MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel
  1. Home
  2. Meeting Program
  3. All
  4. Automated Fiducial Tracking During VMAT Using Beam's-Eye-View Images

Automated Fiducial Tracking During VMAT Using Beam's-Eye-View Images

D Ferguson1*, T Harris1, M Shi1,2, M Jacobson1, I Valencia Lozano1, C Williams1, M Myronakis1, P Huber3, P Baturin4, R Fueglistaller3, D Morf3, M Lehmann3, R Berbeco1, (1) Brigham and Women's Hospital & Dana Farber Cancer Institute & Harvard Medical School, Boston, MA, (2) University of Massachusetts Lowell, Lowell, MA, (3) Varian Medical Systems, Baden, Switzerland, (4) Varian Medical Systems, Palo Alto, CA

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: Intrafraction motion may introduce deviations from the prescribed dose to the target and reduce healthy tissue sparing, a challenge that tumor tracking can address. Fiducial markers are frequently used to guide patient setup for external beam treatment of tumors where target visibility on setup imaging is poor, e.g. liver or prostate. These relatively high contrast objects, implanted close to the target, are good surrogates for tumor motion and can be used to perform automated tracking.


Methods: We present an adaptation of an in-house developed tracking algorithm, utilizing the on-board portal imager of the Linac, designed to detect and track fiducial markers. Fiducial detection is performed by examining the contour maps of identified features of interest to verify that the observed characteristics match those expected for fiducials. Tracking on subsequent images is performed by a template matching based algorithm and relies on template stability metrics and local relative orientations to perform multiple feature tracking simultaneously. Only a single image is required to initialize the algorithm and fiducials are automatically added, modified or removed in response to the input images.


Results: The fiducial tracking algorithm was run retrospectively on ciné MV images collected during liver SBRT treatments. Evaluation of the tracker output was performed by comparing the tracked fiducial positions with truth information generated using a motion model extracted from the patient breathing traces in conjunction with a small number of visually tracked points. The tracking error was found to be 1.45mm.


Conclusion: A fiducial detection and tracking algorithm has been developed to perform tumor tracking using MV images collected during VMAT treatments. The proposed algorithm requires no prior training and is shown to be robust in complex tracking environments. The in-treatment tumor location can be used for several clinical applications including in vivo dosimetry and real-time tumor tracking.

Funding Support, Disclosures, and Conflict of Interest: NIH/NCI R01CA188446

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email

Additional Links