Program Information
Assessment of Clinical Accuracy of the CyberKnife Fiducial-Free Lung Tumor Tracking System
M Nakayama1,2*, K Uehara2 , S Tsudou2 , S Tamura2 , H Nishimura2 , H Mayahara2 , R Sasaki1 , (1) Kobe University Graduate School of Medicine, Kobe, Hyogo, (2) Kobe Minimally Invasive Cancer Center, Kobe, Hyogo
Presentations
SU-I-GPD-J-6 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: The CyberKnife Xsight Lung Tracking (XLT) and 1-View tracking systems are able to synchronize beam targeting with the respiratory motion of visible target during delivery without the need for internal fiducials. These systems use a correlation model that relates the external marker positions to tumor positions; a prediction model then provides the future target position. In this study, we investigated the correlation and prediction uncertainties for the CyberKnife fiducial-free tumor tracking system.
Methods: Data of a total of 219 fractions for 43 patients with lung tumors were analyzed. Patients were treated with a dose of 36–66 Gy in 4–8 fractions. The tumor motion amplitudes for each patient were evaluated by using the expiratory and inspiratory phases computed tomography. Log files produced by the XLT or 1-View tracking system were acquired after each treatment, and the correlation and prediction errors for each patient were analyzed.
Results: The overall mean correlation error was 0.69 ± 0.43 mm, 0.36 ± 0.16 mm, and 0.43 ± 0.22 mm for the superior-inferior (SI), left-right (LR), and anterior-posterior (AP) directions, respectively. The overall mean prediction error was 0.13 ± 0.11 mm, 0.03 ± 0.02 mm, and 0.03 ± 0.02 mm for the SI, LR, and AP directions, respectively. There were no significantly differences in these errors between XLT and 1-View tracking. The tumor motion amplitude was significantly related to both the correlation and prediction errors in the SI direction (r = 0.58, and r = 0.80, respectively; p < 0.01). The tracking tumor volume and density were not correlated with these errors.
Conclusion: Our results can be used to determine the necessary margin sizes to compensate for correlation and prediction errors of XLT and 1-View tracking system. The tumor motion amplitude may facilitate margin determination.
Contact Email: