IP Library Granted Patent US 11,707,255
Granted Patent B2
US 11,707,255 · App. 16/372,727 · Granted Jul 25, 2023

Image-based probe positioning

Inventor: Vasant Salgaonkar (Sunnyvale, CA)
Assignee: Siemens Medical Solutions USA, Inc.
A61B8/4245A61B5/061A61B8/0883A61B8/12A61B8/445A61B8/463A61B8/483A61B34/20
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 11,707,255
App. No.
16/372,727
Granted
Jul 25, 2023
Kind
B2
Abstract

A framework for image-based probe positioning is disclosed herein. The framework receives a current image from a probe. The current image is acquired by the probe within a structure of interest. The framework predicts a position of the probe and generates a recommendation of a next maneuver to be performed using the probe by applying the current image to a trained classifier. The framework then outputs the predicted position and the recommendation of the next maneuver.

Claims (31)

1. One or more non-transitory computer-readable media embodying instructions executable by machine to perform operations for intracardiac catheter positioning, comprising:

(i) receiving a current intracardiac echocardiogram (ICE) image acquired by a catheter configured to be inserted into a heart;

(ii) predicting a position of the catheter and generating a recommendation of a next maneuver to be performed using the catheter by applying the current ICE image and a time history of a plurality of views to a trained classifier, wherein the time history of the plurality of views comprises a current sequence of a predetermined number of the views previously generated by the trained classifier and is used by the trained classifier to detect and correct any error in the predicted position, wherein each of the plurality of views is a visualization of a predefined set of one or more anatomic landmarks and is mapped to a relative position of the catheter; and

(iii) displaying the predicted position and the recommendation of the next maneuver;

wherein the operations further comprise training the classifier based on a plurality of ICE training images and using a mapping between the plurality of ICE training images and a plurality of ICE catheter positions from which the plurality of ICE training images were acquired as ground truth.

2. The one or more non-transitory computer-readable media of claim 1 wherein the displaying the predicted position comprises displaying the catheter at the predicted position in a graphical representation of the heart.

3. The one or more non-transitory computer-readable media of claim 1 wherein the next maneuver is represented by navigational instructions.

4. The one or more non-transitory computer-readable media of claim 3 wherein the navigational instructions comprise machine instructions that are executable by a robotic controller to automatically steer the catheter to a desired position.

5. A method of probe positioning, comprising:

(i) receiving a current image acquired by a probe within a structure of interest;

(ii) predicting a position of the probe and generating a recommendation of a next maneuver to be performed using the probe by applying the current image and a time history of a plurality of views to a trained classifier, wherein the time history of the plurality of views comprises a current sequence of a predetermined number of the views previously generated by the trained classifier and is used by the trained classifier to detect and correct any error in the predicted position, wherein each of the plurality of views is a visualization of a predefined set of one or more anatomic landmarks and is mapped to a relative position of the probe; and

(iii) displaying the predicted position and the recommendation of the next maneuvers;

wherein the method further comprises training the classifier based on a plurality of ICE training images and using a mapping between the plurality of ICE training images and a plurality of ICE catheter positions from which the plurality of ICE training images were acquired as ground truth.

6. The method of claim 5 wherein the current image comprises an intracardiac echocardiogram (ICE) image.

7. The method of claim 5 wherein the next maneuver is represented by navigational instructions.

8. The method of claim 7 wherein the navigational instructions comprise machine instructions that are executable by a robotic controller to automatically steer the probe to a desired position.

9. The method of claim 5 wherein the displaying the predicted position comprises displaying the probe at the predicted position in a graphical representation of the structure of interest.

10. The method of claim 9 wherein the graphical representation of the structure of interest comprises a three-dimensional rendering of the structure of interest.

11. The method of claim 9 wherein the graphical representation of the structure of interest comprises an image-derived model of the structure of interest.

12. The method of claim 9 further comprising repeating steps (i), (ii) and (iii) to update the graphical representation in substantially real time as the probe acquires a new current image at a new position.

13. A system for probe positioning, comprising:

a non-transitory memory device for storing computer readable program code; and

a processor in communication with the non-transitory memory device, the processor being operative with the computer readable program code to perform steps including:

(i) receiving a current image acquired by a probe within a structure of interest,

(ii) predicting a position of the probe and generating a recommendation of a next maneuver to be performed using the probe by applying the current image and a time history of a plurality of position predictions views to a trained classifier, wherein the time history of the plurality of position predictions views comprises a current sequence of a predetermined number of the position predictions views previously generated by the trained classifier and is used by the trained classifier to detect and correct any error in the predicted position, wherein each of the plurality of views is a visualization of a predefined set of one or more anatomic landmarks and is mapped to a relative position of the probe, and

(iii) displaying the predicted position and the recommendation of the next maneuvers;

wherein the steps further comprise training the classifier based on a plurality of ICE training images and using a mapping between the plurality of ICE training images and a plurality of ICE catheter positions from which the plurality of ICE training images were acquired as ground truth.

14. The system of claim 13 wherein the probe comprises an intracardiac echocardiogram (ICE) catheter.

15. The system of claim 13 wherein the processor is operative with the computer readable program code to display the predicted position by displaying the probe at the predicted position in a graphical representation of the structure of interest.

16. The system of claim 15 wherein the graphical representation comprises a plane projection of a three-dimensional rendering of the structure of interest.

17. The system of claim 13 wherein the time history of the plurality of views comprises a home view, an aortic valve view, a left ventricle view, a right ventricular outflow tract view, a left atrium view, a left superior pulmonary vein view, a left inferior pulmonary vein view, an esophagus view, a right pulmonary veins view, or a combination thereof.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Apr 2, 2019
From: SALGAONKAR, VASANT
To: SIEMENS MEDICAL SOLUTIONS USA, INC.
Reel/Frame 048766/0351 →
Continuity (1)
Related Publication 20200315572A1 · Oct 8, 2020
Cited By (2)
US 12,525,145 US 12,605,039