representing the heart lesion.

[0023] According to one embodiment, determining at least one piece of data related to the region representing the cardiac lesion includes: - A fused image is generated by computer by replacing the pixel or voxel values ​​of a first reference bright blood image with new values ​​that enable the differentiation between the cardiac lesion representative area and other tissues in the area to be imaged, the first reference bright blood image being generated based on bright blood signals acquired during at least one of the multiple pairs.

[0024] According to one embodiment, processing signals acquired during multiple pairs of events via a computer includes: -At least one second reference bright blood image is generated by computer segmentation of bright blood signals acquired during at least one pair of the plurality of pairs to generate localization data defining a set of at least one wall of the myocardium. - Optionally, the reference black blood image is segmented based on the location data of at least one wall in the set of at least one walls defining the myocardium to determine the region representing the cardiac lesion, wherein the location data of the at least one wall is extracted from the location data of the set of at least one walls defining the myocardium.

[0025] According to one embodiment, the second reference bright blood image is segmented using a learning function trained from multiple sets of training images of the region to be imaged. These multiple sets of training images are generated from training signals acquired during corresponding acquisition steps via bright blood-delayed gadolinium-enhanced magnetic resonance, with each set of training images obtained from bright blood signals acquired by implementing a relaxation measurement sequence.

[0026] In one embodiment, the second segmentation includes selecting pixels in the reference black-blood image with an intensity greater than a predetermined threshold, the pixels being selected only from the pixels of the black-blood image surrounded by a first wall.

[0027] In one embodiment, the set of at least one wall includes a second wall that defines the myocardium and is surrounded by the first wall.

[0028] In one embodiment, the process includes calculating data representing the permeability percentage of the area represented by the cardiac lesion based on data obtained from positioning data from the first wall and the second wall.

[0029] In one embodiment, the method includes displaying on a screen a representation of data determined during processing or an image generated during processing.

[0030] In one embodiment, the black blood image and the bright blood image are two-dimensional.

[0031] The present invention also relates to an imaging system configured to implement the method according to the invention. The system includes a processing and control unit and a magnetic resonance system, the processing and control unit being configured to control the magnetic resonance system such that the magnetic resonance system implements the acquisition sequence.

[0032] The imaging system is, for example, an acquisition system configured to perform the following steps: acquiring, and / or generating images, and / or generating relaxation time maps, and / or processing images, and / or generating relaxation time maps.

[0033] According to one embodiment, the system includes a device for measuring at least one signal representing an individual's cardiac activity, and the processing and control unit is configured to control the magnetic resonance system based on the measured value of the at least one signal representing the cardiac activity generated by the measuring device, such that the magnetic resonance system performs the acquisition sequence.

[0034] According to one embodiment, the system includes an electrocardiograph, and the processing and control unit is configured to control the magnetic resonance system based on measurements obtained from the electrocardiograph, such that the magnetic resonance system performs the acquisition sequence.

[0035] According to one embodiment, the processing and control unit is configured to implement the step of processing signals acquired during a continuous cardiac cycle via a computer.

[0036] The present invention also relates to a computer program product comprising instructions that cause a system according to the invention to perform the steps of the method according to the invention.

[0037] The present invention also relates to a computer-readable medium having a computer program or program product thereon recorded according to the present invention.

[0038] The present invention also relates to a method for processing data via a computer, the data being generated from signals acquired via magnetic resonance imaging for cardiac imaging.

[0039] According to one embodiment, the method includes: - Receive a dark blood image corresponding to the bright blood image generated by the imaging method according to the present invention.

[0040] - Determine at least one piece of data relating to a cardiac lesion-representing region in a reference dark-blood image, which is generated based on at least one of the dark-blood images and bright blood signals acquired during the multiple pairs.

[0041] The method may include any step, combination of steps, or feature of the processing steps of the imaging method or general processing steps described in this patent application.

[0042] According to one embodiment, the method includes: -At least one second reference bright blood image is generated by computer segmentation of bright blood signals acquired during at least one pair of the plurality of pairs to generate localization data defining a set of at least one wall of the myocardium. - Optionally, the reference black blood image is segmented based on the location data of at least one wall in the set of at least one walls defining the myocardium to determine the region representing the cardiac lesion, the location data of the at least one wall being extracted from the location data of the set of at least one walls defining the myocardium.

[0043] According to one embodiment, the method includes: - Generate a data representation related to the region, and optionally display the visual representation.

[0044] According to one embodiment, the method includes: - Receive the tissue relaxation time map of the region to be imaged, generated by the imaging method according to the present invention. - Generate a visual representation of the map and, if possible, display that visual representation.

[0045] According to one embodiment, the visual representation is of the bullseye type and is generated from location data of a set of at least one wall that defines the myocardium.

[0046] According to one embodiment, the method includes the step of propagating walls detected during the step of generating location data of a set of at least one wall defining the myocardium onto a reference black blood image ISN and an atlas CA.

[0047] According to one embodiment, the reference black blood image is a combination of the black blood images.

[0048] According to one embodiment, at least one piece of data is determined from bright blood signals acquired only during one of the multiple pairs.

[0049] According to one embodiment, at least one piece of data is determined from bright blood signals collected during multiple pairs.

[0050] According to one embodiment, at least one piece of data is determined from the bright blood signals collected during the multiple pairs.

[0051] According to one embodiment, the method includes: - The reference black blood image was segmented by computer to determine the area representing the heart lesion.

[0052] According to one embodiment, the method includes determining at least one piece of data related to the region representing the cardiac lesion, including: - A fused image is generated by computer by replacing the pixel or voxel values ​​of a first reference bright blood image with new values ​​that enable the differentiation between the cardiac lesion representative area and other tissues in the area to be imaged, the first reference bright blood image being based on bright blood signals acquired during at least one of the multiple pairs.

[0053] According to one embodiment, the second reference bright blood image is segmented using a learning function trained from multiple sets of training images of the region to be imaged. These multiple sets of training images are generated from training signals acquired during corresponding acquisition steps via bright blood-delayed gadolinium-enhanced magnetic resonance, with each set of training images obtained from bright blood signals acquired by implementing a relaxation measurement sequence.

[0054] In one embodiment, the second segmentation includes selecting pixels in the reference black-blood image with an intensity greater than a predetermined threshold, the pixels being selected only from the pixels of the black-blood image surrounded by a first wall.

[0055] In one embodiment, the set of at least one wall includes a second wall that defines the myocardium and is surrounded by the first wall.

[0056] In one embodiment, the process includes calculating data representing the permeability percentage of the area represented by the cardiac lesion based on data obtained from positioning data from the first wall and the second wall.

[0057] In one embodiment, the method includes displaying on a screen a representation of data determined during processing or an image generated during processing.

[0058] In one embodiment, the black blood image and the bright blood image are two-dimensional.

[0059] The present invention also relates to a processing system configured to implement the method according to the invention, the system comprising a processing or processing and control unit. This unit implements the steps according to the invention.

[0060] Such a system is, for example, the processing and control system described in this patent application.

[0061] According to one embodiment, the processing system is a processing device such as a computer, and an image processing computer program is installed within the computer. The computer and the computer program are configured to receive images generated by an imaging system (such as the imaging system of the present invention).

[0062] The imaging system may include the processing system. Alternatively, they may be two separate systems, such as two separate devices.

[0063] The present invention also relates to a computer program product including instructions that cause the processing system to perform the steps of the processing method.

[0064] The present invention also relates to a computer-readable medium having the computer program or computer program product thereon recorded.

[0065] According to one embodiment of the present invention, the processing method and / or the computer program product form a "plugin" or additional component or extension component of existing image processing software.

[0066] In the latter case, the method or computer program product is configured to receive the first and second images and to generate output data to display them within an interface generated by the existing image processing software. Attached Figure Description

[0067] Other features and advantages of the invention will become clearer from the following detailed description with reference to the accompanying drawings, which illustrate: Figure 1 An example embodiment of the system according to the present invention is shown; Figure 2 : A schematic diagram of the first basic signal acquisition sequence used to generate the black blood image and bright blood image used in the method of the present invention; Figure 3 : Including multiple basic acquisition sequences (including Figure 2 A schematic diagram illustrating an example of black-blood and bright-blood MRI acquisition sequences performed over multiple cardiac cycles (based on the basic acquisition sequence). Figure 4A three-dimensional (3D) schematic diagram of the heart, showing different cross-sections distributed along the long axis of the heart, and an image generated from signals acquired from one of these cross-sections. Figure 5 : A flowchart of a method example according to the present invention, Figure 6 The diagram shows four images, including the bright blood image in the upper left and upper right (representing walls detected during the segmentation step), and the dark blood image in the lower left and a representation of the walls transferred onto the dark blood image. Figure 7 The above is Figure 6 The lower left image (showing sectors above it) shows the size of the lesion area as represented by a bull's eye chart, while the lower right image shows the percentage of transmural lesion as represented by a bull's eye chart. Detailed Implementation

[0068] This invention relates to the fields of cardiac magnetic resonance imaging or MRI, and delayed gadolinium enhancement in black blood and bright blood.

[0069] This invention relates to a cardiac imaging method.

[0070] An imaging method is defined as a method that includes acquiring signals for generating an image and / or processing signals to generate an image and / or other data (e.g., generating a relaxation time map and / or processing the image and / or other data).

[0071] Cardiac lesions refer to areas of tissue abnormality that are near or located within a reference region represented by the myocardium. This region exhibits altered tissue structure compared to the reference tissue structure (the tissue structure that best represents the myocardium).

[0072] Alterations in tissue structure cause contrast agents to accumulate for a longer time in areas with tissue abnormalities than in areas with tissue structures that are most representative of the myocardium.

[0073] Therefore, the characteristic of the area representing cardiac lesions in a black blood image is that the intensity of the pixel or voxel is greater than the intensity of the pixel representing the most representative tissue structure of the myocardium.

[0074] For example, the intensity is greater than a predetermined threshold.

[0075] Pathological lesions are distinguished from benign lesions. For example, some lesions may affect cardiac function. These lesions are referred to here as pathological lesions. In another example, a lesion may not affect cardiac function, such as fat. Therefore, locating representative areas of cardiac lesions can be meaningful to specialists, as they may be indicators of cardiopathological changes affecting cardiac function (such as myocardial scarring) or changes in fat or other types of cardiac or myocardial tissue. This data should be cross-checked with other information to identify the nature of the feature.

[0076] Imaging system Figure 1 An example embodiment of the imaging system S according to the present invention is illustrated schematically. The system includes the hardware and software components required to implement the method according to the present invention.

[0077] Advantageously, the system S includes a set of measuring devices A, which includes a magnetic resonance imaging (MRI) system B and Figure 1 An electrocardiograph marked ECR.

[0078] The system S also includes a processing and control device C, which includes a processing and control unit TC and a human-machine interface INT.

[0079] In a known manner, the MRI magnetic resonance system B includes a static magnetic field generator GEN_B, a gradient generator GEN_GRAD, and a radio frequency (RF) device D_RF.

[0080] The static magnetic field generator GEN_B includes a main polarization magnet for generating a substantially uniform polarized static magnetic field in a polarization region (typically a tunnel) to accommodate the patient's imaging region, which includes the heart.

[0081] The patient is a mammal. In a non-restrictive sense, the mammal is a human.

[0082] The gradient generator GEN_GRAD comprises three gradient coils (or solenoids) arranged and configured to vary the magnetic field strength in the polarization region along corresponding orthogonal axes x, y, and z, which are fixed relative to the polarization region. The strength flowing through these coils is chosen such that a cross section, having a thickness and a plane in which it lies, can be selected from several possibilities for measuring the magnetization of the area of ​​the patient to be imaged within the polarization region.

[0083] The radio frequency system D_RF includes a radio frequency transmitter containing a coil or solenoid and is capable of generating a radio frequency (RF) pulse sequence (also referred to as an acquisition sequence in this patent application). The acquisition sequence includes a preparation sequence and a readout sequence for magnetizing the region to be imaged. In the readout sequence, the radio frequency receiver of the radio frequency system reads (i.e. receives) the RF signal generated by the effect of the radio frequency sequence emitted by the transmitter from the region to be imaged.

[0084] The radio frequency receiver may be a radio frequency transmitter or something else entirely. Each of the preparation sequence and readout sequence includes at least one radio frequency pulse having a predetermined and adjustable frequency, shape, duration, phase, and amplitude.

[0085] The sequence preparation is configured for excitation, which means changing the magnetization direction of the tissue in the region to be imaged.

[0086] The read sequence is configured to allow the receiver of the RF system to measure the magnetization of the region to be imaged, generated by the preparation module. An electrocardiograph (ECR) is used to acquire a patient's electrocardiogram.

[0087] The processing and control unit TC is configured to generate instructions intended for use with the MRI system B (particularly instructions intended for use with the RF device D_RF and gradient generator GEN_GRAD), causing the MRI device to generate a predefined sequence of acquisition signals derived from a predefined volume or cross-section of the region to be imaged. The processing and control unit TC is configured to process the RF signals acquired by the radio frequency system, such processing including, for example, generating a relaxation time map of the region to be imaged according to reconstruction techniques known to those skilled in the art and generating an image of the region to be imaged, and determining data from these images, as will be seen in more detail in the description below.

[0088] Acquisition sequence Figure 2 This illustrates an example of the first basic acquisition sequence Se1 for an RF signal. Se1 is one of the fundamental sequences of the acquisition sequence SE for an RF signal, which enables the generation of an image of the heart, and more precisely, the region to be imaged.

[0089] Figure 2 An electrocardiogram (ECG) E measured by an electrocardiograph ECR during the first basic acquisition sequence Se1 is also shown.

[0090] The areas to be imaged include the patient's myocardium and blood.

[0091] Advantageously, the region to be imaged includes: at least one cross-section of the patient's cardiac myocardium taken in a plane; and a cross-section of at least one chamber surrounded by myocardium and in which blood flows, taken in the same plane; and a cross-section of the region surrounding the myocardium.

[0092] This invention relates to a magnetic resonance signal acquisition method for cardiac imaging.

[0093] In other words, during this method, signals are acquired to generate images of the individual heart.

[0094] As will be seen below, the acquisition sequence SE consists of a sequence of N basic acquisition sequences Sei, where i = 1 to N, and i is an integer.

[0095] For example, N is between 2 and 20, but N can be greater than 20.

[0096] For example, N is 5.

[0097] Figure 2 The first basic sequence Se1 is represented in the figure.

[0098] Figure 2 The lower part shows the change of longitudinal magnetization Mz of the tissue in the region to be imaged as a function of time t during the basic sequence Se1.

[0099] The basic acquisition sequence Se1 includes the so-called black-blood acquisition ACQ11, followed by the so-called bright-blood acquisition ACQ21, which will be described below. The black-blood acquisition ACQ11 enables the acquisition of signals from the region to be imaged to generate a basic black-blood image IM1 with black-blood contrast for the region to be imaged.

[0100] The bright blood acquisition ACQ21 enables the acquisition of signals from the region to be imaged, thereby generating a basic bright blood image IM2 of the region to be imaged with bright blood contrast.

[0101] The collection steps of the black blood collection ACQ11 and bright blood collection ACQ21 each include a sequence of emitted electrical pulses, which includes preparation modules PREP1 and PREP21 and reading modules LE1 and LE2.

[0102] In this patent application, a module refers to a radio frequency pulse or a series of radio frequency pulses.

[0103] It should be noted that during the entire duration of the basic acquisition sequence Se1, and preferably during the entire duration of the MRI acquisition sequence or N basic acquisition steps, the static magnetic field generator GEN_B is controlled by the processing and control unit TC, such that the static magnetic field generator GEN_B generates a fixed static magnetic field along the z-axis.

[0104] The gradient generator GEN_GRAD is advantageously controlled by the processing and control unit TC, such that the radio frequency system D_RF acquires signals from a predetermined cross section with a predetermined thickness during the first basic acquisition sequence Se1 and throughout the entire duration of each of the N basic sequences Sei from i=1 to N.

[0105] In other words, the signals acquired during the SE acquisition sequence originate from the same cross section, which subsequently forms the region to be imaged.

[0106] The first acquisition sequence is a delayed gadolinium-enhanced acquisition sequence performed after an intravenous injection of a gadolinium-based contrast agent into the patient (this injection is performed 10 to 15 minutes before the acquisition sequence is performed) to obtain images with maximum contrast between the lesion and blood, as well as between the lesion and healthy myocardium or the most representative myocardial tissue structure area.

[0107] In the heart, contrast agents are rapidly cleared from healthy myocardium (which has low content in its extracellular or interstitial tissue), but accumulate in cardiomyopathy in a prolonged manner. In fact, gadolinium has an extracellular distribution, meaning it does not cross the myocardial cell membrane.

[0108] Gadolinium has the effect of shortening the relaxation time T1 in the tissue where it accumulates. Therefore, the magnetization relaxation of lesions after a magnetization reversal pulse is faster than that of blood and healthy myocardium.

[0109] Black Blood Collection First, the goal is to generate a basic black blood image IM1. In such an image, the pixel intensity corresponding to blood and healthy myocardium or myocardial reference areas is zero (black pixels), essentially zero, or very low.

[0110] To generate such a basic black blood image IM1, the RF system D_RF performs an inversion recovery black blood acquisition step ACQ11 during the first basic acquisition sequence Se1. This black blood acquisition sequence ACQ1 includes transmissions from the transmitter of the RF system... Figure 2 The longitudinal reversal pulse, marked 180°, reverses the longitudinal magnetization of the tissue in the imaging region, i.e., it reverses the longitudinal magnetization of the tissue. Figure 2 In the diagram, it is noted that under the influence of the inverting pulse, the magnetization of the region to be imaged changes from Mz to -Mz. Due to longitudinal relaxation, the longitudinal magnetization of different tissues present in the region to be imaged increases to recover to its initial value and passes through zero. Naturally, the relaxation dynamics of different tissues are different.

[0111] In a manner known per se, the black blood collection ACQ11 also includes a preparation module PREP1 implemented after a 180° longitudinal reversal pulse, such as a T1ρ adiabatic module with a duration labeled TSL (“spin-locked time”), or a T2 weighted or MTC (“magnetization transfer contrast”) type module, or a combination of two or three of these modules.

[0112] The preparation module PREP1 is configured such that the longitudinal magnetization of blood A(B) and the longitudinal magnetization of healthy myocardium A(M), or more generally the longitudinal magnetization of the myocardial reference region, cancel each other out at the same time te.

[0113] At the same time te, the longitudinal magnetization of lesion A(C) is significantly greater than zero. By acquiring signals from the region to be imaged at this time te, an image is obtained that has very high contrast between pixels or voxels corresponding to blood (which are generally black) and pixels or voxels corresponding to lesions (which are generally bright).

[0114] Then, the first acquisition step ACQ11 in the inversion recovery includes a readout sequence LE1, which comprises a 90° pulse applied at time te and a readout gradient for reading the lateral magnetization of the region to be imaged. The inversion time TI is the duration that separates the 180° pulse from the readout sequence LE1 of the basic black-blood acquisition ACQ11. To obtain optimal contrast between myocardial lesions and blood, and between myocardial lesions and healthy myocardium, the readout sequence LE1 is advantageously initiated at time te, where the longitudinal magnetization of blood and myocardium cancels out, to generate an image exhibiting optimal contrast.

[0115] exist Figure 2 In the example, the reading module LE1 for black blood collection is time-separated from the preparation module PREP1 for black blood collection. Alternatively, the reading module LE1 begins at the same time as the preparation module PREP1 ends. The same applies to the relative time positioning between the preparation module PREP21 and the reading module LE2 for bright blood collection.

[0116] exist Figure 2 In the example, the inverted pulse IMP1 is generated before the preparation module PREP1. Alternatively, the preparation module PREP1 is generated before the inverted pulse IMP1.

[0117] Bright blood collection The basic bright blood image IM21 of the region to be imaged is generated by the signal acquired by performing the bright blood acquisition step ACQ21 (including the preparation module PREP21 and the subsequent readout module LE2) during the first basic acquisition sequence Se1.

[0118] Then, the bright blood acquisition step ACQ21 includes a readout module LE2, which includes a readout gradient for reading the lateral magnetization of the region to be imaged. This readout module LE2 can be implemented using gradient echo or spin echo methods, similar to the readout module LE1 in the dark blood acquisition step LE1. Readout modules LE1 and LE2 can be the same or different.

[0119] The duration D2 during which the reading module LE2 is separated from the preparation module is defined such that the longitudinal magnetization of blood A(B) is greater than that of myocardium A(M). This results in the generation of an image in which the pixels or voxels of blood are bright (i.e., have high brightness or intensity), and the pixels of myocardium tissue are slightly less bright than the pixels of blood, as shown in the image. Figure 2 The curves shown can be inferred from this. These images allow for perfect visualization of cardiac anatomy, enabling the delineation of the myocardium on the image, which is impossible on dark blood images.

[0120] It is understandable that by using two images, IM11 and IM21, gadolinium accumulation areas or lesions can be detected, and the areas or lesions can be accurately located relative to the myocardium, as well as their size can be determined relative to the myocardium.

[0121] synchronous Preferably, the processing and control unit TC is configured to synchronize the acquisition sequence SE with the electrocardiogram E.

[0122] For this purpose, the processing and control unit TC advantageously uses the electrocardiogram E to generate instructions for triggering (i.e. implementing) acquisition sequences intended for use with RF devices, gradient generators, and possibly main magnetic field generators.

[0123] The acquired sequence SE includes (e.g.) Figure 3 As shown, there are multiple basic collection sequences Sei (i=1 to N, where N is greater than 1 and i is the index of the basic sequence), and the black blood collection step ACQ1i of the basic collection sequence is the same. Figure 2 In this case, i=1. As will be seen below, the bright blood collection steps ACQ2i are different for different basic acquisition sequences Sei.

[0124] Each basic acquisition sequence Sei is acquired in two consecutive cardiac cycles C1i, C2i (in... Figure 2 Implemented during the period marked (where i=1) to constitute Figure 3 A pair of CBis for two consecutive cardiac cycles marked in the middle. Advantageously, each basic acquisition sequence Sei is performed during two consecutive RR intervals.

[0125] Advantageously, the two basic acquisition sequences Sei are performed during the two consecutive interbeats that constitute a pair of consecutive interbeats.

[0126] One advantage is that it minimizes acquisition time, thereby minimizing cardiac motion between different acquisitions and the spatial offset between images IM1i and IM2i generated from signals acquired during the corresponding basic acquisition sequence SEi. In the remainder of this paper, a beat refers to the QRS complex, and the interval between beats refers to the phase of the cardiac cycle between two consecutive beats.

[0127] Advantageously, the continuous basic acquisition sequence SEi is performed during the continuous interval between heartbeats on CBi, such as Figure 3 As shown in the example.

[0128] Each basic acquisition sequence SEi includes: -During the first cardiac cycle C1i of the cardiac cycle to CBi, the black blood collection step ACQ1i; - The bright blood collection step ACQ2i is performed during the second cardiac cycle C2i of the cardiac cycle to CBi.

[0129] One advantage is that it minimizes acquisition time, thereby minimizing cardiac motion between different acquisitions and spatial offset between images IM1 and IM2 acquired during different basic sequence SEi.

[0130] Advantageously, such as Figure 2 As shown, the acquisition sequence SE and the electrocardiogram E are synchronized so that the reading modules LE1 and LE2 are implemented during the same phase of two consecutive cardiac cycles C1i and C2i.

[0131] Favorably, this phase is the intercardiac interval.

[0132] Favorably, this phase is the diastolic phase.

[0133] Advantageously, the acquisition sequence SE and the electrocardiogram E are synchronized so that the reading modules LE1 and LE2 of the black blood acquisition step ACQ1i and the bright blood acquisition step ACQ2i are implemented at the same time during these corresponding cardiac cycles C1i and C2i.

[0134] These moments are defined relative to the same time references for cardiac cycles C1i and C2i. A time reference, for example, is the peak value of the QRS complex.

[0135] exist Figure 2 In the example, these moments are separated from the R-wave peak by the same duration D2'.

[0136] The synchronization of the reading modules LE1 and LE2 of the black blood and bright blood acquisition sequences ACQ1i and ACQ2i (and as will be seen below, the synchronization between reading modules of different bright blood acquisition steps and the synchronization between reading modules of different black blood acquisition steps) enables the generation of heart images at the moment when the heart occupies the same position in the fixed reference frame relative to the main magnet. This allows the images to be superimposed without registration or by performing simple registration.

[0137] like Figure 3 As shown, the continuous basic acquisition sequence Sei is implemented during multiple consecutive pairs of CBi (i=1 to N) consisting of two consecutive cardiac cycles.

[0138] exist Figure 3 In the unrestricted example, N=3.

[0139] Advantageously, the acquisition sequence SE includes multiple basic acquisition sequences Sei implemented during multiple pairs of CBi consisting of two consecutive cardiac cycles.

[0140] In one particular embodiment, the basic acquisition sequence is performed over multiple consecutive pairs consisting of two consecutive cardiac cycles.

[0141] One advantage is that it minimizes the spatial differences between images generated from signals acquired during different basic acquisition sequences and the acquisition duration.

[0142] One advantage is that, in the case of 2D acquisition, all signals corresponding to a cross section can be acquired in the shortest possible time. This makes it possible to generate data associated with that cross section and, for example, display at least one image associated with that cross section, i.e., to generate usable results even if the acquisition is interrupted (e.g., in the case of breath-holding cessation).

[0143] Relaxation measurement According to the present invention, bright blood signals acquired during different cardiac cycle pairs are acquired by implementing a relaxation measurement sequence.

[0144] In other words, the acquisition sequence consisting of N bright blood acquisition sequences ACQ2i (i=1 to N) is a relaxation measurement sequence.

[0145] In this way, a relaxation time map of the tissue in the imaging region can be generated based on the bright blood signal acquired during N cardiac cycles.

[0146] One advantage is that signals are acquired in a single acquisition sequence and therefore for a limited duration, which not only enables the detection and localization of representative areas of cardiac lesions but also allows for the quantification of diffuse changes in myocardial tissue.

[0147] This reduces the workload for technicians and, by limiting the spatial misalignment between images generated from signals acquired during the acquisition sequence, yields high-quality and consistent data.

[0148] A relaxation measurement sequence is a sequence of consecutive acquisition sequences that are different from each other by a predetermined characteristic duration, in order to acquire signals that can generate a relaxation time map.

[0149] exist Figure 2 and Figure 3 In the example, each preparation module PREP2i is, for example, a T1rho adiabatic sequence with a duration of TSLi.

[0150] The continuous bright blood collection sequences ACQ2i differ only in the spin-lock time (TSL).

[0151] The sequence i in each acquisition sequence ACQ2i represents the time order in which the sequence is implemented. Therefore, acquisition sequence 1 (Se1) of sequence 1 is implemented first, and acquisition sequence (SEN) of sequence N is implemented last.

[0152] For example, the spin-lock time TSLi of different basic acquisition sequences is distributed in the range of 0ms to 50ms.

[0153] Typically, the lock-in time is advantageously between 0ms and 50ms or between 0ms and 100ms.

[0154] For example, with N=5, the spin-locking time TSLi includes 0ms, 10ms, 20ms, 35ms, and 50ms.

[0155] exist Figure 3 In the example, the spin-locking time TSLi of the basic bright blood acquisition sequence ACQ2i increases monotonically with the sequence i of acquisition sequence ACQ2i.

[0156] Alternatively, the spin-locking time varies in different ways depending on the sequence i.

[0157] Therefore, by implementing different basic bright blood acquisition sequences ACQ2i, it is possible to generate a spectrum of the relaxation time T1ρ or T1rho (also known as spin-lattice relaxation time) of the tissue to be imaged from the bright blood signal acquired during N cardiac cycle pairs.

[0158] Alternatively, the preparation module PREP2i for each bright blood collection sequence ACQ2i is a T2-weighted module.

[0159] In this case, different preparation modules for different acquisitions are configured such that these sequences are different from each other through the duration of the T2 preparation module.

[0160] Therefore, by implementing different basic bright blood acquisition sequences ACQ2i, it is possible to generate a spectrum of the relaxation time T2 of the tissue to be imaged from the bright blood signals acquired during N cardiac cycle pairs.

[0161] Advantageously, the basic black blood collection steps ACQ1i are the same for different basic collection sequences Sei from i=1 to N.

[0162] Advantageously, the duration of the T2 preparation module is between 0ms and 100ms.

[0163] Section acquisition Advantageously, the processing and control unit TC is configured to generate instructions intended for use with the MRI system to acquire signals from corresponding sections distributed along a predetermined axis of the heart.

[0164] Therefore, the acquisition steps of this method include multiple acquisition sequences (SEs) (e.g., the acquisition sequences described above for acquiring signals from multiple cross sections). It should be understood that during each acquisition sequence (SE), a signal is acquired from a single cross section.

[0165] The acquisition sequences specifically designed for different cross-sections are advantageously set up continuously.

[0166] Alternatively, the acquisition sequences of the cross-sections are interleaved.

[0167] In other words, the acquisition steps include a basic acquisition sequence for acquiring the sequence of the first cross section, followed by a basic acquisition sequence for acquiring the sequence of the second cross section, and then another basic acquisition sequence for acquiring the sequence of the first cross section.

[0168] Advantageously, this axis is the long axis of the heart. The resulting cross-section is called a short axis section. One advantage is that it allows for excellent visualization of both ventricles. However, the invention is also applicable to cases where the cross-sections are distributed along another axis of the heart, such as the axis of a two-chamber heart (vertical long axis) or the axis of a four-chamber heart (horizontal long axis). Each cross-section has a defined thickness along this axis and is centered on a predetermined cross-sectional plane perpendicular to this axis. In this application, a cross-section refers to a slice or layer perpendicular to axis g and having a predetermined thickness along this axis.

[0169] Advantageously, the cross section is continuous along this axis.

[0170] This is achieved by selecting the instructions generated for the gradient generator GEN_GRAD, which is done by synchronizing the instructions intended for the gradient generator GEN_GRAD and those intended for the RF device D_RF.

[0171] Advantageously, signals are acquired along adjacent or partially overlapping sections. This allows for imaging of the entire heart.

[0172] Advantageously, the gradient generator GEN_GRAD is controlled to generate multiple two-dimensional images for each cross section from the signal acquired during the acquisition sequence SE.

[0173] Figure 4 The image shows a three-dimensional (3D) view of the heart, including multiple cross-sectional planes PC distributed along the major axis g. k Where k = 1 to K, and K is an integer greater than 1. The cross-sectional plane extends from the apex AP of the heart (CO) to the base BA. Therefore, the cross-section is called the "minor axis section".

[0174] Figure 4 Only the cross-sectional plane PC is shown in the figure. k The image derived from the cross section centered on it.

[0175] Preferably, for at least one cross-sectional plane PC (e.g., for each cross-sectional plane PC), multiple basic black blood images IM1i and / or multiple basic bright blood images IM2i are generated, where i = 1 to N, and N is an integer greater than or equal to 2. This makes the analysis of the obtained images more robust.

[0176] Alternatively, the system can be configured to acquire signals from a volume (e.g., from the entire heart) in order to generate a three-dimensional image.

[0177] Advantageously, the SE acquisition sequence is performed during the patient's breath-holding. One advantage is obtaining fully registered images, which allows for limiting, simplifying, or eliminating image registration.

[0178] In other words, the patient was in a breath-holding state throughout the entire collection sequence.

[0179] Alternatively, the SE acquisition sequence can be performed during free breathing. Free breathing acquisition has a time advantage. In practice, breath-hold acquisition must be rapid, which necessitates reducing acquisition time from seconds to minutes and often involves narrowing the imaging area, such as limiting it to a 3D portion. However, because the SE acquisition sequence is synchronized with the ECG, allowing readouts to be performed at the same time markers across different cardiac cycles, breath-hold acquisition results in a limited number of images.

[0180] Image generation The system is configured such that the processing and control unit TC receives the signals acquired during the acquisition step SE.

[0181] More generally, the system is configured such that the processing and control unit TC receives the signals acquired during the acquisition step.

[0182] The processing and control unit TC is configured to perform the processing (TRA) steps on the signals acquired during the acquisition sequence SE via a computer, such as... Figure 5 As shown.

[0183] The processing and control unit TC can be configured to implement a general processing step TRAG, which may include one or more processing steps TRAG applied to signals acquired during different acquisition sequences SE of the acquisition step.

[0184] The basic processing steps TRA include generating an image of the region to be imaged using the reconstruction technique GEN, which is known to those skilled in the art.

[0185] For example, the GRAPPA algorithm or the SENSE algorithm (and its iterative versions) can be used.

[0186] The generation step (GEN) includes the steps of generating a two-dimensional (2D) or three-dimensional (3D) basic image of the region to be imaged from the signal acquired during the acquisition sequence (SE).

[0187] In the 2D case, advantageously, for each black blood acquisition step ACQ1, a black blood image IM1i is generated from the signal acquired during the black blood acquisition step ACQ1i, and for each bright blood acquisition step ACQ2i, a bright blood image IM2i is generated from the signal acquired during the bright blood acquisition step ACQ2i.

[0188] The basic images IM1i and IM2i are advantageously generated in grayscale. Each image comprises a set of pixel or voxel-type unit elements, each unit element being characterized by an intensity I, which can take a set of N values ​​(N being a finite integer greater than 1), corresponding to N grayscale levels from 0 to N-1. For example, this value can take 256 values ​​between 0 and 255, but N is not limited to 256. Advantageously, this value can take 4096 values ​​between 0 and 4095.

[0189] The processing and control unit CT is also advantageously configured to use the acquired basic images to generate atlases and / or determine at least one piece of data relating to the area representing the cardiac lesion, as will be seen in more detail in the remainder of this article.

[0190] The basic image generated by the processing unit TC can be displayed on the screen of the human-machine interface INT.

[0191] Registration Advantageously, this method does not require an image registration step.

[0192] Alternatively, the image generation step GEN may include: registering basic dark-blood images to each other, and / or registering basic bright-blood images to each other, and / or registering basic dark-blood images and bright-blood images to each other. This allows for limiting the influence of breathing on heart position, particularly when the patient is breathing freely during the SE acquisition sequence, thereby avoiding image spatial shifts caused by breathing that could affect the accuracy and reliability of the analysis of these images or combinations thereof. In practice, respiratory rhythms are fundamentally different from heart rhythms, but even if there is a correlation between the two rhythms, it is possible for breathing to accelerate while the heart rate remains stable, or vice versa.

[0193] Advantageously, the method includes: registering basic black-blood images of the same cross section.

[0194] Advantageously, registration is performed by implementing a non-rigid image registration algorithm.

[0195] Advantageously, the method includes registering basic bright blood images of the same cross section.

[0196] Advantageously, registration is performed by implementing a non-rigid image registration algorithm.

[0197] Advantageously, the method includes: registering the basic black blood IM1i and bright blood IM2i images to each other.

[0198] Advantageously, this registration is performed by implementing a non-rigid image registration algorithm.

[0199] Advantageously, the method includes: registering a basic bright blood image and a basic dark blood image of the same cross section.

[0200] Advantageously, this registration is performed by implementing a non-rigid image registration algorithm.

[0201] Advantageously, these algorithms are identical. Different algorithms can be chosen to process the basic black-blood image and the basic bright-blood image, but for ease of implementation, the same algorithm is preferred. Registration significantly improves the quality of images derived from multiple basic images of the same cross-section, especially in terms of contrast. Furthermore, it reduces breathing-related artifacts.

[0202] According to the first example, at least one of these non-rigid algorithms is based on an image-to-image mutual information method, which is based on statistical relationships. The function to be optimized can be implemented using statistical similarity criteria. One advantage of this method is that pairings between identical attributes of images at the same cross-section are independent of their geometric location. Furthermore, this method is particularly effective for registering base images with different contrasts, such as black-blood images and bright-blood images.

[0203] Based on a second example compatible with the first example, implement at least one non-rigid algorithm based on a transformation model. The transformation model enables the determination of a function that minimizes the deviation between two images. The deviation can be translated into a geometric error to be minimized. Different methods can be used, such as those based on extracting geometric primitives or shape descriptors (e.g., salient points, shape singularities, or contours) from each image. Parametric or non-parametric methods can be used.

[0204] Based on the optimization examples of the transformation model or similarity criterion, the least squares method can be used.

[0205] Other optimization methods, such as gradient descent, can be implemented. However, the latter method is more specifically applied to image intensity and is not optimal within the scope of this invention, as the purpose of this invention is to optimize the sharpness and contrast of the fused image. Nevertheless, this embodiment is included in the invention.

[0206] Registration can be performed by selecting a reference image and determining the transformation function of other images of the same cross-section relative to that image. Each image is then registered by optimizing the transformation to obtain a reference image based on geometric criteria from the images under consideration.

[0207] When acquiring 3D images, multiple 3D images of the heart can be acquired, and these images can be registered.

[0208] Relaxation Time Graph Advantageously, the processing step TRA includes the following steps: generating a relaxation time map CA of the region to be imaged in CTO from bright blood signals acquired during a corresponding pair of CBi (where i=1 to N) of the cardiac cycle, the corresponding pair of CBi belonging to the acquisition sequence SE.

[0209] Advantageously, for this purpose, the bright blood image IM2i (i=1 to N) is used.

[0210] More precisely, the steps for generating the CTO relaxation time map (CA) include: - Determine the relaxation time associated with the tissue in the region to be imaged.

[0211] In the case of T1rho relaxation measurement, the relaxation time is T1rho.

[0212] In the case of T2 relaxation measurement, the relaxation time is T2.

[0213] Therefore, the relaxation time associated with at least one pixel or voxel of the bright blood image IM2i is advantageously determined based on the signal acquired for the pixel or voxel during different basic acquisition steps Sei (i=1 to N) of the acquisition sequence SE.

[0214] Advantageously, the intensity of the pixel or voxel in the bright blood image IM2i (where i = 1 to N) is used for this purpose, regardless of whether the bright blood image has been registered.

[0215] It is assumed that pixels or voxels occupying the same position in bright blood image IM2i (i=1 to N) come from the same region (pixel or voxel) of the region to be imaged.

[0216] More precisely, for that pixel or voxel, a relaxation curve is determined that is associated with the signal acquired for that pixel or voxel during the relaxation measurement sequence.

[0217] The relaxation curve is defined by the following relationship: SI(TCARACT) = M0e -TCARACT / TRELAX Where M0 is the equilibrium magnetization (dimensionless), and TCARACT is the characteristic duration in milliseconds. This duration is the spin-lock time or TSL for the T1rho relaxation measurement sequence, and the duration of the T2 preparation module for the T2 relaxation measurement sequence.

[0218] TRELAX is the relaxation time associated with a sequence, measured in milliseconds. For a T1rho relaxation measurement sequence, TRELAX is T1rho; for a T2 relaxation measurement sequence, TRELAX is T2.

[0219] SI (TCARACT) is the intensity of a pixel or voxel measured for the characteristic duration TCARACT.

[0220] This step includes, for example, determining, for each pixel or voxel, a relaxation curve from a set of relaxation curves that is closest (in the sense of a predetermined criterion) to the intensity (or signal) associated with that pixel or voxel for different feature durations.

[0221] The relaxation time associated with this decay curve is the relaxation time of the tissue associated with this pixel.

[0222] This step is implemented, for example, by linear regression (e.g., least squares or by a Levenberg-Marquardt type optimization method).

[0223] Alternatively, the curve can be obtained using the "dictionary matching" method.

[0224] This step includes, for example, identifying, for a pixel or voxel, a relaxation curve from a set of simulation-obtained relaxation curves that most closely approximates (in the sense of predetermined criteria) the intensity (or signal) associated with that pixel or voxel for different characteristic durations. This set of relaxation curves is obtained through block simulation or using an EPG (Extended Phase Map) framework. These tools enable simulation of the provided relaxation measurement sequence for a large number of T1rho or T2 values. The simulated signal is compared to the acquired signal. The T1rho (or T2) value corresponding to the curve that most closely approximates the MRI signal is retained. This method is performed on all pixels in the image.

[0225] The CTO step of generating the map CA enables the acquisition of the relaxation time map of all pixels or voxels in the region to be imaged, i.e., the bright blood image.

[0226] In other words, this step involves generating an associated pixel or voxel / relaxation time pair for each pixel or voxel in the region to be imaged. In other words, the relaxation time is associated with each pixel or voxel in the region to be imaged.

[0227] Alternatively, the atlas may include the relaxation times of all or part of the pixels or voxels of the bright blood image IM2i.

[0228] The processing and control unit TC is advantageously configured to generate at least one representation of the data determined during the overall processing step TRAG or the processing step TRA during the generation step GENR.

[0229] Advantageously, the processing and control unit TC is configured to control the display of at least one representation or image generated by the processing and control unit TC during the display step AFFD, the display being performed automatically or when a user sends a command to the processing unit via the input interface INTE of the human-machine interface INT.

[0230] The method advantageously includes a display step AFFD, in which at least one visual representation of the map CA is displayed on the screen of the system's output interface INTS.

[0231] One advantage is that it provides experts with indicators of diffuse changes in the tissue (especially the myocardium) of the area to be imaged.

[0232] For example, a color image can be displayed where the color of each pixel or voxel is defined by the relaxation time of that pixel or voxel. To do this, different and non-overlapping ranges of relaxation times are associated with different colors.

[0233] Alternatively or additionally, the visual representation of the map can be displayed in the form of a bullseye diagram, which will be described below.

[0234] Image combination Since multiple basic images of the same cross-section are generated, operations designed to combine (i.e. merge) these images can be performed to generate a single combined image for each cross-section.

[0235] Advantageously, the image generation step GEN includes: combining the bright blood images IM2i (i=1 to N) of the cross section to obtain the combined bright blood image ISB of the cross section.

[0236] Advantageously, the method includes: combining the black-blood images IM1i (e.g., i=1 to N) of the cross section to obtain the combined black-blood image ISN of the cross section.

[0237] This makes it possible to reduce noise and improve the signal-to-noise ratio (SNR).

[0238] Alternatively, combine M black blood images IM1i to obtain a combined black blood image (where M is an integer less than N and greater than 1), the M images being taken from N black blood images generated for the cross section; and / or combine P bright blood images IM2i (where P is an integer less than N and greater than 1), the P images being taken from N bright blood images generated for the cross section, to obtain a combined bright blood image.

[0239] Image compositing can be performed before, during, or after the step of generating a GEN image.

[0240] As an example, combination is averaging. For instance, combination is averaging. Averaging is performed, for example, in image space or in Fourier space (i.e., in the frequency domain before reconstructing the image). These solutions are fast and computationally inexpensive.

[0241] Averaging has the advantage of preserving image detail because it increases the surface-to-noise ratio (SNR). This technique enables noise smoothing to reduce residual image artifacts. Furthermore, averaging allows for increasing the bit depth of digital images beyond what can be achieved with a single image.

[0242] One advantage of averaging images from the same cross-section is noise reduction. The noise amplitude decreases with the square root of the number of images used; that is, with just four images, the noise amplitude can be reduced by half. Based on a two-minute free-breathing acquisition example, four to five images can be acquired for each cross-sectional plane, resulting in good noise reduction performance.

[0243] In one example embodiment, image combination can alternatively be implemented via motion-compensated iterative reconstruction. In other words, this type of combination is performed during image reconstruction. Compensated MRI reconstruction techniques are described in particular in the following articles: Odille F et al., “Coupled System Inversion Generalized Reconstruction (GRICS) for Free Breathing MRI,” *Medical Magnetic Resonance Imaging*, 2008; and Bustin A et al., “3D Whole-Heart Isotropic Submillimeter Resolution Coronary Magnetic Resonance Angiography Using Non-Rigid Motion-Compensated PROST,” *Journal of Cardiovascular Magnetic Resonance Imaging*, 2020.

[0244] Advantageously, the basic black blood image and the basic bright blood image are combined separately to generate a combined black blood image ISN and a combined bright blood image ISB for each cross section.

[0245] When acquiring three-dimensional images, multiple three-dimensional images of the heart can be acquired, and these multiple three-dimensional images can be combined (e.g., averaged) to obtain a single three-dimensional image of the heart.

[0246] Determination of data related to the area represented by cardiac lesions Advantageously, the processing steps TRA include: - Determine at least one data point relating to DET and at least one representative region ZI of cardiac lesions, the determination being made based on a reference black-blood image obtained from at least one of a basic black-blood image IM1i (where i = 1 to N) and at least one of a bright-blood image IM2i (where i = 1 to N), the basic black-blood image IM1i being generated from a black-blood signal acquired during the acquisition sequence SE, and the bright-blood image IM2i being generated from a bright-blood signal acquired during the acquisition sequence SE.

[0247] One advantage is that from the same acquisition sequence, not only relaxation time maps are determined, but also data related to the regions of cardiac lesions identified on dark blood images, which are enriched by the anatomical features of bright blood images.

[0248] Advantageously, this data is used to locate the ZI region, which represents cardiac lesions, or as an indicator of the location of the ZI region.

[0249] The cardiac lesion region ZI is advantageously obtained by segmenting a reference black blood image, as will be seen below.

[0250] According to one embodiment, the cardiac lesion representation region ZI is the region where the pixel intensity on the reference black blood image is greater than a predetermined intensity threshold.

[0251] This threshold can be defined before the method is implemented, or it can be defined based on the intensity range of a reference black blood image.

[0252] This threshold can be global, meaning it is the same for all pixels or voxels of the reference black blood image.

[0253] Alternatively, the threshold may be local. In other words, it is defined as varying based on the pixels or voxels of a reference black blood image.

[0254] According to one embodiment, the reference black-blood image is a combined black-blood image ISN obtained by combining N black-blood images IM1i (i=1 to N). One advantage is that it takes advantage of the combination of black-blood images, especially the improvement in signal-to-noise ratio.

[0255] Advantageously, the combination is average.

[0256] Alternatively, the combination may be one of the other types described previously.

[0257] Alternatively, the reference black blood image is one of the black blood images IM1i (i is between 1 and N), or a combined image obtained by combining M black blood images selected from N black blood images IM1i (i = 1 to N).

[0258] In one particular embodiment, the data is determined from N bright blood images IM2i (i=1 to N).

[0259] For example, the data is determined from a reference bright blood image derived from N bright blood images (i=1 to N).

[0260] The inventors discovered that combined bright blood image ISBs, especially averaged combined bright blood images, enable the acquisition of substantially the same myocardial anatomy information as combined bright blood images (e.g., averaged combined bright blood images) obtained from N bright blood images generated by signals acquired through the same basic acquisition sequence.

[0261] Alternatively, the data can be determined based on only one of the bright blood images IM2i (for a given index i).

[0262] One advantage is that it avoids the impact of variations in feature duration (TSL or echo duration) on the quality of the generated data. In fact, changing the feature duration has the effect of slightly altering the contrast of bright blood images.

[0263] Advantageously, when the relaxation measurement sequence is a T1rho relaxation measurement sequence, a bright blood image is generated using a bright blood signal acquired based on a spin-locked time TSL of greater than or equal to 20 ms.

[0264] For example, a bright blood image is generated using signals acquired by implementing a basic acquisition sequence Sei with the longest spin-locked time TSL, or a basic acquisition sequence Sei with the longest T2 preparation module duration (where i = 1 to N). This bright blood image is the image with the highest contrast between blood and myocardium, which allows for the acquisition of a wealth of anatomical information about the myocardium.

[0265] Advantageously, when the relaxation measurement sequence is a T1rho or T1ρ relaxation measurement sequence, a bright blood image is generated using a bright blood signal acquired based on a spin-locked time TSL of ≥20 ms.

[0266] According to another embodiment, the data is determined based on G bright blood images IM2i selected from N bright blood images, where G is an integer greater than 1 and less than N.

[0267] For example, G bright blood images IM2i are generated based on the bright blood signal acquired from the largest spin-lock time TSLi among N spin-lock times TSLi (where i = 1 to N), or G bright blood images IM2i are generated based on the bright blood signal acquired from the largest T2 preparation module duration among N T2 preparation module durations (where i = 1 to N).

[0268] Determining the representative area of ​​cardiac lesions In a particular embodiment of the invention, each bright blood collection step ACQ2 is different from the inversion recovery sequence.

[0269] In other words, the sequence does not have a reversal pulse of longitudinal magnetization of the region to be imaged.

[0270] Therefore, unlike PSIR imaging during bright blood collection, the longitudinal magnetization of the myocardium is not canceled out, resulting in images with higher contrast between the myocardium and blood, which is beneficial for image processing.

[0271] Alternatively, each bright blood collection step is an inversion recovery sequence. This inversion recovery sequence is, for example, a PSIR ("phase-sensitive inversion recovery") sequence.

[0272] Advantageously, the blood collection configuration is such that, when the reading module LE2 is implemented, the corresponding longitudinal magnetizations of healthy myocardium, blood, and lesions are positive, and the longitudinal magnetization of the lesions lies between the magnetizations of the healthy myocardium and the blood. This is achieved through the configuration of the preparation module PREP2i, the configuration of the reading module LE2, and the relative time positioning between the two modules.

[0273] like Figure 5 As shown, the steps for determining data relating DET to the ZI region, a representative area of ​​cardiac lesions, advantageously include the following steps: - By performing a first segmentation SEG1 on at least one reference bright blood image derived from at least one bright blood base image IM2i (i=1 to N) using a computer, localization data defining the first wall L1 of the myocardium is generated. - A second segmentation, SEG2, is performed on the reference black blood image using the localization data of the first wall L1 obtained from the data calculated during the segmentation step SEG1, in order to determine the representative area ZI of the cardiac lesion.

[0274] These steps are implemented by the processing and control unit TC.

[0275] In the remainder of this article, such as Figure 5 As shown in the example, each bright blood image ISB is considered to be a combined bright blood image ISB, and each dark blood image ISN is a combined dark blood image ISN.

[0276] One advantage is that it can automatically, repeatably, reliably, and accurately segment the ZI region, which represents cardiac lesions.

[0277] In fact, segmenting the bright blood images (ISBs) generated from signals measured during the bright blood acquisition step ACQ2i (especially when they differ from inversion recovery sequences) allows for automatic, robust, reliable, and accurate localization of the myocardial wall, as these images exhibit significant contrast between the myocardium and blood. Dark blood images, due to the lack of contrast between healthy myocardium and blood, do not yield the same good results.

[0278] Determining (i.e., identifying, extracting, or delineating) representative areas of cardiac lesions by segmenting reference dark blood images can yield excellent results that cannot be obtained from bright blood images containing little or no lesion information alone.

[0279] First section According to one example, at least one bright blood image ISB is used to perform a first segmentation SEG1 to generate location data for a first wall L1 and a second wall L2 that define the myocardium M and surround and define the chambers of the heart, with the first wall L1 surrounding the second wall L2.

[0280] Location data associated with the wall, such as identifiers corresponding to the pixels that make up the wall.

[0281] The result of this segmentation can be seen in Figure 6 It schematically shows a bright blood image ISB of the heart section in the upper left and a bright blood image of the same section in the upper right, on which the first wall L1 and the second wall L2 obtained during the first segmentation step SEG1 are indicated by thick black lines.

[0282] The generated image is, for example, a grayscale image. Figure 6 The white areas in the diagram represent areas brighter than the dotted areas, the dotted areas represent areas brighter than the cross-shaded areas, and the cross-shaded areas represent areas brighter than the brick pattern areas.

[0283] exist Figure 6 In the example, the cardiac chamber is the left ventricle, and the first segmentation SEG1 is implemented to depict the walls L1, L2 of the myocardial portion surrounding and defining the left ventricle.

[0284] It should be noted that in this specification, the invention is described with the left ventricle as the cardiac chamber, but the invention is applicable to any cardiac chamber, such as the right ventricle and atrium, which are also surrounded and defined by myocardium and are also susceptible to cardiac lesions.

[0285] The second wall, L2, defines the myocardium and the left ventricular ventricle (VG). The first wall, L1, surrounding the second wall, is the outer wall of the myocardial portion surrounding the left ventricular VG, i.e., the wall facing outwards from the left ventricular VG. This first wall, L1, is the epicardium.

[0286] The second wall, L2, is the myocardial wall that defines the left ventricle. This second wall, L2, is the endocardium.

[0287] exist Figure 6 In the image (which is a cross-sectional image of the heart), these walls L1, L2 form closed curves because they completely surround the left ventricle VG in the short-axis section.

[0288] It's easy to understand that in 3D, these walls form surfaces.

[0289] Alternatively, the first segmentation SEG1 is implemented to generate localization data for only one of these two walls, such as the localization data for the outer wall of the myocardium.

[0290] The first segmentation SEG1 is performed by implementing a learning function or algorithm (such as an artificial neural network) to segment the bright blood image, thereby depicting at least one wall of the myocardium surrounding and defining the ventricles of the heart.

[0291] In the unrestricted example, the learning function is a neural network.

[0292] Convolutional neural networks are advantageous for use in segmentation.

[0293] Convolutional neural networks are, for example, of the U-Net type.

[0294] Alternatively, artificial neural networks can be transformer-type, also known as self-attention-based models, such as the type commonly referred to as "swin transformer".

[0295] A neural network (or more generally a learning function) is implemented on two-dimensional (2D) and / or three-dimensional (3D) images. In other words, it is trained to perform the desired segmentation by receiving 2D and / or 3D images as input.

[0296] Advantageously, the learning function is a trained learning function.

[0297] Advantageously, prior to implementing the method according to the invention, the learning function is trained using a bright blood image of the heart (more precisely, a bright blood image of the region to be imaged), which is generated from signals acquired during the corresponding bright blood acquisition step and labeled (i.e. segmented by an expert), such that the trained learning function, upon receiving input data including a bright blood image of the heart (more precisely, a bright blood image of the region to be imaged), is able to segment to depict at least one wall of the myocardium surrounding and defining the heart chamber.

[0298] When the bright blood acquisition sequence ACQ2i is different from the inversion recovery sequence, the learning function is performed using bright blood images of the heart generated by implementing an acquisition sequence different from the inversion recovery sequence.

[0299] When the reference bright blood image is a composite image ISB, the bright blood image used for learning is also a composite image obtained from the signal acquired during the relaxation measurement sequence.

[0300] Preferably, the relaxation measurement sequence is the same as the acquisition sequence consisting of the bright blood acquisition sequence ACQ2i (i=1 to N).

[0301] In one particular embodiment, prior to implementing the method according to the invention, the learning function is trained from multiple sets of training images of the region to be imaged, the multiple sets of training images being generated from training signals acquired during the corresponding acquisition steps of bright blood delayed gadolinium enhanced magnetic resonance.

[0302] The images in the image group are labeled by experts (i.e. segmented by experts) so that the trained learning function, after receiving input data including bright blood images IM2i (i=1 to N), can segment the bright blood images in a manner that generates localization data defining at least one wall of the myocardium.

[0303] Each training image was obtained from bright blood signals acquired by performing a relaxation measurement sequence.

[0304] Preferably, the relaxation measurement sequence is the same as the acquisition sequence consisting of the bright blood acquisition sequence ACQ2i (i=1 to N).

[0305] In this particular embodiment, the learning function is configured to generate localization data from input data, which includes N bright blood images IM2i generated from bright blood signals acquired during the acquisition sequence SE.

[0306] The generation step GENR may include generating a bright blood output image corresponding to a bright blood image (e.g., a reference image), in which the intensity of pixels or voxels corresponding to walls or contours L1 and L2 is replaced with a predetermined intensity or color, thereby enabling the wall or contour to be distinguished from other tissues in the image.

[0307] The method may include the step of displaying the output image.

[0308] spread The method may include the step of propagating the walls detected during the first segmentation step SEG1 onto the reference black-blood image ISN and / or the atlas CA.

[0309] In other words, the method may include a step of transferring REP (i.e., propagation), which involves identifying pixels or voxels on the black blood image ISN or atlas that correspond to walls L1 and L2 identified during the first segmentation SEG1.

[0310] exist Figure 6 In the image, a black blood image ISN of a cross-section of the heart is schematically shown in the lower left corner, and a black blood image is shown in the lower right corner, on which the walls L1 and L2 detected during the first segmentation step SEG1 are represented by thick black lines.

[0311] Identifying pixels or voxels corresponding to the first wall L1 and the second wall L2 on the black blood image ISN or map CA is determined based on the positions of pixels or voxels corresponding to these walls on the bright blood image ISB.

[0312] When these images are considered to be spatially registered and because they have the same size and resolution, these pixels or voxels can have the same position on the bright blood image and the dark blood image or atlas.

[0313] When it is believed that there is a spatial offset between these images, a predetermined or calculated spatial offset can be applied to these pixels or voxels instead.

[0314] The transfer may include annotating or coloring the pixels or voxels corresponding to walls L1 and L2.

[0315] Second division The second segmentation is performed by the processing unit CT.

[0316] The steps for determining data related to the DET and the ZI region, which represents the cardiac lesion, may include detecting the DET in the ZI region.

[0317] Advantageously, the second segmentation SEG2 is performed when a representative area of ​​cardiac lesions, ZI, is detected. In other words, the second segmentation SEG2 can only be performed if a representative area of ​​cardiac lesions, ZI, is detected during detection step DE.

[0318] Alternatively, the detection step DE can be performed after the second segmentation step SEG2.

[0319] Alternatively, it can be determined that step DET did not detect step DE.

[0320] The testing steps will be described below.

[0321] The second segment SEG2 uses the black blood image ISN and the localization data of the first wall L1 derived from the first segment SEG1, as well as the possible localization data of the second wall L2.

[0322] This location data can be generated during the first segmentation step SEG1, or it can originate from the location data in the transfer step REP.

[0323] Alternatively, the second segmentation step SEG2 may include a transfer step.

[0324] The second segmentation, SEG2, enables the location of one or more representative areas of cardiac lesions, i.e., the generation of location data for these areas.

[0325] These location data include, for example, the identifiers or locations of pixels or voxels corresponding to the areas represented by these cardiac lesions.

[0326] In other words, the second segmentation SEG2 enables the extraction (i.e., depiction) of representative areas of cardiac lesions on a reference black blood image.

[0327] The second segmentation, SEG2, is implemented, for example, through thresholding.

[0328] The second segmentation SEG2 advantageously includes identifying pixels or voxels with an intensity greater than or equal to a predetermined intensity threshold only within a predetermined search region Z of the reference black-blood image ISN defined by the first wall L1 and / or the second wall L2.

[0329] In fact, from Figure 2It can be inferred that, in the black blood image, the area representing cardiac lesions has a high intensity compared to healthy myocardium and blood.

[0330] The search area Z is determined based on the location data of the first wall L1 derived from the segmented SEG and the possible location data of the second wall L2.

[0331] The search area is, for example, the region of the reference black-blooded image ISN defined by pixels or voxels transferred to the first wall L1 of the black-blooded image ISN and / or pixels or voxels transferred to the second wall L2 of the black-blooded image ISN.

[0332] Advantageously, the search area for the black blood image is the area surrounded and defined by the first wall L1.

[0333] In other words, the second segmentation step SEG2 involves searching only for pixels or voxels with an intensity greater than or equal to a predetermined intensity threshold within a search region defined and surrounded by the first wall L1 on the reference black-blood image ISN. In other words, these pixels or voxels are selected only from pixels or voxels within the region of the reference black-blood image surrounded and defined by the first wall L1. This avoids detecting lesion-representing regions outside the epicardium.

[0334] Alternatively, the search area Z is the region of the black blood image ISN defined by the first wall L1 and the second wall L2.

[0335] In other words, the second segmentation SEG2 involves searching only for pixels with an intensity greater than or equal to a predetermined intensity threshold within the region defined by two walls L1 and L2 of the reference black-blood image ISN. This alternative has the advantage of identifying only pixels or voxels of the myocardium.

[0336] Alternatively and / or additionally, the second segmentation SEG2 includes searching only for pixels with an intensity greater than or equal to a predetermined intensity threshold within the region surrounded by wall L2. This step enables the identification of only pixels or voxels of the papillary muscle.

[0337] As an alternative, the second segmentation is implemented using a neural network, such as a trained convolutional neural network, to segment a representative region of cardiac lesions with predetermined features in a region defined by walls L1 and / or L2, the neural network receiving localization data of the corresponding walls and black blood images as input; or the second segmentation is implemented using at least one active contour segmentation algorithm, i.e., a segmentation algorithm using an active contour model.

[0338] Alternatively, it can be determined that DET does not have a first segmentation step.

[0339] The determination process includes segmenting a reference dark blood image to identify the ZI region representing the cardiac lesion.

[0340] This step is performed, for example, by thresholding or by using a neural network as described above.

[0341] The detection step DE includes using the black blood image ISN and localization data of at least one wall (e.g., the second wall L2) to detect the presence or absence of the area representing the cardiac lesion.

[0342] This step outputs an indication at the output terminal whether the area representing the heart lesion is present or not.

[0343] The detection step DE can be performed by thresholding or using a neural network, such as in the segmentation step. This step may include: determining, within a region defined by wall L1 and / or wall L2 (the location of which is defined for the myocardium during the first segmentation step SEG1), whether the number of consecutive pixels or voxels is greater than a predetermined threshold and whether the intensity of the consecutive pixels or voxels is greater than a predetermined threshold. If this condition is met, a region representing a cardiac lesion is detected; if this condition is not met, a region representing a cardiac lesion is detected as non-existent.

[0344] Neural networks, for example, are convolutional neural networks.

[0345] The neural network, for example, is trained to detect the presence of a cardiac lesion representative area in the region defined by walls L1 and / or L2 when it receives location data of the corresponding wall and a reference black blood image as input.

[0346] exist Figure 7 The image shows a black blood image ISN, on which boundaries L1 and L2 and pixels identified as representing the cardiac lesion region ZI are shown in thick lines. Boundaries L1 and L2 were identified during segmentation and transferred (i.e. propagated) onto the black blood image ISN, for example, to define the search region Z.

[0347] The DET process may optionally include at least one of the following steps: -Calculate the size TA of the CTA cardiac lesion representative region ZI based on the location data of the first wall L1 and possibly the second wall L2 derived from the first segmentation step SEG. -Calculate at least one transmural degree TR of the CTT cardiac lesion representative region ZI based on the first wall L1 and the second wall L2 derived from the first segmentation step SEG1.

[0348] These calculations are performed using a set of at least one black-blooded image.

[0349] The size of the ZI region representing a cardiac lesion should be understood as data representing the size of that region, such as volume or area, or the number of pixels or voxels.

[0350] Size calculation Determining DET advantageously includes the step of calculating the size of the ZI region, which represents the cardiac lesion in CTA.

[0351] The computational step CTA includes determining at least one basic data point representing the size of at least one cardiac lesion representative region (e.g., a lesion representative region of the myocardium) by using location data of the first wall and / or the second wall L1, L2. This location data is, for example, location data directly derived from the first segmentation SEG1 of the myocardium, or data derived from these data, such as data derived from the transfer step, or location data of the cardiac lesion representative region ZI obtained during the second segmentation step SEG2.

[0352] Data items representing the size of region ZI can be the percentage of myocardial surface area occupied by region ZI on sector SEC of the reference black-blood image ISN, or the volume or mass of the cardiac lesion representing region ZI in sector SEC, starting from axis l parallel to axis p and substantially passing through the center of the heart chamber on the black-blood image ISN, and defined by two rays R emanating from axis l, as visible on the black-blood image ISN.

[0353] The percentage of the myocardial area represented by the cardiac lesion region ZI in sector SEC can be calculated based on the ratio of the number of pixels in sector SEC corresponding to the cardiac lesion region ZI to the number of pixels in sector SEC corresponding to the myocardium.

[0354] The number of pixels in the sector SEC corresponding to the cardiac lesion representative region ZI can be calculated based on the location data obtained during the second segmentation step SEG2, or it can be calculated directly during the calculation step CTA, for example by thresholding to select the number of pixels with an intensity greater than a predetermined threshold in the sector SEC portion defined by walls L1 and L2 or by wall L1.

[0355] The CTA step may include: calculating a data item representing the size of the cardiac lesion representative region ZI in the sector SEC based on multiple basic data representing the size of the cardiac lesion representative region ZI calculated for multiple black blood image ISNs distributed along axis p in sector SEC.

[0356] For example, calculating combinations or averages of basic data.

[0357] When the image is two-dimensional, the volume of the cardiac lesion representative region ZI on sector SEC can be calculated based on the ratio of the number of pixels on that sector corresponding to the cardiac lesion representative region ZI to the number of pixels on that sector corresponding to the myocardium, and the thickness of the cross section corresponding to the reference black blood image ISN.

[0358] Size and / or volume are also advantageously calculated from the predetermined resolution of the image.

[0359] It should be noted that the myocardial density is 1.06 g / mL. Therefore, the mass of the ZI representing the cardiac lesion is considered to be substantially equal to its volume, which makes it possible to assess the mass of the ZI representing the cardiac lesion. The calculation steps of CTA may include, for example, dividing the reference black-blood image ISN into a first predetermined number (in... Figure 1 In a non-restrictive example, sectors SEC with the same subtended angle α1 and pointing towards axis l (equal to 12) are used, and the percentage of myocardial area occupied by the cardiac lesion-represented region ZI in different sectors SEC is calculated.

[0360] The first number of sectors and the subtended angle α1 can be determined based on the cross-sectional plane PC. k And change. For example, with the cross-sectional plane PC k Along the long axis closer to the apex, the number of sectors decreases and the opening angle increases.

[0361] The overall processing steps of TRAG advantageously include the following steps: generating by the processing unit via a computer at least one set of representations of GENR data related to the ZI region of cardiac lesions; and displaying at least one representation from the set of at least one representations of AFFD on the screen of the human-machine interface.

[0362] The generation step GENR includes, for example, generating a data item representing the result of the detection step DE (i.e., the presence or absence of the ZI region representing the cardiac lesion), and the display step includes displaying the data item on the screen of the human-machine interface.

[0363] The set of at least one representation advantageously includes a first representation REPT of data (or data items) relating to a cardiac lesion representative region ZI, which is calculated during a calculation step CTA associated with at least one processing step TRA.

[0364] like Figure 7 As shown, a bullseye-type representation can be generated, for example, representing a percentage or data representing that percentage, where the percentage is the percentage of myocardial surface area occupied by the cardiac lesion-representing region ZI in different sectors of a black blood image ISN acquired along the corresponding cross-sectional plane PC.

[0365] Bullseye diagrams are defined by the American Heart Association (AHA) and described in the following article: Manuel D. Cerqueira et al., “Standardized myocardial segmentation and nomenclature for cardiac tomography: A statement to healthcare professionals from the American Heart Association Committee on Cardiac Imaging, Circulation, 2002; 105:539-42.

[0366] A bullseye diagram representation consists of multiple concentric circles CE, separated pairwise by annular bands CO. Each annular band CO is assigned to a section or a set of consecutive sections, where the closer the annular band CO is to the section or section group corresponding to the apex, the closer it is to the center of the circle. Each annular band CO is divided into sectors PSE, and the diagram is displayed within the sectors PSE (e.g., ...). Figure 7 (As shown in the example) The percentage of myocardial area represented by the cardiac lesion ZI is calculated for the corresponding sector of the black blood image ISN for the corresponding cross section, or a combination of percentages (e.g., average) calculated based on the percentage of myocardial area represented by the cardiac lesion ZI calculated for the sector of the black blood image for the corresponding cross section group.

[0367] Alternatively and / or additionally, the pixel intensity of different annular portions depends on the calculated percentage. The lower the percentage, the higher the intensity of the corresponding annular portion.

[0368] For example, in Figure 7 In the figure, the corresponding average of the size percentage of the representative region ZI of cardiac lesions calculated in each sector defined on three sets of consecutive black-blood images distributed along axis p is shown in the form of a known bullseye diagram. These three sets of consecutive black-blood images are associated with three rings corresponding to the apical section (inner ring), the middle of the heart (middle ring), and the basal region (outer ring), respectively.

[0369] Sector portions associated with a percentage greater than 80% are indicated by dashed lines, while sector portions associated with a percentage less than or equal to 80% are indicated by white lines.

[0370] When generating a 3D image of the heart, it is advantageous to divide the image into multiple layers along axis p, and calculate the same data based on these different layers as the data calculated based on the two-dimensional image.

[0371] According to one embodiment, the generation step includes generating a bullseye-type representation from a relaxation time map.

[0372] Advantageously, the spectra calculated for different cross sections are used.

[0373] This representation is generated based on the relaxation time map CA and the positioning data of walls L1 and L2 generated during the first segmentation SEG1.

[0374] Then, the PSE sector portion is correlated with the corresponding relaxation time value or the median or standard deviation of the relaxation time.

[0375] Permeability The determination process for DET advantageously includes the step of calculating the transmural percentage of the representative area of ​​cardiac lesions (ZI). The transmural percentage refers to the percentage of myocardial thickness occupied by the representative area of ​​cardiac lesions (ZI).

[0376] The calculation step CTT includes determining at least one piece of data, using lesion location data and localization data derived from the first and / or second walls L1, L2 of the segmentation step, to determine the percentage of transmural permeability representing at least one cardiac lesion-representing region ZI.

[0377] This data could be the percentage of myocardial thickness occupied by the cardiac lesion-represented region ZI on the sector starting from axis l of the black blood image ISN.

[0378] The percentage of myocardial thickness occupied by lesions in a sector can be calculated based on the ratio of the number of pixels in that sector corresponding to the ZI thickness of the representative area of ​​the cardiac lesion to the number of pixels in that sector corresponding to the myocardial thickness. The number of pixels corresponding to the ZI thickness of the representative area of ​​the cardiac lesion can be a value obtained from the result of the lesion segmentation step SC, or it can be calculated during the calculation step CTT based on the L1 and possibly L2 localization data derived from the first segmentation step SEG1 through, for example, thresholding.

[0379] The number of pixels corresponding to the ZI thickness of the heart lesion representative area in a sector can be the average or maximum number of pixels corresponding to the ZI thickness of the heart lesion representative area, calculated at different radii of that sector.

[0380] The number of pixels corresponding to myocardial thickness in a sector can be the average or maximum number of pixels corresponding to myocardial thickness calculated at different radii of that sector. These values ​​are calculated based on the positioning data of walls L1 and L2.

[0381] The calculation step CTT may include, for example, dividing the black blood image ISN into a second predetermined number of sectors with the same opening angle pointing to the center of the heart chamber, and calculating the percentage of myocardial area occupied by the lesion in the different sectors.

[0382] The second number of sectors and the resulting subtending angle can vary depending on the cross-sectional plane PC. For example, as the cross-sectional plane PC is closer to the apex, the number of sectors decreases and the subtending angle increases.

[0383] Advantageously, the second quantity is greater than the first quantity.

[0384] The generation step GEN may include: calculating the data items representing the permeability of the sector based on the multiple basic data items representing permeability calculated in the sector from the multiple black blood image ISNs distributed along the axis p.

[0385] For example, calculating combinations or averages of basic data.

[0386] The GENR step advantageously includes generating a representation of the percentage of permeability, REPTR. The AFFD display step advantageously includes displaying this representation.

[0387] For example, a bullseye-type representation can be generated, which represents a combination (e.g., an average) of the percentage of transmural lesions in sectors of multiple sets of consecutive black-blood images ISNs acquired along various cross-sectional planes PC.

[0388] The pixel intensity of this image advantageously, but not necessarily, represents the percentage of translucency.

[0389] For example, in Figure 7 The content is shown in the form of a bullseye-type representation REPTR, which represents the average percentage of transmural penetration of the cardiac lesion representative region ZI calculated in defined sectors on multiple sets of consecutive black blood images acquired along the axis p.

[0390] As mentioned earlier, a bullseye diagram represents multiple sector segments whose intensity corresponds to a combination of percentages of permeability calculated for that sector segment.

[0391] The lower the percentage, the higher the strength of the corresponding band. Alternatively and / or additionally, the percentage can be displayed in the sector section.

[0392] fused images In one particular embodiment, the determination step DET includes generating a fused image by replacing the pixel or voxel values ​​of a first reference bright blood image derived from bright blood signals acquired during at least one of a plurality of pairs, the new values ​​enabling the differentiation of the cardiac lesion representative area from other tissues in the area to be imaged.

[0393] This step can be one of the steps in the generation process.

[0394] This fused image is an indicator of the location of the ZI region, representing the cardiac lesion.

[0395] For example, the new pixel value is color.

[0396] The method advantageously includes the step of displaying the fused image.

[0397] One advantage is that it allows experts to easily locate representative areas of cardiac lesions relative to the anatomy of the myocardium.

[0398] advantage The solution provided enables the acquisition of images with sufficient resolution and contrast to reliably and repeatably detect and identify data associated with representative areas of cardiac lesions.

[0399] Furthermore, by separating two crucial pieces of information—cardiac anatomy and the region representing the lesion—onto two distinct images (a bright blood image and a dark blood image, respectively), an automated method for characterizing the region representing the lesion can be implemented. Compared to existing technological solutions, this automation significantly saves time and improves repeatability.

[0400] The method's black-blood and bright-blood acquisition sequences require relatively short acquisition times (especially when acquiring signals to generate 2D images involving minimal computation). This advantageously allows for the acquisition sequence to be performed during breath-holding and limits cardiac movement between images and the resulting corrections, thus enabling computational resource constraints and real-time implementation of the method. It also helps limit artifacts that impair image readability. These artifacts exacerbate the difficulty of reconstructing clear and accurate images to locate and detect areas representing lesions. Furthermore, long MRI acquisition times are uncomfortable for patients. Durations of 10 to 20 minutes are considered quite long, and it is difficult for patients to remain still during MRI.

[0401] Furthermore, in the case of generating 2D images, artifacts from images extracted from cross-sectional planes of 3D images are avoided. These artifacts can lead to situations where it is impossible to distinguish the presence of an area representing a lesion from the presence of blood located near muscle. In fact, in some cases, lesions are so close to blood (called subendocardial) that it is difficult to tell in an image with artifacts whether it represents an area representing a lesion, blood, or an image artifact.

[0402] Furthermore, the provided solution enables the generation of relaxation time maps to quantify diffuse tissue changes, potentially allowing experts to identify cardiomyopathy (e.g., diffuse and acute cardiomyopathy). The solution eliminates the need for specific acquisition sequences for this purpose, avoiding numerous drawbacks such as extended examination time, additional breath-holding for the patient, extra workload for radiology operators, and complex analysis due to the lack of spatial recalibration between the two sequences.

[0403] hardware From a hardware perspective, the processing and control unit (TC) or processing unit can be viewed as a calculator that interacts with computer programs.

[0404] The processing and control or processing unit TC includes, for example, a calculator that includes at least one set of processors and may include memory operatively coupled to the calculator.

[0405] The memory may include, for example, a computer-readable medium. This computer-readable medium is a tangible device that can be read by a reader of the processing unit, is capable of storing electronic instructions, and is coupled to a communication unit.

[0406] In other words, a machine-readable medium is a tangible medium. That is, it is not itself a transient signal, such as radio waves or other freely propagating electromagnetic waves (e.g., light pulses or electronic signals). Such computer-readable storage media include, for example, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any combination thereof.

[0407] As examples, readable media include optical discs, magneto-optical discs, ROM (read-only memory), EPROM (erasable programmable read-only memory), EEPROM (electrically erasable programmable read-only memory), RAM (random access memory), magnetic cards, or optical cards.

[0408] The readable medium may include an operating system and load the program according to the invention. The readable medium includes registers adapted to record parameter variables created and modified during the execution of the program. A computer program including software instructions is then stored on the readable medium.

[0409] Alternatively, the program instructions may originate from an external source and be downloaded over a network. This is especially true for applications.

[0410] The processing and control or processing unit TC includes a calculator, i.e., at least one data processing electronic circuit, which is designed to manipulate and / or convert data represented by electronic or physical quantities in registers of an evaluation system and / or memory into other similar data corresponding to physical data in register memory or other types of display devices, transmission devices or storage devices.

[0411] The processing and control or processing unit CT includes, for example, a memory for storing data, which is operatively coupled to data processing circuitry, and a driver adapted to read computer-readable media.

[0412] The memory is advantageously configured to store acquired signals and / or data generated during processing, such as black blood images and / or bright blood images and / or at least one relaxation time map and / or data related to regional ZI and / or one or more representations and / or at least one other image (e.g., a fused image) generated during processing.

[0413] The processing and control or processing unit TC includes at least one computer, such as a microcomputer, computer network, electronic components, tablet computer, smartphone, or personal digital assistant (PDA).

[0414] The steps of the method according to the invention are performed, for example, by causing the processing circuitry of the CT processing and control unit to read a predetermined program stored on hardware such as a memory, causing its data processing circuitry to perform calculations, control the communication and / or components of the imaging system, and read and / or write data in the memory.

[0415] Processing steps or overall processing steps are performed, for example, on a processing device (such as a single computer) or on a system distributed among multiple computers (especially by using cloud computing).

[0416] The processing and control or processing unit CT includes at least one calculator, which includes at least the following elements: a collection of one or more processors (e.g., at least one central processing unit (CPU), and / or at least one graphics processing unit (GPU), and / or a microcontroller and / or a digital signal processor (DSP) ASIC capable of interpreting instructions in the form of a computational program) and / or a hardware collection, such as an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a programmable logic device (PLD), a programmable logic array (PLA), a system-on-a-chip (SoC), and / or an electronic board in which the steps of the method according to the invention are implemented in hardware elements.

[0417] The present invention relates to a computer program product comprising a computer-readable medium containing instructions which, when executed by processing circuitry, cause a system S to perform the steps of the method according to the present invention.

[0418] Program products may include computer-readable recording media.

[0419] The present invention also relates to a computer-readable medium having a computer program, i.e., a computer program product, recorded thereon according to the present invention.

[0420] Alternatively, the program instructions may originate from an external source and be downloaded over a network. This is especially true for applications. In this case, the computer program product includes a computer-readable data medium on which program instructions are stored, or a data medium signal on which program instructions are encoded.

[0421] Program instructions can take the form of source code, computer-executable code, or any intermediate form between source code and computer-executable code, such as the form obtained by transforming source code through an interpreter, assembler, compiler, linker, or locator. Alternatively, program instructions can be microcode, firmware instructions, state definition data, integrated circuit configuration data (e.g., VHDL), or object code. Program instructions are written in any combination of one or more programming languages, such as object-oriented programming languages ​​(C++, Java, Python) or procedural programming languages ​​(e.g., C).

[0422] The communication unit includes at least one communication device that allows communication between system components and may allow communication between at least one component of the system and external devices. The communication system can establish physical links and / or remote (wireless) communication links between system components and / or between system components and external devices.

[0423] Communication devices may include any hardware, firmware, and / or software adapted to transmit information between components of the device to which the communication device belongs (e.g., via a data bus) or to components outside the device. In order to enable data communication between different devices to which the communication device belongs, such devices may include, where applicable, firmware and / or software for establishing wired or wireless communication links (e.g., Wi-Fi, Bluetooth, cellular, or Ethernet) between them.

[0424] The user interface INT allows users to input data or instructions to interact with the program according to the invention.

[0425] User interfaces (INTs) include, for example, output interfaces (INTS) and input interfaces (INTE).

[0426] Input interfaces include, for example, keyboards or pointing interfaces, such as mice, light pens, touchpads, remote controls, voice recognition devices, and haptic devices.

[0427] Output interfaces (INTS) are designed to present information to the user in a sensory or electrical manner (e.g., visually or audibly). Output interfaces include, for example, displays or screens. Display steps (AFFDs) can be steps that present information through means other than displays or screens.

[0428] The output interface INTS can be the input device INTE, such as in the case of a touch tablet.

[0429] Advantageously, the system S or C is configured such that a user can select one or more representations or images generated during the processing step TRAG or TRA by transmitting instructions to the processing and control or processing unit TC via the input interface INTE, such that each selected image or representation is displayed on one or more screens of the output interface INTS during the display step AFFD.

[0430] For example, users can select one or more representations or images to be generated for a predetermined section through the input interface.

[0431] Advantageously, the user can select a predetermined section from multiple sections via the input interface, such that one or more images and / or one or more representations generated during the processing step TRAG are automatically displayed on one or more screens of the output interface, or the user can select one or more representations and / or one or more images generated for the predetermined section.

[0432] Alternatively, the processing and control unit, or the processing unit, may be configured to automatically display only one or more images and / or one or more representations generated for a predetermined section during the processing step TRAG, during the display step.

Claims

1. A method for acquiring magnetic resonance signals for cardiac imaging, the method comprising an acquisition sequence SE, the acquisition sequence comprising: - Acquiring signals, said acquired signals being included during each of multiple consecutive pairs consisting of two consecutive cardiac cycles: - Acquiring (ACQ1i) black blood signals from the target area of ​​an individual's heart using black blood-delayed gadolinium-enhanced magnetic resonance imaging, which enables the generation of a basic black blood image of the target area. - Bright blood signals (ACQ2i) are acquired from the region to be imaged using bright blood-delayed gadolinium-enhanced magnetic resonance imaging (fMRI), and the dark blood signals enable the generation of a basic dark blood image of the region to be imaged. The bright blood signals acquired during the consecutive multiple pairs are acquired by performing a relaxation measurement sequence.

2. The data acquisition method according to claim 1, wherein, The black blood signal is acquired by performing the same black blood acquisition sequence during the consecutive pairs of events.

3. A cardiac imaging method, comprising the acquisition method according to any one of the preceding claims, the cardiac imaging method comprising: - Signals acquired during continuous cardiac cycles are processed by computer, including: - Generate (CTO) tissue relaxation time maps (CA) of the region to be imaged from bright blood signals acquired during the consecutive cardiac cycles of the multiple pairs of said regions.

4. The cardiac imaging method according to any one of the preceding claims, wherein, Processing the signals acquired during the multiple pairs of events by a computer includes: - Generate a black blood image (IM1i) from the black blood signals acquired during the period of the consecutive multiple pairs. - Generate a bright blood image (IM2i) from the bright blood signals acquired during the consecutive multiple pairs.

5. The cardiac imaging method according to the preceding claim, wherein, Processing the signals acquired during the multiple pairs of events by a computer includes: - Determine (DET) at least one data point from at least one of the said dark blood images and at least one of the said bright blood images that is associated with a cardiac lesion-representing area in the reference dark blood image.

6. The cardiac imaging method according to the preceding claim, wherein, The reference black blood image is a combination of the black blood images.

7. The cardiac imaging method according to any one of claims 5 to 6, wherein, At least one data point is determined from the bright blood signal collected during only one of the multiple pairs.

8. The cardiac imaging method according to any one of claims 5 to 6, wherein, At least one data point is determined from the bright blood signals collected during the multiple pairs.

9. The cardiac imaging method according to any one of claims 5 to 9, wherein, Processing the signals acquired during the multiple pairs of pairs by computer includes: - The reference black blood image is segmented by computer to determine the representative area (ZI) of the cardiac lesion.

10. The cardiac imaging method according to claim 9, wherein, Determining at least one piece of data related to the lesion area (ZI) includes: - A fused image is generated by computer by replacing the pixel or voxel values ​​of a first reference bright blood image with new values ​​that enable the differentiation of the cardiac lesion representation area (ZI) from other tissues in the area to be imaged, the first reference bright blood image being generated based on bright blood signals acquired during at least one of the plurality of pairs.

11. The cardiac imaging method according to any one of claims 9 to 10, wherein, Processing the signals acquired during the multiple pairs of events by a computer includes: - At least one second reference bright blood image is generated from bright blood signals acquired during at least one of the plurality of pairs via computer segmentation (SEG1) to generate localization data defining a set of at least one wall of the myocardium. - The reference black blood image is segmented (SEG2) based on the location data of at least one wall (L1, L2) in the set of at least one wall that defines the myocardium to determine the representative area (ZI) of the cardiac lesion, wherein the location data of the at least one wall (L1, L2) is extracted from the location data of the set of at least one wall that defines the myocardium.

12. The method according to claim 11, wherein, The second reference bright blood image is segmented using a learning function trained from multiple sets of training images of the region to be imaged, the multiple sets of training images being generated from training signals acquired during corresponding bright blood-delayed gadolinium-enhanced magnetic resonance acquisition steps, each set of training images being obtained from bright blood signals acquired by implementing a relaxation measurement sequence.

13. A method for processing data via a computer, the data being generated from signals acquired via magnetic resonance imaging for cardiac imaging, the method comprising: - Receive the black blood image (IM1i) and bright blood image (IM2i) generated by the method according to claim 4. - Determine (DET) at least one piece of data relating to a representative area of ​​cardiac lesions in a reference dark blood image, said reference dark blood image being generated based on at least one of said dark blood images and bright blood signals acquired during said plurality of pairs.

14. The method according to the preceding claim, comprising: - At least one second reference bright blood image is generated from bright blood signals acquired during at least one of the plurality of pairs via computer segmentation (SEG1) to generate localization data defining a set of at least one wall of the myocardium. - Optionally, the reference black blood image is segmented (SEG2) based on the location data of at least one wall (L1, L2) in the set of at least one wall defining the myocardium to determine the cardiac lesion representation area (ZI), wherein the location data of the at least one wall (L1, L2) is extracted from the location data of the set of at least one wall defining the myocardium.

15. The method according to any one of the preceding claims, comprising: - Generate a representation of the data associated with the region, and optionally display the visual representation.

16. The method according to any one of claims 13 to 15, comprising: - Receive (CTO) the tissue relaxation time map (CA) of the region to be imaged generated by the method according to claim 3. - Generate a visual representation of the map (CA) and, if possible, display the visual representation.

17. The method according to the preceding claims and claim 16, wherein, The visual representation is of the bullseye type and is generated from location data of a set of at least one wall that defines the myocardium.

18. The method according to any one of claims 16 to 17 and claim 14, comprising the following steps: The walls detected during the step of generating the localization of at least one set of walls used to define the myocardium are propagated onto the reference black blood image ISN and the atlas CA.

19. An imaging system configured to implement the method according to any one of claims 1 to 12, the system comprising a processing and control unit (TC) and a magnetic resonance system (A), the processing and control unit (TC) being configured to control the magnetic resonance system (A) such that the magnetic resonance system (A) implements the acquisition sequence.

20. The imaging system according to the preceding claim, wherein, The processing and control unit (TC) is configured to perform the steps of processing signals acquired during the continuous cardiac cycles.

21. A computer program product comprising instructions that cause a system according to any one of claims 19 to 20 to perform the steps of the method according to any one of claims 1 to 12.

22. A computer-readable medium having a computer program according to claim 21 recorded thereon.

23. A processing system configured to implement the method according to any one of claims 13 to 18, the system comprising a processing and control unit (TC).

24. A computer program product comprising instructions that cause a system according to any one of claims 23 to perform the steps of the method according to any one of claims 13 to 18.

25. A computer-readable medium having a computer program as described in claim 24 recorded thereon.

Images