Evaluation of 4D Cone Beam Computed Tomography for Target Localization in Stereotactic Body Radiation Therapy for Early Lung Cancer

Abstract

Results: For the 59 images analyzed, the template-based matching algorithm was able to localize the GTV on only 74.6% (44/59) of the SE images vs. 94.9% (56/59) of the DE images (p Z 0.004). For 3 cases, the algorithm was unable to localize the GTVon both the SE and DE images. In the 12 cases in which the algorithm localized the GTVon the DE image but not on the SE image, 8 of the GTVs were < 5 cc in volume (pZ 0.39). Median vector differences in the position of the GTV centroids for the template-based match and the “ground truth” images were 0.5 mm (95% CIZ 0.3 mm, 1.0 mm) for SE images, versus 0.6 mm (95% CIZ 0.4 mm, 1.0 mm) for DE images (p Z 0.63). Conclusions: While the accuracy is comparable, DE imaging provides a higher tracking rate than SE imaging using a template-based matching algorithm on kV planar images for patients with lung tumors. By combining such an algorithm with DE imaging, this approach has the potential to allow for accurate and reliable real-time, markerless-motion tracking of lung tumors. Author Disclosure: A.M. Block: None. R. Patel: E. Research Grant; Varian Medical Systems. J. Panfil: E. Research Grant; Varian Medical Systems. J. Breunig: None. M. Surucu: None. M.M. Harkenrider: None. J.C. Roeske: E. Research Grant; Varian Medical Systems.