Automated Gamma Knife dose planning using polygon clipping and adaptive simulated annealing

Area of Science:

• Neurosurgery
• Radiation Oncology
• Medical Physics

Background:

• The Gamma Knife is a neurosurgical device for treating brain anomalies with focused radiation.
• Current manual dose planning is laborious, subjective, and often results in suboptimal target coverage.
• Automated methods are needed to improve the efficiency and accuracy of Gamma Knife treatment planning.

Purpose of the Study:

• To present an automated dose planning method for Gamma Knife radiosurgery using patient-specific kernels.
• To evaluate the performance of the automated method in generating conformal dose plans for complex cases.

Main Methods:

• Utilized a forward-direct-method with adaptive simulated annealing and Nelder-Mead simplex optimizers.
• Employed a 2 1/2-D approach, incorporating 2-D contouring and polygon clipping for computational tractability.
• Integrated patient-specific kernels from Elekta's Leksell Gamma Plan software.

Main Results:

• Generated highly conformal and competitive dose plans for two historical clinical datasets.
• The automated system optimized the number of shots, weight, collimator size, and 3-D location.
• Achieved superior target coverage compared to traditional manual planning methods.

Conclusions:

• The automated dose planning approach significantly enhances Gamma Knife treatment precision.
• This method offers a more objective and efficient alternative to manual planning, leading to potentially improved clinical outcomes.
• Further validation on diverse clinical cases is warranted to establish broader applicability.