RESUMO
PURPOSE: The aim of our study is to investigate the impact of iodine quantification on image reconstruction when employing a vascular-specific contrast media phantom with varying iodine concentrations. MATERIALS AND METHODS: A 30-cm phantom simulating arterial and venous blood vessel diameters was manufactured. Small (9 mm) and medium (12 mm) cylinders contained iodine concentrations from 10 to 100% while large (21 mm) cylinders were in quartiles from 25 to 100% diluted in blood equivalent medium. Each phantom was filled with either iohexol 350 mgI/mL (Group A) or iodixanol 320 mgI/mL (Group B) and then scanned separately. For each group, tube potential (80-140 kVp) and current (50-400 mAs) were changed and all image series were reconstructed with filtered back projection (FBP), hybrid-based iterative reconstruction (HBIR) and model-based iterative reconstruction (MBIR). Mean opacification was measured in all groups. All data were compared employing an independent t test and Pearson's correlation. Visual grading characteristic (VGC) and Cohens' kappa analyses were performed. RESULTS: At 80 kVp, mean opacification using HBIR was significantly higher in Group B (2165 ± 1108 HU) than in Group A (2040 ± 1036 HU) (p < 0.009). At 140 kVp, MBIR and HBIR were greater in Group A (1704 ± 1033 HU and 1685 ± 1023 HU) versus Group B (1567 ± 1036 HU and 1567 ± 1034 HU) (p < 0.022). CNR using FBP, HBIR and MBIR was higher in Group B (46 ± 42 HU, 70 ± 163 HU and 83 ± 74 HU, respectively) than in Group A (43 ± 39 HU, 174 ± 130 HU and 80 ± 65 HU, respectively) (p < 0.0001-0.035). Qualitative image analysis demonstrated no difference in Cohen's kappa analysis. VGC was higher in Group A at all image reconstruction groups. CONCLUSION: Iohexol outperforms iodixanol in observer performance when assessing image reconstruction techniques and iodine concentrations in a vascular-specific contrast media phantom.