Poster — Thur Eve — 05: Objective phantom‐based and porcine model comparison of filtered back projection, adaptive statistical iterative reconstruction and model based iterative reconstruction algorithms

• Published in: Medical physics (Lancaster) 2014-08, Vol.41 (8Part2), p.7-7
• Main Authors: Omotayo, Azeez, Elbakri, Idris
• Format: Article
• Language: English
• Subjects: Computed tomography; Germanium; Image reconstruction; Medical image noise; Medical image reconstruction
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4894991

We performed objective image quality evaluation of three commercial reconstruction algorithms available on a GE DiscoveryCT750HD multi‐detector CT (MDCT) scanner at different dose levels. Using a Catphan500 phantom and a freshly euthanized pig carcass (∼ 36kg), we evaluated the noise and contrast‐to‐noise ratio (CNR) for the three reconstruction algorithms: filtered back‐projection (FBP), adaptive statistical iterative reconstruction (ASIR™) and model based iterative reconstruction (code named VEO™). At a dose of ∼ 1.5 mGy, ASIR™ offers noise reduction up to 54% with the phantom, while VEO™ reduces noise by up to 70% with the porcine model. At low doses, VEO™ offers better noise reduction capabilities over FBP and ASIR™, depending on helical pitch. The level of noise reduction increases as pitch increases. Similar trend was observed with CNR increases as VEO™ offered significantly better CNR over FBP and ASIR™. Compared to FBP, iterative algorithms reduce image noise and increase CNR. The model‐based iterative algorithm (VEO™) produced better noise and contrast resolution in both phantom and porcine images compared to FBP and ASIR™.