Method and system for acquiring and reconstructing cardiac images in free breath

Abstract

The present invention relates to a method of reconstructing an image of a patient's heart using a magnetic resonance imaging apparatus (MRI 1) generating a magnetic field, the method comprising: acquiring electrical activity of the patient's heart; generating an MRI sequence and collecting a 2D dark blood image; generating an MRI sequence and collecting a 2D bright blood image; applying a non-rigid registration algorithm to each subset of the dark blood image of each profile of the heart; applying a non-rigid registration algorithm to each subset of the bright blood image of each profile of the heart; a set of 2D images is generated for each profile, each image corresponding to a combination generated between each pair of dark and bright blood images.

Description

Technical Field

[0001] The present invention relates to the field of magnetic resonance imaging, in particular to cardiac magnetic resonance imaging. More specifically, it relates to late gadolinium enhancement (LGE) magnetic resonance imaging (MRI). It is applicable to cardiac imaging and also to angiography. The field of the present invention more specifically relates to imaging methods and devices for detecting and locating scars in myocardial tissue. Background Art

[0002] The reference technique for assessing local scarring and myocardial fibrosis is bright blood gadolinium delayed enhancement or BR-LGE (Bright-Blood LGE) imaging. In this type of imaging, an inversion-recovery pulse is used to eliminate the viable myocardial signal, which allows visualization of the scar with high contrast between healthy myocardial tissue and the scar. However, for small scars or myocardial scars near the cardiac blood cavity, the high-intensity signal from the blood hinders clear visualization and precise demarcation of the scar, especially of the subendocardial scar. To solve this problem, BL-LGE (Black-Blood LGE) imaging techniques have been proposed. They make it possible to cancel the signals from healthy myocardium and blood simultaneously, thereby providing high contrast between scar and blood and between scar and healthy myocardium.

[0003] To obtain "black blood" contrast, elimination of the blood signal is obtained by applying a radio frequency (RF) inversion-recovery sequence with 180° pulses to the patient's area to be imaged, followed by a preparation module and a readout module. The time between this preparation module and the readout sequence, which occurs immediately after the emission of the magnetization reversal RF pulse, is called the inversion time TI.

[0004] The purpose is to collect signals to generate an image of the area to be imaged, thereby locating and detecting scars present in the myocardial tissue.

[0005] A disadvantage of prior art solutions known as black blood-bright blood acquisition is that, since the 3D acquisition involves many calculations, they require a significant acquisition time. The patient can breathe freely in this case. One consequence is the need to apply image correction algorithms that require a lot of computing resources, including prolonging the calculation time of the acquisition sequence and the difficulty of applying them in real time. However, when the acquisition time is prolonged, the variability of the consequences of breathing leads to the generation of image processing artifacts, affecting the clarity of the latter. These artifacts increase the difficulty of reconstructing clear and precise images to locate and detect scars. In addition, long MRI acquisitions are uncomfortable for patients. A duration of 10 to 20 minutes is considered to be a very long duration and it is difficult for the patient to remain inside the MRI without moving.

[0006] Today, these sequences require acquisition times of more than 15 minutes. In many cases, the acquired images are unsatisfactory because artifacts in the images extracted from the cross-sections of the 3D images make it impossible to distinguish the presence of a potential scar from the presence of blood located near the muscle. In fact, in some cases, the scar is so close to the blood, known as subendocardial scar, that it is difficult to know whether it is scar, blood or an image artifact. One problem is making clinical decisions based on the presence or absence of a scar, leading to an incorrect diagnosis.

[0007] Another solution consists in performing a fast acquisition during apnea, but this solution leads to the need to reduce the acquisition time from a few seconds to a few minutes and often requires reducing the area to be imaged, for example by limiting it to the 3D part.

[0008] However, it is still necessary to obtain a minimum number of images to reconstruct the 3D part. One problem is that the heartbeat also affects the measurements and involves synchronizing the acquisitions with the heartbeat. This limitation forces a reduction in the number of potential acquisitions in a given period of time. In fact, it is necessary to schedule the acquisition sequence according to the heart rhythm and therefore it is necessary to spread these acquisitions over a minimum number of heartbeats. Acquisitions during apnea significantly limit the acquisition volume.

[0009] Finally, there are 2D acquisition methods, but these methods do not achieve sufficient resolution and contrast to detect and localize scars. It is necessary to perform multiple 2D acquisitions of the same area to unambiguously identify certain cases, but the temporal changes in breathing and heartbeat do not allow 2D images to be taken over time on the same section of an organ. Summary of the invention

[0010] According to a first aspect, the present invention relates to a method for reconstructing a patient's heart image, the method being performed in a magnetic resonance imaging device that generates a magnetic field, the method comprising:

[0011] The first stage of image generation includes:

[0012] Collect the electrical activity of the patient's heart;

[0013] generating a first 180° inversion radio frequency signal to switch the longitudinal magnetization of the tissue of the imaged region, the first inversion radio frequency signal being generated between two QRS complexes of the acquired electrical activity, i.e., the so-called first inter-heartbeat phase;

[0014] After generating the first radio frequency inversion signal, generating a first magnetization preparation including a group of pulses in the same first inter-heartbeat phase;

[0015] Acquiring a first 2D image by magnetization measurement of the imaged region, the acquisition being performed after a first predetermined time duration after generating the first magnetization preparation, the acquisition being synchronized with the heart rhythm at a first time mark of the same inter-heartbeat phase;

[0016] generating a second magnetization preparation after the first inter-heartbeat phase, the second magnetization preparation comprising a set of pulses in a second inter-heartbeat phase;

[0017] acquiring a second 2D image by magnetization measurement of the imaged region, the acquisition being performed after a second predetermined time duration after generating the second magnetization preparation, at a first time mark of the second inter-heartbeat phase that is the same as the first mark of the first inter-heartbeat phase, the acquisition being synchronized with acquisition of the heart rhythm;

[0018] Repeating the first stage to generate a plurality of first images and a plurality of second images for a plurality of cross sections of the heart in an interbeat phase subsequent to the first interbeat phase and the second interbeat phase, each pair of the first image and the second image being associated with one cross section;

[0019] A set of third 2D images is generated for each section, each third image corresponding to a first combination performed between each pair of the first merged image and the second merged image for each section.

[0020] This aspect of the invention relates to the case where an image of each section is acquired. This possibility is provided for a rapid acquisition which requires thirty or so heartbeats to cover fifteen or so sections, each with one black blood image and one bright blood image. However, the sequence can be improved by acquisition in free breathing, by acquiring multiple images of each section of black blood on the one hand and of bright blood on the other hand.

[0021] According to another aspect, the present invention relates to a method for reconstructing an image of a patient's heart, the method being performed in a magnetic resonance imaging device that generates a magnetic field, the method comprising:

[0022] The first stage of image generation includes:

[0023] Collect the electrical activity of the patient's heart;

[0024] generating a first 180° inversion radio frequency module in order to switch the longitudinal magnetization of the imaged region, the first inversion radio frequency signal being generated between two QRS complexes of the acquired electrical activity, i.e., the so-called first inter-heartbeat phase;

[0025] After generating the first radio frequency inversion module, generating a first magnetization preparation including a group of pulses in the same first inter-heartbeat phase;

[0026] Acquiring a first 2D image by magnetization measurement of the imaged region, the acquisition being performed after a first predetermined time duration after generating the first magnetization preparation, the acquisition being synchronized with the heart rhythm at a first time mark of the same inter-heartbeat phase;

[0027] generating a second magnetization preparation after the first inter-heartbeat phase, the second magnetization preparation comprising a set of pulses in a second inter-heartbeat phase;

[0028] acquiring a second 2D image by magnetization measurement of the imaged region, the acquisition being performed after a second predetermined time duration after generating the second magnetization preparation, the acquisition being synchronized with acquisition of the heart rhythm at a first time mark of the second inter-heartbeat phase that is the same as the first mark of the first inter-heartbeat phase;

[0029] Repeating the first stage to generate a plurality of first images and a plurality of second images for a plurality of cross sections of the heart in an interbeat phase subsequent to the first interbeat phase and the second interbeat phase, each pair of the first image and the second image being associated with one cross section;

[0030] The first phase is repeated to generate a plurality of first images and a plurality of second images of the same cross-section of the heart in an interbeat phase subsequent to the first interbeat phase and the second interbeat phase.

[0031] In other words, the method includes repetition of the first phase to generate a subset of the plurality of first images and a subset of the plurality of second images of each of the plurality of cross-sections of the heart in an interbeat phase subsequent to the first interbeat phase and the second interbeat phase.

[0032] The method further includes:

[0033] applying a non-rigid registration algorithm to each subset of the plurality of first images of the same section to register the first images together;

[0034] applying a non-rigid registration algorithm to each subset of the plurality of second images of the same section to register the second images together;

[0035] On the one hand, merging the plurality of first images registered together to generate a first merged image, and on the other hand, merging the plurality of second images registered together to generate a second merged image;

[0036] A set of third 2D images is generated for each section, each third image corresponding to a first combination performed between each pair of the first merged image and the second merged image for each section.

[0037] One advantage is that it allows acquisitions to be made during free breathing of short duration without the computational time required for a 3D acquisition.

[0038] Advantageously, the method comprises the following steps:

[0039] For each section, on the one hand, merging a plurality of first images registered together to generate a first merged image, and on the other hand, merging a plurality of second images registered together to generate a second merged image;

[0040] A set of third 2D images is generated for the plurality of sections, each third image corresponding to a first combination, the first combination being performed between the first merged image and the second merged image for one section among the plurality of sections.

[0041] Advantages:

[0042] For each section, applying a non-rigid registration algorithm to the plurality of first images of the section to register together the plurality of first images of a subset of the plurality of first images;

[0043] For each section, applying a non-rigid registration algorithm to the plurality of first images of the section to register together the plurality of first images of a subset of the plurality of first images;

[0044] According to one embodiment, the first duration is determined so that the acquired first image is an image with a first contrast that enables display of a dark blood image, and the second duration is determined so that the acquired second image is an image with a second contrast that enables display of a bright blood image.

[0045] According to one embodiment, the registration step is followed by the following steps:

[0046] averaging the first registered acquired images to generate a first merged image for each section;

[0047] The second registered acquired images are processed to generate a first merged image for each section.

[0048] One advantage is an increased signal-to-noise ratio, also known as SNR.

[0049] According to one embodiment, the first duration is a reversal time in reversal-recovery.

[0050] Advantageously, the first duration is defined such that during acquisition of the first 2D image, the longitudinal magnetization of the blood and the longitudinal magnetization of the healthy myocardium are zero at the same time.

[0051] According to an embodiment, the determination of the first duration is automatically calculated by a computer-implemented method of calculating the optimal inversion time, the computer-implemented method being performed by processing at least one image acquired from a pre-MRI sequence.

[0052] This calculation is performed, for example, by a computer-implemented method described in patent application FR 2203766 for determining the optimal inversion time. The MRI pre-sequence comprises the acquisition of a plurality of MRI images by an acquisition sequence with different respective inversion times. The inversion time is calculated from these images.

[0053] According to an embodiment, the method comprises a step of coloring the pixels of each first merged image whose brightness is greater than a given threshold, the coloring step being performed before the first combination. In an exemplary case, the colored image corresponds to a black blood image.

[0054] One advantage is that it allows for safe (ie, unambiguous) localization and identification of scars in the vicinity of the myocardium and blood volume(s).

[0055] According to one embodiment, the first combination is a superposition of the first image on the second merged image, that is, the colored black blood image is superimposed on the colored bright blood image.

[0056] In a particular embodiment, the first image is a first merged image.

[0057] According to an embodiment, each new generation of a plurality of first images and a plurality of second images of the same section is performed after each repetition of the first stage to generate a plurality of first images and a plurality of second images of sections of the heart.

[0058] According to an embodiment, each generation of a new first image and a new second image in a new section of the heart is performed after the set of repetitions of the first stage to generate a plurality of first images and second images of previous sections of the heart.

[0059] One advantage is to acquire all images of the same section within a short time interval to avoid image offset drift in the same section when they are intertwined after different images of other sections have been acquired.

[0060] According to one embodiment, the method comprises:

[0061] displaying a set of first merged images for each section;

[0062] displaying a set of second merged images for each section;

[0063] displaying a set of third 2D images for each cross section;

[0064] Each first image of a given section is displayed in close proximity to the second image and the third image of the same given section.

[0065] Advantageously, the first image and the second image displayed for the cross section are respectively a first merged image and a second merged image.

[0066] According to one embodiment, the first merged image of each section corresponds to a black blood image of the section of the heart, the second merged image of each section corresponds to a bright blood image of the section of the heart, and the third image of each section corresponds to an image of the organ section, on which scars, if any, are displayed in color.

[0067] According to an embodiment, the first magnetization preparation and the second magnetization preparation are T1rho-weighted T1, which are performed during a delayed enhancement after the injection of gadolinium.

[0068] According to an embodiment, the first magnetization preparation is a preparation of the T1rho type defined by a sequence of magnet configurations defined as follows: 90x-SLy-180y-SL-y-SLy-180-y-SL-y-90-x.

[0069] According to another exemplary preparation of the T1rho type, another magnetization configuration may be, for example, 90x-Sly-90-x.

[0070] According to an embodiment, the first magnetization preparation is a T2prep type preparation or an MTC type preparation.

[0071] According to one embodiment, the step following the generation of the second magnetization preparation comprising a set of pulses of inter-heartbeat phases is filtering the acquired images corresponding to fat regions of the imaged organ.

[0072] According to an embodiment, the first image and the second image are electrical activities of the heart acquired continuously in a synchronized manner between two QRS complexes during a single examination, over several minutes of free breathing.

[0073] According to one embodiment, multiple first images and multiple second images are acquired on a set of 8 to 20 sections of the organ, and the stage of repeatedly acquiring multiple first images and second images on the same section enables 2 to 10 first images and 2 to 10 second images to be acquired for each section.

[0074] According to another aspect, the invention relates to a magnetic resonance imaging system comprising a magnetic field generator and radiofrequency equipment, an electrocardiograph and processing equipment, a display for displaying the generated images, the system being configured to implement the method of the invention.

[0075] One advantage is the ability to generate images in which radiologists can locate and detect scars in myocardial tissue in a reliable manner.

[0076] One advantage is that regions of interest, such as scars, can be displayed in color.

[0077] Another advantage is that images can be acquired quickly during free breathing, in particular without the need to image the patient while they are apneic.

[0078] Furthermore, the method according to the invention is computationally simple, fast and cheap. It can be easily integrated into existing MRI systems without modification.

[0079] One advantage is that for each given portion of the heart, multiple images are collected, which makes it possible to reduce noise, or increase the signal-to-noise ratio and reduce residual artifacts.

[0080] One advantage of using a non-rigid algorithm is the mitigation of motion artifacts associated with respiratory changes. The implementation of such an algorithm can improve the signal-to-noise ratio.

[0081] Finally, 2D acquisition is faster and more robust than 3D acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Other features and advantages of the present invention will become more apparent by reading the following detailed description with reference to the accompanying drawings.

[0083] Figure 1 : An exemplary embodiment of the system of the present invention comprises an MRI device, an electrocardiograph, a computer for performing processing operations of acquired images, and a human-computer interface for viewing the images;

[0084] Figure 2 : Schematic diagram of a portion of a sequence generated by an MRI device over two heart beats;

[0085] Figure 3 : Schematic diagram of the black blood and bright blood MRI acquisition phases performed over multiple heart beats,

[0086] Figure 4 : Schematic diagram of a 3D heart showing different sections and images acquired in each section;

[0087] Figure 5 : Flow chart of method steps according to one embodiment of the present invention. DETAILED DESCRIPTION

[0088] MRI System

[0089] Figure 4 Schematically presented is an imaging system S1 comprising a magnetic resonance apparatus MRI1 according to the invention comprising a set of excitation and measurement devices, a computer K1 performing processing of acquired images and a display IHM1 showing the generated images.

[0090] In a manner known per se, the set of excitation and measurement devices comprises an MRI imaging device comprising a static magnetic field generator GEN_B0 comprising a polarized main magnet, a gradient generator GEN_GRAD and a radio frequency device DISPO_RF comprising a radio frequency signal generator and a radio frequency antenna.

[0091] The static magnetic field generator GEN_B0 includes a magnet for generating a substantially uniform static polarized magnetic field in a polarized region (usually a tunnel) including a region of the patient to be imaged, such as the heart.

[0092] The gradient generator GEN_GRAD comprises three gradient coils or solenoids arranged and configured to vary the magnetic field strength in the polarization zone along respective orthogonal axes fixed relative to the polarization zone, the orthogonal axes usually being labeled x, y and z. The choice of the strengths circulating in these coils makes it possible to select a thickness and a profile from among several possible thicknesses and profiles in which the magnetization of the region to be imaged of the patient received in the polarization zone is measured.

[0093] The radio frequency device GEN_RF comprises a coil or a solenoid and is capable of generating an MRI acquisition sequence, the MRI acquisition sequence comprising a magnetization preparation sequence of the region to be imaged and a signal reading sequence from the region to be imaged. Each acquisition sequence comprises at least one radio frequency pulse consisting of a predetermined and controllable frequency, shape, duration, phase, amplitude.

[0094] The preparation module, also called preparation or preparation phase, is configured to excite, ie change the magnetization direction of the tissues of the area to be imaged, ie modulate their magnetization. TAprep or TBprep or T1rho is mentioned in the rest of the description.

[0095] The reading module is configured to measure the magnetization of the area to be imaged resulting from the preparation sequence. The reading module is also called an acquisition module or a reading phase or a reading window.

[0096] In other words, the image is acquired by measuring the magnetization of the area to be imaged.

[0097] The assembly advantageously comprises an electrocardiograph ELC1 for acquiring an electrocardiogram ECG1 of the patient.

[0098] The imaging system comprises a processing device K1 which is configured to perform calculations, in particular registration of images and averaging of these images. The processing device may also colorize the images.

[0099] The imaging system S1 further comprises a display for displaying a two-dimensional image of the area to be imaged in grayscale and / or color according to the signal measured by the RF magnetic field generator and processed by the processing device.

[0100] According to one embodiment, the reconstructed image is a color and grayscale matrix image. It includes a set of pixels, each pixel is characterized by an intensity I, which can take a set of M values ​​(M is a finite integer greater than 1), corresponding to M gray levels ranging from 0 to M-1. For example, the value can take 256 values ​​between 0 and 255. In other embodiments, the value can be higher. According to one embodiment, for the selection of pixels whose brightness is greater than a given threshold, the image can be partially colored.

[0101] The representation of the image generated by the processing device K1 is intended for display on a screen of a human-machine interface IHM1 of the imaging system S1 .

[0102] Acquisition sequence

[0103] Figure 2 An example of a portion of a black blood and bright blood MRI acquisition sequence SEQ1 of a myocardial image and electrical activity ECG1 represented over two consecutive heartbeats is shown. The electrocardiogram ECG1 acquired by electrocardiograph ECL1 during two heartbeats and the acquisition sequence SEQ1 applied to the region to be imaged are overlapped to better represent the time-dependent characteristics of each of these curves. ECG acquisition is a step, labeled ACQ0. Figure 2 The lower part of shows the variation of the longitudinal magnetization Mz of the tissue of the area to be imaged as a function of time t, and two images IM1 and IM2 reconstructed from the signals measured during the sequence.

[0104] Preferably, the invention allows synchronizing the MRI excitation and acquisition sequences on the electrocardiogram ECG1. This synchronization allows, on the one hand, measurements to be taken at the most stable moments of the heart, and on the other hand, measurements to be taken at the same moment between two identical positions of the heart, in order to compare images of the same section. A method for synchronizing these excitations and acquisitions and triggering the measuring device on the QRS complex.

[0105] In the rest of the description, the first image IM1 will refer to the image acquired with black blood, and the second image IM2 will refer to the image acquired with bright blood. According to one example, throughout the sequence, starting from the first heartbeat in which image acquisition is performed, odd heartbeats are used to acquire black blood images, and even heartbeats are used to acquire bright blood images. This example will be described in more detail in the description. However, other configurations for performing the method of the present invention can be envisioned, such as associating odd heartbeats with the acquisition of bright blood images, and associating even heartbeats with the acquisition of black blood images.

[0106] The present invention is also applicable to the case where a bright blood image is acquired before a black blood image. The present invention is also applicable to the case where the acquired black blood and bright blood images are not acquired immediately on two consecutive heartbeats, but are, for example, separated by one or more heartbeats without measurement. However, the benefit of acquiring these images on two consecutive heartbeats is to minimize the acquisition time, thus minimizing the effects of breathing, and thus minimizing image artifacts on the time scale of two heartbeats.

[0107] For both black and bright blood MRI acquisitions, a gadolinium-based contrast product is advantageously injected intravenously into the patient 10 to 15 minutes before the acquisition sequence is applied in order to obtain images showing maximum contrast between scar and healthy tissue and blood. In the heart, the contrast agent is rapidly eliminated from healthy myocardium, is present at low levels in interstitial tissue, but continuously accumulates in myocardial scars. The effect of gadolinium is to shorten the relaxation time T1 of tissues in which gadolinium accumulates. As a result, the magnetization of the scar relaxes faster than that of blood and healthy myocardium after the magnetization pulse.

[0108] In the first case, the goal is to generate an image showing black blood contrast. In this type of image, the intensity of the pixels corresponding to blood and healthy muscle is zero (black pixels) or approximately zero because it was acquired when the longitudinal magnetization of the blood was zero.

[0109] To generate such an image IM1, the device DISPO_RF generates an inversion-recovery acquisition sequence.

[0110] The acquisition sequence includes Figure 2 The 180° inversion pulse, labeled Rfi in FIG. , shifts the longitudinal magnetization of the tissue in the imaged region in the opposite direction, ie, reverses the longitudinal magnetization. Figure 2 It is shown that under the influence of the inversion pulse Rfi, the magnetization of the region to be imaged changes from Mz to -Mz. Due to longitudinal relaxation, the longitudinal magnetization of the different tissues present in the region to be imaged increases and thus returns to its initial value, passing through the zero value. Naturally, the return rate of the magnetization of different tissues is different.

[0111] In a manner known per se, the preparation sequence then comprises a preparation module, for example an adiabatic T1rho preparation, a TSL (standing for “spin lock time”) duration Figure 2 Labeled as T1rho or TAprep.

[0112] Other preparations can be used, such as T2prep, magnetization transfer prep MT.

[0113] Note that in this application, this preparation may sometimes be defined as T1rho preparation, or sometimes as T1rho preparation prior to a 180° Rfi inversion sequence. The second image IM2 is acquired without a 180° inversion pulse, but with only T1rho preparation, in which case the preparation is limited to T1rho preparation.

[0114] The preparation sequence (Rfi(180°), T1rho) is configured so that the longitudinal magnetization of the blood and the longitudinal magnetization of the healthy myocardium are zero at the same instant TI, see the position of the NP point of the A(blood) and A(Musc) curves. The inversion time TI corresponds to the time measurement between the emission of the Rf signal at 180° and the acquisition module ACQ1. The instant TI corresponds to the duration marked D1 in the present application. In fact, the invention is also applicable to the case where the measurement of magnetization is not necessarily calculated at the time of inversion, but may be calculated at another instant, depending on the desired result.

[0115] Regardless of whether each heartbeat is labeled as an acquisition window or a read window, for read windows associated with dark blood acquisition, they are labeled ACQ1, and for read windows associated with bright blood acquisition, they are labeled ACQ2.

[0116] Note that Figure 2 In the same instant TI, the longitudinal magnetization of scar A (Cica) is significantly greater than zero, in fact the PP point is greater than zero. By acquiring the signal from this area to be imaged at the instant TI, an image is obtained with very high contrast between the pixels corresponding to blood and healthy myocardium, where blood and healthy myocardium are black, while scars are usually bright.

[0117] Then, the inversion recovery sequence includes a read sequence ACQ1 including a 90° pulse and a read gradient to read the transverse magnetization of the region to be imaged.

[0118] The inversion time TI is the duration separating the 180° pulse and the acquisition window ACQ1. The time stamp of the acquisition window ACQ1 is applied upstream of the ECG, usually in diastole. This moment is chosen as the moment of cancellation of the muscle and blood signals, i.e. the moment of cancellation of the longitudinal magnetization of the blood and myocardium, in order to generate an image presenting the best contrast. In order to obtain the best contrast between myocardial scar and blood and healthy myocardium, it is sought to start the inversion Rf signal at an early moment to eliminate the signal within the duration TI.

[0119] This sequence is followed by a bright blood acquisition phase performed on the heartbeat after the dark blood acquisition. In this phase, the second image IM2 is acquired. Figure 2 In the present invention, the bright blood collection phase is labeled as TBprep. According to an embodiment, the preparation TBprep may be different from the preparation TAprep for black blood imaging of the cardiac region, however it is preferred that the black blood preparation TAprep and the bright blood preparation TBprep are the same. Therefore, the present invention relates to other preparations. However, in the rest of the description, the bright blood collection phase is described with a bright blood preparation that is the same as the black blood preparation, for example a T1rho preparation.

[0120] The duration D2 separating the time marker from the acquisition window AQ2 and the time marker corresponding to the bright blood preparation TBprep can be adjusted, wherein the duration D2 is set for each heartbeat to optimize the bright blood contrast to best image the muscle contour relative to the blood.

[0121] The measurement of the magnetization gradient is configured to acquire the bright blood image ACQ2 at the same inter-heartbeat markers as the black blood image. Figure 2 In , the reading phase ACQ2 is located at the same duration of the previous QRS complex as the duration between the QRS of the heartbeat preceding the reading phase ACQ1.

[0122] When the magnetization is read at the same moment, an image can be obtained in which the pixels of the blood areas are represented in white, i.e. with high brightness, and in which the pixels of the myocardial tissue are slightly darker than the pixels of the blood. It is thus possible to obtain a 2D topology of the section, allowing visualization of the blood and myocardial areas in relation to each other.

[0123] It will be appreciated that by combining images IM1 and IM2 it will be possible to locate and detect a scar, if present.

[0124] To this end, the present invention makes it possible to perform continuous measurements of the sequence of these two black blood IM1 and bright blood IM2 images in multiple sections of the heart to study the entire heart and to make the image analysis more robust by considering multiple images per section.

[0125] Figure 3An example of the ordering of these phases, denoted Phi, over a plurality of consecutive heartbeat pairs is shown, where i=1 to N. Phase PHi comprises, over two heartbeats at a time:

[0126] First heartbeat: Generates GEN1 inverted RF signal, generates the preparation phase and is labeled GEN2, and is labeled ACQ1

[0127] The reading phase;

[0128] Second heartbeat: generates the preparation phase GEN3 and the reading phase marked ACQ2.

[0129] Figure 4 A 3D view of a heart is shown, including a plurality of (K) sections, labeled PC k , where k=1 to K. Figure 4 A plurality of sections may be represented, each section comprising a plurality of images acquired according to the method of the invention. k images, Figure 4 Other images are not shown. k The first image collected is marked as IM1 k (i), where i=1 to N.

[0130] According to one embodiment, the first acquisition of the heart image can evaluate the number of sections and therefore the number of repeated acquisitions. If the heart is considered to be small, 7 to 9 sections can be defined, and if the heart is large, 14 to 20 sections can be defined. If the heart is of medium size, the number of sections can be defined between 10 and 13 sections.

[0131] This estimation may be performed at the beginning of the examination, such as during a pre-sequence.

[0132] According to another example, during the first acquisition of the method of the invention, during the first heartbeat, an estimate of the heart size is automatically determined. This size may allow automatic dynamic calculations to be performed to acquire images in a manner according to a given number of sections.

[0133] According to the example of a complete sequence of 3 minutes with a pulse of 60 beats per minute, considering a configuration of 15 sections of an organ, we will have a total of 3×60=180 images, of which 90 are black blood images and 90 are bright blood images. Such a configuration makes it possible to acquire 6 images {IM1 k (1),IM1 k (2),IM1 k (3),IM1 k (4),IM1 k (5),IM1 k(6)}, and collected 6 images {IM2 k (1),IM2 k (2),IM2 k (3),IM2 k (4),IM2 k (5),IM2 k (6)}.

[0134] One advantage of the method of the invention is that it is performed during free breathing. For this purpose, it is necessary to process the images acquired in the same section so that they are not affected by free breathing. The respiratory rhythm is a priori different from the heart rhythm. Even if there is a correlation between the two rhythms, it is possible that breathing can accelerate while the heartbeat remains stable, or vice versa.

[0135] The problem is that during the second reading of ACQ2, breathing can significantly change the position of the heart, with the result that the image acquired at another heartbeat in the same section may be offset relative to the image previously acquired in the same section.

[0136] According to one embodiment, the images of each section acquired from the black blood are registered by a non-rigid image registration algorithm. The bright blood images acquired for each section are registered using the same method as the selected black blood method. Different algorithms can be selected to process the black blood images and the bright blood images, but for ease of implementation, it is preferred to select the same algorithm. The registration significantly improves the image merged from multiple images acquired in the same section because the contrast is better and the non-rigid algorithm has been able to compensate for breathing artifacts.

[0137] According to a first example, the non-rigid algorithm is an algorithm based on the mutual information method between images of statistical relationships. The function to be optimized can be implemented by a statistical similarity criterion. One benefit of this method is that the pairing between homologous attributes of images of the same section is independent of their geometric position.

[0138] According to a second example, a non-rigid algorithm based on a transformation model is implemented. The transformation model can determine a function that minimizes the difference between the two images. The deviation can be converted into a geometric error to be minimized. Different methods can be used, such as methods based on extracting geometric primitives or shape descriptors (such as salient points, shape singularities or contours) from each image. Parametric or non-parametric methods can be used.

[0139] According to one example of optimization of the transformation model or the similarity criterion, the least squares method may be used.

[0140] Other optimization methods may be implemented, such as gradient descent. However, the latter method applies more specifically to image intensity and is not optimal in the context of the present invention, as it attempts to optimize the sharpness and contrast of the merged image. Nevertheless, the present invention includes this embodiment.

[0141] The registration can be performed by selecting a reference image and determining the transformation function of the other images of the same section with respect to this reference image. Each image is then registered by optimizing the transformation to obtain the reference image from the image under consideration according to geometric criteria.

[0142] When multiple images are registered together in a section, operations may be performed aimed at merging these images to generate a single merged image for each section plane.

[0143] According to one embodiment, a step of averaging the images of the same section is performed in order to reduce noise and increase the signal-to-noise ratio.

[0144] The advantage of averaging is that it preserves the details of the image because it increases the signal-to-noise ratio (SNR). This technique can remove noise and reduce residual image artifacts. In addition, averaging can increase the bit depth of the digital image and may exceed the bit depth of a single image.

[0145] One benefit of the step of averaging the images taken from the same section is the reduction of the maximum deviation. The amplitude of the noise decreases with the square root of the number of images used, i.e. using only 4 images, the amplitude of the noise can be reduced by half. Based on the example acquired during free breathing lasting 2 minutes, 4 to 5 images can be collected per section, which makes it possible to obtain good noise reduction performance.

[0146] Each set of images acquired in dark blood was merged on the one hand, and each set of images acquired in bright blood on the other hand to generate a PC for each section. k Single black blood image IM 1F k and a single bright blood image IM 2F k .

[0147] In other words, Figure 5 As shown, for each section PC k Generate a merged black blood image IM 1F k and a merged bright blood image IM 2F k .

[0148] According to one embodiment, the method of the invention comprises a step of coloring the brightest pixels of the black blood image. Assuming that the signals from muscle and blood are zeroed and represented by pixels of low intensity, i.e. black, the brightest pixels correspond specifically to scars. The coloring of scars allows the scar contour to be identified and represented.

[0149] One benefit of colorization is that it allows scars to be clearly identified and allows this representation to be used in conjunction with a bright blood image representing muscle contours. This combination can therefore identify and localize scars, particularly with respect to muscle and blood near the scar.

[0150] Each section PC k Execute Black Blood IM 1F k and bright blood IM 2F k Combination of images COMB1. Thus, the same combination operation enables to end up with a plurality of combined black blood-bright blood images to generate each section PC k A single image of IM3 k .

[0151] The combination COMB1 of the images can be performed in different ways. According to a first example, the images are superimposed on each other. The superposition can include taking into account the variation of the opacity of the black blood image in order to properly visualize the outline of the bright blood image. The opacity of the black blood image can be configured to have a value comprised between 40% and 90%, for example 75%. According to a second example, scar pixels having a brightness greater than a given threshold are extracted to be integrated into the bright blood image.

[0152] Other image combination methods may be used to generate a single image from the black blood image and the bright blood image for each section.

[0153] In other words, for each profile PC k , from the merged black blood image IM 1F k and the merged bright blood image IM 2F k Generate a single combined image IM3 k .

[0154] One benefit of the method of the present invention is that multiple reconstructed images of each section are generated from the black blood image and the bright blood image. Depending on the size of the heart, 8 to 20 sections can be selected. The method of the present invention can display the reconstructed image of each section to the user, especially the radiologist. Each image allows to show the presence or absence of scar by accurate positioning relative to the cardiac anatomy, myocardium and blood volume surrounding the imaged area.

Claims

1. A method for reconstructing a patient's heart image, the method being performed in a magnetic resonance imaging device (MRI1) that generates a magnetic field (B0), the method comprising: The first stage of image generation (PH1) includes: collecting (ACQ0) electrical activity (ECG1) of the patient's heart; Generate (GEN1) the first 180° inverted RF signal (Rf i ) to reverse the longitudinal magnetization of the tissue in the imaged area, the first inversion radio frequency signal (Rf i ) occurs between two QRS complexes of the acquired electrical activity (ECG1), the so-called first interbeat phase (Tc1); Generate the first inverted radio frequency signal (Rf i ), then, in the same first inter-heartbeat phase (Tc1), generating (GEN2) a group of pulses ({p i }) first magnetization preparation (TA prep ); A first 2D image (IM1) is acquired (ACQ1) by measuring the magnetization of the imaged area, the acquisition (ACQ1) generates the first magnetization preparation (T prep ), said acquisition (ACQ1) being synchronized with said heart rhythm (ECG1) at a first time mark (M1) of the same said inter-beat phase (Tc1); generating (GEN3) a set of pulses ({p i }) second magnetization preparation; A second 2D image (IM2) is acquired (ACQ2) by measuring the magnetization of the imaged area, wherein the acquisition (ACQ2) in the second magnetization preparation (TB prep ), the acquisition (ACQ2) being synchronized with the acquisition of the heart rhythm (ECG1) at the first time mark (M1) of the second inter-heartbeat phase (Tc2) being the same as the first mark (M1) of the first inter-heartbeat phase (Tc1); The first phase (PH1) is repeated to generate each of the plurality of cross-sections of the heart (PC) in an inter-heartbeat phase (Tcik) after the first inter-heartbeat phase (Tc1) and the second inter-heartbeat phase (Tc2). k ) a subset of the plurality of first images and a subset of the plurality of second images; A non-rigid registration algorithm (ALG1) is applied to the plurality of sections (PC k ) of a plurality of first images (IM1 k (i)) to convert the plurality of first images (IM1) of the subset of the plurality of first images into k (i)) are registered together; A non-rigid registration algorithm (ALG1) is applied to the plurality of sections (PC k ) of a cross section (IM2 k (i)) to convert the plurality of second images (IM2) of the subset of the plurality of second images into k (i)) are registered together; For each of the plurality of sections, on the one hand, the plurality of first images registered together are merged to generate a first merged image (IM 1F k ), and on the other hand, merging the plurality of second images registered together to generate a second merged image (IM 2F k ); A set of third 2D images is generated for the plurality of cross sections, each third image (IM3 k ) corresponds to a first combination (COMB1), which is performed between the first merged image and the second merged image for one of the multiple sections.

2. The method according to claim 1, wherein the first duration (D1) is determined so that the first acquired image is an image with a first contrast, which enables the display of a black blood image, and the second duration (D2) is determined so that the second acquired image is an image with a second contrast, which enables the display of a bright blood image.

3. The method according to any one of claims 1 to 2, wherein the registration step is followed by the following steps: The first registered images are averaged to generate each section (PC k ) of the first merged image (IM 1F k ); The second registered images are averaged to generate a first merged image (IM 2F k ).

4. The method according to any one of claims 1 to 3, wherein: The first duration (D1) is defined in such a way that, at the same instant TI of acquisition (ACQ1) of the first 2D image, the longitudinal magnetization of the blood and the longitudinal magnetization of the healthy myocardium are zero.

5. The method according to any one of claims 1 to 4, wherein: The determination of the first duration (D1) is automatically calculated by a computer-implemented method of calculating an optimal inversion time, the computer-implemented method being performed by processing at least one image acquired from a pre-MRI sequence.

6. The method according to any one of claims 1 to 5, comprising for each first merged image (IM 1F k ) is subjected to a step of coloring pixels, the coloring step being performed before the first combination (COMB1).

7. A method according to any one of claims 1 to 6, wherein the first combination (COMB1) is the first image (IM 1F k ) in the second merged image (IM 2F k ) on top.

8. The method according to any one of claims 1 to 7, wherein: - After the following steps, a new first image and a new second image are each generated in a new cardiac section plane: - Generate a previous cross section of the heart (PC k ) of a plurality of first images ({IM1 k (i)} i[1-N] ) and a plurality of second images ({IM2 k (i)} i[1-N] ) of all repetitions of the first phase (PH1).

9. The method according to any one of claims 1 to 8, comprising: Display for each section (PC k ) of a set of first merged images (IM 1F k ); Display for each section (PC k ) of a set of second merged images (IM 2F k ); Display for each section (PC k ) a set of third 2D images; Given a profile (PC k ) of each first image (IM 1F k ) are displayed in the same given section (PC k ) of the second image and the third image (IM 3F k ) in the immediate vicinity of.

10. The method according to any one of claims 1 to 9, wherein: Each section (PC k ) of the first merged image (IM 1F k ) corresponds to a black blood image of a cross section of the heart, each cross section (PC k ) of the second merged image (IM 2F k ) corresponds to the bright blood image of the cross section of the heart, each cross section (PC k ) of the third image ((IM 3F k ) corresponds to an image of a cross-section of the organ, on which scars, if any, are displayed in color.

11. The method for reconstructing an image according to any one of claims 1 to 10, wherein: The first magnetization preparation and the second magnetization preparation (TAprep, TBprep) are T1rho-weighted.

12. The method for reconstructing an image according to any one of claims 1 to 11, wherein: The first magnetization preparation and / or the second magnetization preparation (TAprep, TBprep) is a T2prep type preparation or a magnetization transfer preparation MT.

13. The method for reconstructing an image according to any one of claims 1 to 12, wherein: The step following said generating (GEN3) a second magnetization preparation (TBprep) is filtering the acquired images corresponding to the fat regions of said imaged organ, said second magnetization preparation (TBprep) comprising an inter-cardiac phase (Tc N ) in a group of pulses ({p i }).

14. The method for reconstructing an image according to any one of claims 1 to 13, wherein: The plurality of first images (IM1) and the plurality of second images (IM2) are acquired continuously in synchronization with the electrical activity of the heart between two QRS complexes during a single examination, over a plurality of minutes of free breathing.

15. The method for reconstructing an image according to any one of claims 1 to 14, wherein the plurality of first images and the plurality of second images are acquired on a group of 8 to 20 sections of the organ, and the plurality of first images and the plurality of second images are acquired on the same section (PC k ) repeatedly collects multiple first images (IM1 k (i)} i[1-N] ) and the second image (IM2 k (i)} i[1-N] ) stage enables acquisition of a plurality of first images (IM1 k (i)} i[1-N] ) and a plurality of second images (IM2 k (i)} i[1-N] ) of 2 to 10 images.

16. A magnetic resonance imaging system (S1), comprising a magnetic field generator (GEN_B0, GEN_GRAD) and a radio frequency device (DISPO_RF), an electrocardiograph (ELC1) and a processing device (K1), and a display (IHM1) for displaying the generated image, the system being configured to implement the method according to any one of claims 1 to 15.

Images