2D myocardial deformation imaging based on RF-based non-rigid image registration.

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Cardiovascular Mechanics

Background:

• Myocardial deformation imaging assesses cardiac function using echocardiography.
• Current methods like speckle tracking and block matching face challenges in lateral motion estimation due to ultrasound limitations.
• Existing non-rigid image registration (NRIR) methods for B-mode images sacrifice the benefits of radio-frequency (RF) data.

Purpose of the Study:

• To develop a novel non-rigid image registration (NRIR) motion estimator specifically designed for radio-frequency (RF) ultrasound data.
• To evaluate the accuracy of the developed RF-based NRIR estimator.
• To compare the performance of the RF-based NRIR estimator against a state-of-the-art block matching (BM) solution.

Main Methods:

• Development of a new non-rigid image registration (NRIR) algorithm tailored for radio-frequency (RF) ultrasound datasets.
• Quantification of the estimator's accuracy using synthetic ultrasound data.
• Clinical application of the developed RF-based NRIR algorithm for in-vivo myocardial motion estimation.

Main Results:

• The developed RF-based NRIR estimator demonstrated superior tracking accuracy compared to the block matching (BM) method.
• The improvement in accuracy was particularly significant in the lateral direction, as hypothesized.
• The algorithm successfully estimated both in-plane velocity components in clinical, in-vivo applications.

Conclusions:

• RF-based NRIR is a promising advancement for myocardial deformation imaging, overcoming limitations of previous methods.
• This technique offers enhanced accuracy in estimating myocardial motion, especially in the challenging lateral direction.
• The developed algorithm has potential for improved clinical assessment of cardiac function through echocardiography.