Purpose: To compare the accuracy of dose distributions obtained using single energy CT (SECT) and Dual Energy (DECT)-based DirectSPR datasets in proton therapy.
Methods: Perspex boxes (10x10x15 cm³, 0.25 inch walls) were filled with water, ground sirloin (representing muscle), ground bovine fat trimmings and ground bovine heart. A Mevion S250i Hyperscan system was used to deliver a uniform 5x5 cm2 field of 227.2 MeV and 164.8 MeV protons (2.5mm spacing) through the center of each box. A multilayer ion chamber (MLIC) was used to measure the range of the protons exiting each box. SECT and DECT of each box was obtained and the latter was used to obtain a DirectSPR dataset (Siemens Healthineers). DirectSPR datasets contain the relative stopping power ratio within each voxel and offer the promise of more accurate calculation of proton dose deposition. Both SECT and DirectSPR data were imported into RayStation (version 10A) and each image set was used in separate plans created to calculate the dose distribution from each of the uniform fields investigated. The dose was sampled using regions of interest that were designed to mimic the individual ion chambers in the MLIC. The distal 90% range (D90) calculated for each beam/scan were compared to the MLIC-measured value.
Results: The average difference between range measurements and calculations based on DirectSPR and SECT, respectively, was 0 mm versus 1.7 mm for the 164.8 MeV beam and 0.1 mm versus 2.6 mm for 227.2 MeV beam.
Conclusion: We demonstrated that DirectSPR datasets can be successfully used for better predicting proton dose distributions and these predictions agree much more closely to measurements than those based on the more traditional use of SECT data.
Not Applicable / None Entered.