(Thursday, 7/18/2019) 7:30 AM - 9:30 AM
Room: 221AB
Purpose: 90% of 4DCT scans contain one or more imaging errors >4mm. These errors impact multiple steps in the radiotherapy process, ultimately effecting treatment delivery efficacy. To reduce these errors we developed Respiratory Adaptive Computed Tomography (REACT) to pause CT imaging during irregular breathing. For the first time, REACT was experimentally realized on clinical hardware and comparisons of imaging errors compared to conventional 4DCT were made.
Methods: Five patient-measured respiratory traces were simulated on a Standard Imaging phantom with lung and tumor insert on a motion platform. Motion was observed with the Varian RPM system and was sent to our in-house REACT software where breathing irregularity (based on displacement and phase) was assessed. If breathing was irregular, a beam off signal was automatically sent to a Siemens Somatom CT scanner, allowing for prospective gating during imaging. A comparison of error magnitude and frequency between REACT and conventional cine 4DCT images was made. Error magnitude was defined as the region of difference per sagittal slice to a static-phantom ground truth image and frequency as the total number of errors per image.
Results: Using REACT, the frequency of ≥4mm errors was reduced for all five traces for the lung (overall reduction 84%) and for all but one trace for the tumor where errors were tripled (overall reduction 69%). The greatest reduction in ≥4mm errors around the lung (93%) was seen for a trace with a 2cm amplitude spike. The greatest reduction for errors around the tumor (82%) was seen for a trace with consistent baseline variation.
Conclusion: For the first time, REACT was successfully realized on clinical hardware and reduced both the magnitude and frequency of imaging errors compared to conventional 4DCT. These experimental results provide compelling evidence for clinical trials of the REACT system, providing clearer patient images for clinical use.
Funding Support, Disclosures, and Conflict of Interest: This research was supported by a research grant from Siemens Healthineers. Ricky OBrien acknowledges the support of a Cancer Institute NSW Career Development fellowship. Paul Keall acknowledges the support of an NHMRC Senior Principal Research Fellowship.
Keywords
CT, Motion Artifacts, Respiration
Taxonomy
Contact Email
