Scholarlink

Purpose: <p>Computed tomography (CT)-guided percutaneous thermal ablation is widely used for minimally invasive treatment of liver tumors where needle path planning critically affects treatment efficacy and safety. This study explores the feasibility of deep reinforcement learning (DRL) for automating needle path planning, framing it as a one-shot anatomical target search problem.</p> Methods: <p>A DRL framework was developed comprising a cylindrical conformal environment derived from CT-based three-dimensional models of key anatomical structures (skin, bones, liver, tumor, vessels, spleen), a continuous action space for entry parameters <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\((z, \alpha )\)</EquationSource> </InlineEquation>, and a clinically informed reward function integrating five anatomical safety constraints and terminal rewards. The proximal policy optimization (PPO) algorithm was adapted to this non-sequential, constraint-rich search space and trained on clinically annotated cases from a public CT dataset, with the remaining cases for validation.</p> Results: <p>Compared with the conventional rapidly-exploring random tree (RRT) method, the proposed approach achieved a 21.8% median improvement in path quality and resulted in reduced inference time (0.12&#xa0;±0.03 s vs. 3.45±&#xa0;0.21 s).</p> Conclusion: <p>These results demonstrate the feasibility and efficiency of DRL-based one-shot anatomical target search for CT-guided percutaneous liver tumor ablation.</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

Deep Reinforcement Learning Based Automatic Needle Path Planning for CT-guided Percutaneous Thermal Ablation of Liver Tumors: A Feasibility Study

  • 期刊论文

  • 发表日期：

    2026年6月22日

查看全文
  • Feifei Ding,
  • Wujun Jiang,
  • Shuicai Wu,
  • Honghai Zhang,
  • Zhuhuang Zhou,
  • Weiwei Wu

摘要

Purpose:

Computed tomography (CT)-guided percutaneous thermal ablation is widely used for minimally invasive treatment of liver tumors where needle path planning critically affects treatment efficacy and safety. This study explores the feasibility of deep reinforcement learning (DRL) for automating needle path planning, framing it as a one-shot anatomical target search problem.

Methods:

A DRL framework was developed comprising a cylindrical conformal environment derived from CT-based three-dimensional models of key anatomical structures (skin, bones, liver, tumor, vessels, spleen), a continuous action space for entry parameters \((z, \alpha )\) , and a clinically informed reward function integrating five anatomical safety constraints and terminal rewards. The proximal policy optimization (PPO) algorithm was adapted to this non-sequential, constraint-rich search space and trained on clinically annotated cases from a public CT dataset, with the remaining cases for validation.

Results:

Compared with the conventional rapidly-exploring random tree (RRT) method, the proposed approach achieved a 21.8% median improvement in path quality and resulted in reduced inference time (0.12 ±0.03 s vs. 3.45± 0.21 s).

Conclusion:

These results demonstrate the feasibility and efficiency of DRL-based one-shot anatomical target search for CT-guided percutaneous liver tumor ablation.