Impact of iterative reconstructions on image noise and low-contrast object detection in low kVp simulated abdominal CT: a phantom study. (September 2016)

• Record Type: Journal Article
• Authors: Holmquist, Fredrik; Nyman, Ulf; Siemund, Roger; Geijer, Mats; Söderberg, Marcus
• Is Part Of: Acta radiologica. Volume 57:Number 9(2016:Sep.)
• Page Start: 1079
• Page End: 1088
• Publication Date: 2016-09
• Subjects: Computed tomography (CT) -- filtered back-projection -- image noise -- iterative reconstruction -- low tube voltage -- noise reduction -- radiation; Radiology, Medical -- Periodicals; Radiography, Medical -- Periodicals; Radiotherapy -- Periodicals; 616.0757
• DOI: 10.1177/0284185115617347
• Languages: English
• ISSNs: 0284-1851

Abstract

Background: Low kilovoltage (kVp) computed tomography (CT) may be used to reduce contrast medium dose in patients at risk of contrast nephropathy, at the cost of increased image noise. Purpose: To evaluate: (i) the impact of iterative reconstructions (Siemens SAFIRE) on low-contrast object detection to compensate for increased noise instead of increased tube loading when decreasing tube potential; and (ii) the change in iodine attenuation in simulated abdominal CT. Material and Methods: A phantom was scanned at 70, 80, 100, and 120 kVp at fixed effective tube loading (170 mAsEFF ) and fixed radiation dose (CTDIVOL 10 mGy). Images were reconstructed with filtered back-projection (FBP) and SAFIRE strengths S1–S5. Iodine attenuation, objective image noise, contrast-to-noise ratio (CNR), noise power spectrum (NPS), spatial resolution, and subjective detectability of low-contrast objects were evaluated. Results: Compared with 120 kVp iodine attenuation increased by a factor 1.6 and 2.0, and image noise increased by a factor 1.9 and 2.5 at 80 and 70 kVp, respectively. Compared with FBP, SAFIRE showed objective reduction in image noise and increased CNR without loss of spatial resolution or any significant NPS alteration, with general tendency to improve subjective detectability of low-contrast objects. At 170 mAsEFF the number of discernible 1.0% contrast objects at 70 kVp/S5 and 80 kVp/S5 was similar to that at 120 kVp/FBP. Conclusion: With the SAFIRE algorithm image noise, CNR and detectability of low-contrast objects may be kept unchanged without increased tube loading when using low kVp settings to reduce contrast medium dose in azotemic patients.