Deconvolution approaches for energy-preserving spatiotemporal cavitation dosimetry using passive acoustic mapping

Area of Science:

• Medical Imaging
• Acoustic Physics
• Biomedical Engineering

Background:

• Passive acoustic mapping (PAM) is crucial for quantifying cavitation treatments.
• Current PAM methods suffer from imaging artifacts due to point spread functions (PSFs), limiting spatial resolution.
• Existing deconvolution techniques like Lucy-Richardson deconvolution (LRD) reduce artifacts but obscure source distribution and introduce biases.

Purpose of the Study:

• To investigate errors in PAM-derived cavitation dosimetry caused by PSF variations.
• To develop and evaluate novel deconvolution algorithms for improved spatial accuracy and source localization in PAM.
• To present a shift-variant LRD (svLRD-PAM) for enhanced cavitation dosimetry.

Main Methods:

• Developed four deconvolution algorithms to address PAM imaging artifacts.
• Implemented a shift-variant LRD (svLRD-PAM) accounting for spatial PSF variations.
• Validated algorithms using numerical simulations and experimental data (in vitro and ex vivo porcine liver).

Main Results:

• svLRD-PAM significantly improved spatial accuracy and reduced imaging artifacts by 4-5 fold compared to standard methods.
• The shift-variant approach maintained accuracy while only modestly increasing computation time (36%).
• Tested algorithms demonstrated varied performance in accuracy, robustness, and computational load.

Conclusions:

• Shift-variant LRD (svLRD-PAM) effectively preserves source localization and improves spatial accuracy in PAM-based cavitation dosimetry.
• svLRD-PAM offers a promising advancement for accurate, spatially resolved cavitation dose measurements.
• This method addresses key limitations in current PAM techniques for clinical applications.