Click here to

Session: Multi-Disciplinary General ePoster Viewing [Return to Session]

End-To-End Validation of Fiducial Marker Tracking Accuracy in Robotic Radiosurgery Utilizing MRI-Only Simulation

K Singhrao1*, T Nano1, J Scholey1, M Descovich1, (1) University of California, San Francisco, San Francisco, CA


PO-GePV-M-267 (Sunday, 7/10/2022)   [Eastern Time (GMT-4)]

ePoster Forums

Purpose: MRI-only simulation for robotic radiosurgery (RR) could potentially improve treatment accuracy and reduce planning margins. Fiducial markers (FMs) are used in many radiosurgery treatments for setup and tracking however, RR fiducial tracking algorithms are unable to accurately identify FM signal-voids in MRI. In this study we develop a novel method of adding FMs to synthetic-CT (sCT) images and quantitatively evaluate FM detection and targeting accuracy in RR.

Methods: The MRI-only workflow was tested on a homogenous carrageenan-based phantom and anthropomorphic head phantom containing a hidden target and orthogonal film pair insert. Each phantom had at least 4 FMs implanted with a minimum spacing of 2cm. For each phantom a bulk-density sCT was generated, and artificial FMs were inserted at the implantation location. Two methods of fiducial insertion were tested: 1) replacing HU with a fixed value (7000 HU) (voxel-burned); 2) using a fiducial template derived from a linear combination of 20 CT FM (composite-fiducial). All tests were performed on a Cyberknife system (Accuray, Sunnyvale, CA). Treatment plans/digital-reconstructed-radiographs were generated from the original CT and sCTs with embedded fiducials, and used to align the phantom on the treatment couch. Differences in initial phantom alignment (3D translations/rotations) and tracking parameters between CT-based plans and sCT-based plans were analyzed. End-to-end plans for both scenarios were generated and analyzed following our clinical protocol.

Results: For all plans, the fiducial tracking algorithm was able to identify the fiducial locations. The mean FM-extraction uncertainty for the composite-FM was 27+13%, meeting the clinical uncertainty threshold of 40%. The total targeting error was within tolerance (<0.95 mm) for all end-to-end tests (0.6 mm, 0.9mm and 0.3 mm for voxel-burned/composite fiducial sCTs, and original CT respectively.

Conclusion: MRI-only simulation setups with composite-derived FMs provides clinically acceptable setup accuracy in line with CT-based standards for FM-based robotic radiosurgery.


Not Applicable / None Entered.


IM- MRI : Multi-modality MRI-CT

Contact Email