TH‐D‐332‐01: The Use of Dual‐Energy CT Images for Monte Carlo Treatment Planning: Material Extraction and Metal Artifact Reduction

Purpose: To investigate the feasibility of dual‐energy CT‐based material extraction (DECT) for Monte Carlo (MC) dose calculations, to evaluate metal streaking artifact reduction using DECT and to propose a novel scanning technique for motion artifacts reduction. Method and Materials: A phantom with 17 tissue equivalent materials was scanned at two tube voltages. The accuracy of mass density (ρ) extraction using DECT (ρ dual ) compared to the standard single‐energy CT technique (ρ single ) and the accuracy of the effective atomic number ( Z ) extraction was evaluated. A comparison of ρ single and ρ dual for a canine subject was made. The DECT extraction was also used on a prostate phantom and a patient with brachytherapy seeds in order to evaluate metal streaking artifact reduction. Moreover, a novel scanning technique with a moving copper filter is proposed. An optimal filter thickness is found and a MC simulation of DECT is performed. Results: The mean error on the relative electron density (ρ e ) and Z extraction is 1.8% and 3.2%. The improvement in ρ dual compared to the standard ρ single is significant, especially in soft bone materials where the ρ assignment error decreases from 8–10% to less than 2% when the DECT technique is used. Similar conclusions are drawn from the canine study where the comparison showed large differences (up to 12%) in bone marrow. The metal streaking artifacts caused by the brachytherapy seeds in the prostate phantom and the patient were significantly reduced. A 2 mm copper filter thickness was found to be optimal for the novel scanning technique. MC simulations of a CT scanner confirmed that with this filtration materials can be extracted with an accuracy of 1.8% in ρ e and 3.6% in Z . Conclusion: The dual‐energy CT technique results in improved tissue extraction and reduced metal artifacts, and is therefore a useful tool for MC dose calculations.