Room: Karl Dean Ballroom B1
Purpose: Compton scatter imaging is an emerging technique in which photons scattered during radiation therapy are collected to image the irradiated volume. To assess the viability of scatter imaging, the image quality and tumor tracking accuracy were characterized with anthropomorphic and breathing-motion phantoms.
Methods: Using a pinhole camera and a flat panel detector, scatter images were collected during the 6 MV irradiation of two phantoms: the anthropomorphic LUNGMAN phantom (Kyoto Kagaku) and the QUASAR respiratory motion phantom (Modus Medical). To characterize scatter image quality, spherical tumors of various sizes (2.1-2.8 cm) were placed in the lung section of the LUNGMAN. To understand the lung tumor tracking capability, scatter images of a moving spherical tumor (3 cm diameter) were collected during sinusoidal breathing motion (15 breaths-per-minute, ±1.5-2 cm). The position of the tumor was identified in the scatter images, which were collected during open field, conformal dynamic arc, and VMAT irradiations.
Results: The 2.1 cm diameter LUNGMAN tumor was identifiable on scatter images collected with 10 MU (contrast-to-noise ratio = 4.4). For an open field 1200 MU/minute irradiation, the moving QUASAR tumor was tracked with a root-mean-squared error (RMSE) of 0.7-0.9 mm and 0.6 mm for 300 ms (6 MU) and 600 ms (12 MU) frame integrations, respectively. For dynamic conformal arc irradiations, RMSE = 1.0 mm was measured (QUASAR, 500 ms frame, 10 MU). For VMAT deliveries, modulation reduced the tumor tracking accuracy (RMSE = 2.8 mm, 500 ms frame, max 10 MU).
Conclusion: Based on phantom studies, spherical 2-3 cm diameter lung tumors are identifiable on scatter images collected with 10 MU. Scatter imaging is capable of providing real-time, sub-millimeter lung tumor tracking capabilities without additional imaging dose. Further studies will employ comparison to simulated images to improve the accuracy of scatter imaging in the presence of modulation.
Not Applicable / None Entered.