Source encoding for viscoacoustic ultrasound computed tomography.

Area of Science:

• Medical Imaging
• Computational Physics
• Biomedical Engineering

Background:

• Ultrasound computed tomography (USCT) is a valuable noninvasive imaging technique for breast cancer diagnostics.
• Waveform-based inversions in USCT utilize full ultrasound data for higher resolution compared to travel-time methods.
• Full-physics wave propagation modeling in 3D is computationally intensive, limiting practical application.

Purpose of the Study:

• To propose a computationally efficient method for full-physics 3D waveform inversions in Ultrasound Computed Tomography.
• To mitigate the significant computational cost associated with detailed wave propagation modeling.
• To enable faster and more practical implementation of high-resolution USCT for clinical diagnostics.

Main Methods:

• A crosstalk-free source encoding technique for explicit time-domain solvers was developed.
• Gradient computation was optimized to require only two numerical wave simulations, irrespective of source/receiver count.
• Orthogonal frequencies per source and double-difference measurements were employed to eliminate crosstalk and the need for source time function knowledge.

Main Results:

• The proposed method enables gradient computation using only two wave simulations.
• Crosstalk is eliminated by using orthogonal frequencies for each source.
• Full-physics 3D waveform inversions can be performed rapidly, within minutes.
• The method achieves this efficiency using reasonable computational resources.

Conclusions:

• The developed crosstalk-free source encoding method significantly reduces computational cost for USCT waveform inversions.
• This advancement makes full-physics 3D USCT imaging feasible within clinical time constraints.
• The technique holds promise for improving the resolution and diagnostic capabilities of breast cancer imaging with USCT.