Realtime Automatic Assessment of Cardiac Function in Echocardiography

Area of Science:

• Cardiology
• Biomedical Engineering
• Medical Imaging

Background:

• Echocardiography assessment of cardiac function is difficult for nonexperts.
• Mitral annular excursion (MAE) and velocities are crucial for diagnosing heart failure in patients with dyspnea.

Purpose of the Study:

• To develop an automatic algorithm for estimating MAE and peak velocities.
• To assist nonexperts in providing preliminary diagnoses of heart failure.

Main Methods:

• An automatic algorithm was developed to localize mitral annular points using a deformable ventricle model and tissue Doppler data.
• MAE and peak systolic/diastolic velocities were estimated automatically.
• Algorithm performance was validated against cardiologist measurements in 367 tissue Doppler recordings.

Main Results:

• The algorithm achieved an estimation error standard deviation of 2.1 mm for MAE, comparable to expert agreement.
• Classification of reduced MAE (<10 mm) was accurate in 90% of cases, with 83% sensitivity and 92% specificity.
• Estimation errors for peak velocities were approximately 1 cm/s.

Conclusions:

• The developed algorithm demonstrates good accuracy for assessing cardiac function via MAE and velocities.
• This automated approach can aid nonexperts in preliminary heart failure diagnosis.
• The algorithm's accuracy is comparable to interobserver variability in clinical practice.