Acceleration of high resolution temperature based optimization for hyperthermia treatment planning using element...

Area of Science:

• Medical Physics
• Biomedical Engineering
• Computational Science

Background:

• Regional hyperthermia requires optimized applicator settings for effective treatment.
• Previous high-resolution temperature-based optimization methods were computationally intensive.

Purpose of the Study:

• To propose and evaluate a speed-up method for high-resolution temperature-based optimization in regional hyperthermia.
• To assess the effectiveness of element grouping in reducing computation time while maintaining treatment accuracy.

Main Methods:

• Element grouping technique using eigenvalues/eigenvectors of precomputed temperature matrices.
• Estimation of high-resolution temperatures and eigenvalues using low-resolution computations and zooming.
• Application of four different grouping criteria with constraints on average group temperatures for five patients using the AMC-8 system.

Main Results:

• Element grouping reduced computation time significantly, with speed-up factors of 1.8-3.5 for 5-10 runs, converging to 4.3-8.5 for many runs.
• Strict grouping criteria were necessary to avoid exceeding normal tissue constraints (approx. 2°C).
• Tumor temperatures remained comparable, with acceptable maximum exceeding of constraints (0.09-0.34°C).

Conclusions:

• High-resolution temperature-based optimization using element grouping offers a substantial speed-up (4-8x) for regional hyperthermia treatment planning.
• This method achieves this efficiency without significant deviations from conventional approaches, making it a viable enhancement.