Validation of a CBCT-Based Online Evaluation Workflow for Proton Therapy
C Chang1, R Nilsson2, S Andersson2, D Bohannon3, L Lin1, S Patel1, P Patel1, K Langen1, J Bradley1, T Liu1, X Yang1, J Zhou1*, (1) Emory University School of Medicine, Atlanta, GA, (2) RaySearch Laboratories AB, Stockholm, SE (3) Georgia Institute of Technology, Atlanta, GA
Presentations
PO-GePV-T-168 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: To validate the efficiency and accuracy of cone-beam CT (CBCT) based online evaluation workflows for different treatment sites in proton therapy.
Methods: Twenty-three patients (eight head and neck (HN), seven lung, and eight prostate plus pelvic nodes) were evaluated. The online CBCT evaluation workflow based on a research version of RayStation 11B planning system includes two methods for image generation: corrected CBCT (cCBCT) and deformed planning CT (vCT). The same day deformed quality assurance CT (dQACT) to CBCT was used as the reference. Process time for the workflows was analyzed. The gamma index (1%/1 mm) was evaluated to quantify the dosimetry accuracy with the generated cCBCT and vCT. The accuracy of beam path water equivalent thickness (WET) for targets (CTV) was also evaluated for ΔWET_cCBCT and ΔWET_vCT, which denote the differences of CTV WET between cCBCT/vCT and dQACT.
Results: Mean online evaluation time among all patients were 98.9±18.0 s for cCBCT workflow, plus 16.2±1.9 s for the vCT. Monte Carlo dose calculation and deformable image registration accounted for ~64% of the cCBCT workflow time. The generation of cCBCT and region of interest mapping accounted for ~17% and 8% of the evaluation time, respectively. The mean gamma passing rates for HN, lung, and prostate patients using cCBCT were 95.5%±2.2%, 91.9%±4.0%, and 97.9%±1.4%, respectively; while the passing rates using vCT were 94.3%±1.8%, 94.0%±4.2%, and 98.3%±1.0%, respectively. The median ΔWET_cCBCT vs ΔWET_vCT was 0.5mm vs 0.8mm (HN), -2.8mm vs 0.5mm (lung), and 0.1mm vs 0.3mm (pelvis). The 95% CIs of ΔWET_cCBCT and ΔWET_vCT were: -0.1~1.1mm and 0.2~1.4mm (HN); -4.9~-0.7mm and -0.3~1.3mm (lung); -0.6~0.8mm and -0.5~1.1mm (pelvis).
Conclusion: The cCBCT-based online evaluation can be achieved within two minutes for most cases. Overall, the vCT workflow achieves better gamma passing rates, and may be suitable for lung patients due to its higher WET accuracy.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email