Prospective image planning in radiation therapy for optimization of image quality and reduction of patient dose

Introduction: CT simulation data in image-guided radiation therapy (IGRT) provides patient-specific subject contrast. This information can be exploited to establish, a priori, a suitable imaging goal and to select patient-specific imaging acquisition parameters that optimize the similarity between reference and daily set-up images and reduce imaging dose. This study aims to describe and clinically validate a computerized algorithm designed to provide such optimization. Material and methods: An image planning system (IPS) was developed to assist in planar kV imaging technique selection for radiation therapy. The system's patient-specific image quality and dose reduction capabilities were validated herein. Anthropomorphic phantom and clinical data were acquired. Mutual information (MI) was used to compare simulated and measured images in both phantom and clinical tests. Variations in contrast resolution resulting from imaging panel underexposure, saturation and a contrast plateau were investigated. For evaluation of patient-specific imaging dose reduction, the IPS was used to modify acquisition settings for six patients. Results: Phantom data confirmed the IPS's predictive capability regarding image contrast. Measured and simulated images showed similar progressions from under-exposure, image quality peak, and loss of contrast due to detector saturation. Clinical data demonstrated that contrast resolution and imaging dose could be prospectively improved without loss of image contrast. The algorithm reduced imaging dose by an average of 47%, and a maximum of 80%. Conclusions: Loss of image contrast resulting from under-exposure or over-exposure, as well as a contrast plateau can be predicted by use of a prospective image planning algorithm. Image acquisition parameters can be predicted that reduce patient dose without loss of useful contrast.