Objective: The clinical application of an opto-electronic system for real-time three-dimensional (3D) control of patient position in breast cancer radiotherapy is described. The specific features of the motion analysis technology (shape recognition of passive markers) are detailed, and the outcomes of its clinical use for quantitative position control and immobility verification of the thoracic irradiation field during breast cancer treatment are reported.
Materials and methods: The position control system is based on the ELITEtrade mark opto-electronic motion analyzer, which provides in real time the 3D coordinates of a set of passive markers (plastic hemispheres 3 mm in diameter) previously placed on selected landmarks on the patient's skin. The system-dedicated hardware performs marker recognition by means of 2D correlation of shape with a predefined marker modeling mask. This feature ensures a high accuracy, even with small marker dimensions, and successful analysis in a noisy environment (due to room light, reflexes, etc.). The patient repositioning control was based on a comparison between the current positions of the markers and a corresponding reference configuration. The resulting marker displacements were graphically displayed in real time for immediate control. This information was not provided to the operator as a repositioning tool. Instead, the kinematic data was stored for subsequent off-line analysis aimed at quantifying the different factors contributing to patient mis-positioning (initial repositioning errors, patient's breathing, and random movements) when conventional means for patient alignment (laser centering) and immobilization (casting techniques) are used.
Results: Clinical application of the system revealed median 3D localization errors for the directly controlled anatomical landmarks of around 4.5 mm. This value is proposed to represent the intrinsic accuracy of conventional laser-centering techniques in breast cancer radiotherapy, including the effects of patient body deformations. When the positional inaccuracies introduced by patients' respiration were also considered, the extent of the resulting 3D mis-positioning of the control points increased to median values of up to 8 mm.
Conclusions: The reported clinical trial confirms the significant role that real-time opto-electronic motion analysis based on passive markers can have in augmenting the accuracy of patient repositioning and immobility verification in the radiotherapy of a non-rigid body area while also accounting for physiological movements. Evaluation of the data collected during each irradiation session for five patients provided valuable information concerning the optimization of the efficacy of traditional methods for patient centering and immobilization.
Copyright 2000 Wiley-Liss, Inc.