Temporal diffeomorphic free-form deformation for strain quantification in 3D-US images

Area of Science:

• Medical imaging
• Biomechanical analysis
• Computational anatomy

Background:

• Accurate quantification of cardiac motion and strain is crucial for diagnosing cardiovascular diseases.
• Existing temporal registration methods can be sensitive to noise and may lack time consistency.
• 3D+t Ultrasound (US) imaging provides rich spatio-temporal data for cardiac analysis.

Purpose of the Study:

• To develop and validate a novel diffeomorphic temporal registration algorithm for motion and strain quantification in 3D+t cardiac US images.
• To assess the algorithm's robustness to noise and its clinical applicability in healthy subjects and Cardiac Resynchronization Therapy (CRT) patients.

Main Methods:

• A new diffeomorphic temporal registration algorithm using forward Eulerian integration of a non-stationary velocity field.
• Representation of the velocity field as a sum of continuous spatiotemporal B-Spline kernels for time consistency.
• Validation on synthetic US images with known motion and noise, followed by application to clinical 3D+t US datasets of the left ventricle.

Main Results:

• The proposed algorithm exhibited greater robustness to noise compared to a simple pairwise registration strategy.
• In healthy cases, the algorithm produced uniform regional strain patterns consistent with clinical literature.
• In CRT patients, the post-treatment strain patterns normalized in agreement with clinical outcomes.

Conclusions:

• The developed diffeomorphic temporal registration algorithm accurately quantifies motion and strain from 3D+t cardiac US images.
• The method is robust to noise and shows clinical relevance for assessing cardiac function in healthy and CRT patient populations.
• This technique offers a promising tool for non-invasive cardiac assessment and treatment monitoring.