Scholarlink

<p>The goal of this study is to propose a new intensity-modulated radiation therapy (IMRT) fluence map optimization method that incorporates dose-volume constraints (DVCs). Instead of optimizing the fluence map by treating DVCs as hard constraints, which makes them difficult to incorporate into gradient-based optimization, we approximate the DVCs using a sigmoid function. This function is differentiable and has a gradient across its entire domain, thereby allowing us to perform fluence map optimization through gradient-based methods. Additionally, we propose a three-stage optimization scheme that minimizes the impact of the sigmoid function’s non-convexity and enables finding a feasible solution. By applying proposed method to the TG119 phantom cases, we aimed to assess its suitability for various clinical cases. The optimization of four TG119 Phantom cases successfully produce fluence maps that satisfy the DVCs in every case.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fluence map optimization with dose-volume constraints in intensity modulated radiation therapy by using sigmoid function

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Sang Won Kang,
  • Jihun Kim,
  • Gil Ho Yoon

摘要

The goal of this study is to propose a new intensity-modulated radiation therapy (IMRT) fluence map optimization method that incorporates dose-volume constraints (DVCs). Instead of optimizing the fluence map by treating DVCs as hard constraints, which makes them difficult to incorporate into gradient-based optimization, we approximate the DVCs using a sigmoid function. This function is differentiable and has a gradient across its entire domain, thereby allowing us to perform fluence map optimization through gradient-based methods. Additionally, we propose a three-stage optimization scheme that minimizes the impact of the sigmoid function’s non-convexity and enables finding a feasible solution. By applying proposed method to the TG119 phantom cases, we aimed to assess its suitability for various clinical cases. The optimization of four TG119 Phantom cases successfully produce fluence maps that satisfy the DVCs in every case.