Mitral valve detection for transthoracic echocardiography
View Patent ↗A mitral valve is detected in transthoracic echocardiography. The ultrasound transducer is positioned against the chest of the patient rather than being inserted within the patient. While data acquired from such scanning may be noisier or have less resolution, the mitral valve may still be automatically detected. Using both B-mode data representing tissue as well as flow data representing the regurgitant jet, the mitral valve may be detected automatically with a machine-learnt classifier. A series of classifiers may be used, such as determining a position and orientation of a valve region with one classifier, determining a regurgitant orifice with another classifier, and locating mitral valve anatomy with a third classifier. One or more features for some of the classifiers may be calculated based on the orientation of the valve region.
1. A method for detecting a mitral valve in transthoracic echocardiography, the method comprising the acts of:
scanning, with a transducer adjacent a patient, a cardiac region of a patient with ultrasound;
detecting, with a B-mode detector and in response to the scanning, B-mode data representing tissue in the cardiac region;
estimating, with a flow estimator and in response to the scanning, flow data representing fluid in the cardiac region, the flow data comprising energy, velocity, or energy and velocity;
calculating, with a processor, feature values from both the B-mode data and the flow data;
applying, with the processor, the feature values from both the B-mode data and the flow data to a first machine-learnt classifier, the first machine-learnt classifier indicating a global region of the mitral valve, the mitral valve comprising a tissue structure;
applying, with the processor, the feature values from both the B-mode data and the flow data to a second machine-learnt classifier, the second machine learnt classifier indicating a regurgitant orifice, the regurgitant orifice constrained to be within the global region of the mitral valve;
identifying, with the processor, a mitral annulus and closure line as a function of a third machine-learnt classifier and orientation and scale of the global region;
transforming, with the processor, a statistical shape model as a function of the mitral annulus and closure line; and
outputting to a display device a valve image, the valve image being a function of the transformed statistical shape model.
2. The method of claim 1 wherein scanning comprises performing transthoracic echocardiography, wherein detecting comprises detecting intensities of acoustic echoes from tissue, and wherein estimating comprises estimating Doppler velocity.
3. The method of claim 1 wherein calculating the feature values comprises calculating Haar, steerable, or Haar and steerable feature values for the B-mode data and for the flow-data.
4. The method of claim 1 wherein applying to the first machine-learnt classifier comprises applying to a marginal space learnt classifier, the global region indicated with sequentially determined translation, rotation, and scale along three-dimensions.
5. The method of claim 1 wherein applying to the second machine-learnt classifier comprises automatically locating the regurgitant orifice free of user indication of location in a cardiac image.
6. The method of claim 1 wherein the regurgitant orifice has a joint probability that is a function of the first and second machine-learnt classifiers.
7. The method of claim 1 further comprising:
segmenting a mitral regurgitation jet with the regurgitant orifice as a seed; and
quantifying the mitral regurgitation jet from the segmenting.
8. The method of claim 1 wherein identifying the mitral annulus and closure line comprises sequentially classifying points in the global region with input features of the third machine-learnt classifier being determined as a function of the orientation and the scale, the mitral annulus and closure line comprising discrete ones of the points with relatively greater probabilities.
9. The method of claim 1 wherein transforming comprises fitting modeled annulus and modeled line from the statistical shape model to the mitral annulus and closure line with a thin plate spline.
10. The method of claim 1 wherein outputting comprises outputting the valve image as a three-dimensional rendering with at least one anatomical location of the mitral valve highlighted relative to other tissue, the at least one anatomical location determined from the transformed statistical shape model.
11. The method of claim 1 wherein outputting comprises outputting the valve image with a regurgitant jet in the valve image.
12. The method of claim 1 wherein the acts are performed prior to a mitral clipping procedure and are repeated after the mitral clipping procedure.