Room: Exhibit Hall | Forum 6
Purpose: To study the accuracy with which proton stopping power ratio (SPR) in lung and sinus tissues can be determined from dual energy CT (DECT) and single energy (SECT) CT scans.
Methods: Polylactic acid (PLA ) cylinders (3 cm diameter, 5 cm height) were 3D printed with volumetric fill densities ranging from 1.0 (100% PLA) down to 0.1 (10% PLA, 90% air) to approximate lung and sinus cavities. The cylinders were inserted individually into a high density polyethylene phantom and DECT and SECT scans were obtained using a Siemens SOMATON Definition Edge CT scanner. Effective atomic number (zeff) and electron density (Ï?â‚‘) images were reconstructed from the DECT and used to determine SPR of each plug. SPR from SECT images was calculated using a clinical HU-to-SPR calibration table. Next, the zeff and Ï?â‚‘ for each plug were calculated using the known elemental composition and physical density of PLA (Ï? = 1.22 g/cc), and the SPR for each plug was measured using a clinical proton beam. DECT determined zeff, Ï?â‚‘, and SPR values were compared to calculated and measured values.
Results: The difference between DECT determined and calculated zeff and Ï?â‚‘ was less than 5% for cylinder fill densities greater than 0.5 with the percent difference increasing steadily as fill density decreased to 0.1. DECT, SECT, and measured SPR values all increased linearly as fill percentage increased. DECT-SPR values were within 5% of measured values for fill densities > 0.2. SECT-SPR was accurate within 5% for fill densities > 0.7, but multiplying the HU-to-SPR calibration curve by 1.1 provided 5% SECT-SPR accuracy for fill densities > 0.5.
Conclusion: For regions of the body, (e.g. lung and sinus), that contain a mixture of tissue and air, the accuracy with which SPR can be determined can be improved by using DECT.