Room: Exhibit Hall | Forum 9
Purpose: To develop and validate a real-time position monitoring system that enables liver tumour tracking on a general purpose linear accelerator.
Methods: A software tool, SeedTracker, was developed in-house to enable online position verification of liver tumours by monitoring the position of embolization coil based radiopaque markers. The developed system estimates the 3D position of the tumour using monoscopic x-ray images acquired using the Elekta XVI system by following two main steps:1. Calculate the amplitude binned 3D positions of markers and plan isocentre using Cone Beam Computed Tomographic (CBCT) projection images acquired for patient positioning. 2. Estimate the 3D position of isocentre based on the marker positions in the projection image acquired during treatment and learned 3D position information. The performance of SeedTracker was evaluated using retrospective analysis of two sets of CBCT projection images,acquired in the same treatment session, of five liver SBRT patients. The first set of projection images were used for learning the 3D position. The 3D position of markers and isocentre during the acquisition of second set of monoscopic images was estimated using the binned data and markers position in the images. The accuracy of estimated 3D positions were compared with the XVI reconstructed 4D CBCT data of the second image set.
Results: The SeedTracker system successfully segmented the embolization markers in the projection images acquired at gantry angles ranging from 0Ëš to 360Ëš and showed a True Positive Rate of 94%. The 3D position of markers and plan isocentre estimated by SeedTracker agreed with the XVI reconstructed data with a mean (Ïƒ) difference of 0.3(1.6) mm.
Conclusion: The developed software has been shown to estimate the 3D position of the markers and planned isocentre using the monoscopic x-ray images. This system has the potential application of monitoring tumor position during hypofractionated liver radiotherapy.