System and method for detecting and scoring motion artifacts in magnetic resonance imaging

Abstract

A method for detecting and scoring motion artifacts in magnetic resonance (MR) images of a subject is provided. The method includes calculating a k-space difference map based, at least in part, on first MR signals of the subject acquired using a first coil and second MR signals of the subject acquired using a second coil. The method also includes generating a difference graph based on the k-space difference map, the difference graph including a curve. The method further includes calculating a motion score based on the curve in the difference graph, wherein the motion score indicates a level of motion artifact in the image caused by motion of the subject during acquisition of the first MR signal and the second MR signal, and the motion score includes an area under the curve. The method also includes outputting the motion score.

Description

Background Art

[0001] The field of the present disclosure relates generally to systems and methods for detecting and scoring motion artifacts, and more particularly to systems and methods for automatically detecting and scoring motion artifacts in magnetic resonance (MR) images.

[0002] Magnetic resonance imaging (MRI) has proven useful in the diagnosis of many diseases. MRI provides detailed images of soft tissue, abnormal tissue (such as tumors), and other structures that cannot be easily imaged by other imaging modalities such as computed tomography (CT). In addition, MRI operates without exposing the patient to the ionizing radiation experienced in modalities such as CT and X-rays.

[0003] Patient motion is one of the biggest causes of inefficiency in clinical MRI, often requiring patients to be rescanned or even undergo a second visit. Specifically, patient motion can cause blurring, artifacts, and other inconsistencies in MR images. Known methods for detecting motion require either additional hardware for monitoring motion, which increases cost and patient setup time, or require a navigator sequence, which takes time away from the imaging sequence. Summary of the Invention

[0004] In one aspect, a method for detecting and scoring motion artifacts in magnetic resonance (MR) images of a subject is provided. The method includes calculating a k-space difference map based at least in part on first MR signals of the subject acquired using a first coil and second MR signals of the subject acquired simultaneously using a second coil. The method also includes generating a difference graph based on the k-space difference map, the difference graph including a curve. The method further includes calculating a motion score based on the curve in the difference graph, wherein the motion score indicates a level of motion artifact in the image caused by motion of the subject during acquisition of the first MR signal and the second MR signal, and the motion score includes an area under the curve. Furthermore, the method includes outputting the motion score.

[0005] In another aspect, a motion detection and scoring calculation device is provided, comprising at least one processor in communication with at least one memory device. The at least one processor is programmed to calculate a k-space difference map based, at least in part, on first MR signals of a subject acquired using a first coil and second MR signals of the subject acquired simultaneously using a second coil. The at least one processor is further programmed to generate a difference curve graph based on the k-space difference map, the difference curve graph comprising a curve. The at least one processor is further programmed to calculate a motion score based on the curve in the difference curve graph, wherein the motion score indicates a level of motion artifact in an image caused by motion of the subject during acquisition of the first MR signal and the second MR signal, and the motion score comprises an area under the curve. Furthermore, the at least one processor is programmed to output the motion score. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 is a schematic diagram of an exemplary magnetic resonance imaging (MRI) system.

[0007] Figure 2A is a flow chart of an exemplary method for detecting and scoring motion artifacts in magnetic resonance (MR) images.

[0008] Figure 2B is an MR image with motion artifacts.

[0009] Figure 3A is a pulse sequence diagram of an exemplary fast spin echo sequence.

[0010] Figure 3B It is an application Figure 3A An exemplary scan order for a turbo spin echo sequence is shown.

[0011] Figure 4A Figure 2 is the composite k-space data from two different coils in the receiver array, where the magnitude of the composite k-space data is shown.

[0012] Figure 4B It is used for collection Figure 4A The composite coil sensitivity map of the coil for the k-space data is shown, where the magnitude of the composite coil sensitivity map is displayed.

[0013] Figure 4C is based on Figure 4A The k-space data shown and Figure 4B An intensity-corrected composite image of the coil sensitivity map is shown, where the magnitude of the intensity-corrected composite image is displayed.

[0014] Figure 4D By subtracting Figure 4CThe intensity-corrected composite image is shown, the result is Fourier transformed, and then the magnitude is taken to generate the k-space difference map.

[0015] Figure 4E is based on Figure 4D The projected difference curve plot of the k-space difference map is shown.

[0016] Figure 4F is based on Figure 4E The difference plots shown are the average difference plots.

[0017] Figure 4G yes Figure 4F Normalized difference plots of the difference plots shown.

[0018] Figure 4H is based on Figure 4G The summed difference plot of the difference plots shown.

[0019] Figure 4I It is from Figure 4H The difference curve graph shown is derived from the difference curve graph.

[0020] Figure 5A is a graph showing the calculated motion scores and user-rated motion scores for images in a development dataset.

[0021] Figure 5B is a graph showing the calculated motion scores and user-rated motion scores for images in a test dataset.

[0022] Figure 5C It shows Figure 5A The plots shown are for the development dataset and Figure 5B The graph shown is a graph of the calculated motion scores and user-rated motion scores for the combined images of the test dataset.

[0023] Figure 6A This is an MR image with a discrete motion fraction of 1.3.

[0024] Figure 6B is an MR image with a discrete motion score of 0.

[0025] Figure 7A is a set of MR images and their corresponding global motion scores.

[0026] Figure 7B is another set of MR images and their corresponding global motion scores.

[0027] Figure 8 is a block diagram of an exemplary computing device. DETAILED DESCRIPTION

[0028] The present disclosure includes systems and methods for detecting and scoring motion artifacts in magnetic resonance (MR) images of an object using MR signals from different coils. As used herein, an object is an object placed within the bore of an MR imaging (MRI) system and imaged by the MRI system. The object can be a human subject, an animal, or a phantom. The systems and methods disclosed herein provide objective indicators of motion artifacts without requiring input from a user. Neither additional hardware nor navigator pulses are required. In addition, compared to artificial intelligence scoring methods for motion, the systems and methods disclosed herein require minimal training image data sets and limited human involvement for scoring motion. Methodological aspects will be apparent in part and discussed explicitly in part in the following description.

[0029] In magnetic resonance imaging (MRI), an object is placed in a magnet. When the object is in the magnetic field generated by the magnet, the magnetic moments of the nuclei (such as protons) attempt to align with the magnetic field, but precess around the magnetic field in a random order at the Larmor frequency of the nuclei. The magnetic field of the magnet is called B0 and extends in the longitudinal or z-direction. During the acquisition of MRI images, a magnetic field in the xy plane and close to the Larmor frequency (called the excitation field B1) is generated by a radio frequency (RF) coil and can be used to rotate or "tilt" the net magnetic moment Mz of the nuclei from the z-direction toward the transverse or xy plane. After the excitation signal B1 terminates, the nuclei emit a signal, which is called the MR signal. In order to use the MR signal to generate an image of the object, magnetic field gradient pulses (G x , G y and G z Gradient pulses are used to scan through k-space, the space of spatial frequencies or distances. A Fourier relationship exists between the acquired MR signals and the image of the object, so an image of the object can be derived by reconstructing the MR signals.

[0030] Figure 1 A schematic diagram of an exemplary MRI system 10 is shown. In the exemplary embodiment, the MRI system 10 includes a workstation 12 having a display 14 and a keyboard 16. The workstation 12 includes a processor 18, such as a commercially available programmable machine running a commercially available operating system. The workstation 12 provides an operator interface that allows scanning plans to be entered into the MRI system 10. The workstation 12 is coupled to a pulse sequence server 20, a data acquisition server 22, a data processing server 24, and a data storage server 26. The workstation 12 and each of the servers 20, 22, 24, and 26 communicate with each other.

[0031] In the exemplary embodiment, the pulse sequence server 20 operates the gradient system 28 and the radio frequency ("RF") system 30 in response to instructions downloaded from the workstation 12. The instructions are used to generate gradient waveforms and RF waveforms in an MR pulse sequence. The RF coil 38 and the gradient coil assembly 32 are used to execute the prescribed MR pulse sequence. The RF coil 38 is shown as a whole-body RF coil. The RF coil 38 can also be a local coil that can be placed near the anatomical structure to be imaged, or a coil array including multiple coils.

[0032] In the exemplary embodiment, gradient waveforms for performing a prescribed scan are generated and applied to the gradient system 28, which energizes the gradient coils in the gradient coil assembly 32 to generate magnetic field gradients G for position encoding the MR signals. y and G z The gradient coil assembly 32 forms part of a magnet assembly 34, which also includes a polarizing magnet 36 and an RF coil 38. The magnet assembly 34 forms a bore 35 in which a subject 37, such as a patient, is received and scanned.

[0033] In an exemplary embodiment, the RF system 30 includes an RF transmitter for generating RF pulses used in MR pulse sequences. The RF transmitter responds to the scan scheme and direction from the pulse sequence server 20 to generate RF pulses with a desired frequency, phase, and pulse amplitude waveform. The generated RF pulses are applied by the RF system 30 to an RF coil 38. The responsive MR signals detected by the RF coil 38 are received by the RF system 30 and amplified, demodulated, filtered, and digitized according to commands generated by the pulse sequence server 20. The RF coil 38 is depicted as both a transmitter and receiver coil, such that the RF coil 38 transmits RF pulses and detects MR signals. In one embodiment, the MRI system 10 may include a transmitter RF coil for transmitting RF pulses and a separate receiver coil for detecting MR signals. The transmit channels of the RF system 30 may be connected to the RF transmit coil, and the receiver channels may be connected to separate RF receiver coils. Typically, the transmit channels are connected to the whole-body RF coil 38, and each receiver segment is connected to a separate local RF coil.

[0034] In the exemplary embodiment, the RF system 30 also includes one or more RF receiver channels. Each RF receiver channel includes an RF amplifier that amplifies the MR signals received by the RF coil 38 to which the channel is connected; and a detector that detects and digitizes the I and Q quadrature components of the received MR signals. The magnitude of the received MR signal can then be determined as the square root of the sum of the squares of the I and Q components, as shown in the following equation (1):

[0035]

[0036] And the phase of the received MR signal can also be determined as shown in the following equation (2):

[0037]

[0038] In an exemplary embodiment, the digitized MR signal samples generated by the RF system 30 are received by the data acquisition server 22. The data acquisition server 22 can operate in response to instructions downloaded from the workstation 12 to receive real-time MR data and provide buffer memory so that no data is lost due to data overflow. In some scans, the data acquisition server 22 only passes the acquired MR data to the data processing server 24. However, in scans where information derived from the acquired MR data is needed to control further execution of the scan, the data acquisition server 22 is programmed to generate the required information and transmit it to the pulse sequence server 20. For example, during a pre-scan, MR data is acquired and used to calibrate the pulse sequence executed by the pulse sequence server 20. In addition, navigator signals can be acquired during the scan and used to adjust operating parameters of the RF system 30 or gradient system 28, or to control the order in which views are sampled in k-space.

[0039] In an exemplary embodiment, the data processing server 24 receives MR data from the data acquisition server 22 and processes the MR data according to instructions downloaded from the workstation 12. Such processing may include, for example, performing Fourier transforms on the raw k-space MR data to produce two-dimensional or three-dimensional images, applying filters to reconstructed images, performing backprojection image reconstruction on the acquired MR data, generating functional MR images, and computing motion or flow images.

[0040] In an exemplary embodiment, the image reconstructed by the data processing server 24 is transmitted back to the workstation 12 and stored there. In some embodiments, the real-time image is stored in a database memory cache ( Figure 1 (not shown), real-time images can be output from the database memory cache to the operator display 14 or a display 46 located near the magnet assembly 34 for use by the attending physician. Batch mode images or selected real-time images can be stored on a disk storage device 48 or in a host database on the cloud. When such images have been reconstructed and transferred to the storage device, the data processing server 24 notifies the data storage server 26. The operator can use the workstation 12 to archive images, generate films, or send images to other facilities via the network.

[0041] During a scan, it is desirable to keep the subject still for the duration of the scan, as movement of the subject during the scan will produce motion artifacts 203 that will degrade the quality of the MR images of the subject. Sometimes, the motion artifacts are so severe that the acquired images are blurred by the image artifacts and cannot provide a meaningful medical interpretation and diagnosis. Therefore, the subject must be rescanned. By this time, the subject may have already passed away. Rescanning not only adds time, cost, and inconvenience to the subject, but also, due to the time lag between the original scan and the rescan and the different imaging settings associated with the two scanning events, the medical information and images may lack consistency and alignment between the original scan and the rescan. Therefore, systems and methods that notify the operator of motion and / or motion levels are desirable. If the motion level is at a level suitable for rescanning, the operator can rescan the subject. The operator can also activate a motion correction program to correct the image.

[0042] Known motion detection methods require additional specialized hardware or additional navigator pulses or sequences to detect motion. The additional hardware increases scan setup time and requires additional system design and software to combine the motion signals acquired by the additional hardware with the MR signals acquired by the MRI system 10. The additional navigator pulses or sequences have the same associated challenges and increased scan time. In contrast, the systems and methods disclosed herein do not require additional hardware or navigator sequences to produce enhanced motion detection.

[0043] Figure 2A is a flow chart of an exemplary MRI method 200 for detecting and scoring motion artifacts in MR images of a subject 37 . Figure 2B The method 200 may be implemented on a motion detection and scoring computing device. The motion detection and scoring computing device may be the workstation 12, or may be a computing device separate from the workstation 12 and communicating with the workstation 12 via wired or wireless communication.

[0044] In an exemplary embodiment, the RF coil 38 of the MRI system 10 includes a plurality of RF coils for acquiring MR signals emitted from the subject. Method 200 includes receiving 202 a first composite image of the subject reconstructed based on MR signals acquired using a first coil. Method 200 further includes receiving 204 a second composite image of the subject reconstructed based on MR signals acquired simultaneously using a second coil. The first composite image and the second composite image can be represented by complex numbers, pairs of real and imaginary numbers, or phasors. The first and second images are acquired simultaneously using the same pulse sequence. In MR, a pulse sequence is a sequence of RF pulses, gradient pulses, and data acquisitions applied by the MRI system 10 when acquiring MR signals. The pulse sequence can be a fast spin echo sequence.

[0045] In an exemplary embodiment, a k-space difference map is calculated 206 between the first and second composite images based on the first and second images. In one embodiment, the k-space difference map is generated by subtracting the first and second composite images from each other; applying a two-dimensional (2D) Fourier transform of the difference; and then taking the magnitude of the Fourier transformed difference. The first and second images are acquired using a pulse sequence. Method 200 further includes generating 208 a difference graph based on the k-space difference map. Method 200 also includes calculating 210 a motion score based on a curve in the difference graph. The motion score indicates the level of motion of the subject during imaging. The motion score may be the area under the curve in the difference graph. Furthermore, method 200 includes outputting 212 the motion score. If the motion score is above a predetermined level, an alarm may be generated. The predetermined level may be set to a level above which the acquired image is unusable for medical diagnostic purposes and the subject needs to be rescanned using a pulse sequence. The alarm may be displayed on the display 14 of the MRI system 10. Once the alert has been received, the operator decides whether to rescan only or partially the slice with motion artifacts, or to rescan using only the pulse sequence. The operator can also choose not to rescan but to activate a motion correction process to correct motion artifacts in the image.

[0046] Figures 3A to 3B A turbo spin echo sequence 302 and a scan order 304 for acquiring MR images using the turbo spin echo sequence 302 are shown. Figure 3A 302 is a pulse sequence diagram of a fast spin echo sequence. Figure 3B is a diagram of an exemplary scan order 304 .

[0047] A fast spin echo sequence 302 includes RF pulses 308 and gradient pulses 310. The gradient pulses 310 can be along a readout direction 312 or a phase encoding direction 315. In a three-dimensional (3D) pulse sequence, the gradient pulses 310 can be along a slice encoding direction (not shown) or a second phase encoding direction (not shown). The diagram of the fast spin echo sequence 302 also shows MR signals in a signal channel 320. In the fast spin echo sequence 302, the RF pulses 308 include an excitation pulse 314 and a plurality of refocusing pulses 316. The excitation pulse 314 excites the magnetization and rotates it to the xy plane. The refocusing pulse 316 refocuses the dephased magnetization, forming an echo 321. The signal channel 320 includes a series of echoes 321 or an echo train 322. The k-space position of the echoes in the ky direction or phase encoding direction is determined by the phase encoding gradient 317. The time between repetitions of the pulse sequence 302 is called the repetition time (TR). The number of echoes in one TR of the turbo spin echo sequence 302 is called the echo train length (ETL). The ETL can be any number between two and the image matrix size in the phase encoding direction.

[0048] In operation, a slice in the subject is selected and excited by an excitation pulse 314 and refocused by a refocusing pulse 316. K-space is scanned by varying gradient pulses 310. Echoes 321 corresponding to multiple ky lines in k-space are acquired in one TR. Sequence 302 is repeated to scan through k-space, acquiring MR signals at additional ky lines. The MR signals are used to reconstruct an MR image.

[0049] Figure 3B is an exemplary scan order 304 in the ky direction. Each point 324 represents an echo 321 or ky line at a phase encoding order, number, or index 318. In the example shown, the matrix size of the image in the y direction is 256. The matrix size can be any other number. Figure 3B The ETL shown is 8. That is, in the first TR, an echo train 322-1 having 8 echoes 321 is acquired, where the 8 echoes 321 correspond to ky lines at phase encoding indices 318-1 to 318-8. In the next TR, the next set of ky lines corresponding to phase encoding indices 318'-1 to 318'-8 are acquired. The pulse sequence 302 is repeated until the last echo train 322-n of a slice or slice encoding step (for 3D acquisition) is acquired, where the last echo train 322-n corresponds to phase encoding indices 318"-1 to 318"-8. The scan order 304 is for one slice or one slice encoding step. To acquire multiple slices of an object, the scan order 304 is repeated over the slices or slice encoding steps, and the ky values ​​associated with a single echo train can be acquired on all slices before continuing to the next echo train. The pulse sequence 302 can be applied together with signals 320 read simultaneously from multiple coils of the RF coil 38. The MR signals acquired by the coils are combined to reconstruct an image of a slice in the object.

[0050] Figures 4A to 4I The generation of a motion score that indicates the level of motion of the subject during imaging is depicted. In an exemplary embodiment, a first MR signal or first composite k-space data 402 acquired by a first coil and a second MR signal or second composite k-space data 404 acquired by a second coil are provided. Figure 4A ). Figure 4B A coil sensitivity map 406 for a first coil and a coil sensitivity map 408 for a second coil are shown. Intensity-corrected composite images 410, 412 are derived by taking the inverse Fourier transform of the first k-space data 402 or the second k-space data 404, multiplying by the complex conjugate of the corresponding coil sensitivity map 406, 408, and dividing the product by the squared magnitude of the corresponding coil sensitivity map 406, 408. Figure 4CA k-space difference map 414 is generated by subtracting the intensity-corrected composite images 410, 412 from each other and then Fourier transforming the difference, or by Fourier transforming the intensity-corrected composite images 410, 412 and subtracting the Fourier transformed images. Figure 4D ).

[0051] In an exemplary embodiment, the k-space difference map 414 is projected along the kx direction or readout direction, thereby deriving a projected k-space difference curve map 416 ( Figure 4E ). The projected k-space difference graph 416 depicts the k-space difference as a function of the phase encoding index 318, such as Figure 3B 318-1 to 318-8, 318'-1 to 318'-8 and 318"-1 to 318"-8 are shown. Figure 4E The projected k-space difference graph 416 is shown for one slice. Figure 4F In FIG, the projected k-space difference plots 416 are averaged across the slices by averaging the projected k-space difference plots 416 for all slices along the slice direction to derive an average difference plot 418. Averaging across slices increases the signal-to-noise ratio (SNR) and improves the robustness and accuracy of motion detection and scoring. The average difference plot 418 is normalized by subtracting the baseline value at each phase encode index 318 and dividing the difference by the baseline value to derive a normalized difference plot 420 ( Figure 4G If the difference between the average difference plot 418 and the baseline value at a phase encode index 318 is zero, then no division by the baseline value is performed at that phase encode index. The baseline value can be derived by linear regression modeling of the average difference plot 418. To increase the SNR in the difference plot 420, the normalized difference plot 420 can be summed over each echo train 322 to derive a summed difference plot 421 ( Figure 4H ). For example, the values ​​422 in the normalized difference graph 420 corresponding to the phase encoding indices 318-1 to 318-8 in the echo train 322-1 are summed together, and the summing operation is repeated for the remaining echo trains 322 (such as 322-2 and 322-n). The area under the curve 423 in the summed difference graph 421 is the total motion score. The area under the curve is the integral of the curve from the starting point to the end point. For example, the area under the curve 423 is the integral of the curve 423 from point 424 to point 426. The integral of the curve can be calculated using any method known to those skilled in the art, for example, by numerical integration. The total motion score indicates the level of motion of the object during imaging using the pulse sequence 302. The motion indicated by the total motion score includes discrete motion and continuous motion. Discrete motion is motion that occurs occasionally. Continuous motion refers to motion that occurs continuously over a period of time, as opposed to discrete motion.

[0052] In an exemplary embodiment, Figure 3B In the scan sequence 304 shown, ky lines 325 in the central region of k-space are acquired with the early echo train 322-e, while ky lines 327 in the peripheral region of k-space are acquired in the late echo train 322-1. Figure 4H , curve 423 can be divided into two parts, namely, a first curve 432 and a second curve 434. The baseline value 428 of the sum difference graph 421 of the echo train 322-e in the first curve 432 is higher than the baseline value 430 of the sum difference graph 421 of the echo train 322-1 in the second curve 434. For the echo train 322-e, the baseline value 428 is subtracted from the sum difference graph 421. For the echo train 322-1, the baseline value 430 that is smaller than the baseline value 428 is subtracted from the sum difference graph 421. Therefore, a modified difference graph 436 ( Figure 4I ). The area under the curve 438 of the modified difference graph 436 is the discrete motion score.

[0053] During the data formation and acquisition process, for each line with a given ky value and for each coil, the instantaneous MRI image of the object is multiplied by the coil's sensitivity function and transformed into k-space, and the current ky line is read out. This multiplication of the coil sensitivity maps causes the readout ky line of each coil to contain information from adjacent ky lines in a manner that differs between different coils. If motion occurs, there is a new object pose, so the readout ky of each coil mixes the new k-space and the old k-space in a coil-specific manner. At the boundaries of motion, each coil contains a different mixture of the two underlying k-spaces. When there is no motion, the difference remains consistent. When there is continuous motion, there is a continuous response in the difference between the coils. Therefore, discrete peaks are associated with discrete motion, and offsets are associated with continuous motion.

[0054] Compared to the full motion score, which captures both continuous and discrete motion, the discrete motion score captures discrete motion. Compared to continuous motion or a combination of discrete and continuous motion, the discrete motion score provides better separation of images with small motion artifacts from images with no motion artifacts, but does not capture the significance of continuous motion.

[0055] A fast spin echo sequence is used as an example only. The systems and methods disclosed herein can be applied to MR signals and images acquired using other pulse sequences.

[0056] The full motion score and discrete motion scores disclosed herein provide an objective measure of motion artifacts because the calculation does not rely on input from or determinations made by a user.The full motion score and discrete motion scores may be referred to as calculated motion scores.

[0057] In some embodiments, a combined motion score can be generated. The development dataset is rated to derive a combined motion score based on the calculated motion score and the user-rated motion scores rated by multiple observers. A linear regression model is applied to fit the calculated motion score to the user-rated motion score. The combined motion score is calculated as a linear transformation of the calculated motion score using parameters derived from the linear regression modeling (such as the baseline value and / or the slope in the derived linear regression model). In one embodiment, different linear regression models are used between the calculated motion score and the user-rated motion score depending on whether the calculated full motion score is above or below a predetermined value. When the calculated full motion score is greater than or equal to the predetermined value, a combined score is derived by linearly fitting the calculated full motion score to the user-rated motion score. When the calculated full motion score is equal to or less than the predetermined value, the combined score is derived as a linear fit of the calculated discrete motion score to the user-rated motion score. This approach can be used to distinguish fine motion because the discrete motion score provides a better indication of fine motion than the full motion score.

[0058] Figures 5A to 5C The verification results are shown. Figure 5A The results using the development dataset are shown. The development dataset was used to derive a linear relationship between the calculated motion scores 502 and the user-rated motion scores. Figures 4A to 4I The difference curve plots described above derive calculated motion scores 502 for images in the development dataset. The development dataset is also rated by multiple human observers to derive user-rated motion scores 504. The calculated motion scores 502 and the user-rated motion scores 504 are plotted against each other and linearly fitted to derive a linear relationship 506 between the calculated motion scores 502 and the user-rated motion scores 504. The inverse of the linear relationship 506 is used to transform the calculated motion scores into a combined motion score 508. This approach is tested with a test dataset. Figure 5B As shown, the combined motion score 508 matches the user-rated motion score 504, where the exemplary goodness-of-fit indicator R 2 is 0.9134, where R 2 A value closer to 1 indicates a better fit of the linear regression model. The combined motion score 508 was validated with the development and test datasets, and the results are shown in Figure 5C. A combined motion score 508 is calculated by transforming the calculated motion scores by the inverse of the relationship 506 derived from the development data set. In one embodiment, when the calculated full motion score 502 is below a predetermined threshold level, the combined motion score 508 is calculated as the calculated discrete motion scores transformed by the inverse of the relationship 506. When the calculated full motion score 502 is above the predetermined threshold level, the combined motion score 508 is calculated as the calculated full motion score 502 transformed by the inverse of the relationship 506. When the calculated full motion score 502 is equal to the predetermined threshold level, the combined motion score 508 may be based on the calculated full motion score 502 or the calculated discrete motion scores. Thus, the use of the calculated discrete motion scores further distinguishes fine motion indicated by relatively small calculated full motion scores 502.

[0059] Figure 6A and Figure 6B Discrete movement scores indicating fine movements are shown. Figure 6A This is a brain image 602 with a discrete motion score of 1.3. Figure 6B 604 is a brain image with a discrete motion score of 0. The brain image 602 has subtle motion artifacts 606 that are not easily discernible.

[0060] 7A to 7B Images and their calculated motion scores 502 using the systems and methods disclosed herein are shown. The calculated motion score 502 accurately reflects the level of motion artifacts present in the images. For example, the calculated motion scores for images 702 and 704 are 0 and do not include discernible motion artifacts, while the calculated motion scores for images 706 and 708 are 7.5 and 10, respectively, and include severe motion artifacts that render images 706 and 708 unusable.

[0061] In some embodiments, the calculated motion score 502 is used to determine whether motion correction of the image should be performed. For example, if the calculated motion score 502 is above a predetermined level or within a specified range, motion correction can be performed on the acquired image. The motion-corrected image can be reconstructed by jointly estimating the motion parameters of the motion-corrected image and optimally predicting the motion parameters of the acquired k-space data. Correction can be performed using a neural network model or an iterative optimization method. In methods using a neural network model, the neural network model is trained using images with motion artifacts and images with motion artifacts corrected. The training data can be a simulated training data set consisting of images without motion artifacts and images without motion artifacts but with simulated motion artifacts added.

[0062] In iterative optimization methods, motion correction is formulated as an optimization problem in which a cost function representing the data fit is minimized with respect to the motion-corrected image and motion parameters as optimization variables. The optimization problem can be solved by a numerical algorithm that iteratively updates the motion-corrected image and motion parameters.

[0063] Assume that M discrete motions are detected, and their time sequence is t1,…,t M Then, M+1 poses can be defined such that pose i corresponds to time t i and t i+1 Data collection between 0≤i≤M, where t0 and t M+1 The scan start time and scan end time are respectively. Without loss of generality, posture 0 can be considered as the reference posture, and the motion-corrected image will be reconstructed relative to this reference posture. The multi-coil k-space data y acquired corresponding to posture i i can be modeled as

[0064] y i =A i FST(θ i )x,

[0065] where x is the motion-free image corresponding to the reference pose, T(θ i ) is given by θ i is the parameterized operator that transforms the reference pose to pose i, S applies the coil sensitivities to the image, F is the Fourier transform operator, and A i is the sampling matrix containing the sampling pattern for data collection for pose i. Assuming rigid body motion, each θ i There are 3 parameters for 2D (i.e., 2 parameters for translation and 1 parameter for rotation), and 6 parameters for 3D (i.e., 3 parameters for translation and 3 parameters for rotation). The total number of unknown motion parameters is 3M for 2D and 6M for 3D (excluding the reference pose).

[0066] The motion-corrected image can be reconstructed by solving the following optimization problem:

[0067]

[0068] in is the reconstructed motion-corrected image, θ is θ i cascade, and represents the estimated motion parameters. The L2 norm may be used for the least squares cost function. Alternatively, other types of cost functions may be used, such as those based on the L1 norm or weighted least squares. In some embodiments, the cost function may be regularized, for example, by adding a regularization function to the motion-corrected image or the motion parameters. For example, the total variation penalty function used in compressed sensing may be used as a regularization function for the motion-corrected image.

[0069] To solve the above optimization problem, a numerical algorithm such as gradient descent, conjugate gradient, or Newton's algorithm can be used. The numerical algorithm iteratively updates the motion-corrected image and the motion parameters. The motion-corrected image and the motion parameters can be updated simultaneously or alternatively. In order to use an iterative numerical algorithm, an initial estimate of the motion-corrected image and the motion parameters should be available. For example, a motion-corrupted image obtained without any motion correction and parameters corresponding to no motion can be used for the initial estimate. Alternatively, a grid search can be used for the initialization step. For example, for predetermined values ​​of the motion parameters, the goodness of fit is calculated, and the best parameters in terms of goodness of fit are selected and used for the initial estimate of the motion parameters. The grid search method is computationally expensive but can help avoid local minima.

[0070] The workstation 12 and motion detection and scoring computing devices described herein may be any suitable computing device 800 and software implemented therein. Figure 8 800 is a block diagram of an exemplary computing device 800. In an exemplary embodiment, the computing device 800 includes a user interface 804 that receives at least one input from a user. The user interface 804 may include a keyboard 806 that enables the user to enter relevant information. The user interface 804 may also include, for example, a pointing device, a mouse, a stylus, a touch-sensitive panel (e.g., a touchpad and a touch screen), a gyroscope, an accelerometer, a position detector, and / or an audio input interface (e.g., including a microphone).

[0071] Furthermore, in an exemplary embodiment, the computing device 800 includes a display interface 817 for presenting information (such as input events and / or verification results) to a user. The display interface 817 may also include a display adapter 808 coupled to at least one display device 810. More specifically, in an exemplary embodiment, the display device 810 may be a visual display device such as a cathode ray tube (CRT), a liquid crystal display (LCD), a light emitting diode (LED) display, and / or an "electronic ink" display. Alternatively, the display interface 817 may include an audio output device (e.g., an audio adapter and / or speakers) and / or a printer.

[0072] The computing device 800 also includes a processor 814 and a memory device 818. The processor 814 is coupled to the user interface 804, the display interface 817, and the memory device 818 via a system bus 820. In an exemplary embodiment, the processor 814 communicates with the user, such as by prompting the user via the display interface 817 and / or by receiving user input via the user interface 804. The term "processor" generally refers to any programmable system, including system and microcontrollers, reduced instruction set computers (RISCs), complex instruction set computers (CISCs), application specific integrated circuits (ASICs), programmable logic circuits (PLCs), and any other circuit or processor capable of performing the functions described herein. The above examples are exemplary only and are therefore not intended to limit the definition and / or meaning of the term "processor" in any way.

[0073] In an exemplary embodiment, the memory device 818 includes one or more devices that enable information (such as executable instructions and / or other data) to be stored and retrieved. In addition, the memory device 818 includes one or more computer-readable media, such as, but not limited to, dynamic random access memory (DRAM), static random access memory (SRAM), a solid-state drive, and / or a hard disk. In an exemplary embodiment, the memory device 818 stores, but is not limited to, application source code, application object code, configuration data, additional input events, application state, assertion statements, verification results, and / or any other type of data. In an exemplary embodiment, the computing device 800 may also include a communication interface 830 coupled to the processor 814 via the system bus 820. In addition, the communication interface 830 is communicatively coupled to a data acquisition device.

[0074] In an exemplary embodiment, processor 814 may be programmed by encoding operations using one or more executable instructions and providing the executable instructions in memory device 818. In an exemplary embodiment, processor 814 is programmed to select a plurality of measurements received from a data collection device.

[0075] In operation, a computer executes computer-executable instructions embodied in one or more computer-executable components stored on one or more computer-readable media to implement various aspects of the present invention described and / or illustrated herein. Unless otherwise indicated, the order in which the operations in the embodiments of the present invention shown and described herein are performed or implemented is not required. That is, unless otherwise indicated, the operations may be performed in any order, and embodiments of the present invention may include more or fewer operations than those disclosed herein. For example, it is contemplated that performing or implementing a particular operation before, simultaneously with, or after another operation is within the scope of various aspects of the present invention.

[0076] At least one technical effect of the systems and methods described herein includes (a) automatic scoring of motion artifacts; (b) discrete motion scores that indicate relatively fine motion; (c) motion scores that provide an objective indication of motion artifacts; and (d) motion correction of images once motion artifacts are detected.

[0077] Exemplary embodiments of systems and methods for detecting and scoring motion artifacts are described above in detail. These systems and methods are not limited to the specific embodiments described herein, but rather, components of the systems and / or operations of the methods can be used independently and separately from other components and / or operations described herein. Furthermore, the components and / or operations described can also be defined in or used in conjunction with other systems, methods, and / or devices, and are not limited to practice with only the systems described herein.

[0078] Although specific features of various embodiments of the present invention may be shown in some drawings and not in others, this is for convenience only. In accordance with the principles of the invention, any feature of a drawing may be referenced and / or claimed in combination with any feature of any other drawing.

[0079] This written description uses examples to disclose the invention, including the best mode, and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any included methods. The patentable scope of the invention is defined by the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insignificant differences from the literal language of the claims.

Claims

1. A method for detecting and scoring motion artifacts in a magnetic resonance image of a subject, the method comprising: calculating a k-space difference map based at least in part on first MR signals of the subject acquired using a first coil and second MR signals of the subject concurrently acquired using a second coil; generating a difference curve map based on the k-space difference map, comprising projecting the k-space difference map along a readout direction to derive the difference curve map, the difference curve map depicting k-space difference as a function of phase encoding index, the difference curve map comprising a curve; and calculating a motion score based on the curve in the difference graph, wherein the motion score indicates a level of motion artifacts in the magnetic resonance image caused by motion of the subject during acquisition of the first MR signal and the second MR signal, and the motion score comprises an area under the curve; and The motion score is output.

2. The method of claim 1, wherein the motion score is a full motion score, wherein The motion indicated by the full motion score includes discrete motion and continuous motion.

3. The method of claim 1 , wherein generating a difference curve graph further comprises: The curve in the difference curve map is divided into a first curve and a second curve, wherein the first curve corresponds to a central region of k-space and the second curve corresponds to a peripheral region of the k-space.

4. The method according to claim 3, further comprising: normalizing the first curve by the baseline value of the first curve; normalizing the second curve by the baseline value of the second curve; as well as Calculating the motion score further includes deriving the motion score by calculating an area under a normalized first curve and an area under a normalized second curve, wherein the motion score includes a discrete motion score.

5. The method of claim 1 , wherein the first MR signal and the second MR signal are MR signals of a plurality of slices in the object, and generating the difference curve map further comprises averaging a plurality of difference curve maps over the plurality of slices to derive the difference curve map, wherein each difference curve map corresponds to one slice in the plurality of slices.

6. The method of claim 1 , wherein the first MR signal and the second MR signal are MR signals of the object acquired using a plurality of echo trains, and generating a difference curve map further comprises summing the difference curve map over each of the plurality of echo trains to derive the difference curve map as a function of an index of the plurality of echo trains. The method of claim 1 , wherein generating a difference curve plot further comprises normalizing the difference curve plot by a baseline value of the curve. 8 . The method of claim 1 , wherein the motion score comprises a full motion score and a discrete motion score, the full motion score indicating levels of both discrete motion and continuous motion, and the discrete motion score indicating levels of the discrete motion.

9. The method of claim 1 , wherein calculating the motion score further comprises: A combined athletic score is generated based on the calculated athletic score and the user-rated athletic score.

10. The method of claim 1 , wherein the motion score comprises a full motion score and a discrete motion score, and calculating the motion score further comprises: generating a combined motion score based on the discrete motion scores if the overall motion score is less than a predetermined level; The motion indicated by the full motion score includes discrete motion and continuous motion.

11. The method of claim 1 , wherein the method further comprises generating an alert if the motion score is above a predetermined level.

12. The method of claim 1, further comprising: When the motion score is within a specified range, motion correction is performed on the image of the object to derive a motion-corrected image.

13. A motion detection and scoring computation device comprising at least one processor in communication with at least one memory device, and wherein the at least one processor is programmed to: calculating a k-space difference map based at least in part on first MR signals of the subject acquired using a first coil and second MR signals of the subject concurrently acquired using a second coil; generating a difference curve map based on the k-space difference map, comprising projecting the k-space difference map along a readout direction to derive the difference curve map, the difference curve map depicting k-space difference as a function of phase encoding index, the difference curve map comprising a curve; and calculating a motion score based on the curve in the difference graph, wherein the motion score indicates a level of motion artifacts in a magnetic resonance image of the subject caused by motion of the subject during acquisition of the first MR signal and the second MR signal, and the motion score comprises an area under the curve; and The motion score is output.

14. The computing device of claim 13, wherein the at least one processor is further programmed to: dividing the curve in the difference curve map into a first curve and a second curve, wherein the first curve corresponds to a central region of k-space and the second curve corresponds to a peripheral region of k-space; normalizing the first curve by the baseline value of the first curve; normalizing the second curve by the baseline value of the second curve; as well as The motion score is derived by calculating an area under a normalized first curve and an area under a normalized second curve, wherein the motion score comprises a discrete motion score.

15. The computing device of claim 13 , wherein the first MR signal and the second MR signal are MR signals of a plurality of slices in the object, and the at least one processor is further programmed to average a plurality of difference curve maps over the plurality of slices to derive the difference curve map, wherein each difference curve map corresponds to one slice in the plurality of slices.

16. The computing device of claim 13 , wherein the first MR signal and the second MR signal are MR signals of the object acquired using a plurality of echo trains, and the at least one processor is further programmed to sum the difference curve map over each of the plurality of echo trains to derive the difference curve map as a function of an index of the plurality of echo trains.

17. The computing device of claim 13, wherein the motion score comprises a full motion score and a discrete motion score, and the at least one processor is further programmed to: generating a combined motion score based on the discrete motion scores if the overall motion score is less than a predetermined level; in, The motion indicated by the full motion score includes discrete motion and continuous motion.

18. The computing device of claim 13, wherein the at least one processor is further programmed to generate an alert if the motion score is above a predetermined level.

19. The computing device of claim 13, wherein the at least one processor is further programmed to perform motion correction on the magnetic resonance image of the subject to derive a motion-corrected image when the motion score is within a specified range.

Images