ABSTRACT
To acquire reference data for setting an appropriate compressed sensitivity encoding (CS) for brain lesion detectability, the effects of contrast and noise on contrast-enhanced magnetic resonance imaging (MRI) were evaluated. Gadobutrol at various concentrations and manganese chloride tetrahydrate were used as a phantom. Various CS factors (0-10) and denoising levels (weak, medium, and strong) were assessed. The contrast amount decreased from CS7 in non-denoised images for 0.5-2 mmol/L solutions but slightly decreased from CS7 with denoising. The noise amount significantly increased with an increasing CS factor. Generally, there was a significant difference in the denoising level and rate across all CS factors in the case of the 2 and 0 mmol/L solutions. When the CS factor was increased without denoising, the integrated noise power spectrum (NPS) increased and decreased in the high-frequency and low-frequency areas, respectively. These data can be used to establish settings based on the degree of denoising.
Subject(s)
Algorithms , Magnetic Resonance Imaging , Phantoms, Imaging , Signal-To-Noise RatioABSTRACT
Fat suppression is a technique used to suppress the signals from adipose tissues, during clinical evaluation of the tissues near the fat-tissue boundary. However, in cases where the scan area has a complicated shape, the effect of fat suppression may demonstrate poor uniformity, resulting in diagnosis-related difficulties. To improve the uniformity of fat suppression, phantom studies are more suitable than volunteer studies. In this study, we evaluated the reliability of the region of interest (ROI) dependency using an unevenness phantom, to develop a method to assess the uniformity of fat suppression while using whole magnetic resonance imaging by masking the surrounding phantom. We modulated different ROI sizes, which were eroded from 100% to approximately 50%, and observed that the normalized absolute average deviation and error increased with decreased ROI. Using our method, more objective, concrete, and accurate data could be obtained by including the whole-body phantom (whole poor uniformity area).
