US20210161422A1

US20210161422A1 - Automatic imaging plane planning and following for mri using artificial intelligence

Info

Publication number: US20210161422A1
Application number: US17/060,860
Authority: US
Inventors: Xiao Chen; Shanhui Sun; Zhang Chen; Terrence Chen
Original assignee: Shanghai United Imaging Intelligence Co Ltd
Current assignee: Shanghai United Imaging Intelligence Co Ltd; Uii America Inc
Priority date: 2019-11-29
Filing date: 2020-10-01
Publication date: 2021-06-03
Also published as: US11992289B2; US20210165064A1

Abstract

A method includes acquiring initial scout images of a patient's heart, using a neural network to establish a patient specific heart model, and automatically plan imaging planes of the patient specific heart model, performing an accelerated scan of the patient's heart, using the neural network to determine a current location and pose of the patient's heart from the accelerated scan, and to reposition the imaging planes to correspond to the current location and pose of the patient's heart, and using the repositioned imaging planes to perform an acquisition scan and generate an image of the patient's heart from the acquisition scan according to a selected imaging protocol.

Description

CROSS REFERENCE TO RELATED APPLICATION

This application claims the benefit of U.S. Provisional Application No. 62/941,904, filed 29 Nov. 2019, which is incorporated by reference herein in its entirety.

BACKGROUND

The aspects of the present disclosure relate generally to Magnetic Resonance Imaging (MRI), and in particular to predicting cardiac signals from MRI data. MRI is a widely used medical technique which produces images of a region of interest using magnetic and radio frequency energy. During an MRI scan, volume coils (for example, body coils) and local coils (for example, surface coils) may acquire MR signals produced by nuclear relaxation inside the object being examined. Cardiac MR imaging is widely regarded as one of the most complex examinations utilizing magnetic resonance due to the patient's respiratory and cardiac motions. In a conventional scanning workflow, obtaining target heart views is typically performed through a multi-step approach 100 as illustrated in FIG. 1. As shown in block 102, scout images and multi-slice localizer images, typically in the form of a set of three plane, low resolution, large field of view images, may be initially acquired to determine an approximate location of the heart in the patient's body, including the standard heart views: the short axis view corresponding to the echocardiographic parasternal short axis plane; the horizontal long axis view corresponding to the echocardiographic apical 4-chamber plane; and the vertical long axis view corresponding to the echocardiographic apical 2-chamber plane. The localizer images may be used to manually plan slices through the standard heart views, as shown in block 104.
The image plane planning of the subsequently performed cardiac MR acquisition scans, for example, cine or functional scans, relies on the views determined above, and the planning is generally accomplished by performing an acquisition scan, as shown in block 106, and referring or copying the slice locations from the pre-determined standard heart views to the acquisition scan, as shown in block 108. However, patient motion and inconsistency of breath-holding positions in between scans may introduce mis-registrations of the slices from scan to scan, and may introduce difficulties in interpreting the images. In order to overcome the mis-registration, an additional acquisition scan may be performed, as shown in block 110, and a technician may reposition the slices manually, as shown in block 112, and an acquisition scan for the selected imaging protocol may be performed as shown in block 114. Severe mis-registration may even require patient repositioning, additional repeated scans, or additional post-processing to register the images, any one of which may add additional cost in the form of additional labor, time, computation, etc. to the MR scanning processes.
Navigation techniques have been used to monitor respiratory motion of objects being imaged, however, most navigation techniques aim at compensating for the respiratory motion by using a brief MR scan limited to the patient's diaphragm between k-space lines. Thus, the technique may compensate for motion within one data acquisition but cannot address patient motion between multiple data acquisitions. Furthermore, the navigation signal is usually a beam perpendicular to the diaphragm which has a one dimensional limited view and description of the motion of the diaphragm, which may lead to an erroneous respiratory motion estimation.
As a result, image quality and usability for diagnosis depends on additional scans between acquisition scans for the selected imaging protocol and on an operators' skills and experience in relocating the slices between scans. This represents one of the major barriers to cardiac MRI being widely applied in clinical procedures.

SUMMARY

It would be advantageous to provide a method and system that may automatically acquire and adjust planned image planes to compensate for changes in a pose of a heart throughout an entire scanning process without manual intervention.
According to the present disclosure, the method and system may exploit artificial intelligence, for example, a neural network, to: automatically estimate a pose of the heart; automatically provide imaging slice planning; reconstruct highly accelerated inter-acquisition scout imaging; and monitor and follow patient motion to maintain consistency of planned slice locations for each acquisition. This may advantageously allow for a more automatic and efficient scanning workflow for cardiac MRI and facilitate implementation in most clinical settings for cardiac diagnosis.
The disclosed embodiments are directed to a method including acquiring initial scout images of a patient's heart, using a neural network to establish a patient specific heart model, and automatically plan imaging planes of the patient specific heart model, performing an accelerated scan of the patient's heart, using the neural network to determine a current location and pose of the patient's heart from the accelerated scan, and to reposition the imaging planes to correspond to the current location and pose of the patient's heart, and using the repositioned imaging planes to perform an acquisition scan and generate an image of the patient's heart from the acquisition scan according to a selected imaging protocol.
The method may include acquiring the initial scout images from standard MRI body views.
The initial scout images may include 2D or 3D multi-slice images from one or more of axial sagittal and coronal views.
Using the neural network to determine a current location and pose of the patient's heart from the accelerated scan may include reconstructing an image from the accelerated scan and comparing the reconstructed image from the accelerated scan to the patient specific heart model.
The method may include comparing the current location and pose of the patient's heart to the location and pose of the patient specific heart model, and repositioning the imaging planes obtained from the patient specific heart model to correspond to the current location and pose of the patient's heart.
The neural network may include one or more of a combination of CNN and RNN models, a GRU model, an LSTM model, a fully convolutional neural network model, a generative adversarial network, a back propagation neural network model, a radial basis function neural network model, a deep belief nets neural network model, an Elman neural network model.
The accelerated scan may include one or more of compressed sensing; parallel imaging; or fast spin echo techniques to allow for acquisition in less time of a reduced amount of data than required to support a higher resolution or larger field of view.
The selected imaging protocol may include one or more of obtaining anatomic images of the heart, determining cardiac function, or determining myocardial viability.
The disclosed embodiments are further directed to a system including an MRI scanner, and a processing engine coupled to the MRI scanner, the processing engine comprising a processor and a memory comprising computer readable program code, wherein the processor under control of the computer readable program code is operable to acquire initial scout images of a patient's heart, use a neural network to establish a patient specific heart model, and automatically plan imaging planes of the patient specific heart model, perform an accelerated scan of the patient's heart, use the neural network to determine a current location and pose of the patient's heart from the accelerated scan, and to reposition the imaging planes to correspond to the current location and pose of the patient's heart, and cause the MRI scanner to use the repositioned imaging planes to perform an acquisition scan and generate an image of the patient's heart from the acquisition scan according to a selected imaging protocol.
These and other aspects, implementation forms, and advantages of the exemplary embodiments will become apparent from the embodiments described herein considered in conjunction with the accompanying drawings. It is to be understood, however, that the description and drawings are designed solely for purposes of illustration and not as a definition of the limits of the disclosed invention, for which reference should be made to the appended claims. Additional aspects and advantages of the invention will be set forth in the description that follows, and in part will be obvious from the description, or may be learned by practice of the invention. Moreover, the aspects and advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following detailed portion of the present disclosure, the invention will be explained in more detail with reference to the example embodiments shown in the drawings. These embodiments are non-limiting exemplary embodiments, in which like reference numerals represent similar structures throughout the several views of the drawings, wherein:

FIG. 1 illustrates a conventional scanning workflow, where obtaining target heart views is performed through a multi-step approach including multiple scans and manual repositioning of planned slices.

FIG. 2 illustrates an exemplary MRI apparatus according to aspects of the disclosed embodiments;

FIG. 3 illustrates an exemplary architecture of a processing engine according to the disclosed embodiments;

FIG. 4 illustrates an exemplary process flow according to aspects of the disclosed embodiments;

FIGS. 5A-5C schematically illustrate usage of a patient specific heart model according to aspects of the disclosed embodiments; and

FIG. 6 depicts an exemplary simple neural network that may be utilized to implement the disclosed embodiments.

DETAILED DESCRIPTION

In the following detailed description, numerous specific details are set forth by way of examples in order to provide a thorough understanding of the relevant disclosure. However, it should be apparent to those skilled in the art that the present disclosure may be practiced without such details. In other instances, well known methods, procedures, systems, components, and/or circuitry have been described at a relatively high-level, without detail, in order to avoid unnecessarily obscuring aspects of the present disclosure. Various modifications to the disclosed embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirits and scope of the present disclosure. Thus, the present disclosure is not limited to the embodiments shown, but to be accorded the widest scope consistent with the claims.
It will be understood that the term “system,” “unit,” “module,” and/or “block” used herein are one method to distinguish different components, elements, parts, section or assembly of different level in ascending order. However, the terms may be displaced by other expressions if they may achieve the same purpose.
It will be understood that when a unit, module or block is referred to as being “on,” “connected to” or “coupled to” another unit, module, or block, it may be directly on, connected or coupled to the other unit, module, or block, or intervening unit, module, or block may be present, unless the context clearly indicates otherwise. As used herein, the term “and/or” includes any and all combinations of one or more of the associated listed items.
Generally, the word “module,” “unit,” or “block,” as used herein, refers to logic embodied in hardware or firmware, or to a collection of software instructions. A module, a unit, or a block described herein may be implemented as software and/or hardware and may be stored in any type of non-transitory computer-readable medium or another storage device. In some embodiments, a software module/unit/block may be compiled and linked into an executable program. It will be appreciated that software modules can be callable from other modules/units/blocks or from themselves, and/or may be invoked in response to detected events or interrupts. Software modules/units/blocks configured for execution on computing devices may be provided on a computer-readable medium, such as a compact disc, a digital video disc, a flash drive, a magnetic disc, or any other tangible medium, or as a digital download (and can be originally stored in a compressed or installable format that needs installation, decompression, or decryption prior to execution). Such software code may be stored, partially or fully, on a storage device of the executing computing device, for execution by the computing device. Software instructions may be embedded in firmware, such as an Erasable Programmable Read Only Memory (EPROM). It will be further appreciated that hardware modules/units/blocks may be included in connected logic components, such as gates and flip-flops, and/or can be included of programmable units, such as programmable gate arrays or processors. The modules/units/blocks or computing device functionality described herein may be implemented as software modules/units/blocks, but may be represented in hardware or firmware. In general, the modules/units/blocks described herein refer to logical modules/units/blocks that may be combined with other modules/units/blocks or divided into sub-modules/sub-units/sub-blocks despite their physical organization or storage. The description may be applicable to a system, an engine, or a portion thereof.
The terminology used herein is for the purposes of describing particular examples and embodiments only, and is not intended to be limiting. As used herein, the singular forms “a,” “an,” and “the” may be intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “include,” and/or “comprise,” when used in this disclosure, specify the presence of integers, devices, behaviors, stated features, steps, elements, operations, and/or components, but do not exclude the presence or addition of one or more other integers, devices, behaviors, features, steps, elements, operations, components, and/or groups thereof.
These and other features, and characteristics of the present disclosure, as well as the methods of operation and functions of the related elements of structure and the combination of parts and economies of manufacture, may become more apparent upon consideration of the following description with reference to the accompanying drawings, all of which form a part of this disclosure. It is to be expressly understood, however, that the drawings are for the purpose of illustration and description only and are not intended to limit the scope of the present disclosure. It is understood that the drawings are not to scale.
FIG. 2 shows a schematic block diagram of an exemplary MRI apparatus 202 for providing MRI data according to the disclosed embodiments. The MRI apparatus 202 may include an MRI scanner 204, receive and control circuitry 206 and a display 208. The MRI scanner 204 may include, as shown in cross section in FIG. 2, a magnetic field generator 210, a gradient magnetic field generator 212, and a Radio Frequency (RF) generator 214, all surrounding a table 216 on which subjects under study may be positioned. The MRI scanner 204 may also include an ECG signal sensor 218 for capturing MRI data in the form of ECG signals from the subject under study during MRI scanning, a camera 220 for capturing MRI data in the form of video images of the subject under study during MRI scanning, and a pulse detector 222, for capturing MRI data in the form of a subject's pulse during MRI scanning. In some embodiments, the MRI scanner 204 may perform a scan on a subject or a region of the subject. The subject may be, for example, a human body or other animal body. For example, the subject may be a patient. The region of the subject may include part of the subject. For example, the region of the subject may include a tissue of the patient. The tissue may include, for example, lung, prostate, breast, colon, rectum, bladder, ovary, skin, liver, spine, bone, pancreas, cervix, lymph, thyroid, spleen, adrenal gland, salivary gland, sebaceous gland, testis, thymus gland, penis, uterus, trachea, skeletal muscle, smooth muscle, heart, etc. In some embodiments, the scan may be a pre-scan for calibrating an imaging scan. In some embodiments, the scan may be an imaging scan for generating an image.
The main magnetic field generator 210 may create a static magnetic field B0 and may include, for example, a permanent magnet, a superconducting electromagnet, a resistive electromagnet, or any magnetic field generation device suitable for generating a static magnetic field. The gradient magnet field generator 212 may use coils to generate a magnetic field in the same direction as B0 but with a gradient in one or more directions, for example, along X, Y, or Z axes in a coordinate system of the MRI scanner 204.
In some embodiments, the RF generator 214 may use RF coils to transmit RF energy through the subject, or region of interest of the subject, to induce electrical signals in the region of interest. The resulting RF field is typically referred to as the B1 field and combines with the B0 field to generate MR signals that are spatially localized and encoded by the gradient magnetic field. The MRI scanner 204 may further include an RF detector 224 implemented using, for example, an RF coil, where the RF detector operates to sense the RF field and convey a corresponding output to the receive and control circuitry 206. The RF detector may also include one or more coil arrays for parallel imaging. The function, size, type, geometry, position, amount, or magnitude of the MRI scanner 204 may be determined or changed according to one or more specific conditions. For example, the MRI scanner 204 may be designed to surround a subject (or a region of the subject) to form a tunnel type MRI scanner, referred to as a closed bore MRI scanner, or an open MRI scanner, referred to as an open-bore MRI scanner. As another example, the MRI scanner may be portable and transportable down hallways and through doorways to a patient, providing MR scanning services to the patient as opposed to transporting the patient to the MRI scanner. In some examples, a portable MRI scanner may be configured to scan a region of interest of a subject, for example, the subject's brain, spinal cord, limbs, heart, blood vessels, and internal organs.
The ECG signal sensor 218 may operate to capture ECG signals from the subject under study during MRI scanning for use in subsequently identifying cardiac cycles and cardiac phases of the subject. The camera 220 may operate to capture video images of the subject under study during MRI scanning for use in subsequently identifying cardiac cycles and cardiac phases of the subject. During MRI scanning the subject may be requested to hold their breath and to stay still in order to provide accurate MRI cardiac data while scanning. However, this may be difficult for any number of reasons, and video images of the subject may be used to compensate for subject movement or breathing patterns during scanning that may adversely affect the acquired MRI data. The pulse detector 222 may provide pulse data from the subject during MRI scanning which may also be used to enhance cardiac cycle and phase predictions.
The receive and control circuitry 206 may control overall operations of the MRI scanner 204, in particular, the magnetic field generator 210, the gradient magnetic field generator 212, the RF generator 214, and the RF detector 224. For example, the receive and control circuitry 206 may control the magnet field gradient generator to produce gradient fields along one or more of the X, Y, and Z axes, and the RF generator to generate the RF field. In some embodiments, the receive and control circuitry 206 may receive commands from, for example, a user or another system, and control the magnetic field generator 210, the gradient magnetic field generator 212, the RF generator 214, and the RF detector 224 accordingly. The receive and control circuitry 206 may be connected to the MRI scanner 204 through a network 226. The network 226 may include any suitable network that can facilitate the exchange of information and/or data for the MRI scanner 204. The network 226 may include one or more of a public network (e.g., the Internet), a private network (e.g., a local area network (LAN), a wide area network (WAN)), etc.), a wired network (e.g., an Ethernet network), a wireless network (e.g., an 802.11 network, a Wi-Fi network, etc.), a cellular network (e.g., a Long Term Evolution (LTE) network), a frame relay network, a virtual private network (“VPN”), a satellite network, a telephone network, routers, hubs, switches, server computers, and/or any combination thereof. Merely by way of example, the network 418 may include a cable network, a wireline network, a fiber-optic network, a telecommunications network, an intranet, a wireless local area network (WLAN), a metropolitan area network (MAN), a public telephone switched network (PSTN), a Bluetooth® network, a ZigBee® network, a near field communication (NFC) network, or the like, or any combination thereof. In some embodiments, the network 226 may include one or more network access points. For example, the network 226 may include wired and/or wireless network access points such as base stations and/or internet exchange points through which one or more components of the MRI scanner 204 may be connected with the network 226 to exchange data and/or information.
According to some embodiments, the receive and control circuitry 206 may operate the MRI scanner 204 to perform operations according to the disclosed embodiments, including automatically estimating a pose of the heart; automatically providing imaging slice planning; performing a highly accelerated scout scan between acquisition scans, and automatically adjusting the image slices to maintain consistency of planned slice locations for each acquisition in spite of movement which may arise as a result of heart movement, breathing, patient movement, or other factors that cause changes in heart position between acquisition scans. The receive and control circuitry 206 may include a processing engine 300 for operating the MRI scanner 204 to perform the operations and workflows according to the disclosed embodiments.
FIG. 3 illustrates an example implementation of the processing engine 300 according to the disclosed embodiments. The processing engine 300 may include computer readable program code stored on at least one computer readable medium 302 for carrying out and executing the process steps described herein. The computer readable program code for carrying out operations for aspects of the present disclosure may be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB. NET, Python or the like, conventional procedural programming languages, such as the “C” programming language, Visual Basic, Fortran 2103, Perl, COBOL 2102, PHP, ABAP, dynamic programming languages such as Python, Ruby, and Groovy, or other programming languages. The computer readable program code may execute entirely on the processing engine 300, partly on the processing engine 300, as a stand-alone software package, partly on the processing engine 300 and partly on a remote computer or server or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the processing engine 300 through any type of network, including those mentioned above with respect to network 226.
The computer readable medium 302 may be a memory of the processing engine 300. In alternate aspects, the computer readable program code may be stored in a memory external to, or remote from, the processing engine 300. The memory may include magnetic media, semiconductor media, optical media, or any media which is readable and executable by a computer. The processing engine 300 may also include a computer processor 304 for executing the computer readable program code stored on the at least one computer readable medium 302. In at least one aspect, the processing engine 300 may include one or more input or output devices, generally referred to as a user interface 306 which may operate to allow input to the processing engine 306 or to provide output from the processing engine 300, respectively. The processing engine 300 may be implemented in hardware, software or a combination of hardware and software. According to one or more embodiments, the processing engine 300 may be part of the receive and control circuitry 206, while in other embodiments the processing engine 300 may be located remotely from the receive and control circuitry 206.
FIG. 4 shows an exemplary work flow that may be implemented using the exemplary MRI apparatus 202. As shown in block 402, initial scout images may be acquired before performing subsequent scans. The scout images may be 2D multi-slice images from all three standard body views (axial, sagittal and coronal) at a low spatial resolution. The scout images may also be low resolution true 3D image volumes. As shown in block 404, from the scout images, the location, pose, shape, and other aspects of a patient's heart may be estimated using a neural network, an example of which is illustrated as item 600 in FIG. 6. Referring to block 406, the neural network 600 may utilize the aspects the heart to establish a patient specific heart model, as schematically illustrated in FIG. 5A. Referring to block 408, the neural network 600 may be used to estimate standard heart views according to clinical standards from the patient specific heart model, including the short axis view, horizontal long axis view, and vertical long axis view. Referring to block 410, the estimated standard heart views may optionally be updated and refined by one or more MR technicians and the updates and refinements may be used to update the patient specific heart model. Referring to block 412, the standard heart views may be used to automatically plan imaging planes, as schematically illustrated in FIG. 5B. Blocks 414 and 416 represent operations that may be referred to as an Artificial Intelligence (AI) scout scan. As shown in block 414, an accelerated scan, for example, one or more of a multi-slice, multi-view, 2D or 3D, scan may be performed to acquire cardiac positioning data to determine the location of the heart before the next acquisition scan. The accelerated scan techniques may include using compressed sensing where data is undersampled in the K-space, parallel imaging where data is individually obtained from multiple receiver coils, and fast spin echo where multiple echoes are acquired during each sequence pulse to allow for acquisition in less time of a reduced amount of data than required to support a higher resolution or larger field of view.
The neural network 600 may be used to reconstruct the highly accelerated data as shown in block 416, and may be used to compare the heart location and pose to those of the patient specific heart model from the initial scouting, as shown in block 418. Referring to block 420, the prescribed imaging planes may then be automatically adjusted to correspond to the heart location and pose, as illustrated in FIG. 5C, and used for an acquisition scan for a selected imaging protocol, as shown in block 422. Thus, the AI scout scan 414, 416 may be used to determine the location and pose of the heart immediately before the acquisition scan 422 and may be used to represent the location and pose of the heart for that particular acquisition scan.
Cardiac MRI imaging protocols may be generally tailored to specific clinical indications, for example, anatomic images of the heart and great vessels, including axial, coronal, sagittal, long axis, and short axis views, and views of coronary arteries, and valves. Other cardiac MRI imaging protocols may be directed to cardiac function, for example, motion of the ventricular walls during systole and diastole, turbulence created by valvular stenosis, and cine studies obtained by repeatedly imaging the heart at a single slice location throughout the cardiac cycle. Still other cardiac MRI imaging protocols may be directed to myocardial viability, utilizing for example, segmented, T1-weighted, inversion-prepared fast gradient echo sequences.
While the AI scout scan 414, reconstruction of AI scout scan data 416, comparison of the heart location and pose 418, automatic repositioning 420, and acquisition scan 422, are described in the context of being performed by a single neural network 600, it should be understood that the scan 414, reconstruction 416, comparison 418, automatic repositioning 420, and acquisition scan 422, may be performed individually by different neural networks or performed in groups by different neural networks.
It should be noted that utilizing the neural network 600 advantageously ensures reconstruction quality and reduces the time required for establishing the patient specific heart model, planning and repositioning the image planes, computing the reconstructions, and repositioning the image planes. For example, because the position, pose, and short and long axes are defined by the patient specific heart model, the neural network may utilize this information to automatically plan the imaging planes, instead of having a technician manually plan the imaging planes. Furthermore, because the disclosed embodiments establish a patient specific heart model, a technician is no longer required to perform additional scans to relocate the heart position, pose, and short and long axes, because the position, pose, and short and long axes are defined by the patient specific heart model. Thus, the desired slice location relative to the structure of the heart as defined during the initial scouting may be maintained regardless of changes in the heart location and pose throughout the imaging protocol scans. Still further, use of the neural network 600 enables completion of the AI scout scan and in particular, reconstruction of the accelerated data, in significantly less time than technician controlled rescans, reducing the time required for the patient to stop breathing or remain immobile, or both. It should also be noted that while the disclosed embodiments are described in the context of utilizing a neural network, other computational methods that meet the speed and accuracy requirements may also be utilized.
FIG. 6 depicts an example of the neural network 600 that may be utilized to implement the disclosed embodiments. While a simple neural network is shown, it should be understood that the disclosed embodiments may be implemented utilizing a deep learning method or deep learning model including one or more gated recurrent units (GRUs), long short term memory (LSTM) networks, fully convolutional neural network (FCN) models, generative adversarial networks (GANs), back propagation (BP) neural network models, radial basis function (RBF) neural network models, deep belief nets (DBN) neural network models, Elman neural network models, or any deep learning or machine learning model capable of performing the operations described herein.
Techniques that train to learn or to select a particular neural network structure can be used to learn the hyperparameter of the neural network 600 for optimal performance. One example following a reinforcement learning framework can be a searching neural network that can act on the tested neural network by changing the hyperparameters and observing the resulting performance. The searching network can continuously perform trials of acting and observing, and accumulate experiences through the trials. The target of the searching network is to maximize some reward, which can be defined as achieving better performance. The searching network will eventually reach an optimal performance point, at which the operations of the searching network may be terminated.
Thus, while there have been shown, described and pointed out, fundamental novel features of the invention as applied to the exemplary embodiments thereof, it will be understood that various omissions, substitutions and changes in the form and details of devices and methods illustrated, and in their operation, may be made by those skilled in the art without departing from the spirit and scope of the presently disclosed invention. Further, it is expressly intended that all combinations of those elements, which perform substantially the same function in substantially the same way to achieve the same results, are within the scope of the invention. Moreover, it should be recognized that structures and/or elements shown and/or described in connection with any disclosed form or embodiment of the invention may be incorporated in any other disclosed or described or suggested form or embodiment as a general matter of design choice. It is the intention, therefore, to be limited only as indicated by the scope of the claims appended hereto.

Claims

What is claimed is:

1. A method comprising:

acquiring initial scout images of a patient's heart;

using a neural network to establish a patient specific heart model, and automatically plan imaging planes of the patient specific heart model;

performing an accelerated scan of the patient's heart;

using the neural network to determine a current location and pose of the patient's heart from the accelerated scan, and to reposition the imaging planes to correspond to the current location and pose of the patient's heart; and

using the repositioned imaging planes to perform an acquisition scan and generate an image of the patient's heart from the acquisition scan according to a selected imaging protocol.

2. The method of claim 1, comprising acquiring the initial scout images from standard MRI body views.

3. The method of claim 1, wherein the initial scout images comprise 2D or 3D images from one or more of axial sagittal and coronal views.

4. The method of claim 1, wherein using the neural network to determine a current location and pose of the patient's heart from the accelerated scan comprises reconstructing an image from the accelerated scan and comparing the reconstructed image from the accelerated scan to the patient specific heart model.

5. The method of claim 1, further comprising comparing the current location and pose of the patient's heart to the location and pose of the patient specific heart model, and repositioning the imaging planes obtained from the patient specific heart model to correspond to the current location and pose of the patient's heart.

6. The method of claim 1, wherein the neural network comprises one or more of a combination of CNN and RNN models, a GRU model, an LSTM model, a fully convolutional neural network model, a generative adversarial network, a back propagation neural network model, a radial basis function neural network model, a deep belief nets neural network model, an Elman neural network model.

7. The method of claim 1, wherein the accelerated scan comprises one or more of compressed sensing; parallel imaging; or fast spin echo techniques to allow for acquisition in less time of a reduced amount of data than required to support a higher resolution or larger field of view.

8. The method of claim 1, wherein the selected imaging protocol comprises one or more of obtaining anatomic images of the heart, determining cardiac function, or determining myocardial viability.

9. A system comprising:

an MRI scanner; and

a processing engine coupled to the MRI scanner, the processing engine comprising a processor and a memory comprising computer readable program code, wherein the processor under control of the computer readable program code is operable to:

acquire initial scout images of a patient's heart;

use a neural network to establish a patient specific heart model, and automatically plan imaging planes of the patient specific heart model;

perform an accelerated scan of the patient's heart;

use the neural network to determine a current location and pose of the patient's heart from the accelerated scan, and to reposition the imaging planes to correspond to the current location and pose of the patient's heart; and

cause the MRI scanner to use the repositioned imaging planes to perform an acquisition scan and generate an image of the patient's heart from the acquisition scan according to a selected imaging protocol.

10. The system of claim 9, wherein the processor under control of the computer readable program code is operable to cause the MRI scanner to acquire the initial scout images from standard MRI body views.

11. The system of claim 9, wherein the initial scout images comprise 2D or 3D multi-slice images from one or more of axial sagittal and coronal views.

12. The system of claim 9, wherein the processor under control of the computer readable program code is operable to use the neural network to determine a current location and pose of the patient's heart from the accelerated scan by reconstructing an image from the accelerated scan and comparing the reconstructed image from the accelerated scan to the patient specific heart model.

13. The system of claim 9, wherein the processor under control of the computer readable program code is operable to use the neural network to compare the current location and pose of the patient's heart to the location and pose of the patient specific heart model, and reposition the imaging planes obtained from the patient specific heart model to correspond to the current location and pose of the patient's heart.

14. The system of claim 9, wherein the neural network comprises one or more of a combination of CNN and RNN models, a GRU model, an LSTM model, a fully convolutional neural network model, a generative adversarial network, a back propagation neural network model, a radial basis function neural network model, a deep belief nets neural network model, an Elman neural network model.

15. The system of claim 9, wherein the accelerated scan comprises one or more of compressed sensing; parallel imaging; or fast spin echo techniques.

16. The system of claim 9, wherein the selected imaging protocol comprises one or more of obtaining anatomic images of the heart, determining cardiac function, or determining myocardial viability.