Irreversible electroporation (IRE) is a promising non-thermal tumor ablation method, but achieving optimal outcomes is challenging due to complex patient-specific 3D anatomies. Effective treatment requires synergistic optimization of both electrode placement and electrical parameters. To address this, this study proposes a multi-objective sparrow search algorithm based on evolutionary reference-points (ERM-SSA), which utilizes a 3D electric field distribution model incorporating heterogeneous tissue conductivity. Key optimization variables include the number of electrode needles, needles’ position and voltage. ERM-SSA establishes a multi-objective optimization model aiming to maximize tumor ablation rate while minimizing damage to the surrounding healthy tissues. To improve the optimization performance, ERM-SSA introduces several key improvements based on the standard sparrow search algorithm (SSA). First, Poisson disk sampling is used for population initialization to enhance the diversity of solutions. Secondly, an adaptive reference point strategy based on the electrode needle position is designed to improve the search guidance ability in high-dimensional space. Furthermore, its position update formula is optimized to make it more suitable for solving multi-objective problems. Finally, the improved sinusoidal perturbation strategy is integrated to effectively enhance its ability to jump out of the local optimum. Experimental results demonstrate that ERM-SSA achieves remarkable optimization performance across various tumor sizes and clinical constraints. This study introduces a fully automated, high-precision 3D optimization tool, offering a significant advancement for clinical IRE treatment planning. Moreover, its dynamic search mechanism provides novel and valuable insights for the clinical application implementation of multi-objective optimization algorithms.
原文链接
分享
引用
