Systems and methods for generating localizer scan settings from calibration images

Patent metadata

Field	Value
Publication number	US-11506739-B2
Application number	US-201916573955-A
Country	US
Kind code	B2
Filing date	Sep 17, 2019
Priority date	Sep 17, 2019
Publication date	Nov 22, 2022
Grant date	Nov 22, 2022

Abstract

Methods and systems are provided for determining scan settings for a localizer scan based on a magnetic resonance (MR) calibration image. In one example, a method for magnetic resonance imaging (MRI) includes acquiring an MR calibration image of an imaging subject, mapping, by a trained deep neural network, the MR calibration image to a corresponding anatomical region of interest (ROI) attribute map for an anatomical ROI of the imaging subject, adjusting one or more localizer scan parameters based on the anatomical ROI attribute map, and acquiring one or more localizer images of the anatomical ROI according to the one or more localizer scan parameters.

First claim

The invention claimed is:

1. A method for magnetic resonance imaging (MRI), the method comprising:
acquiring a magnetic resonance (MR) calibration image of an imaging subject with an MRI system;
mapping, by a trained deep neural network, the MR calibration image to a corresponding anatomical region of interest (ROI) attribute map for an anatomical ROI of the imaging subject, the anatomical ROI attribute map defining attributes of the anatomical ROI within a set field of view (FOV) used to obtain the MR calibration image;
adjusting one or more localizer scan parameters based on the anatomical ROI attribute map; and
acquiring, with the MRI system, one or more localizer images of the anatomical ROI according to the one or more localizer scan parameters.

2. A method, comprising:
training a deep neural network to map magnetic resonance (MR) calibration images to corresponding anatomical region of interest (ROI) attribute maps, wherein the training includes:
feeding a plurality of training data pairs to the deep neural network, wherein each training data pair includes an MR calibration image and a corresponding ground truth anatomical ROI attribute map with embedded attribute parameters;
correlating the MR calibration image in a training data pair to a predicted anatomical ROI attribute map using the deep neural network;
calculating a difference between the predicted anatomical ROI attribute map and the ground truth anatomical ROI attribute map; and
adjusting parameters of the deep neural network via backpropagation based on the difference between the predicted anatomical ROI attribute map and the ground truth anatomical ROI attribute map;

receiving a particular MR calibration image; and
determining, from the particular MR calibration image, a corresponding anatomical ROI attribute map using the trained deep neural network.

3. A magnetic resonance imaging (MRI) system, comprising:
a memory storing a trained deep neural network and instructions; and
a processor communicably coupled to the memory and when executing the instructions, configured to:
instruct the MRI system to acquire an MR calibration image of an imaging subject;
determine, with the MR calibration image and using the trained deep neural network, an anatomical region of interest (ROI) attribute map, the anatomical ROI attribute map defining one or more attribute parameters of an anatomical ROI of the imaging subject, the one or more attribute parameters comprising one or more of a center point and/or extent of the anatomical ROI within an imaging bore of the MRI system, an orientation of the anatomical ROI, and an identification of the anatomical ROI;
determine one or more localizer scan settings for a localizer scan of the anatomical ROI based on the anatomical ROI attribute map; and
instruct the MRI system to acquire a localizer image using the one or more localizer scan settings.

4. The method of claim 1 , further comprising:
generating a graphical prescription using the localizer images; and
performing a diagnostic scan of the anatomical ROI of the imaging subject with the MRI system according to the graphical prescription, and wherein the MR calibration image is acquired with a first MR protocol and the one or more localizer images are acquired with a second, different MR protocol.

5. The method of claim 1 , wherein adjusting the one or more localizer scan parameters comprises adjusting a localizer FOV, and wherein acquiring the one or more localizer images comprises acquiring the one or more localizer images with the adjusted localizer FOV, wherein the localizer FOV is smaller than the set FOV.

6. The method of claim 1 , wherein adjusting the one or more localizer scan parameters comprises adjusting a location of a table supporting the imaging subject within an imaging bore of the MRI system, and wherein acquiring the one or more localizer images comprises acquiring the one or more localizer images with the table at the adjusted position.

7. The method of claim 1 , wherein adjusting the one or more localizer scan parameters comprises adjusting a no phase wrap factor, and wherein acquiring the one or more localizer images comprises acquiring the one or more localizer images with the adjusted no phase wrap factor.

8. The method of claim 1 , wherein adjusting the one or more localizer scan parameters comprises adjusting one or more of a number of slices, slice thickness, slice gap, orientation, and number of lines of k-space, and wherein acquiring the one or more localizer images comprises acquiring the one or more localizer images with the adjusted number of slices, slice thickness, slice gap, orientation, and/or number of lines of k-space.

9. The method of claim 1 , further comprising determining one or more attribute parameters based on the anatomical ROI attribute map, wherein the one or more attribute parameters comprise one or more of a center point and/or extent of the anatomical ROI, an orientation of the anatomical ROI, and an identification of the anatomical ROI, and wherein adjusting one or more localizer scan parameters comprises adjusting one or more localizer scan parameters based on the one or more attribute parameters.

10. The method of claim 1 , wherein the anatomical ROI is determined based on a scanning protocol selected by a user.

11. The method of claim 10 , further comprising determining, based on the anatomical ROI attribute map, whether the anatomical ROI matches an anatomical ROI specified by the scanning protocol, and if the anatomical ROI determined does not match the anatomical ROI specified by the scanning protocol, outputting a notification alerting an operator that the anatomical ROI does not match the anatomical ROI specified by the scanning protocol.

12. The method of claim 10 , further comprising determining, based on the anatomical ROI attribute map, whether a receive radio frequency (RF) coil is positioned at a target position, and if the receive RF coil is not positioned at the target position, outputting a notification alerting an operator of the MRI system that the receive RF coil is not positioned at the target position.

13. The method of claim 10 , further comprising determining, based on the anatomical ROI attribute map, whether the set FOV of the MR calibration image matches a target FOV specified by the scanning protocol, and if the set FOV does not match the target FOV, outputting a notification alerting an operator that the set FOV does not match the target FOV.

14. The method of claim 2 , wherein the anatomical ROI attribute map includes one or more of a center point and/or extent of the anatomical ROI within an imaging bore of the MRI system, an orientation of the anatomical ROI, and an identification of the anatomical ROI.

15. The MRI system of claim 3 , wherein the memory further stores instructions for training the deep neural network.

16. The MRI system of claim 15 , wherein the processor, when executing the instructions for training the deep neural network, is configured to:
feed a plurality of training data pairs to the deep neural network, each training data pair includes an MR calibration image and a corresponding ground truth anatomical ROI attribute map;
map the MR calibration image in a training data pair to a predicted anatomical ROI attribute map using the deep neural network;
calculate a difference between the predicted anatomical ROI attribute map and the corresponding ground truth anatomical ROI attribute map; and
adjust parameters of the deep neural network via backpropagation based on the difference between the predicted anatomical ROI attribute map and

Assignees

• Ge Prec Healthcare Llc

Inventors

• Gui Dawei
• Shanbhag Dattesh Dayanand
• Bhushan Chitresh
• Maximo André De Almeida

Classifications

• A61B5/055

  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06F9/542

  Event management; Broadcasting; Multicasting; Notifications · CPC title

• A61B5/7264

  Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems · CPC title

• A61B5/0037

  Performing a preliminary scan, e.g. a prescan for identifying a region of interest · CPC title