Feasibility of markerless tumor tracking by sequential dual-energy fluoroscopy on a clinical tumor tracking system

Abstract

Real-time tumor tracking (RTTT) through adaptation of the treatment beam has been reported to significantly reduce the treatment uncertainties related to respiratory induced motion in thoracic and abdominal tumors. However, the only two systems clinically available that are capable of performing this type of RTTT, both rely on implanted fiducials for accurate target localization on kV images, for the majority of lesions. The use of such fiducials is known to cause severe risks for complications, and the presence of possible marker migration with respect to the tumor. In this study, we evaluated the feasibility of markerless tumor tracking through the implementation of dual-energy imaging into the current clinical dynamic tracking (DT) workflow of the Vero SBRT system. As fast-switching kV generators are not readily available, a novel approach to dual-energy imaging was proposed, consisting of using two sequential fluoroscopic sequences. Incorporated at the beginning of the workflow of each DT treatment fraction on the Vero is a 20 s high-frequency (11 Hz) fluoroscopic sequence, using two orthogonal on-board kV imagers at ±45° from the MV beam. To enable dual-energy (DE) subtraction, another orthogonal fluoroscopic sequence, at the same gantry angle but al lower energy, was added into the workflow. Offline, through cross-correlation of the imaging sets from which the fiducial was removed, the best matching high- (HE) and low-energy couples per breathing phase were extracted for DE subtraction. A contrast analysis based on contrast-to-noise ratio was carried-out to evaluate improved tumor visibility between HE and DE images. The results, based on images from 7 lung cancer patients treated with DT on the Vero, suggest that DE subtraction can be used on the Vero system to enhance tumor visibility on kV images for markerless tumor localization, provided that all DE imaging parameters, such as imager settings and imaging angles, are optimized.