Analysis of patient repositioning accuracy in precision radiation therapy using automated image fusion

This work describes a rapid and objective method of determining repositioning error during the course of precision radiation therapy using off‐line CT imaging and automated mutual‐information image fusion. The technique eliminates the variability associated with manual identification of anatomical landmarks by observers. A phantom study was conducted to quantify the accuracy of the image co‐registration‐based analysis itself. For CT voxel dimensions of [Formula: see text] , the method is shown to detect translations with an accuracy of 0.5 mm in the anterior‐posterior and lateral dimensions and 0.8 mm in the superior‐inferior dimension. Phantom rotation in the coronal plane was detected to within 0.5° of expected values. The analysis has been applied to eight radiotherapy patients at two independent clinics, each immobilized by the same system for cranial stereotactic radiotherapy and CT‐imaged once per week over the five‐ to six‐week course of treatment. Among all patients, the ranges of translation in the anterior‐posterior, lateral, and superior‐inferior dimensions were [Formula: see text] , and [Formula: see text] , respectively. Considering all patients and CT scans, the standard deviations of translation were 0.42 mm, 0.47 mm, and 1.36 mm in the anterior‐posterior, lateral, and superior‐inferior dimensions, respectively. The ranges of patient rotation about the superior‐inferior, left‐right, and anterior‐posterior axes were [Formula: see text] , and [Formula: see text] , respectively. PACS numbers: 87.53.‐j, 87.53.Kn, 87.53.Ly, 87.53.Xd