Methods and system for reconstructing four-dimensional magnetic resonance image

Abstract

The embodiments of the present description provide methods and system for constructing four-dimensional magnetic resonance images. A method comprises: determining a data acquisition mode of each phase on the basis of a real-time motion signal of an object; acquiring original data of the object according to the data acquisition mode; classifying the original data to obtain data of multiple categories, wherein data of each category corresponds to a different phase; and performing image reconstruction on the data of the multiple categories to obtain a four-dimensional magnetic resonance image.

Description

Method and system for reconstructing four-dimensional magnetic resonance images Technical Field

[0001] The present application relates to the field of imaging technology, and in particular to a four-dimensional magnetic resonance image reconstruction method and system. Background Art

[0002] In the field of radiotherapy, imaging has become a key technology for improving treatment outcomes, and image quality often significantly impacts the accuracy of radiotherapy beams. However, due to the patient's breathing and heartbeat during treatment, static images are often inadequate for therapeutic reference. Four-dimensional magnetic resonance imaging (4D MRI), with its superior soft tissue contrast and lack of ionizing radiation, enables more efficient optimization of radiotherapy treatment plans, improves radiotherapy accuracy, and provides superior imaging reference for the field.

[0003] Summary of the Invention

[0004] One embodiment of the present specification provides a four-dimensional magnetic resonance image reconstruction method, comprising: determining a data acquisition method for each phase based on a real-time motion signal of an object; acquiring original data of the object according to the data acquisition method; classifying the original data to obtain multiple categories of data, each category of data corresponding to a different phase; and obtaining a four-dimensional magnetic resonance image by reconstructing the multiple categories of data.

[0005] One of the embodiments of the present specification also provides a four-dimensional magnetic resonance image reconstruction method, comprising: determining the full K-space data of each phase based on the acquired original data; performing image reconstruction based on the full K-space data for each phase to obtain a corresponding reconstructed image; and combining the reconstructed images of each phase to obtain a four-dimensional magnetic resonance image; wherein, determining the full K-space data of each phase comprises: for each undersampled target phase, supplementing the uncollected data in the target phase based on the sampling data of the motion stationary period and / or the sampling data of the candidate phase with more sampling data than the target phase, so as to obtain the full K-space data of the target phase.

[0006] One of the embodiments of the present specification also provides a four-dimensional magnetic resonance image reconstruction system, comprising: at least one storage medium, comprising a set of instructions; and one or more processors communicating with the at least one storage medium, wherein, when executing the instructions, the one or more processors are used to: determine the data acquisition method for each phase based on the real-time motion signal of the object; acquire the original data of the object according to the data acquisition method; classify the original data to obtain multiple categories of data, each category of data corresponding to a different phase; and obtain a four-dimensional magnetic resonance image by reconstructing the multiple categories of data.

[0007] One of the embodiments of this specification also provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the following method: determining a data acquisition method for each phase based on the real-time motion signal of the object; acquiring original data of the object according to the data acquisition method; classifying the original data to obtain multiple categories of data, each category of data corresponding to a different phase; and obtaining a four-dimensional magnetic resonance image by reconstructing the multiple categories of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0009] FIG1 is a schematic diagram of an exemplary four-dimensional magnetic resonance image reconstruction system according to some embodiments of the present specification;

[0010] FIG2 is a schematic diagram of hardware and / or software of an exemplary computing device according to some embodiments of the present specification;

[0011] FIG3 is a schematic diagram of modules of an exemplary four-dimensional magnetic resonance image reconstruction apparatus according to some embodiments of the present specification;

[0012] FIG4 is a flowchart of an exemplary four-dimensional magnetic resonance image reconstruction method according to some embodiments of the present specification;

[0013] FIG5( a ) is a schematic diagram illustrating the phase division of an exemplary respiratory signal according to some embodiments of the present specification;

[0014] FIG5( b ) is a schematic diagram illustrating the phase division of an exemplary respiratory signal according to some embodiments of the present specification;

[0015] FIG6 is a phase diagram of an exemplary respiratory signal according to some embodiments of the present specification;

[0016] FIG7( a ) is a schematic diagram of K-space filling of an exemplary Cartesian acquisition according to some embodiments of the present specification;

[0017] FIG7( b ) is a schematic diagram of K-space filling of an exemplary Cartesian acquisition according to some embodiments of the present specification;

[0018] FIG8( a ) is a schematic diagram of K-space filling of an exemplary Zig-Zag acquisition according to some embodiments of the present specification;

[0019] FIG8( b ) is a schematic diagram of K-space filling of an exemplary Zig-Zag acquisition according to some embodiments of the present specification;

[0020] FIG9( a ) is a schematic diagram of K-space filling of an exemplary radial acquisition according to some embodiments of the present specification;

[0021] FIG9( b ) is a schematic diagram of K-space filling of an exemplary radial acquisition according to some embodiments of the present specification;

[0022] FIG10( a ) is a schematic diagram of K-space filling of an exemplary spiral acquisition according to some embodiments of the present specification;

[0023] FIG10( b ) is a schematic diagram of K-space filling of an exemplary spiral acquisition according to some embodiments of the present specification;

[0024] FIG11( a ) is a schematic diagram of K-space filling of an exemplary windmill acquisition according to some embodiments of this specification;

[0025] FIG11( b ) is a schematic diagram of K-space filling of an exemplary windmill acquisition according to some embodiments of this specification;

[0026] FIG12( a ) is a schematic diagram of K-space filling of an exemplary radial 3D acquisition according to some embodiments of the present specification;

[0027] FIG12( b ) is a schematic diagram of K-space filling of an exemplary radial 3D acquisition according to some embodiments of the present specification;

[0028] FIG13 is a schematic diagram of an exemplary under-sampling K-space filling according to some embodiments of the present specification;

[0029] FIG14( a ) is a schematic diagram of exemplary under-sampling K-space filling according to other embodiments of this specification;

[0030] FIG14( b ) is a schematic diagram of exemplary under-sampling K-space filling according to other embodiments of this specification; and

[0031] FIG15 is a schematic diagram of an exemplary full-sampling K-space filling according to some embodiments of the present specification. DETAILED DESCRIPTION

[0032] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0033] It should be understood that the terms "system," "device," "unit," and / or "module" used herein are a method for distinguishing different components, elements, parts, portions, or assemblies at different levels. However, if other terms can achieve the same purpose, the terms may be replaced by other expressions.

[0034] As used in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not refer to the singular but also include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0035] Flowcharts are used throughout this specification to illustrate the operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations do not necessarily need to be performed in exact order. Instead, the steps may be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0036] This specification provides systems and methods for non-invasive imaging and / or treatment, for example, for disease diagnosis, treatment or research purposes. In some embodiments, the system may include a radiation therapy (RT) system, a computed tomography (CT) system, an emission computed tomography (ECT) system, an X-ray imaging system, a positron emission tomography (PET) system, a magnetic resonance imaging (MRI) system, or the like, or any combination thereof. For illustrative purposes, this specification describes a system and method for four-dimensional magnetic resonance image reconstruction. In this specification, the term "image" may refer to a 2D image, a 3D image, or a 4D image. In some embodiments, the term "image" may refer to an image of a region of a patient (e.g., a region of interest (ROI)).

[0037] As mentioned above, 4D MRI images provide superior imaging references for radiotherapy due to their superior soft tissue contrast and lack of ionizing radiation. However, due to the specificity of magnetic resonance imaging technology, the actual image data acquisition speed is limited, resulting in motion artifacts and low temporal resolution.

[0038] In some embodiments, respiratory amplitude information at the time of acquisition can be assigned to the acquired signal data based on the respiratory signal, assisting in combining the K-space data under the same respiratory state to form a complete and reconstructible magnetic resonance raw data. The combined data is then inverse Fourier transformed to obtain a 4D MRI image. However, this method has the problems of long scanning time and large data redundancy.

[0039] The present invention provides a four-dimensional magnetic resonance image reconstruction system and method based on real-time motion information of an object. When the object is in a stable motion state, more data is collected to ensure that high-frequency regions of K-space (such as edges) are fully populated. When the object's motion amplitude fluctuates significantly, data from the central region of K-space is collected, while data from high-frequency regions is discarded, maintaining the capture of image contrast information. This results in images with high temporal resolution. This approach can further reduce image acquisition time and improve scanning efficiency without compromising or even improving image quality.

[0040] FIG1 is a schematic diagram of an exemplary four-dimensional magnetic resonance image reconstruction system according to some embodiments of the present specification.

[0041] As shown in FIG1 , a magnetic resonance image reconstruction system 100 may include an imaging device 110, a motion signal acquisition device 112, a network 120, a processing device 130, a terminal device 140, and a storage device 150. In some embodiments, the various components in the magnetic resonance image reconstruction system 100 may be interconnected via the network 120 or may be directly connected without going through the network 120. For example, the imaging device 110 and the terminal device 140 may be connected via the network 120. For another example, the imaging device 110 and the processing device 130 may be connected via the network 120 or directly connected. For another example, the processing device 130 and the terminal device 140 may be connected via the network 120 or directly connected to the terminal device 140.

[0042] The imaging device 110 can scan an object within a detection area or a scanning area to obtain scan data of the object (e.g., raw data, reconstructed images, etc.). For example, the imaging device 110 can collect raw data of the object according to a determined data collection method. The object can include biological or non-biological objects. As just one example, the object can include a patient, an artificial object (e.g., an artificial phantom), etc. For another example, the object can include a specific part, organ, and / or tissue of a patient (e.g., the head, ear, nose, mouth, neck, chest, abdomen, liver, gallbladder, pancreas, spleen, kidney, spine, heart, or tumor tissue, etc.).

[0043] In some embodiments, the imaging device 110 may include one or a combination of X-ray equipment, computed tomography (CT) equipment, ultrasound imaging components, fluoroscopic imaging components, magnetic resonance imaging (MRI) equipment, single photon emission computed tomography (SPECT) equipment, positron emission tomography (PET) equipment, etc.

[0044] In some embodiments, the imaging device 110 may be an MRI device. The MRI device uses a strong magnetic field, a gradient magnetic field, and radio waves to generate magnetic resonance images (such as two-dimensional magnetic resonance images, three-dimensional magnetic resonance images, four-dimensional magnetic resonance images, etc.) of the object to be scanned. In some embodiments, the MRI device may include a magnet assembly, a gradient coil assembly, and a radio frequency (RF) coil assembly. The magnet assembly may generate a first magnetic field (also referred to as a main magnetic field) for polarizing the object. The gradient coil assembly may be used to generate a second magnetic field (also referred to as a gradient magnetic field).

[0045] In some embodiments, the gradient coil assembly can generate one or more magnetic field gradient pulses in the X direction (Gx), Y direction (Gy) and Z direction (Gz) for the main magnetic field to encode the spatial information of the object. In some embodiments, the X direction can be designated as the frequency encoding direction, and the Y direction can be designated as the phase encoding direction. In some embodiments, Gx can be used for frequency encoding or signal readout, which is generally referred to as frequency encoding or readout (RO) gradient. In some embodiments, Gy can be used for phase encoding, which is generally referred to as phase encoding (PE, Phase-Encoding) gradient. In some embodiments, Gz can be used for slice selection to obtain two-dimensional k-space data, which is generally referred to as slice selection (SS) gradient. In some embodiments, Gz can be used for phase encoding to obtain three-dimensional k-space data.

[0046] In some embodiments, the RF coil may include one or more transmitting coils and / or one or more receiving coils. The RF transmitting coil may transmit RF pulses to the subject. Under the coordinated action of the main magnetic field / gradient magnetic field and the RF pulses, magnetic resonance signals related to the subject may be generated according to a pulse sequence. The RF receiving coil may acquire magnetic resonance signals from the subject according to the pulse sequence. Magnetic resonance signals may also be referred to as echo signals. The magnetic resonance signals may be processed using a transform operation (e.g., Fourier transform) to fill in K-space data, thereby generating a magnetic resonance image of the subject.

[0047] A pulse sequence can be defined by imaging parameters and the arrangement of imaging parameters in time. For example, a pulse sequence can be defined by one or more parameters related to time (e.g., acquisition time (TA), echo time (TE), etc.). Exemplary pulse sequences may include spin echo sequences, gradient echo sequences, diffusion sequences, inversion recovery sequences, etc., or any combination thereof. For example, a spin echo sequence may include fast spin echo (FSE), eddy current spin echo (TSE), rapid acquisition with relaxation enhancement (RARE), half-Fourier acquisition single shot spin echo (HASTE), eddy current gradient spin echo (TGSE), etc., or any combination thereof. In some embodiments, a pulse sequence may include a diffusion sequence. A diffusion sequence may include at least one diffusion block associated with one or more parameters of a diffusion gradient and at least one imaging block associated with one or more scanning parameters (e.g., parameters associated with one or more encoding gradients applied in the scan). In some embodiments, the pulse sequence may include a single-shot pulse sequence (e.g., single-shot FSE, single-shot EPI, etc.) or a multi-shot pulse sequence (e.g., multi-shot FSE, multi-shot EPI, etc.). In some embodiments, for a single-shot pulse sequence (e.g., single-shot FSE, single-shot EPI, etc.), an RF pulse may be applied once within one TR (repetition time, TR), and one or more echo signals may be sampled within the TR. For a multi-shot pulse sequence (e.g., multi-shot FSE, multi-shot EPI, etc.), an RF pulse may be applied multiple times within one TR, and one or more echo signals may be sampled during each excitation or stimulation.

[0048] The motion signal acquisition device 112 can be used to acquire real-time motion signals of the object (such as breathing signals, heartbeat signals, etc.). In some embodiments, the motion signal acquisition device 112 may include a breathing signal acquisition device or a heartbeat signal acquisition device. In some embodiments, the motion signal acquisition device 112 may include a contact motion signal acquisition device and / or a non-contact motion signal acquisition device. The contact motion signal acquisition device can calculate the motion signal from the electrocardiogram chest lead or collect the motion signal from the carbon dioxide sensor. The non-contact motion signal acquisition device can collect motion signals through bioradar, a special mattress-type sensor-based method, optical signal acquisition, etc. or a combination thereof. In some embodiments, the motion signal acquisition device 112 can collect real-time motion signals through a variety of methods. The acquisition method may include feature recognition technology based on real-time images, optical surface imaging technology, infrared marking technology, respiratory strap technology, etc. or any combination thereof.

[0049] In some embodiments, the motion signal acquisition device 112 may further process the acquired real-time motion signal, for example, by filtering, noise reduction, smoothing, etc.

[0050] In some embodiments, the motion signal collected by the motion signal collection device 112 may be a motion signal curve with time as the horizontal axis and amplitude as the vertical axis. In some embodiments, the real-time motion signal collected by the motion signal collection device 112 may include one or more signal cycles. In some embodiments, the motion signal collection device 112 may collect real-time motion signals of the subject before, during, and after the scan. In some embodiments, the motion signal collection device 112 may synchronously collect real-time motion signals during the imaging process.

[0051] In some embodiments, the motion signal acquisition device 112 may be a part of the imaging device 110 , or a device independent of the imaging device 110 .

[0052] The network 120 may include any suitable network capable of facilitating information and / or data exchange within the magnetic resonance image reconstruction system 100. In some embodiments, one or more components of the magnetic resonance image reconstruction system 100 (e.g., the imaging device 110, the motion signal acquisition device 112, the processing device 130, the terminal device 140, and the storage device 150) may exchange information and / or data with one or more components of the magnetic resonance image reconstruction system 100 via the network 120. For example, the processing device 130 may obtain real-time motion signals of a subject from the motion signal acquisition device 112 via the network 120. In some embodiments, the network 120 may include 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., Ethernet), a wireless network (e.g., an 802.11 network, a wireless 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, a router, a hub, a server computer, or a combination thereof. For example, the network 120 may include a wired network, an optical fiber network, a telecommunication network, a local area network, a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth TM Network, ZigBee TM In some embodiments, the network 120 may include one or more network access points. For example, the network 120 may include wired and / or wireless network access points, such as base stations and / or Internet exchange points. Through the access points, one or more components of the magnetic resonance image reconstruction system 100 may connect to the network 120 to exchange data and / or information.

[0053] The processing device 130 can process data and / or information obtained from the imaging device 110, the motion signal acquisition device 112, the terminal device 140, the storage device 150, or other components of the magnetic resonance image reconstruction system 100. For example, the processing device 130 can obtain a real-time motion signal of an object from the motion signal acquisition device 112 and determine a data acquisition method for each phase based on the real-time motion signal of the object. In some embodiments, the processing device 130 can be local or remote. For example, the processing device 130 can access information and / or data from the imaging device 110, the motion signal acquisition device 112, the terminal device 140, and / or the storage device 150 via the network 120.

[0054] In some embodiments, the processing device 130 and the imaging device 110 may be integrated into one. In some embodiments, the processing device 130 and the imaging device 110 may be directly or indirectly connected to work together to implement the methods and / or functions described in this specification.

[0055] In some embodiments, the processing device 130 and the motion signal acquisition device 112 can be integrated into one. In some embodiments, the processing device 130 and the motion signal acquisition device 112 can be directly or indirectly connected to work together to implement the methods and / or functions described herein. In some embodiments, the imaging device 110, the motion signal acquisition device 112, and the processing device 130 can be integrated into one. In some embodiments, the imaging device 110, the motion signal acquisition device 112, and the processing device 130 can be directly or indirectly connected to work together to implement the methods and / or functions described herein.

[0056] In some embodiments, the processing device 130 may include an input device and / or an output device. The input device and / or the output device may enable interaction with a user (e.g., displaying a real-time motion signal, displaying a four-dimensional magnetic resonance image, etc.). In some embodiments, the input device and / or the output device may include a display screen, a keyboard, a mouse, a microphone, or any combination thereof.

[0057] The terminal device 140 can be connected and / or communicate with the imaging device 110, the motion signal acquisition device 112, the processing device 130, and / or the storage device 150. For example, the terminal device 140 can obtain and display a reconstructed four-dimensional magnetic resonance image from the processing device 130. For another example, the terminal device 140 can obtain and display a data acquisition method determined based on a real-time motion signal from the processing device 130. In some embodiments, the terminal device 140 may include a mobile device 141, a tablet computer 142, a laptop computer 143, or the like, or any combination thereof. In some embodiments, the terminal device 140 (or all or part of its functions) can be integrated into the processing device 130.

[0058] The storage device 150 may store data, instructions, and / or any other information. In some embodiments, the storage device 150 may store data acquired from the imaging device 110, the motion signal acquisition device 112, and / or the processing device 130 (e.g., real-time motion signals, data acquisition methods, raw data with markers, four-dimensional magnetic resonance images, etc.). In some embodiments, the storage device 150 may store computer instructions for implementing a four-dimensional magnetic resonance image reconstruction method, etc.

[0059] In some embodiments, the storage device 150 may include one or more storage components, each of which may be a standalone device or part of another device. In some embodiments, the storage device 150 may include random access memory (RAM), read-only memory (ROM), mass storage, removable memory, volatile read-write memory, or any combination thereof. Exemplary mass storage may include a magnetic disk, an optical disk, a solid-state disk, or the like. RAM may include dynamic RAM (DRAM), double data rate synchronous dynamic RAM (DDR SDRAM), static RAM (SRAM), thyristor RAM (T-RAM), and zero-capacitance RAM (Z-RAM). ROM may include mask ROM (MROM), programmable ROM (PROM), erasable programmable ROM (PEROM), electrically erasable programmable ROM (EEPROM), compact disc ROM (CD-ROM), and digital versatile disk ROM. In some embodiments, the storage device 150 may be implemented on a cloud platform.

[0060] It should be noted that the magnetic resonance image reconstruction system 100 is provided for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art will appreciate that various modifications and variations can be made based on the description of this specification. For example, the magnetic resonance image reconstruction system 100 may include one or more medical devices. In some embodiments, one of the one or more medical devices may be used for both imaging and treatment. In some embodiments, the imaging and treatment processes may also be performed by multiple different medical devices. However, such variations and modifications do not deviate from the scope of this specification.

[0061] FIG. 2 is a schematic diagram of hardware and / or software of an exemplary computing device according to some embodiments of the present specification.

[0062] As shown in FIG. 2 , computing device 200 may include a processor 210 , memory 220 , input / output 230 , and communication port 240 .

[0063] The processor 210 can execute computing instructions (program code) and perform the functions of the magnetic resonance image reconstruction system 100 described in this application. The computing instructions may include programs, components, data structures, processes, modules and functions (the functions refer to specific functions described in this application). For example, the processor 210 can process data obtained from any component of the magnetic resonance image reconstruction system 100. For example, the processor 210 can analyze the real-time motion signal of the object obtained from the motion signal acquisition device 112 to determine the data acquisition method for each phase. For another example, the processor 210 can classify the labeled raw data obtained from the imaging device 110 to obtain multiple categories of data, and obtain a four-dimensional magnetic resonance image of the object by performing image reconstruction on the multiple categories of data. In some embodiments, the processor 210 may include a microcontroller, a microprocessor, a reduced instruction set computer (RISC), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a central processing unit (CPU), a graphics processing unit (GPU), a physical processing unit (PPU), a microcontroller unit, a digital signal processor (DSP), a field programmable gate array (FPGA), an advanced RISC machine (ARM), a programmable logic device, and any circuit and processor capable of performing one or more functions, or any combination thereof. For illustrative purposes only, the computing device 200 in FIG. 2 depicts only one processor, but it should be noted that the computing device 200 in this application may also include multiple processors.

[0064] The memory 220 may store data / information obtained from any other components of the magnetic resonance image reconstruction system 100. In some embodiments, the memory 220 may include mass storage, removable storage, volatile read and write memory, ROM, etc., or any combination thereof.

[0065] The input / output 230 can be used to input or output signals, data, or information. In some embodiments, the input / output 230 can allow a user to communicate with the magnetic resonance image reconstruction system 100. In some embodiments, the input / output 230 can include an input device and an output device. The communication port 240 can be connected to a network for data communication. The connection can be a wired connection, a wireless connection, or a combination of the two. The wired connection can include an electrical cable, an optical cable, a telephone line, or any combination thereof. The wireless connection can include Bluetooth TM , Wi-Fi, WiMax, WLAN, ZigBee TM, mobile networks (e.g., 3G, 4G, or 5G, etc.), or any combination thereof. In some embodiments, the communication port 240 may be a standardized port, such as RS232, RS485, etc. In some embodiments, the communication port 240 may be a specially designed port. For example, the communication port 240 may be designed according to the Digital Imaging and Medicine Communications Protocol (DICOM).

[0066] FIG3 is a schematic diagram of modules of an exemplary four-dimensional magnetic resonance image reconstruction apparatus according to some embodiments of the present specification.

[0067] As shown in FIG3 , in some embodiments, a four-dimensional magnetic resonance image reconstruction apparatus 300 may include an acquisition mode determination module 310, a data acquisition module 320, a classification module 330, and a reconstruction module 340. In some embodiments, the corresponding functions of the four-dimensional magnetic resonance image reconstruction apparatus 300 may be executed and implemented by the processing device 130 or the processor 210.

[0068] The acquisition mode determination module 310 can be configured to determine a data acquisition mode for each phase based on the real-time motion signal of the object. In some embodiments, the data acquisition mode may include full sampling or edge sampling during a stationary phase of the real-time motion signal, and center sampling during all other phases. In some embodiments, a signal cycle of the real-time motion signal may include multiple different phases, and at least two of the multiple different phases may have different echo train lengths or phase encoding steps.

[0069] In some embodiments, the acquisition mode determination module 310 may be configured to determine a data acquisition mode for each phase based on the real-time motion signal and the signal analysis model.

[0070] The data acquisition module 320 can be used to collect raw data of the object according to the data acquisition method. In some embodiments, the data acquisition module 320 can mark the motion phase or amplitude corresponding to each part of the raw data during the data acquisition process to obtain marked raw data.

[0071] The classification module 330 can be used to classify the original data to obtain multiple categories of data, each category of data corresponding to a different phase. In some embodiments, the classification module 330 can classify the original data based on the label to obtain multiple categories of data.

[0072] The reconstruction module 340 may be configured to obtain a four-dimensional magnetic resonance image by performing image reconstruction on the data of the multiple categories.

[0073] In some embodiments, the reconstruction module 340 can be used to perform K-space stitching on multiple categories of data to obtain K-space data for each phase; perform image reconstruction based on the K-space data for each phase to obtain a corresponding reconstructed image; and combine the reconstructed images of each phase to obtain a four-dimensional magnetic resonance image.

[0074] In some embodiments, for each undersampled target phase, the reconstruction module 340 may supplement the uncollected data in the target phase based on the sampled data in the motion stationary phase to obtain the full K-space data of the target phase. In some embodiments, for each undersampled target phase, the reconstruction module 340 may supplement the uncollected data in the target phase based on the sampled data of each phase between the target phase and the motion stationary phase to obtain the full K-space data of the target phase.

[0075] For more information about the acquisition mode determination module 310 , the data acquisition module 320 , the classification module 330 and the reconstruction module 340 , please refer to other places in this specification, such as FIG. 4 and its related description.

[0076] It should be understood that the system and its modules shown in Figure 3 can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware.

[0077] It should be noted that the above description of the four-dimensional magnetic resonance image reconstruction apparatus 300 and its modules is for illustrative purposes only and is not intended to limit this specification to the exemplary embodiments described. It is understood that those skilled in the art, after understanding the principles of the system, may arbitrarily combine the modules or construct subsystems connected to other modules without departing from these principles. Such variations are within the scope of this specification.

[0078] FIG4 is a flowchart illustrating an exemplary four-dimensional magnetic resonance image reconstruction method according to some embodiments of the present disclosure. In some embodiments, process 400 can be executed by the magnetic resonance image reconstruction system 100 (e.g., the processing device 130 in the magnetic resonance image reconstruction system 100) or the four-dimensional magnetic resonance image reconstruction apparatus 300. For example, process 400 can be stored in the form of a program or instructions in a storage device (e.g., the storage device 150 or a storage unit of the system). When a processor or the module shown in FIG3 executes the program or instructions, process 400 can be implemented. As shown in FIG4 , in some embodiments, process 400 can include the following steps.

[0079] In step 410 , based on the real-time motion signal of the object, a data collection method for each phase is determined. In some embodiments, step 410 may be performed by the processing device 130 or the collection method determination module 310 .

[0080] A motion signal refers to a signal that reflects motion information such as the breathing or heartbeat of an object, and can represent changes in the breathing amplitude or heartbeat amplitude of the object over a period of time. In some embodiments, the motion signal can be presented in the form of a curve. For example, the respiratory signal can be a sine wave or cosine wave curve. In some embodiments, the trough in the respiratory signal curve can correspond to the end of the exhalation phase, and the peak can correspond to the end of the inspiration phase. In some embodiments, the trough in the respiratory signal curve can correspond to the end of the inspiration phase, and the peak can correspond to the end of the exhalation phase.

[0081] In some embodiments, the real-time motion signal of the object can be acquired by a motion signal acquisition device (e.g., motion signal acquisition device 112). In some embodiments, the acquisition of the real-time motion signal of the object can be performed synchronously with the scanning. For example, while the imaging device 110 is scanning the object, the motion signal acquisition device 112 simultaneously acquires the real-time motion signal of the object.

[0082] Phases refer to different states during exercise. For example, the phases corresponding to a respiratory signal can include early inspiration, late inspiration, early expiration, and late expiration; or early inspiration, mid-inspiration, late inspiration, early expiration, mid-expiration, and late expiration.

[0083] In some embodiments, the processing device 130 can divide at least a portion of the motion signal (for example, one signal period, multiple signal periods, half a signal period, a portion of one or more signal periods, etc.) into multiple phases based on the signal period of the real-time motion signal.

[0084] Taking the respiratory signal as an example, the signal period can refer to the signal interval between two adjacent peaks in the respiratory signal curve, the signal interval between two adjacent troughs, the signal interval containing a pair of adjacent peaks and troughs, the signal interval between two non-adjacent peaks or troughs, and any other signal interval that can reflect the changing period of the respiratory signal.

[0085] It is understandable that for the same patient, the signal period of the motion signal is a time value of constant or substantially constant length. In some embodiments, the signal period of the motion signal is affected by factors such as age, gender, body shape, respiratory muscle strength, lung and thorax elasticity, etc.

[0086] In some embodiments, the processing device 130 may divide at least a portion of the motion signal into multiple periods based on phase. For example, as shown in FIG5(a), taking a signal interval containing a pair of adjacent peaks and troughs as one signal period of the respiratory signal, the processing device 130 may divide one signal period of the respiratory signal into N periods based on phase.

[0087] In some embodiments, processing device 130 may divide at least a portion of the motion signal into multiple segments based on the motion amplitude, and determine multiple phases corresponding to each of the multiple segments. For example, as shown in FIG5(b), where the signal interval between two adjacent peaks is considered a signal cycle of the respiratory signal, processing device 130 may classify the portion between each two dashed lines as the same phase based on the respiratory amplitude, thereby obtaining N phases.

[0088] In some embodiments, multiple segments of signals with the same amplitude change within a signal cycle can be the same phase. For example, as shown in FIG5(b), taking the signal interval between two adjacent peaks as a signal cycle of the respiratory signal as an example, the signals of two time periods that are symmetrical about the midline A within a signal cycle of the curve (i.e., two signal segments that are symmetrical about the midline A between every two adjacent dotted lines) can be two identical phases. In some embodiments, multiple segments of signals with the same amplitude change within a signal cycle can be different phases. For example, as shown in FIG5(a), taking the signal interval containing a pair of adjacent peaks and troughs as a signal cycle of the respiratory signal as an example, the two places with the same amplitude change within a signal cycle of the curve, i.e., two time periods that are symmetrical about the midline at the peak or trough, such as mid-phase 3 and phase 4 in FIG5(a), can be two different phases, such as phase 3 being the end of inspiration and phase 4 being the beginning of expiration.

[0089] In some embodiments, the processing device 130 may divide at least a portion of the real-time motion signal equally, regularly, or randomly. For example, the processing device 130 may evenly divide a signal cycle of the motion signal into N segments, where each of the N segments corresponds to N phases. For another example, the processing device 130 may divide a signal cycle of the motion signal into N phases in a gradually increasing manner according to a certain regularity. For another example, the processing device 130 may randomly divide a signal cycle of the motion signal into N phases.

[0090] In combination with the above, under the synergistic effect of the main magnetic field / gradient magnetic field and RF pulses, magnetic resonance signals related to the object can be generated according to the pulse sequence. An MRI pulse sequence is actually the temporal arrangement of changes in the RF pulses and gradient fields. Therefore, each pulse sequence will have time-related concepts, mainly including repetition time (TR), echo time (TE), effective echo time (effective TE), echo train length (ETL), echo spacing (ES), number of excitations (NEX), acquisition time (TA), etc.

[0091] Repetition time refers to the time interval between two consecutive executions of a pulse sequence. For example, in an SE sequence, TR is the time interval between the midpoints of two adjacent 90° pulses; in a gradient echo sequence, TR is the time interval between the midpoints of two adjacent low-angle pulses; and in inversion recovery and rapid inversion recovery sequences, TR is the time interval between the midpoints of two adjacent 180° inversion prepulses.

[0092] Echo time is the interval from the midpoint of the pulse that generates the macroscopic transverse magnetization vector to the midpoint of the echo. For example, in an SE sequence, TE refers to the interval from the midpoint of the 90° pulse to the midpoint of the spin echo, while in a gradient echo sequence, TE refers to the interval from the midpoint of the small-angle pulse to the midpoint of the gradient echo.

[0093] In fast spin echo (FSE) or echo planar imaging (EPI) sequences, a single radiofrequency pulse generates multiple echoes, each filling different locations in k-space. Each echo has a different TE. In these sequences, the time interval from the midpoint of the radiofrequency pulse to the midpoint of the echo filling the center of k-space is called the effective TE.

[0094] The echo train length refers to the number of echoes generated, acquired, and populated within a single k-space region after a single 90° pulse excitation. The presence of an echo train proportionally reduces the number of TR repetitions. Each echo in the echo train applies a different phase encoding gradient field and is populated at a different location in the phase encoding direction of k-space.

[0095] Echo gap refers to the time gap between the midpoints of two adjacent echoes in an echo chain. The smaller the ES, the less time is required to acquire the entire echo chain.

[0096] The number of excitations, also known as the number of signal averaged (NSA) or the number of acquisitions (NA), refers to the number of repetitions of each phase encoding step in the pulse sequence.

[0097] Acquisition time (TA), also known as scan time, refers to the time required for the entire pulse sequence to complete signal acquisition.

[0098] In combination with the above, the data acquisition method can reflect the above time-related parameter information of the corresponding pulse sequence when data is acquired in each phase.

[0099] In some embodiments, the data collection method may include performing full sampling during a stationary phase of the real-time motion signal and performing center sampling during other phases.

[0100] The motion stability period may refer to a phase in which the motion is stable or substantially stable. Stable motion means that the amplitude change of the motion signal within a certain time period is 0 or less than a preset value, such as the difference between the motion amplitude at the start time and the motion amplitude at the end time within the time period is 0 or the absolute value is less than a first preset value (such as the first preset value is in the range of 0.1 to 1.5), or the absolute value of the difference between the maximum amplitude and the minimum amplitude within the time period is less than a second preset value (such as the second preset value is in the range of 1.5 to 3). Accordingly, other phases refer to phases in which the respiratory amplitude changes greatly.

[0101] FIG6 is a schematic diagram of an exemplary respiratory signal phase according to some embodiments of the present specification, wherein the abscissa represents time, the ordinate represents the amplitude of the respiratory signal, and the curve in the figure represents a portion of the respiratory signal. As shown in FIG6 , if the portion between each two adjacent vertical dashed lines (including dotted and dashed lines) corresponds to a phase, then the phase corresponding to the dashed portion has a relatively small change in respiratory amplitude and belongs to a stable phase, while the phase corresponding to the dotted portion has a relatively large change in respiratory amplitude and belongs to another phase.

[0102] In some embodiments, a signal cycle may include one or more stationary periods, and each stationary period may include one or more adjacent phases. For example, the processing device 130 may determine a portion of the real-time respiratory signal with relatively small amplitude changes (such as the dashed portion in FIG6 ) as a stationary period; or divide the portion with relatively small amplitude changes into two phases (such as the end-expiration and the beginning-inspiration phase, or the end-inspiration and the beginning-expiration phase) based on the expiratory phase and the inspiratory phase, both of which are stationary periods.

[0103] Full sampling means that the acquired data can completely fill the K-space. Center sampling means that the acquired data can fill the central area of ​​the K-space, and no data is collected in other areas.

[0104] The central area of ​​the K space refers to the area that covers the center line of the K space parallel to the horizontal axis and does not reach the periphery of the K space, and may include the area above and / or below the center line. In some embodiments, after the data obtained by the central sampling is filled into the central area of ​​the K space, the filled area may account for 30% to 60% of the entire area of ​​the K space. In some embodiments, the central sampling data corresponding to different phases may have the same or different proportions of the filled area in the K space to the entire area of ​​the K space. For example, after the central sampling data corresponding to phase 1 is filled into the central area of ​​the K space, a filled area as shown in Figure 13 can be obtained, and after the central sampling data corresponding to phase 2 is filled into the central area of ​​the K space, a filled area as shown in Figure 14(a) can be obtained.

[0105] Figures 7(a) and 7(b) are schematic diagrams of K-space filling according to exemplary Cartesian acquisitions shown in some embodiments of the present specification, wherein the ordinate (Ky) represents the phase encoding PE direction, the abscissa represents the echo time TE, and the lines with arrows represent the acquired data. Figure 7(a) shows the K-space corresponding to full sampling, and Figure 7(b) shows the K-space corresponding to central sampling.

[0106] It should be noted that Figures 7(a) and 7(b) are only examples. In some embodiments, the number of acquisition lines for center sampling can be adjusted according to actual needs, such as 2, 3, 4, etc. including the center. This specification does not impose specific restrictions on this.

[0107] In some embodiments, the data acquisition method may include edge sampling during the stationary period of the real-time motion signal and center sampling during other phases. Edge sampling means that the collected data fills the edge portion of the K space, and the edge portion corresponds to other areas outside the center area of ​​the K space (such as the blank area in Figure 7(b)). In some embodiments, after the data obtained by edge sampling is filled into the K space, from the periphery of the K space to the center line parallel to the horizontal coordinate, it can occupy 20% to 35% of the entire area of ​​the K space. In some embodiments, the data obtained by edge sampling can be filled into the upper and lower edges of the K space respectively. In some embodiments, the data filled into the upper and lower edges of the K space can occupy 40% to 70% of the entire area of ​​the K space. For example, after the data obtained by edge sampling is filled into the K space, the data at the upper edge of the K space accounts for 20% of the entire area of ​​the K space, and the data at the lower edge of the K space accounts for 20% of the entire area of ​​the K space.

[0108] In some embodiments, the real-time motion signal may include multiple different phases, and at least two of the multiple different phases may correspond to different echo train lengths. For example, the early exhalation phase and the end exhalation phase may correspond to different echo train lengths.

[0109] In some embodiments, the echo train length of each phase is related to the length of the initial echo train. In some embodiments, the echo train lengths corresponding to the at least two phases may each account for a different proportion of the length of the initial echo train. For example, if the initial echo train in a determined pulse sequence includes seven echo signals, the echo train corresponding to the early expiration phase may be the first three echo signals in the initial echo train, and the echo train corresponding to the late expiration phase may be the first five or all echo signals in the initial echo train.

[0110] In some embodiments, the echo train length during a stationary phase of movement can be greater than the echo train lengths during other phases. For example, the echo train corresponding to mid-expiration may contain three echo signals, and the echo train corresponding to end-expiration may contain five echo signals. Generally, near the peaks or troughs of the signal curve, the amplitude of the respiratory signal varies little, indicating stationary breathing. The peaks or troughs of the signal curve may correspond to end-expiration or end-inspiration.

[0111] In some embodiments, when the echo train length of the motion stationary phase is greater than the echo train lengths of other phases, the echo train lengths of each of the other phases may be the same. For example, the echo train of the respiratory stationary phase may include four echo signals, while the echo trains of the other phases may each include two echo signals.

[0112] In some embodiments, the echo train lengths of each of the multiple different phases can be different, and the echo train length of the motion stationary phase is the largest. In some embodiments, the echo train lengths of the multiple phases before or after the motion stationary phase can gradually increase or decrease according to the temporal sequence of the motion signal, and the echo train length of the motion stationary phase is the largest. For example, in conjunction with FIG6 , the echo train lengths of the five phases before the short dash can gradually increase, that is, the echo train length of phase 5> the echo train length of phase 4> the echo train length of phase 3> the echo train length of phase 2> the echo train length of phase 1, and the echo train length of the long dashed dotted line portion is greater than the echo train length of phase 5. In some embodiments, the echo train lengths of the multiple phases before or after the motion stationary phase can be randomly distributed. For example, in conjunction with FIG6 , the echo train length of phase 1 is 2, the echo train length of phase 2 is 5, the echo train length of phase 3 is 4, the echo train length of phase 4 is 3, and so on.

[0113] In combination with the above, one signal cycle of the real-time motion signal may contain two or more identical phases, such as two parts or more signals with the same amplitude change. In some embodiments, when one signal cycle of the real-time motion signal contains the same phase, the echo chain lengths corresponding to the two identical phases in different time periods may be the same or different. Exemplarily, as shown in FIG6 , assuming that phase 5 and phase M on both sides of the dashed line belong to the same phase, such as both are mid-respiration, the echo chains corresponding to phase 5 and phase M may both contain 3 echo signals; or the echo chain of phase 5 contains 4 echo signals, and the echo chain of phase M contains 3 echo signals.

[0114] In some embodiments, the phase encoding step corresponding to at least two phases in a signal cycle of a real-time motion signal is different. The phase encoding step refers to how many phase encoding gradients of different current intensities and durations are executed to complete the K-space data filling. In some embodiments, the phase encoding step of the motion stationary period can be greater than the phase encoding step of other phases, and the phase encoding steps of other phases can be the same. In some embodiments, the phase encoding step of each phase in a plurality of different phases can be different, and the phase encoding step of the motion stationary period is the largest.

[0115] For example, as shown in Figure 6, if the vertical dotted lines (including dotted and dashed lines) in the figure represent phase encoding steps, within the same time length, the distribution of the dashed lines is denser, and the distribution of the dotted lines is looser. That is, the dashed line portion corresponds to the phase encoding step when the respiratory amplitude variation is low (i.e., steady breathing), that is, the phase encoding step at this stage is larger, and more data at the edge of the K space can be collected, thereby obtaining more image details and increasing image information redundancy; the dotted line portion corresponds to the phase encoding step when the respiratory amplitude variation is large, and the phase encoding step at this stage is smaller, which can quickly collect data in the center area of ​​the K space, thereby obtaining higher image contrast.

[0116] The determination of the phase encoding step corresponding to each phase can be similar to the determination of the echo train length. For more details, see above.

[0117] In some embodiments, the processing device 130 can determine the data acquisition method for each phase based on the real-time motion signal and the signal analysis model. The input of the signal analysis model is the real-time motion signal, and the output is the data acquisition method. For example, the processing device 130 can input the real-time motion signal of the collected object into a trained signal analysis model. After analysis and processing, the signal analysis model outputs the number of phases contained in the real-time motion signal, the time information and / or amplitude information of each phase, and the echo train length (and / or phase encoding step) corresponding to each phase.

[0118] In some embodiments, the signal analysis model can be obtained based on training of pre-collected sample motion signals. The training labels include the sample period phase of the sample motion signal and / or the sample echo chain length corresponding to each sample period (or the sample phase encoding step). For example, the sample motion signal can be used as a training input, and the initial model is input. The sample period phase of the sample motion signal and the sample echo chain length corresponding to each sample period are used as training labels, and the trained signal analysis model is obtained by iteratively training the initial model. In some embodiments, during the training process, the model can be optimized by constructing a loss function. For example, a loss function is constructed based on the prediction results of the initial model and the training labels. When the loss function value converges to the target value, the training ends and a trained signal analysis model is obtained.

[0119] In some embodiments, the signal analysis model may include a neural network. For example, the neural network may include a convolutional neural network (CNN), a U-Net, a recurrent neural network (RNN), a graph neural network (GNN), etc.

[0120] In step 420 , raw data of the object is collected according to the data collection method. In some embodiments, step 420 may be performed by the processing device 130 or the data collection module 320 .

[0121] In some embodiments, the processing device 130 may collect data according to a determined echo train length (or phase encoding step) in each phase of the real-time motion signal to obtain raw data of the object.

[0122] In some embodiments, the processing device 130 can synchronously mark the motion phase or amplitude corresponding to each portion of the raw data during the data acquisition process to obtain labels for the raw data. For example, when the processing device 130 acquires data at each respiratory phase according to a determined echo train length (or phase encoding step), it can synchronously mark the respiratory signal amplitude or phase corresponding to the currently acquired data.

[0123] In some embodiments, the processing device 130 may mark the collected raw data after completing the collection of all raw data of the object. In some embodiments, the processing device 130 may mark the collected portion after completing the collection of part of the raw data of the object, and then collect other parts of the data.

[0124] In some embodiments, the processing device 130 can use a high-speed sequence to acquire the original data of the object. For example, the high-speed sequence can include a GRE sequence, an EPI sequence, etc. Data acquisition based on the EPI sequence can also be called a Zig-Zag acquisition mode. Figures 8(a) and 8(b) are schematic diagrams of K-space filling according to exemplary Zig-Zag acquisitions shown in some embodiments of this specification. For example, as shown in Figures 8(a) and 8(b), the lines in the shape of a Chinese character with an arrow in the figure represent acquisition lines, and the rectangular box represents the entire K space. The acquisition line shown in Figure 8(a) can correspond to the data that needs to be acquired during the stable period of motion, and the acquisition line shown in Figure 8(b) can correspond to the amount of data that needs to be acquired during other phases (such as phases with large changes in motion amplitude). In some embodiments, the processing device 130 can use other sequences to acquire the original data of the object, and this specification does not impose specific restrictions on this.

[0125] In some embodiments, the data acquisition method of the object's raw data may include at least one of the following: Cartesian acquisition, Zig-Zag acquisition, radial acquisition, spiral acquisition, or windmill acquisition. The data acquisition method described here is for different objects, and different objects can use the same or different data acquisition methods, but the same acquisition method is generally used throughout the data acquisition process for the same object. For example, the same object can use a Cartesian acquisition method to sample raw data for each phase according to a determined echo train length and / or phase encoding step. In some embodiments, different acquisition methods such as Cartesian, radial, spiral, or windmill can be achieved by changing the encoding.

[0126] It is understood that the data filling method used during K-space filling corresponds to the acquisition method used during raw data sampling. For example, when a Cartesian method is used to sample the center data or edge data of each phase, a Cartesian method will be used to perform K-space filling on the raw data of each phase accordingly.

[0127] Figures 7(a) and 7(b) are schematic diagrams of K-space filling for exemplary Cartesian acquisition according to some embodiments of the present specification. The lines with arrows represent acquired data, and the rectangular box represents the entire K-space. The acquisition line shown in Figure 7(a) corresponds to the data that needs to be acquired when the motion is in a stable state, and the acquisition line shown in Figure 7(b) corresponds to the amount of data that needs to be acquired when the motion amplitude changes greatly.

[0128] Figures 8(a) and 8(b) are schematic diagrams of K-space filling according to exemplary Zig-Zag acquisition shown in some embodiments of the present specification. In the figures, the U-shaped lines with arrows represent acquisition lines, and the rectangular boxes represent the entire K-space. The acquisition lines shown in Figure 8(a) may correspond to data that needs to be acquired when the motion is in a stable state, and the acquisition lines shown in Figure 8(b) may correspond to data that needs to be acquired when the motion amplitude changes greatly.

[0129] Figures 9(a) and 9(b) are schematic diagrams of K-space filling for exemplary radial acquisition according to some embodiments of the present specification, wherein the ordinate (Ky) represents the phase encoding PE direction, the abscissa represents the echo time TE, radial lines with arrows represent acquired data, and dotted circles represent the entire K-space. The acquisition lines shown in Figure 9(a) correspond to the data that needs to be acquired when the motion is in a stable state, and the acquisition lines shown in Figure 9(b) correspond to the amount of data that needs to be acquired when the motion amplitude changes greatly.

[0130] Figures 10(a) and 10(b) are schematic diagrams of K-space filling according to exemplary spiral acquisition shown in some embodiments of the present specification, wherein the ordinate (Ky) represents the phase encoding PE direction, the abscissa represents the echo time TE, the spiral lines represent the acquired data, and the rectangular box represents the entire K-space. The acquisition line shown in Figure 10(a) corresponds to the data that needs to be acquired when the motion is in a stable state, and the acquisition line shown in Figure 10(b) corresponds to the amount of data that needs to be acquired when the motion amplitude changes greatly.

[0131] Figures 11(a) and 11(b) are schematic diagrams of K-space filling of exemplary windmill acquisition according to some embodiments of the present specification, wherein the ordinate (Ky) represents the phase encoding PE direction, the abscissa represents the echo time TE, the rectangular solid line frame represents the entire K-space, the lines with arrows represent acquisition lines, and every three acquisition lines form a blade of the windmill (as shown in the dotted box in the figure), each blade corresponds to one acquisition, and multiple blade layers are star-shaped. The acquisition lines shown in Figure 11(a) correspond to the data that needs to be acquired when the motion is in a stable state, and the acquisition lines shown in Figure 11(b) correspond to the amount of data that needs to be acquired when the motion amplitude changes greatly.

[0132] In some embodiments, the data acquisition method of the raw data of the object includes but is not limited to 2D acquisition and 3D acquisition, for example, 2D Cartesian acquisition, 3D Cartesian acquisition, 2D radial acquisition, 3D radial acquisition, 2D spiral acquisition, 3D spiral acquisition, 2D pinwheel acquisition, 3D pinwheel acquisition, etc.

[0133] Figures 12(a) and 12(b) are schematic diagrams of K-space filling for exemplary radial 3D acquisition according to some embodiments of the present specification. The corresponding K-space is also 3-dimensional, wherein radial lines with arrows represent acquisition data, and dashed circles represent the entire K-space. The acquisition lines shown in Figure 12(a) correspond to the data that needs to be acquired when the motion is stable, and the acquisition lines shown in Figure 12(b) correspond to the amount of data that needs to be acquired when the motion amplitude changes greatly.

[0134] Step 430 , classifying the original data to obtain data of multiple categories. In some embodiments, step 430 may be performed by the processing device 130 or the classification module 330 .

[0135] In the multiple categories obtained by classification, each category of data corresponds to a different phase. In some embodiments, the processing device 130 can classify the motion amplitudes or phases of the markers into the same category, thereby obtaining multiple categories of data.

[0136] In some embodiments, the processing device 130 may classify the raw data according to the data acquisition method. For example, when the echo train lengths corresponding to each period are different, the processing device 130 may classify the raw data acquired using the same echo train length into one category.

[0137] In some embodiments, the processing device 130 may simultaneously perform data classification during the data collection process. For example, while collecting the raw data of the object according to the determined data collection method, the processing device 130 may separate and store the data collected at different phases into bins to obtain raw data of multiple categories.

[0138] In step 440 , a four-dimensional magnetic resonance image is obtained by performing image reconstruction on the data of the multiple categories. In some embodiments, step 440 may be performed by the processing device 130 or the reconstruction module 340 .

[0139] In some embodiments, the processing device 130 can perform K-space stitching on multiple categories of data to obtain K-space data for each phase. Furthermore, for each phase, image reconstruction is performed based on the K-space data of that phase to obtain a corresponding reconstructed image; the reconstructed images of each phase are combined to obtain a four-dimensional magnetic resonance image. In some embodiments, K-space can include, but is not limited to, a 2D K-space (as shown in Figures 7(a)-11(b)) and a 3D K-space (as shown in Figures 12(a) and 12(b)).

[0140] K-space stitching is to fill the K-space of the acquired data, such as obtaining the K-space data shown in Figures 7(a) to 12(b), or part of the K-space data shown in Figures 7(a) to 12(b).

[0141] In some embodiments, for each of the multiple categories obtained, the processing device 130 may fill the data corresponding to the category into the K-space, thereby obtaining the K-space data for the phase corresponding to the category. In some embodiments, the processing device 130 may obtain the full K-space data for each phase by sharing the raw data corresponding to each category.

[0142] In some embodiments, for each undersampled target phase, the processing device 130 may supplement the uncollected data in the target phase based on the sampled data in the motion stationary period to obtain full K-space data of the target phase.

[0143] Undersampled target phases are phases where the acquired data cannot completely fill the K-space. For example, phases with large motion amplitude changes may use a shorter echo train length or phase encoding step size for data acquisition, resulting in a smaller number of sampled data points. This may result in insufficient K-space filling, such as only enough to fill the central region of the K-space.

[0144] In some embodiments, when full sampling is performed during the stationary motion phase, the processing device 130 may supplement the uncollected data in the target phase based on the sampled data during the stationary motion phase to obtain full K-space data for the target phase. In some embodiments, the processing device 130 may determine target data not collected in the target phase, obtain the target data from the collected data during the stationary motion phase, and perform K-space filling based on the sampled data of the target phase and the target data to obtain full K-space data for the target phase.

[0145] Figures 13-14(b) are schematic diagrams of exemplary undersampling K-space filling according to some embodiments of the present specification, and Figure 15 is a schematic diagram of exemplary full-sampling K-space filling according to some embodiments of the present specification. The solid-line box in the figures represents the size of the complete K-space. For example, taking a 2D Cartesian acquisition method as an example, in conjunction with Figures 13 and 15 , when full sampling is performed during a stationary phase, K-space filling based on the acquired raw data (also referred to as sampled data) can produce complete K-space data similar to that shown in Figure 15 . However, for phases with large variations in motion amplitude (i.e., target phases), due to undersampling, if K-space filling is performed directly using the acquired raw data, K-space data similar to that shown in Figures 13 or 14(a) or 14(b) can be obtained. In this case, the processing device 130 can obtain data not acquired in the target phase from the sampled data during the stationary phase, and fill the K-space with the acquired data and the sampled data of the target phase, thereby obtaining complete K-space data for the target phase (e.g., K-space data similar to that shown in Figure 15 ). For example, taking the undersampled target phase shown in Figure 13 as an example, the processing device 130 can obtain the data corresponding to the P1, P2, and P3 rows in Figure 15, and fill the data of the P1, P2, and P3 rows into the dotted lines L1, L2, and L3 shown in Figure 13, respectively, to obtain the full K-space data of the target phase.

[0146] In some embodiments, when edge sampling is performed during a stationary motion period, the processing device 130 may supplement the uncollected data in the target phase based on the edge sampling data during the stationary motion period to obtain the full K-space data for the target phase. Accordingly, the processing device 130 may supplement the uncollected data (e.g., data in the center region of the K-space) during the stationary motion period based on the sampling data of other phases to obtain the full K-space data for the stationary motion period. In some embodiments, the processing device 130 may supplement the uncollected data during the stationary motion period based on the sampling data of one or more other phases adjacent to the stationary motion period.

[0147] The data in the center of the K-space determines the contrast of the reconstructed image, while the data at the edge of the K-space determines the details of the reconstructed image. Full sampling or edge sampling during the stationary phase of motion acquires more high-dimensional data (i.e., image detail information), while the target phase (i.e., the phase with a large change in motion amplitude) during center sampling acquires central region data (i.e., image contrast information). By sharing the raw data collected during the stationary phase of motion with the target phase, scanning efficiency can be improved while the target phase simultaneously possesses both image detail information and central contrast information, resulting in higher image quality for the reconstructed image of each phase.

[0148] In some embodiments, for each undersampled target phase, the processing device 130 may supplement the uncollected data of the target phase based on the motion stationary phase and one or more candidate phases having more sampled data than the target phase, to obtain the full K-space data of the target phase. For example, the processing device 130 may first supplement the sampled data of the phase having more sampled data than the target phase to the target phase, and then obtain the remaining uncollected data of the target phase from the motion stationary phase to supplement the target phase, thereby obtaining the complete data of the target phase.

[0149] In some embodiments, for each undersampled target phase, the processing device 130 can supplement the uncollected data of the target phase based on the sampling data of each phase between the target phase and the motion stationary period to obtain the full K-space data of the target phase.

[0150] In some embodiments, when the echo chain lengths (or phase encoding steps) corresponding to multiple phases before or after the motion steady period gradually increase or decrease over time, the processing device 130 can supplement the uncollected data of the target phase based on the sampling data of each phase between the target phase and the motion steady period.

[0151] Specifically, the processing device 130 can supplement the data of the first part of the target phase in the K space based on the sampling data of the first phase adjacent to the target phase; supplement the data of the second part of the target phase in the K space based on the sampling data of the second phase adjacent to the first phase; repeat the above steps until the data of the target phase in the K space is supplemented based on the sampling data of the stable period of motion.

[0152] For example, using a 2D Cartesian acquisition method as an example, in conjunction with Figures 13, 14(a)-14(b), and 15, it is assumed that the small black dots shown in Figure 13 are sampled data corresponding to the current target phase, the small black dots shown in Figure 14(a) are sampled data corresponding to the first phase adjacent to the target phase, the small black dots shown in Figure 14(b) are sampled data corresponding to the second phase adjacent to the first phase, the first phase and the second phase being the phases between the target phase and the phase of stable breathing, and the small black dots shown in Figure 15 are sampled data corresponding to the stable phase of motion, which is adjacent to the second phase. The small black dots in the figure represent sampled data. For the target phase, the processing device 130 may fill the data corresponding to row P1 in Figure 14(a) with the row corresponding to the dotted line L1 in Figure 13, fill the data corresponding to row P2 in Figure 14(b) with the row corresponding to the dotted line L2 in Figure 13, and fill the data corresponding to row P3 in Figure 15 with the row corresponding to the dotted line L3 in Figure 13, thereby obtaining the full K-space data for the target phase.

[0153] It can be understood that the above description and the undersampled K-space data in Figures 13-14(b) are only examples. In some embodiments, the undersampled K-space data may be larger or smaller than the amount of data shown in the figure, such as including 2 rows in Figure 13, or 1 more row of data based on the data shown in Figure 13; in some embodiments, the target phase may be the phase corresponding to the K-space data shown in Figure 14(a) or Figure 14(b), and this specification does not impose any restrictions on this.

[0154] Since the echo chain lengths (or phase encoding steps) corresponding to the multiple phases before or after the motion steady period gradually increase (or decrease) over time, the sampling data of each phase between the target phase and the motion steady period gradually increases (or decreases), and the uncollected data of the target phase are gradually supplemented from the first phase adjacent to the target phase to the motion steady period, the authenticity of the K-space data of the target phase can be guaranteed.

[0155] In some embodiments, when there are two adjacent first phases of the target phase (e.g., adjacent phases on the left and right sides of the target phase), and the sampled data of the two adjacent first phases is greater than the sampled data of the target phase, the processing device 130 may supplement the data of the first portion of the target phase in K space based on the sampled data of the two adjacent first phases. For example, in conjunction with FIG13 and FIG14(a), the processing device 130 may average the data of row P1 of the first phase adjacent to the left of the target phase with the data of row P1 of the first phase adjacent to the right of the target phase, and fill the averaged data into row L1 of the dotted line shown in FIG13.

[0156] After obtaining the full K-space data for each phase in the above manner, processing device 130 can further perform image reconstruction for each phase based on the full K-space data for that phase to obtain a corresponding reconstructed image. In some embodiments, image reconstruction algorithms may include, but are not limited to, Fourier transform, back-projection reconstruction, iterative reconstruction, and fitting approximation, and are not specifically limited in this specification.

[0157] By combining the reconstructed images of each phase, a four-dimensional magnetic resonance image with time information can be obtained. The four-dimensional magnetic resonance image can be used for lesion diagnosis, scientific research, radiotherapy planning, etc., which is not limited in this specification.

[0158] It should be noted that the above description of process 400 is for illustration and purpose only and does not limit the scope of application of this specification. Those skilled in the art may make various modifications and variations to process 400 under the guidance of this specification. However, such modifications and variations are still within the scope of this specification.

[0159] Some embodiments of the present specification also provide a four-dimensional magnetic resonance image reconstruction method, comprising: determining full K-space data for each phase based on acquired raw data; performing image reconstruction for each phase based on the full K-space data to obtain a corresponding reconstructed image; and combining the reconstructed images of each phase to obtain a four-dimensional magnetic resonance image. For each undersampled target phase, unacquired data in the target phase is supplemented based on sampled data from a stationary phase of motion and / or sampled data from one or more other candidate phases having more sampled data than the target phase, to obtain the full K-space data for the target phase.

[0160] Some embodiments of the present disclosure further provide a four-dimensional magnetic resonance image reconstruction system, comprising at least one storage medium including a set of instructions, and one or more processors in communication with the at least one storage medium. When executing the instructions, the one or more processors are configured to perform the four-dimensional magnetic resonance image reconstruction method described above (e.g., process 400).

[0161] Some embodiments of this specification also provide a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the four-dimensional magnetic resonance image reconstruction method as described above (such as process 400).

[0162] The beneficial effects that may be brought about by the embodiments of this specification include but are not limited to: (1) performing full sampling or edge sampling when the motion amplitude changes slightly can obtain high-dimensional data reflecting image details while reducing the impact of motion artifacts; (2) performing center sampling when the motion amplitude changes significantly can improve scanning efficiency while ensuring image contrast information; (3) sharing sampling data of different phases so that each phase has data with high contrast and image details, thereby improving the quality of the reconstructed image of each phase; (4) marking during the data acquisition process, and dividing the acquired original data into multiple categories based on the marking, so that the data of each category carries phase information, reducing the impact of motion artifacts.

[0163] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed disclosure is merely illustrative and does not limit this specification. Although not explicitly stated herein, various modifications, improvements, and revisions to this specification may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in this specification and remain within the spirit and scope of the exemplary embodiments of this specification.

[0164] This specification also uses specific terms to describe the embodiments of this specification. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different locations in this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of this specification may be appropriately combined.

[0165] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0166] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0167] In some embodiments, numbers are used to describe the quantity of components and attributes. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise stated, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the description and claims are approximate values, which may change according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the settings of such numerical values ​​are as accurate as possible within the feasible range.

[0168] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, and documents, cited in this specification is hereby incorporated by reference in its entirety. This includes application history documents that are inconsistent with or conflict with the content of this specification, as well as documents (currently or subsequently attached to this specification) that limit the broadest scope of the claims of this specification. It should be noted that if the descriptions, definitions, and / or terminology used in the accompanying materials are inconsistent or conflicting with the content of this specification, the descriptions, definitions, and / or terminology used in this specification will control.

[0169] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A four-dimensional magnetic resonance image reconstruction method, characterized in that, Comprising: Determine the data acquisition method for each phase based on the real-time motion signal of the object; Acquire the raw data of the object according to the data acquisition method; Classify the raw data to obtain data of multiple categories, and each category of data corresponds to a different phase; Obtain a four-dimensional magnetic resonance image by performing image reconstruction on the data of the multiple categories.

2. The method according to claim 1, characterized in that The data acquisition method includes full sampling during the motion stable phase of the real-time motion signal and central sampling during other phases.

3. The method according to claim 1, wherein The data acquisition method includes edge sampling during the motion stable phase of the real-time motion signal and central sampling during other phases.

4. The method according to claim 2 or 3, characterized in that, A signal cycle of the real-time motion signal includes multiple different phases, and at least two of the multiple different phases have different echo train lengths corresponding thereto.

5. The method according to claim 2 or 3, characterized in that, A signal cycle of the real-time motion signal includes multiple different phases, and at least two of the multiple different phases have different phase encoding steps corresponding thereto.

6. The method according to claim 4 or 5, characterized in that, The echo train length in the motion stable phase is greater than the echo train length in the other phases; or The phase encoding step in the motion stable phase is greater than the phase encoding step in the other phases.

7. The method according to claim 4 or 5, characterized in that, The echo train lengths of each phase in the multiple different phases are all different, and the echo train length in the motion stable phase is the largest; or The phase encoding steps of each phase in the multiple different phases are all different, and the phase encoding step in the motion stable phase is the largest.

8. The method according to any one of claims 4 to 7, characterized in that For two identical phases located in different time periods within one signal cycle, the corresponding echo train lengths or phase encoding steps are different.

9. The method according to any one of claims 2-8, characterized in that The determining the data acquisition method for each phase based on the real-time motion signal of the object includes: Determine the data acquisition method for each phase based on the real-time motion signal and a signal analysis model, and the signal analysis model includes a trained machine learning model.

10. The method according to claim 9, characterized in that, The signal analysis model is trained based on pre-acquired sample motion signals, and the training labels include the sample phases of the sample motion signals and / or the sample echo train lengths or sample phase encoding steps corresponding to each sample phase.

11. The method according to any one of claims 2-10, characterized in that, The motion stable phase includes a phase with stable or substantially stable motion, and the stable or substantially stable motion means that the amplitude change of the motion signal is 0 or less than a preset value within a certain time period.

12. The method according to any one of claims 1-11, characterized in that, The data acquisition method includes at least one of the following: Cartesian acquisition, Zig-Zag acquisition, radial acquisition, spiral acquisition, or windmill acquisition.

13. The method according to any one of claims 1 to 12, characterized in that, The acquiring the raw data of the object includes: During the data acquisition process, mark the motion phase or amplitude corresponding to each part of the raw data to obtain the mark of the raw data.

14. The method according to any one of claims 1-13, characterized in that, The obtaining a four-dimensional magnetic resonance image by performing image reconstruction on the data of the multiple categories includes: Perform K-space stitching on the data of the multiple categories to obtain the K-space data of each phase; For each phase, perform image reconstruction based on the K-space data to obtain the corresponding reconstructed image; and Combine the reconstructed images of each phase to obtain the four-dimensional magnetic resonance image.

15. The method according to claim 14, characterized in that, The performing K-space stitching on the data of the multiple categories further includes: For each under-sampled target phase, Based on the sampling data in the motion stationary phase, supplement the data that has not been collected in the target phase to obtain the full K-space data of the target phase.

16. The method according to claim 14, wherein The further K-space stitching of the data of the multiple categories includes: For each under-sampled target phase, Based on the sampling data of each phase between the target phase and the motion stationary phase, supplement the data that has not been collected in the target phase respectively to obtain the full K-space data of the target phase.

17. The method according to claim 16, characterized in that, The supplementing of the data that has not been collected in the target phase respectively based on the sampling data of each phase between the target phase and the motion stationary phase includes: Based on the sampling data of the first phase adjacent to the target phase, supplement the data of the first part of the target phase in the K-space; Based on the sampling data of the second phase adjacent to the first phase, supplement the data of the second part of the target phase in the K-space; Repeat the above steps until the data of the target phase in the K-space is completely supplemented based on the sampling data of the motion stationary phase.

18. A four-dimensional magnetic resonance image reconstruction method, characterized in that, including: Based on the collected raw data, determine the full K-space data of each phase; For each phase, perform image reconstruction based on the full K-space data to obtain the corresponding reconstructed image; and Combine the reconstructed images of each phase to obtain a four-dimensional magnetic resonance image; wherein, determining the full K-space data of each phase includes: for each under-sampled target phase, based on the sampling data of the motion stationary phase and / or the sampling data of the candidate phase with more sampling data than the target phase, supplement the data that has not been collected in the target phase to obtain the full K-space data of the target phase.

19. A four-dimensional magnetic resonance image reconstruction system, including: At least one storage medium, including a set of instructions; and One or more processors communicatively coupled to the at least one storage medium, wherein when the instructions are executed, the one or more processors are configured to: Based on the real-time motion signal of the object, determine the data acquisition method for each phase; According to the data acquisition method, acquire the raw data of the object; Classify the raw data to obtain data of multiple categories, and each category of data corresponds to a different phase respectively; Obtain a four-dimensional magnetic resonance image by performing image reconstruction on the data of the multiple categories.

20. A computer-readable storage medium, the storage medium stores computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the following method: Based on the real-time motion signal of the object, determine the data acquisition method for each phase; According to the data acquisition method, acquire the raw data of the object; Classify the raw data to obtain data of multiple categories, and each category of data corresponds to a different phase respectively; Obtain a four-dimensional magnetic resonance image by performing image reconstruction on the data of the multiple categories.

Images