Assessing heart parameters using neural networks
An embodiment of the invention provides a system for automatically deriving a parameter of a human heart from ultrasound results. A first neural network is arranged to receive a plurality of echocardiographic images and to classify the images into one of at least a two-chamber view and a four-chamber view. A second neural network is arranged to receive echocardiographic images comprising a two- or four-chamber view and to identify the endocardial border of the left ventricle (LV) for each view. End-systole and end-diastole images are then identified and a parameter such as LV volume, ejection fraction, global longitudinal strain and regional longitudinal strain is calculated.
1. A system for providing a parameter of a human heart, the system comprising:
a first trained neural network arranged to:
receive a plurality of echocardiographic images;
identify at least one of a two-chamber view or a four-chamber view in the plurality of echocardiographic images;
a third responder arranged to:
determine a confidence measure for the given view, wherein the confidence measure is indicative of a degree of confidence that the identified view is correct;
select, based on the confidence measure exceeding a predetermined threshold, a subset of images for the given view; and
generate a first output based on the subset of images;
a second trained neural network arranged to:
receive the first output
contour the first output by identifying the endocardial border of the left ventricle for the given view; and
generate a second output based on the contoured first output;
a first responder arranged to:
receive the second output;
identify end-systole and end-diastole images based on the second output; and
generate a third output based on the end-systole and end-diastole images;
a second responder arranged to:
receive the third output; and
derive at least one first parameter of the heart based on the third output, wherein the at least one first parameter is Global Longitudinal Strain, principal strain, shear strain, or regional strain.
2. A system as claimed in claim 1 , wherein the second responder is arranged to combine the two-chamber and four-chamber endocardial borders of the left ventricle to derive volume-based parameters.
3. A system as claimed in claim 1 , wherein the second responder is further configured to:
derive one first parameter of the heart based on the end-systole and end-diastole images, wherein the one first parameter is Global Longitudinal Strain, principal strain, shear strain, or regional strain; and
derive a second parameter of the heart based on the end-systole and end-diastole images, wherein the second parameter of the heart is the ejection fraction.
4. A system as claimed in claim 1 , wherein the regional strain parameter comprises six regions.
5. A system as claimed claim 1 , wherein the plurality of echocardiographic images comprise a plurality of non-contrast images.
6. A system as claimed in claim 1 , wherein the plurality of echocardiographic images comprise a plurality of contrast echocardiographic images.
7. A method of training a system as claimed in claim 1 , wherein the neural networks are trained on non-contrast data.
8. A method of training a system as claimed in claim 1 , wherein the neural networks are trained on contrast data.
9. A method of diagnosing a pathology of a heart comprising:
receiving a plurality of echocardiographic images from a subject;
using the system according to claim 1 to analyse the plurality of echocardiographic images to derive a parameter of the heart;
comparing the parameter of the heart to at least one predetermined threshold parameter; and
detecting a difference between the parameter of the heart and the at least one predetermined threshold parameter, said difference being indicative of a pathology of the heart.
10. A method of treating a pathology of a heart comprising:
receiving a plurality of echocardiographic images from a subject, using the system according to claim 1 to analyse the plurality of echocardiographic images to derive a parameter of the heart;
comparing the parameter to at least one threshold parameter;
detecting a difference between the parameter of the heart and the at least one predetermined threshold parameter, said difference being indicative of a pathology of the heart, and
administering a therapeutically effective amount of a drug to the subject, the drug alleviating one or more symptoms of the pathology of the heart.
11. A system as claimed in claim 1 , wherein the first responder is further configured to:
identify one or more cardiac cycles, wherein the third output is further based on the identified one or more cardiac cycles.