An Improved 4DCT Reconstruction Algorithm for Patients with Irregular Breathing
A Harnage1*, C Ramsey1,2, M Ratner1, J Bowling2, I Pfeiffer2, (1) University of Tennessee, Knoxville, TN, (2) Thompson Cancer Survival Center, Knoxville, TN
Presentations
PO-GePV-M-98 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: The purpose of this study was to evaluate 4DCT reconstruction using a database of measured patient respiration cycles. The primary objective of this study was to evaluate the respiration motion data output from the Varian RPM system to determine if 4DCT image reconstruction could be improved, especially for patients with irregular breathing.
Methods: The Varian RPM system generates a *.vxp file that contains information relating to the measured respiration cycle. The locations of the respiration peaks are automatically identified by the RPM in amplitude space and are used to convert the motion into a linear phase binning that is stored in the vxp file. This methodology works properly for a periodic breathing pattern, but represents a source of error in 4DCT reconstruction for patients with irregular breathing. A program was created for modifying the vxp files with non-linear binning based on correctly identified inhale peaks and a new predictive filter.
Results: A total of 445 patients’ respiration cycles measured with the RPM system were analyzed and ranked based on the reproducibility of the respiration cycles. Phase images were corrected for the 10 patients with the largest 4DCT reconstruction errors. These patients all had visible artifacts with 4DCT reconstructions using the original RPM generated vxp files. These respiration cycles were then re-binned, a new vxp file was written, and a second set of phase images was created and evaluated. The new predictive filter increased the amount of data for reconstruction. The modified vxp files showed a decrease in imaging artifact in the 4DCT reconstructions.
Conclusion: A tool was developed that allows user control of phase binning, peak identification, and predictive filter selection of data to be included in 4DCT reconstruction. This tool has been shown to improve 4DCT reconstruction in a database of extremely challenging clinical cases.
Keywords
Taxonomy
Contact Email