Room: AAPM ePoster Library
The purpose of this project is to validate the reported real-time stereotactic offsets of a commercial optical surface monitoring system against concurrent target localization with portal imaging.
A custom phantom was utilized with external features for surface monitoring and an internal ball-bearing (BB) for simultaneous tracking using portal imaging and stereotactic analysis software. The ball-bearing was aligned to the kV-CBCT imaging isocenter with submillimeter precision using a 6-DOF couch, and a reference surface was captured for monitoring. A surface imaging system (C-RAD Catalyst) reported offsets relative to the anticipated reference surface position at non-zero couch angles. Translations were applied to the phantom before each set of couch rotations, and real-time surface offsets were recorded. The BB position was simultaneously recorded at each table angle using EPID image analysis.
Real-time surface imaging offsets for a phantom and BB initially co-aligned with imaging isocenter revealed features that coincide with known, reproducible couch axis wobble and rotational hysteresis effects with magnitudes as large as 0.5mm. Furthermore, the spatial orientation of the hysteresis effect is identical in both EPID images and real-time surface offsets. Shifts applied to the phantom are seen to add constructively with these inherent offsets due to walkout or wobble. Real-time surface imaging offsets at each couch angle, when projected into the room reference frame, were within an average of 0.6 mm (max: 0.95 mm) of their corresponding reference BB positions from the EPID images.
We have confirmed that optical surface monitoring can detect deviations from couch axis wobble or rotational hysteresis with submillimeter accuracy at non-zero couch angles. The results suggest that reported real-time surface imaging offsets of a static phantom could be subtracted from live patient monitoring data to provide a more accurate representation of patient motion in the couch frame of reference.