Cardiac flow detection based on morphological modeling in medical diagnostic ultrasound imaging
For cardiac flow detection in echocardiography, by detecting one or more valves, sampling planes or flow regions spaced from the valve and/or based on multiple valves are identified. A confidence of the detection may be used to indicate confidence of calculated quantities and/or to place the sampling planes.
1. A method for detecting cardiac flow in echocardiography, the method comprising:
acquiring, by an ultrasound scanner, one or more volumetric ultrasound scans including B-mode data of a heart of a patient over time;
detecting, by an image processor, two or more heart valves over time from the B-mode data of the one or more volumetric ultrasound scans, wherein detecting comprises detecting with a machine-learnt classifier;
fitting a patient specific valve model to the detected two or more heart valves;
determining, by the image processor, a confidence value indicating an accuracy of the detecting of the mitral annulus and the left ventricle outflow tract in the fit patient specific valve model, wherein determining the confidence comprises determining the confidence with the machine-learnt classifier;
placing a first sampling plane for the mitral annulus and a second sampling plane for the left ventricle outflow tract based on the fit patient specific valve model;
tracking the first and second sampling planes over time based on optical flow, boundary detection, and motion prior;
calculating, by the image processor, a cardiac flow value from flow data of the volumetric ultrasound scans for the first and second sampling planes over time, the calculating of the cardiac flow value limited to avoid flow during a portion of a heart cycle; and
outputting an image of the cardiac flow value;
wherein the image indicates the confidence value for the first and second sampling planes.
2. The method of claim 1 wherein detecting the two or more valves comprises detecting a location, orientation, and scale of each of the mitral annulus and the left ventricle outflow tract represented by the B-mode data.
3. The method of claim 1 wherein placing comprises placing the one or more sampling planes in a cardiac flow region spaced from the mitral annulus and the left ventricle outflow tract.
4. The method of claim 3 wherein placing comprises placing the one or more sampling planes to avoid quantifying flow from other valves, the placing comprising determining intersection with valve models.
5. The method of claim 1 wherein placing comprises placing based on the detecting of the mitral annulus and the left ventricle outflow tract at one time and tracking the one or more sampling planes for other times.
6. The method of claim 1 wherein calculating is limited to avoid regurgitant flow during a portion of a heart cycle.
7. The method of claim 1 wherein calculating comprises calculating the cardiac flow value for as least one of in-flow, out-flow, or stroke volume.
8. The method of claim 1 wherein outputting comprises outputting the image of models of the detected the mitral annulus and the left ventricle outflow tract and a quantity for the cardiac flow value.
9. The method of claim 1 wherein outputting comprises outputting with the image including an indication of the confidence value.
10. The method of claim 1 wherein placing comprises placing based on the confidence value.
11. The method of claim 1 further comprising refining the placement of the first or second sampling planes based on at least one of aliased flow, relative flow, maximum flow, a smoothness of outflow as a function of time, or a smoothness of inflow as a function of time.