Systematic Performance Evaluation of a Cross-Correlation-Based Ultrasound Strain Imaging Method

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Acoustics

Background:

• Accurate estimation of lateral tissue motion is a significant challenge in 2-D ultrasound strain imaging (USI).
• Existing methodologies for in-plane displacement and strain estimation in USI have limitations, with optimal parameter selection scattered in literature.
• A systematic approach is needed to understand and optimize 2-D USI for retrieving the full strain tensor.

Purpose of the Study:

• To systematically investigate a 2-D cross-correlation-based USI method.
• To incorporate additional estimation strategies to enhance in-plane displacement and strain estimation accuracy.
• To determine optimal algorithmic parameters for various tissue deformation scenarios.

Main Methods:

• Revisiting a previously developed cross-correlation-based USI method.
• Investigating performance using different matching kernel sizes (axial: 1λ–14λ, lateral: 1–13 pitches) and data formats (radiofrequency, envelope).
• Utilizing Field II simulations with coherent plane wave compounding for kinematic scenarios (normal, shear, hybrid deformation), validated with phantom and in vivo data.

Main Results:

• For radiofrequency-based USI, optimal kernel sizes varied with strain type: smaller axial/larger lateral for normal strains, larger along shear/smaller orthogonal for shear strains.
• Envelope-based USI showed more relaxed kernel size requirements.
• A compromise in kernel sizes is necessary for accurate strain component estimation in complex deformations.

Conclusions:

• The study provides practical strategies for accurate motion estimation using 2-D cross-correlation-based USI.
• Optimal parameter selection is crucial and dependent on deformation type and data format.
• The findings are validated through phantom and preliminary in vivo studies, advancing USI capabilities.