57. Comparison of organ dose results from different software in CT exams

Purpose During last years the number of CT exams increased causing an increment of collective dose. Our purpose was to estimate organ doses for the main CT scans performed in our hospital to provide data useful for cancer risk estimation. We calculated organ dose using different commercial software and evaluated the influence of exposure parameters. Methods Using the RDIM system Nexo[DOSE]® (Bracco, Italy) we collected exposure information for the main CT protocols: scanner, CTDIvol, voltage, collimation, pitch. The scan region was deduced from the CT images of a patients’ sample under the specific procedure. Using the RadLex Playbook, CT protocols are clustered in macro-areas with the same scan region and phases number. So, starting from 595 protocols, we have obtained only 12 stylized protocols to simulate. For these protocols, we calculated organ dose using CT-Expo, NCICT, NCICTX and Virtual Dose. Each software uses its own phantom family. CT-Expo uses the first generation of computational mathematical phantoms. NCICT, NCICTX and Virtual Dose use a hybrid-voxel phantom, more realistic and, for the last two software, adaptable to the patient habitus. Since the scan length is the primary origin of variability in the dose calculation, for each protocol the organs completely inside the scan region have been considered separately from the partially irradiated and from distributed ones. Results Only using CT-Expo organ doses present dependence from collimation, pitch and scanner model. Discrepancies using different software are within 30% for organs inside the scan region, while they can be more significant, till 150–200%, for the other organs. Conclusions To manage the organ dose information, a data interpretation is needed, which could explain the discrepancies between results from different software. Since the result is strongly related to the algorithm and to the phantom used, the organ dose result for CT exam is to relate to the software.