Knowledge-based model for automated multi-isocenter total marrow and lymphoid irradiation planning across standard and large patient anatomies

基于知识的自动化多中心全骨髓和淋巴组织照射计划模型,适用于标准和大型患者解剖结构

阅读:1

Abstract

BACKGROUND AND PURPOSE: Total marrow and lymphoid irradiation (TMLI) planning is challenging. This study evaluates whether a knowledge-based (KB) model for TMLI delivered using volumetric modulated arc therapy (VMAT) can achieve clinically acceptable dose distributions through fully or semi-automated optimization and whether a single model is effective across varying patient anatomies. MATERIALS AND METHODS: Fifty-one consecutive VMAT-TMLI patients were selected. A KB model was trained using 30 patients treated with standard configurations (5 body isocenters). Validation included two cohorts: 10 standard patients and 11 patients with a larger anatomy treated using separate isocenters for the arms (4 body and 2 arms isocenters). Two planning approaches were explored: fully automated (AutoKB), and KB with manual adjustments (HybridKB) by a planner with no prior experience in TMLI. KB plans were evaluated against clinical plans (CPs) using paired t-tests. RESULTS: The KB model reduced mean doses to major organs-at-risk (OARs). For standard configurations, mean OAR doses were 71% ± 2%, 66% ± 2%, and 66% ± 2% for CP, AutoKB, and HybridKB (both p < 0.01). For larger patients, the corresponding values were 75% ± 3%, 69% ± 2%, and 68% ± 2% (both p < 0.01). D2% of the planning target volume increased in AutoKB, reaching 122% ± 2% (p < 0.001) vs. 117% ± 3% in CP for standard configurations, and 126% ± 2% (p < 0.001) vs. 117% ± 3% in CP for arms configurations. HybridKB was on par with CPs. CONCLUSIONS: A single KB model enabled effective planning for multi-isocenter TMLI, including anatomies requiring separate isocenters for the arms. Fully automated KB provided suboptimal dose distributions. KB with manual refinements reduced planner dependence and improved plan quality.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。