Modeling of soft tissue thermal damage based on GPU acceleration.

Area of Science:

• Biomedical Engineering
• Computational Biology
• Medical Physics

Background:

• Accurate modeling of thermal damage is crucial for effective hyperthermia treatments to target tumors precisely.
• Current computational methods for predicting soft tissue thermal damage are often computationally intensive, limiting real-time application.

Purpose of the Study:

• To develop and validate a GPU-accelerated methodology for modeling bio-heat conduction and thermal-induced tissue damage in soft tissues.
• To improve the computational efficiency of soft tissue thermal damage prediction for hyperthermia therapies like thermal ablation.

Main Methods:

• Combined Arrhenius Burn integration with Pennes' bio-heat transfer model.
• Employed the Galerkin finite element method for spatial discretization and explicit forward finite difference for temporal discretization.
• Implemented GPU acceleration using High-Level Shader Language (HLSL) for compute shaders.

Main Results:

• The GPU-accelerated finite element method accurately predicts temperature distribution and thermal damage in real time.
• Peak temperatures occur at the heat source, with damage rapidly progressing from the source outwards.
• Achieved a maximum reduction of 55.3 times in computation time compared to CPU execution.

Conclusions:

• The proposed GPU-accelerated method offers a significant improvement in computational performance for soft tissue thermal damage prediction.
• This methodology enables faster and more precise thermal ablation planning, potentially improving patient outcomes in hyperthermia treatments.