Performance of a Motion Tracking System During Cyberknife Robotic Radiosurgery

Cyberknife (Accuracy Inc., Ca) is a robotic radio-surgery system that includes a compact 6 MV linac delivering up to 800 cGy per minute, and an automate arm to aim at any part of the body from any angle. An essential tool is the guidance system based on x-ray imaging cameras located on supports around the patient. A Cyberknife system has been operational at the Vicenza (Italy) Hospital for years and is mainly employed for treating benign and malignant tumors, and Arterior-Venous Malformations. In radiation therapy, delivery of high doses to targets that move with respiration is challenging because of possible spatial inaccuracies. The purpose of this work was to estimate the accuracy of the prediction algorithm used to compensate for system latency in a real-time respiratory tracking system. We have analyzed respiratory signals of 30 patients who had lung or liver Cyberknife treatments. The “Synchrony” (Accuracy Inc.) motion tracking system we use is based on the correlation between the position of LED markers, detected in real time, and the position of internal markers, sampled through x-ray imaging. The position of the external LED signals, though read in real time, must be predicted to compensate for a few hundred ms time lag in the feedback loop that redirects the beam to the current target position. The respiratory signals were described employing their frequency power spectrum, as recently proposed by other authors. Prediction errors above 1.5 mm, lasting for periods longer than 5 seconds were observed for irregular breathers. These episodes correlate to the presence of a bimodal distribution in the power spectral density, and of very low frequencies contribution. A more refined approach would include a personalized choice of the prediction algorithm based on the very first minutes of treatment. Patient training aimed at reducing breathing irregularities might also result in improved spatial accuracy.