Laura Marxkors BS, Abstract Co-Author: Employee, General Electric Company

Debashish Pal PhD, Presenter: Employee, General Electric Company

Sandeep Dutta PhD, Abstract Co-Author: Employee, General Electric Company

Yasuhiro Imai MS, Abstract Co-Author: Employee, General Electric Company

Jean-Baptiste Thibault, Abstract Co-Author: Employee, General Electric Company

Masahiro Jinzaki MD, Abstract Co-Author: Nothing to Disclose

Subjective image quality (IQ) comparison using CCTA scans was performed on EKG-gated MBIR and ASiR images. The clinical study was supported by quantitative measurement of resolution using a cardiac phantom.

Clinical CCTA exams (n=20; age: 67 ± 6 yrs; BMI: 25 ± 4) showing pathology such as stents, plaque and acquired using high-resolution step and shoot acquisition mode on Discovery CT750 HD scanner (GE Healthcare, Waukesha, WI) were used in this study. Each CCTA scan was reconstructed with 40% ASiR (HD Stnd. kernel) and MBIR at 0.625 mm slice thickness. The clinical images were reviewed by two radiologists on a 5 point Likert scale (1 = Non-diagnostic, 2 = Sub-optimal, 3 = Acceptable, 4 = Good, 5 = Excellent). Noise and signal-to-noise (SNR) were calculated in the proximal arteries to support the IQ comparison. In addition, a phantom consisting of contrast enhanced vessels with stent, calcified plaque (Hydroxyapatite), and stair-step, non-calcified plaque (ABS resin) was also scanned using a similar protocol. Images at 0.625mm thickness were generated using FBP (HD Stnd. kernel) and MBIR. Contrast dependent resolution was compared between FBP and MBIR using the phantom data. Full-width half maximum (FWHM) of the line spread function (LSF) was used as metric.

MBIR images had superior overall image quality and vessel visualization compared to standard of care ASiR images (5 vs 4, P <0.001). The mean attenuation in the proximal vessels for MBIR images was not different from ASiR (435.6 ± 74.2 HU vs 433.5 ± 71.1 HU; p=0.39). The MBIR images showed a significantly higher SNR (23.23±4.24 vs 11.22±2.28;p<0.001) and significantly lower noise (19.0 ± 3.2 HU vs 39.3 ±4.6 HU; p<0.001). The FWHM of LSF across calcified plaque, stents and contrast enhanced vessels were 0.46 mm, 0.74mm and 0.64mm for MBIR compared to 0.93 mm, 1.17 mm and 0.91 mm for FBP.

MBIR CCTA images were significantly better than ASiR images in overall IQ and vessel visualization. MBIR images also demonstrated superior SNR and lower noise. Improvement in subjective IQ is also supported by significantly lower FWHM of LSF in MBIR compared to FBP images.

MBIR has been demonstrated to lower the radiation dose compared to standard of care images. In addition, the superior SNR and resolution characteristics of MBIR images can lead to improved diagnostic quality of CCTA images.

Marxkors, L, Pal, D, Dutta, S, Imai, Y, Thibault, J, Jinzaki, M, A Subjective and Objective Comparison of Cardiac Computed Tomography Angiography (CCTA) Model-based Iterative Reconstruction (MBIR) with Standard of Care Images .  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14018563.html