(Sunday, 7/29/2018) 4:00 PM - 4:55 PM
Room: Room 202
Purpose: Respiratory-correlated (RC) 4DCT, the standard-of-care for motion-managed thoracic and abdominal radiotherapy (RT), is prone to significant image artifacts, largely due to cycle-to-cycle variations. Such variations occur primarily due to varying levels of engagement of thoracic vs abdominal muscles, and are not adequately captured by the single-point surrogates (e.g., optical markers, abdominal belts) used to sort and bin the RC-4DCT volumes, because the one-dimensional motion of these surrogates often underdetermines the complex three-dimensional motion of the underlying internal anatomy. In this study, we use an externally and internally deformable lung phantom, which uses two independent motion actuators, to induce programmably-variable correlation within the phantom, and characterize the sensitivity of 4DCT sorting to the nature of the surrogate signal.
Methods: Four two-dimensional (superior-inferior and anterior-posterior) patient-recorded lung tumor motion traces were programmed into our in-house dual-actuator lung motion phantom embedded with seven internal radio-opaque markers. The trajectories of internal markers were recorded via kV x-ray fluoroscopy (11 frames/s), and served as ground truth. For each motion trace, two RC-4DCTs were acquired, corresponding to primarily SI and primarily AP breathing motion. The marker trajectories derived from each RC-4DCT were compared with the corresponding ground-truth trajectories obtained via fluoroscopy.
Results: Phase differences in AP and SI components of motion resulted in binning artifacts and phase mis-assignments. Compared to fluoroscopic measurements, 4DCT consistently underestimated marker motion range (~7 mm). Position-dependent variations in marker motion range (~5 mm) were also observed.
Conclusion: Human respiratory motion is a complex phenomenon, produced by several motion actuators within the body such as the intercostal muscles of the chest and the diaphragm. Selection of the AP displacement of the abdominal surface as a breathing metric and tumor displacement is an under-determined problem and cannot accurately capture the complexities of breathing motion.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by NIH R01CA169102, Varian Medical Systems, Vision RT Ltd.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
