Anatomically-informed deep learning on contrast-enhanced cardiac MRI for scar segmentation and clinical feature extraction

Inventors: Natalia A. Trayanova (Baltimore, MD); Haley Gilbert Abramson (Baltimore, MD); Dan Popescu (Baltimore, MD); Mauro Maggioni (Baltimore, MD); Katherine C. Wu (Bel Air, MD)

Assignee: THE JOHNS HOPKINS UNIVERSITY

G06T7/11 G06T7/0012 G06T2207/10088 G06T2207/20081 G06T2207/20084 G06T2207/30048

View Patent ↗

Loading inventors, assignments & file history…

Monitor This Case

Get email alerts when status or documents change.

Order Certified Copies

Most orders are placed with the USPTO same day — all within 24 business hours.

Order via The Patent Place →

Pre-filled with this patent's details

Quick Facts

Patent No.: US 12,670,598
App. No.: 18/249,583
Granted: Jun 30, 2026
Kind: B2

External Links

USPTO Patent Center↗ Full Patent PDF↗

Abstract

Fully automated computer-implemented deep learning techniques of contrast-enhanced cardiac MRI segmentation are provided. The techniques may include providing cardiac MRI data to a first computer-implemented deep learning network trained in order to identify a left ventricle region of interest to generate left ventricle region-of-interest-identified cardiac MRI data. The techniques may also include providing the left ventricle region-of-interest-identified cardiac MRI data to a second computer-implemented deep learning network trained in order to identify myocardium to generate myocardium-identified cardiac MRI data. The techniques may further include providing the myocardium-identified cardiac MRI data to at least one third computer-implemented deep learning network trained to conform data to geometrical anatomical constraints in order to generate anatomical-conforming myocardium-identified cardiac MRI data. The techniques may further include outputting the anatomical-conforming myocardium-identified cardiac MRI data.

Claims (18)

1 . A fully automated computer-implemented deep learning method of contrast-enhanced cardiac MRI segmentation, the method comprising:

providing cardiac MRI data to a first computer-implemented deep learning network trained to identify a left ventricle region of interest, whereby left ventricle region-of-interest-identified cardiac MRI data is produced, wherein the first computer-implemented deep learning network comprises a convolutional neural network with residuals;

providing the left ventricle region-of-interest-identified cardiac MRI data to a second computer-implemented deep learning network trained to identify myocardium, whereby myocardium-identified cardiac MRI data is produced, wherein the second computer-implemented deep learning network comprises a convolutional neural network with residuals;

providing the myocardium-identified cardiac MRI data to at least one third computer-implemented deep learning network trained to conform data to geometrical anatomical constraints to ensure anatomical accuracy, whereby anatomical-conforming myocardium-identified cardiac MRI data is produced, wherein the at least one third computer-implemented deep learning network comprises a convolutional autoencoder coupled to a Gaussian mixture model, wherein the at least one third computer-implemented deep learning network is trained to conform data to geometrical anatomical constraints by: autoencoding the myocardium-identified cardiac MRI data to generate a latent vector space; and statistically modeling the latent vector space; wherein the latent vector space is populated with anatomically-correct samples and allows for nearest-neighbor identification of the anatomical-conforming myocardium-identified cardiac MRI data, whereby the anatomical-conforming myocardium-identified cardiac MRI data is anatomically correct; and

outputting the anatomical-conforming myocardium-identified cardiac MRI data, wherein no manual human intervention is required.

2 . The method of claim 1 , wherein the anatomical-conforming myocardium-identified cardiac MRI data comprises scar segmentation data.

3 . The method of claim 1 , further comprising reducing a background based on the ventricle region-of-interest-identified cardiac MRI data.

4 . The method of claim 1 , wherein the second computer-implemented deep learning network is trained to identify myocardium by delineating endocardium and epicardium.

5 . The method of claim 1 , wherein the outputting comprises displaying on a computer monitor.

6 . A fully automated computer system for deep learning contrast-enhanced cardiac MRI segmentation, the computer system comprising:

a first computer-implemented deep learning network trained to identify a left ventricle region of interest in cardiac MRI data to produce left ventricle region-of-interest-identified cardiac MRI data, wherein the first computer-implemented deep learning network comprises a convolutional neural network with residuals;

a second computer-implemented deep learning network trained to identify myocardium in the left ventricle region-of-interest-identified cardiac MRI data to produce myocardium-identified cardiac MRI data, wherein the second computer-implemented deep learning network comprises a convolutional neural network with residuals;

at least one third computer-implemented deep learning network trained to conform the myocardium-identified cardiac MRI data to geometrical anatomical constraints to produce anatomical-conforming myocardium-identified cardiac MRI data to ensures anatomical accuracy, wherein the at least one third computer-implemented deep learning network comprises a convolutional autoencoder coupled to a Gaussian mixture model, wherein the at least one third computer-implemented deep learning network is trained to conform data to geometrical anatomical constraints by: autoencoding the myocardium-identified cardiac MRI data to generate a latent vector space; and statistically modeling the latent vector space; wherein the latent vector space is populated with anatomically-correct samples and allows for nearest-neighbor identification of the anatomical-conforming myocardium-identified cardiac MRI data, whereby the anatomical-conforming myocardium-identified cardiac MRI data is anatomically correct; and

an output configured to provide the anatomical-conforming myocardium-identified cardiac MRI data, wherein no manual human intervention is required.

7 . The system of claim 6 , wherein the anatomical-conforming myocardium-identified cardiac MRI data comprises scar segmentation data.

8 . The system of claim 6 , wherein the computer system is configured to reduce a background based on the ventricle region-of-interest-identified cardiac MRI data.

9 . The system of claim 6 , wherein the second computer-implemented deep learning network is trained to identify myocardium in the cardiac MRI data by delineating endocardium and epicardium.

10 . The system of claim 6 , wherein the output comprises a computer monitor configured to display the anatomical-conforming myocardium-identified cardiac MRI data.