MRI system and method using neural network for detection of patient motion

Patent metadata

Field	Value
Publication number	US-10996306-B2
Application number	US-201916394791-A
Country	US
Kind code	B2
Filing date	Apr 25, 2019
Priority date	Apr 25, 2019
Publication date	May 4, 2021
Grant date	May 4, 2021

Abstract

A magnetic resonance imaging (MRI) system includes control and analysis circuitry having programming to acquire magnetic resonance (MR) data using coil elements of the MRI system, analyze the MR data, and reconstruct the MR data into MR sub-images. The system also includes a trained neural network associated with the control and analysis circuitry to transform the MR sub-images into a prediction relating to a presence and extent of motion corruption in the MR sub-images. The programming of the control and analysis circuitry includes instructions to control operations of the MRI system based at least in part on the prediction of the trained neural network.

First claim

The invention claimed is:

1. A magnetic resonance imaging (MRI) method comprising:
producing a first sub-image from first magnetic resonance (MR) partial k-space data acquired by an MM system during a first time interval from a patient;
producing a second sub-image from second MR partial k-space data acquired by the MRI system during a second time interval from the patient, wherein the first time interval and the second time interval are adjacent to one another in time;
combining the first sub-image and the second sub-image to generate a combined sub-image;
using a trained neural network to generate a prediction relating to the presence and extent of motion occurring between the first time interval and the second time interval using the combined sub-image as an input; and
performing further operations of the MRI system based at least partially on the prediction generated by the trained neural network, wherein performing further operations of the MRI system based at least partially on the prediction generated by the trained neural network comprises, in response to determining that the prediction indicates that motion has occurred between the first and second time intervals:
aggregating the first MR data with MR data collected before the first time interval as k-space data corresponding to a first motion state as a first aggregate;
aggregating the second MR data with MR data collected after the first time interval as k-space data corresponding to a second motion state as a second aggregate; and
separately reconstructing the first aggregate and the second aggregate to produce a first motion-free sub-image and a second motion-free sub-image, respectively.

2. The method of claim 1 , wherein the partial k-space data from the first time interval and the partial k-space data second time interval are from a single coil within a multi-coil receiver array of the MRI system.

3. The method of claim 1 , wherein the prediction relating to the presence and extent of motion occurring between the first time interval and the second time interval is a motion score, and wherein the motion score is calculated such that as the magnitude of the motion score increases, motion is more likely to have had an effect on the combined sub-image.

4. The method of claim 3 , wherein a motion score is calculated for each coil in a multi-coil receiver array of the MRI system to generate motion scores for the multi-coil receiver array, and the motion scores for the multi-coil receiver array are combined into a net motion score, by taking mean, median, maximum, or minimum of individual coil scores.

5. The method of claim 4 , comprising determining whether motion has occurred by comparing the net motion score to a threshold.

6. The method of claim 3 , wherein the motion score is based on a weighted sum, or a log of a weighted sum, or average, of a per-pixel metric of the difference between the combined sub-image and a combined sub-image obtained if there is no motion between the first and second time intervals.

7. The method of claim 6 , where the per-pixel metric is a per-pixel entropy of the difference, or per-pixel difference, or per-pixel log of the difference, or per-pixel log of the absolute value of the difference.

8. The method of claim 1 , wherein producing the first sub-image comprises using only the first MR data collected during the first time interval, and wherein producing the second sub-image comprises using only the second MR data collected during the second time interval.

9. The method of claim 1 , wherein the first sub-image and the second sub-image are complex, and are combined by addition.

10. The method of claim 1 , wherein the first sub-image and the second sub-image are combined by aggregation of their respective partial k-space data in k-space, before transformation into the image domain.

11. The method of claim 1 , wherein producing the first sub-image using the first MR data collected during the first time interval and producing the second sub-image using the second MR data collected during the second time interval comprises using data collected by all receiving coils of a receiving coil array of the MRI system during the first time interval and the second time interval, respectively.

12. The method of claim 1 , wherein the trained neural network is a trained convolutional neural network having a plurality of convolutional layers, a plurality of max pooling layers, and a plurality of fully connected layers.

13. The method of claim 1 , wherein the trained neural network is trained according to a training process comprising:
generating training data via a training data generation process comprising:
simulating rigid-body motion by applying translation and/or rotation to a motion-free image to generate an offset image;
replacing portions of motion-free k-space data of the motion-free image with k-space data of the offset image to generate motion-corrupted mixed k-space data according to a scan order, wherein the scan order describes how k-space was filled via phase encodes as a function of time step;
simulating partial data collection by applying a mask to the mixed k-space data according to the scan order to generate partial k-space data;
generating a motion-corrupted sub-image from the partial k-space data; and
calculating a motion score for the motion-corrupted sub-image based at least partially on a difference between the motion-corrupted sub-image and a corresponding motion-free sub-image generated using corresponding partial k-space data of the motion-free image;

repeating at least a portion of the training data generation process by applying different translations and/or rotations to the motion-free image or other motion-free images to produce a population of motion-corrupted sub-images and associated motion scores; and
training a neural network with at least some of the population of motion-corrupted sub-images and associated motion scores to generate the trained neural network.

14. The method of claim 1 , wherein performing further operations of the MRI system based at least partially on the prediction generated by the trained neural network further comprises registering and combining the first motion-free sub-image and the second motion-free sub-image.

15. A computer-based method that generates a trained neural network to generate predictions relating to the presence and extent of motion in magnetic resonance (MR) sub-images comprising:
providing training data comprising motion-corrupted sub-images as available inputs and corresponding motion scores as outputs, wherein the training data is generated by a training data generation process comprising:
applying translation and/or rotation to a motion-free image to produce an offset image;
replacing a portion of motion-free k-space data associated with the motion-free image with motion-corrupted k-space data associated with the offset image to generate mixed k-space data; and
simulating partial data collection by applying a mask to the mixed k-space data according to a scan order to generate partial k-space data, wherein the scan order defines how k-space was filled with the motion-free k-space data as a function of time step for the motion-free image, and wherein the mask defines a section of k-space from which to collect the partial k-space data, the section corresponding to adjacent time steps in the scan order;

training, using the training data, a neural network to transform an MR sub-image into a corresponding motion score that indicates whether motion occurred during an MM scan used to obtain data for the MR sub-image;
wherein the motion-corrupted sub-image

Assignees

• Gen Electric

Inventors

• Heukensfeldt Jansen Isabelle
• Ahn Sangtae
• Hardy Christopher Judson
• Malkiel Itzik
• Brada Rafael Shmuel
• Wein Ron
• Rotman Michael

Classifications

• G06V10/993

  Evaluation of the quality of the acquired pattern · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06T12/30

  Image post-processing, e.g. metal artefact correction · CPC title

• G06F18/2148

  characterised by the process organisation or structure, e.g. boosting cascade · CPC title