Matsuo Yukinori, Nakamura Mitsuhiro, Mizowaki Takashi, Hiraoka Masahiro
Department of Radiation Oncology and Image-applied Therapy, Kyoto University, 54 Shogoin-Kawaharacho, Sakyo, Kyoto 606-8507, Japan.
Med Phys. 2016 Sep;43(9):5195. doi: 10.1118/1.4961397.
The purpose of this technical note is to introduce variance component analysis to the estimation of systematic and random components in setup error of radiotherapy.
Balanced data according to the one-factor random effect model were assumed.
Analysis-of-variance (anova)-based computation was applied to estimate the values and their confidence intervals (CIs) for systematic and random errors and the population mean of setup errors. The conventional method overestimates systematic error, especially in hypofractionated settings. The CI for systematic error becomes much wider than that for random error. The anova-based estimation can be extended to a multifactor model considering multiple causes of setup errors (e.g., interpatient, interfraction, and intrafraction).
Variance component analysis may lead to novel applications to setup error analysis in radiotherapy.
本技术说明的目的是将方差分量分析引入放射治疗摆位误差中系统分量和随机分量的估计。
假定数据符合单因素随机效应模型且达到平衡。
采用基于方差分析（anova）的计算方法来估计系统误差和随机误差的值及其置信区间（CI），以及摆位误差的总体均值。传统方法会高估系统误差，尤其是在大分割放疗的情况下。系统误差的CI比随机误差的CI宽得多。基于anova的估计可以扩展到考虑摆位误差多种原因（例如，患者间、分次间和分次内）的多因素模型。
方差分量分析可能会为放射治疗中的摆位误差分析带来新的应用。
Zotero 插件