Technical Note: Introduction of variance component analysis to setup error analysis in radiotherapy

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.