In-silico heart model phantom to validate cardiac strain imaging

Area of Science:

• Cardiovascular Imaging and Mechanics
• Computational Biology
• Biomedical Engineering

Background:

• Quantification of cardiac strains is crucial for diagnosing cardiac function but faces challenges due to complex 4D motion.
• Inconsistent strain estimations across imaging modalities and algorithms limit the clinical utility of cardiac strains as biomarkers.
• A reliable benchmark is needed to validate 4D regional strain quantification methods.

Purpose of the Study:

• To develop and validate an in-silico heart phantom for accurate 4D regional strain quantification.
• To establish a benchmark for assessing the reliability of strain calculation algorithms.
• To investigate the impact of image quality on regional strain calculations.

Main Methods:

• Developed an in-silico heart phantom using finite element (FE) simulations.
• Synthesized dynamic magnetic resonance (MR) images from FE models of cardiac kinematics.
• Calculated 4D regional strains using a non-rigid image registration (NRIR) framework.
• Conducted in-silico experiments to assess the effects of image quality on strain calculations.

Main Results:

• Demonstrated recovery of "ground-truth" twist angles in a proof-of-concept torsion model.
• Successfully synthesized dynamic MR images of mouse cardiac kinematics.
• Quantified 4D regional strains using the NRIR framework in simulated cardiac models.
• Evaluated the influence of image quality on regional strain accuracy.

Conclusions:

• The proposed in-silico heart phantom provides a rigorous and feasible tool for validating 4D regional strain calculations.
• Standardizing regional strain quantification can enhance its clinical impact as a biomarker for cardiac dysfunction.
• This approach facilitates the reliable assessment of cardiac mechanics and dysfunction.