Method and system for automatic detection of coronary stenosis in cardiac computed tomography data

Patent metadata

Field	Value
Publication number	US-9761004-B2
Application number	US-48722309-A
Country	US
Kind code	B2
Filing date	Jun 18, 2009
Priority date	Sep 22, 2008
Publication date	Sep 12, 2017
Grant date	Sep 12, 2017

Abstract

A method and system for automatic coronary stenosis detection in computed tomography (CT) data is disclosed. Coronary artery centerlines are obtained in an input cardiac CT volume. A trained classifier, such as a probabilistic boosting tree (PBT) classifier, is used to detect stenosis regions along the centerlines in the input cardiac CT volume. The classifier classifies each of the control points that define the coronary artery centerlines as a stenosis point or a non-stenosis point.

First claim

The invention claimed is:

1. A method for detecting coronary stenosis in a cardiac computed tomography (CT) volume, comprising:
obtaining at least one coronary artery centerline in said CT volume, said at least one coronary artery centerline comprising a plurality of control points;
detecting stenosis regions along said at least one coronary artery centerline in said CT volume by classifying each of said plurality of control points as a stenosis point or a non-stenosis point using a trained machine learning based classifier; and
displaying CT images with a visual indication of the control points classified as stenosis points.

2. The method of claim 1 , wherein said machine learning based trained classifier is a probabilistic boosting tree (PBT) classifier trained based on training data.

3. The method of claim 2 , wherein said training data comprises a plurality of training cardiac CT volumes having annotated stenosis regions and annotated coronary artery centerlines comprising a plurality of control points, and said PBT classifier is trained by sampling a plurality of sampling points surrounding each control point of the coronary artery centerlines of the training CT volumes and extracting plurality of local features at each sampling point.

4. The method of claim 1 , wherein said step of classifying each of said plurality of control points as a stenosis point or a non-stenosis point using said trained machine learning based classifier comprises:
determining, for each of said plurality of control points, a probability of being in a stenosis region, using said trained machine learning based classifier; and
classifying each of said plurality of control points as a stenosis point or a non-stenosis point based on the determined probability and a threshold.

5. The method of claim 1 , wherein said at least one coronary artery centerline comprises Left Main (LM), Left Anterior Descending (LAD), Left Circumflex (LCX), and Right Coronary Artery (RCA) coronary artery trunk centerlines.

6. The method of claim 5 , wherein each of said LM, LAD, LCX, and RCA coronary artery trunk centerlines comprises a plurality of control points, and said step of detecting stenosis regions along said at least one coronary artery centerline comprises:
classifying each of said plurality of control points of each of said LM, LAD, LCX, and RCA coronary artery trunk centerlines as a stenosis point or a non-stenosis point using said trained machine learning based classifier.

7. The method of claim 6 , wherein said LM coronary artery trunk centerline comprises eight uniformly distributed control points, and said LAD, LCX, and RCA coronary artery centerlines each comprise 64 uniformly distributed control points.

8. The method of claim 1 , wherein said step of obtaining at least one coronary artery centerline in said CT volume comprises:
automatically obtaining said at least one coronary artery centerline in said CT volume using machine learning based coronary artery detection.

9. The method of claim 1 , further comprising:
displaying said CT volume with the detected stenosis regions.

10. An apparatus for detecting coronary stenosis in a cardiac computed tomography (CT) volume, comprising:
a processor; and
a memory storing computer program instruction, which when executed by the processor cause the processor to perform operations comprising:
obtaining at least one coronary artery centerline in said CT volume, said at least one coronary artery centerline comprising a plurality of control points;
detecting stenosis regions along said at least one coronary artery centerline in said CT volume by classifying each of said plurality of control points as a stenosis point or a non-stenosis point using a trained machine learning based classifier; and
displaying CT images with a visual indication of the control points classified as stenosis points.

11. The apparatus of claim 10 , wherein said classifying each of said plurality of control points as a stenosis point or a non-stenosis point using said trained machine learning based classifier comprises:
determining, for each of said plurality of control points, a probability of being in a stenosis region using said trained machine learning based classifier; and
classifying each of said plurality of control points as a stenosis point or a non-stenosis point based on the determined probability and a threshold.

12. The apparatus of claim 10 , wherein said at least one coronary artery centerline comprises Left Main (LM), Left Anterior Descending (LAD), Left Circumflex (LCX), and Right Coronary Artery (RCA) coronary artery trunk centerlines.

13. The apparatus of claim 10 , wherein said obtaining at least one coronary artery centerline in said CT volume comprises:
automatically obtaining said at least one coronary artery centerline in said CT volume using machine learning based coronary artery detection.

14. The apparatus of claim 10 , wherein said operations further comprise:
displaying said CT volume with the detected stenosis regions.

15. A non-transitory computer readable medium encoded with computer executable instructions for detecting coronary stenosis in a cardiac computed tomography (CT) volume, the computer executable instructions defining steps comprising:
obtaining at least one coronary artery centerline in said CT volume, said at least one coronary artery centerline comprising a plurality of control points;
detecting stenosis regions along said at least one coronary artery centerline in said CT volume by classifying each of said plurality of control points as a stenosis point or a non-stenosis point using a trained machine learning based classifier; and
displaying CT images with a visual indication of the control points classified as stenosis points.

16. The non-transitory computer readable medium of claim 15 , wherein said trained machine learning based classifier is a probabilistic boosting tree (PBT) classifier trained based on training data.

17. The non-transitory computer readable medium of claim 15 , wherein the computer executable instructions defining the step of classifying each of said plurality of control points as a stenosis point or a non-stenosis point using said trained classifier comprise computer executable instructions defining the steps of:
determining, for each of said plurality of control points, a probability of being in a stenosis region, using said trained machine learning based classifier; and
classifying each of said plurality of control points as a stenosis point or a non-stenosis point based on the determined probability and a threshold.

18. The non-transitory computer readable medium of claim 15 , wherein said at least one coronary artery centerline comprises Left Main (LM), Left Anterior Descending (LAD), Left Circumflex (LCX), and Right Coronary Artery (RCA) coronary artery trunk centerlines.

19. The non-transitory computer readable medium of claim 18 , wherein each of said LM, LAD, LCX, and RCA coronary artery trunk centerlines comprises a plurality of control points, and the computer executable instructions defining the step of detecting stenosis regions along said at least one coronary artery centerline comprise computer executable instructions defining the step of:
classifying each of said plurality of control points of each of said LM, LAD, LCX, and RCA coronary artery trunk centerlines as a stenosis point or a non-stenosis point using said trained machine learning based classifier.

20. The non-transitory computer readable medium of claim 15 , wher

Assignees

• Mittal Sushil
• Zheng Yefeng
• Georgescu Bogdan
• Vega-Higuera Fernando
• Comaniciu Dorin
• Siemens Healthcare Gmbh

Inventors

• Mittal Sushil
• Zheng Yefeng
• Georgescu Bogdan
• Vega-Higuera Fernando
• Comaniciu Dorin

Classifications

• G06T2207/30101

  Blood vessel; Artery; Vein; Vascular · CPC title

• G06T7/13 Primary

  Edge detection · CPC title

• G06T2207/20044

  Skeletonization; Medial axis transform · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T2207/30048

  Heart; Cardiac · CPC title