Method and apparatus with user guidance and automated image setting selection for mitral regurgitation evaluation
Ultrasound imaging system and method for improving MR quantification is described. An ultrasound imaging system according to some embodiments may be programmed to execute one or more routines and use one or more trained predictive models to provide user guidance and/or automation for positioning the ultrasound probe to acquire a desired cardiac view and tune the imaging settings of the system for each measurement of the MR quantification process.
1 . An ultrasound imaging system for quantifying mitral regurgitation of a heart of a subject, the system comprising:
an ultrasound imaging device including a probe configured to transmit ultrasound into the subject in accordance with acquisition settings and to receive ultrasound echoes for generating ultrasound images of the heart;
a display configured to display one or more of the ultrasound images; and
a processor in communication with the ultrasound imaging device and the display, wherein the processor is configured to:
receive one or more ultrasound images of the heart that show a mitral regurgitation (MR) jet;
provide a graphical user interface (GUI) for an MR exam, wherein the GUI is configured to enable a user to make a selection of an automation setting and a measurement from a plurality of predetermined measurements consisting of vena contracta width (VCW), regurgitant volume (RVol), regurgitant fraction (RF) and effective regurgitant orifice area (EROA);
receive a user input via the GUI, wherein the user input comprises the selected automation setting and the selected measurement; and
for one or more of the selected measurements of the plurality of measurements, and based on the one or more images, provide measurement-specific user guidance on the display for at least one of positioning the probe, adjusting acquisition settings, and determining a target frame from a sequence of ultrasound images of the MR jet for obtaining the selected measurement.
2 . The system of claim 1 , wherein the processor is configured to determine whether the one or more ultrasound images show a target cardiac view, wherein the target cardiac view depends on the selected measurement, and to provide guidance to the user for positioning the probe with respect to the subject to acquire the target cardiac view.
3 . The system of claim 2 , wherein the processor is configured, responsive to selection of the RVol or RF, to:
determine whether the one or more ultrasound images show apical 4-chamber (A4C) view; and
upon determining that the one or more ultrasound images do not show a A4C view, provide user guidance for positioning the probe to acquire the A4C view.
4 . The system of claim 1 , wherein the selection of the automation setting comprises selecting an automated tissue specific preset (auto-TSP) mode, and wherein the processor is configured to:
automatically adjust the acquisition settings to processor-estimated MR-specific TSP settings if the auto-TSP mode is selected; and
provide guidance on the display for manually adjusting acquisition settings if the auto-TSP mode is not selected.
5 . The system of claim 4 , wherein the processor is configured to estimate the MR-specific TSP settings using a predictive model trained to generate a multi-parameter output comprising a set of estimated optimal values for each of a plurality of acquisition settings selected from color gain, Nyquist shift, frame rate and image gain.
6 . The system of claim 4 , wherein the processor is configured to use at least one predictive model to classify at least one acquisition setting associated with the image, wherein the classification indicates suitability of the at least one acquisition setting for MR evaluation, the processor being further configured to display a visual indicator of the suitability concurrently with the image.
7 . The system of claim 1 , wherein the processor is further configured to automatically obtain the selected measurements from the determined target frame.
8 . The system of claim 1 , wherein the processor is configured, responsive to selection of the VCW, to:
a direction of the MR jet from the one or more ultrasound images;
provide user guidance for adjusting the probe to a position in which the probe is substantially parallel to the direction of the MR jet;
acquire the sequence of ultrasound images spanning at least one full phase of a cardiac cycle of the heart; and
the target frame for measuring the VCW from the sequence of ultrasound images spanning at least one full phase of a cardiac cycle of the heart.
9 . The system of claim 8 , wherein the processor is configured, prior to determining the direction of the MR jet, to:
whether the one or more ultrasound images show a parasternal long-axis view; and
upon determining that the one or more ultrasound images do not show a PLAX view, provide user guidance for positioning the probe to acquire the PLAX view.
10 . The system of claim 9 , wherein the processor is configured to provide the one or more ultrasound images to a predictive model trained to identify components of the MR jet from ultrasound images, and wherein the processor determines the direction of the MR jet based on the components of the MR jet identified by the deep learning model.
11 . The system of claim 10 , wherein the processor is configured, when providing user guidance for adjusting the probe, to compare an azimuthal direction of the probe to the direction of the MR jet, and provide a visual indicator of alignment of the probe to the direction of the MR jet.
12 . The system of claim 1 , wherein the processor uses one or more predictive models to provide the measurement-specific user guidance on the display.
13 . The system of claim 12 , wherein the one or more predictive models include a first model configured to identify components of the MR jet in images input to the first model, at least one second predictive model configured to classify one or more imaging settings of the input image.
14 . The system of claim 13 , wherein the at least one second predictive model comprises a trained neural network configured to provide a multi-parameter output comprising a classification for a plurality of image settings as a set, wherein the classification indicates either a quality of each of the image settings or an estimated optimal value for each image setting.
15 . The system of claim 14 , wherein the trained neural network is configured to output, for each input image, a classification for each of color gain, Nyquist shift, frame rate and image gain.