Quantification of aortic stenosis based on the morphology of Doppler ultrasound signals using image processing techniques

Abstract

A method was developed to evaluate the morphologic structure of continuous wave Doppler ultrasound signals in order to quantify aortic valve stenosis in the human heart. The systolic peaks of the Doppler signal were assigned digital images and the stability of shape information was tested for patients with low stroke volume and aortal fibrillation. The results indicate that the shape of the signal peaks remains stable despite variations in amplitude and duration. The shape information was then used to quantify the severity of aortic stenosis by optimally matching Doppler peak images to selected templates representing typical diagnostic patterns. The developed software is capable of automatic extraction of shape information from Doppler ultrasound signals in order to support the clinical decision about valve functioning and replacement.