Program Information
An Investigation Of Multi-Scanner CT Hounsfield Unit Calibration for Pencil Beam Scanning Proton Therapy Using 3D Gamma Analysis
J Zhang*, X Li , G Liu , Q Liu , J Liang , X Ding , Beaumont Health System, Royal Oak, MI
Presentations
SU-F-T-215 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: We compare and investigate the dosimetric impacts on pencil beam scanning (PBS) proton treatment plans generated with CT calibration curves from four different CT scanners and one averaged ‘global’ CT calibration curve.
Methods: The four CT scanners are located at three different hospital locations within the same health system. CT density calibration curves were collected from these scanners using the same CT calibration phantom and acquisition parameters. Mass density to HU value tables were then commissioned in a commercial treatment planning system. Five disease sites were chosen for dosimetric comparisons at brain, lung, head and neck, adrenal, and prostate. Three types of PBS plans were generated at each treatment site using SFUD, IMPT, and robustness optimized IMPT techniques. 3D dose differences were investigated using 3D Gamma analysis.
Results: The CT calibration curves for all four scanners display very similar shapes. Large HU differences were observed at both the high HU and low HU regions of the curves. Large dose differences were generally observed at the distal edges of the beams and they are beam angle dependent. Out of the five treatment sites, lung plans exhibits the most overall range uncertainties and prostate plans have the greatest dose discrepancy. There are no significant differences between the SFUD, IMPT, and the RO-IMPT methods. 3D gamma analysis with 3%, 3 mm criteria showed all plans with greater than 95% passing rate. Two of the scanners with close HU values have negligible dose difference except for lung.
Conclusion: Our study shows that there are more than 5% dosimetric differences between different CT calibration curves. PBS treatment plans generated with SFUD, IMPT, and the robustness optimized IMPT has similar sensitivity to the CT density uncertainty. More patient data and tighter gamma criteria based on structure location and size will be used for further investigation.
Contact Email: