[Development of New Digital Phantom Creation Tool for Evaluation of Low-contrast Detectability Using Iterative Reconstruction]
| Autor: | Yoshimi Katagiri, Shun Hayasaka, Ayaka Chiba, Kanako Numasawa, Yoshiya Ohashi, Tatsuya Imai, Kohei Harada |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Computer science
Iterative reconstruction Radiation Dosage Imaging phantom 030218 nuclear medicine & medical imaging 03 medical and health sciences symbols.namesake 0302 clinical medicine Hounsfield scale Computer vision Radionuclide Imaging Radon transform business.industry Phantoms Imaging Visibility (geometry) General Medicine Gaussian filter 030220 oncology & carcinogenesis symbols Radiographic Image Interpretation Computer-Assisted Tomography Artificial intelligence business Tomography X-Ray Computed Smoothing Algorithms |
| Zdroj: | Nihon Hoshasen Gijutsu Gakkai zasshi. 74(8) |
| ISSN: | 0369-4305 |
| Popis: | PURPOSE We developed a novel digital phantom-creation tool that will help formulate the standard shooting method for a three-phase dynamic liver study. Here, we present data demonstrating the usefulness of this tool in the assessment of low-contrast detectability and visibility. METHODS We performed a visual evaluation by adding a spherical digital phantom with a diameter of 8 mm and a computed tomography (CT) value difference of 10 Hounsfield unit (HU) to images taken using filtered back projection and seven types of adaptive iterative dose reduction 3D (Weak, Mild, eMild, Standard, eStandard, Strong, and eStrong). We also examined the partial-volume effect by drawing a profile curve using a digital phantom with a CT value difference of 30 HU and a diameter of 5 mm. Furthermore, a digital phantom with two kinds of filters (smoothing and Gaussian) was added to the image of the home-made simulated tumor phantom to visual valuate its visibility in the phantom's low-contrast module and the digital phantom. RESULTS Detection sensitivity was significantly decreased in Standard, eStandard, Strong, and eStrong, and the area under the curve also decreased in a similar fashion. We confirmed that the partial-volume effect was due to the different maximum CT values in the profile curve at 4 and 5 mm thickness. The visibility of the low-contrast module and digital phantom was most consistent when using the Gaussian filter. CONCLUSION This tool can be used for low-contrast detection ability evaluation. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|