Room: Davidson Ballroom A
Purpose: Lung motion phantoms are widely used to validate motion management strategies for imaging and dose delivery in thoracic and abdominal radiotherapy. Most such phantoms have fairly simplistic designs rigid exterior, rigid or deformable interior), which do not adequately represent the complexity of human respiration. In this work, we develop a programmable, externally- and internally-deformable lung phantom that concurrently uses two independent motion actuators to simulate variably correlated motion internally as well as externally, thereby emulating variably correlated human respiratory motion caused by cycle-to-cycle differences in abdominal vs thoracic breathing.
Methods: A commercially available externally deformable lung phantom (RSD, CA) was filled with a deformable fast-response latex foam embedded with radiopaque markers. The motion of the foam and the phantom exterior was actuated by two independent, precision linear motion actuators â€“ simulating superior-inferior (SI) and anterior-posterior (AP) motion. Trajectories of the internal and external markers, tracked via kV fluoroscopy, were used to quantify the variability of external-internal and internal-internal correlation in the AP and SI directions. As a demonstration of a potential application, two motion models, correlating internal volumetric motion with (i) a single point on the surface and (ii) the entire optical surface, were implemented and compared.
Results: Repeatability measurements obtained over four days showed internal and external marker motions to be reproducible within 0.5mm. The variable correlation between the displacement of external-internal markers and internal-internal markers was observed with 0.3 <= r <= 1 (external-external) and -0.45<=r<=0.5 internal-internal). Position estimation errors for the single-point motion model were (SI and AP:<5mm) and for the surface model (SI<3mm, AP<2mm).
Conclusion: We have successfully developed and validated a deformable, variable correlation lung motion phantom that represents the closest approximation thus far, to our knowledge, to human thoracic anatomy undergoing respiration.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by NIH R01CA169102, Varian Medical Systems, Vision RT Ltd.
Not Applicable / None Entered.