Real-Time Acoustic Simulation Framework for tFUS: A Feasibility Study Using Navigation System

Area of Science:

• Biomedical Engineering
• Medical Physics
• Neuroscience

Background:

• Transcranial focused ultrasound (tFUS) offers non-invasive brain therapy but faces accuracy issues due to skull-induced acoustic aberrations.
• Current image-guided navigation systems struggle with skull distortions, necessitating computationally intensive acoustic simulations.
• Real-time acoustic simulation is crucial for accurate tFUS treatment but is hindered by high computational costs.

Purpose of the Study:

• To develop and validate a neural network-based framework for real-time acoustic simulation in tFUS.
• To implement a simulation-guided navigation (SGN) system integrating this real-time simulation with existing navigation technologies.
• To assess the accuracy and speed of the proposed SGN system for predicting intracranial acoustic fields.

Main Methods:

• A 3D conditional generative adversarial network (3D-cGAN) with residual blocks and multiple loss functions was employed for real-time acoustic simulation.
• The 3D-cGAN model was trained using data from the conventional k-Wave numerical acoustic simulation program.
• The SGN system integrated the 3D-cGAN simulation with standard image-guided navigation.

Main Results:

• The SGN system achieved a real-time frame rate of 5 Hz (approximately 0.2 seconds per simulation).
• Numerical validation showed average peak intracranial pressure and focus position errors of 6.8 ± 5.5% and 5.3 ± 7.7 mm, respectively.
• Experimental validation with a skull phantom yielded average errors of 4.5% for peak pressure and 6.6 mm for focus position.

Conclusions:

• The proposed neural network-based SGN system enables accurate, real-time prediction of intracranial acoustic fields during tFUS.
• This advancement can significantly improve the precision and efficacy of non-invasive therapeutic ultrasound applications in the brain.
• The SGN system demonstrates feasibility for real-time guidance in tFUS procedures, overcoming limitations of conventional methods.