Clinical Implementation of Dual-Energy CT (DECT) In Radiation Oncology
Abstract
Purpose
This study evaluates the clinical integration of dual-energy CT (DECT) into radiation therapy workflows. The investigation focuses on image quality, the Hounsfield Unit (HU) accuracy of virtual monoenergetic images (VMIs), and the subsequent dosimetric impact on treatment planning.
Methods
Scans were acquired using a Siemens SOMATOM go.Open Pro scanner. Image quality and artifacts were assessed using water with hammer phantom designed to simulate orthopedic implants. Comparisons were made between standard CT, standard CT with iMAR, and DECT combined with iMAR. HU accuracy was validated using a Gammex phantom. To determine the optimal surrogate for conventional CT, HU values for various plugs in different VMIs (generated using iMAR DECT) were benchmarked against standard 120 kVp scan. Image-derived attenuation coefficients were compared against NIST database values for the optimal VMI, in Our case, at 70 keV. Finally, dosimetric robustness was tested with the Eclipse AAA planning algorithm by systematically varying HU values in a water phantom (-100 to 100 HU) and recording the change in dose for a 3 cm target at 5 cm depth.
Results
DECT combined with iMAR provided the most effective artifact reduction, compared to both standard CT and standard CT-iMAR. In the Gammex phantom, VMIs at 70 keV showed clinically negligible HU differences compared to standard 120 kVp scans. Attenuation coefficients for 70 keV VMIs matched NIST values within 3%. HU variations resulted in less than a 2.5% change for Water HU change (-100to100HU) in the calculated dose. While these dosimetric changes were observed for a single broad beam, the impact on small beams still needs to be evaluated.
Conclusion
Intermediate energy VMIs (at 70 keV) provide accurate HU mapping suitable for dose calculation using conventional electron density curves, while higher energy VMIs superiorly mitigate artifacts.