Improved Quantitative Assessment of Left Ventricular Volumes and Chamber Dimensions Using Iterative Model Reconstruction Algorithm

Objectives:
To evaluate whether the use iterative model reconstruction (IMR) in comparison to iterative reconstruction (IR) could improve of the image quality and diagnostic performance of computed tomography (CT) for the assessment of left ventricular (LV) volumes with cardiac magnetic resonance imaging (MRI) serving as a gold standard.

Background:
Cardiac CT became more accurate in the evaluation of LV function parameters, but very little is known regarding assessment of LV function parameters using IMR technique compared to older techniques such as IR.

Methods:
Nineteen patients (men 14, mean age 44.6 ± 12.3 years) underwent retrospective ECG-gated 256-slice CT and 1.5 Tesla MRI. The CT data was reconstructed using IR and IMR and in MRI multiple short-axis slices with complete coverage of the left ventricle were performed. For comparison of IR and IMR, image quality, signal to noise ratio (SNR), contrast to noise ratio (CNR), LV end-diastolic volume, mass, and dimensions were evaluated. LV function parameters obtained by IMR and IR were compared with MRI.

Results:
In comparison to IR the use of IMR improves the image quality of CT for evaluation of LV function parameters and improves SNR (19.2 ± 4.9 vs. 13.9 ± 5.5; retrospectively, P = 0.004) and CNR (24.0 ± 5.3 vs. 17.2 ± 6.1, retrospectively, P < 0.001). Inter-study variability between IMR and MRI was smaller compared to the differences between IR and MRI for the assessment of LV parameters: End-diastolic volume: -0.9 ± 3.8ml vs. -1.8 ± 7.1ml; LV mass: 4.1 ± 10.5g vs. 4.8 ± 13.2g, septum -0.1 ± 0.3mm vs. -0.6 ± 1.4mm, lateral wall -0.2 ± 0.7mm vs. -0.6±1.1mm, and end-diastolic diameter -0.2 ± 1.6 vs. -0.9 ± 2.3mm; respectively.

Conclusions: The use of IMR algorithm improves the image quality and diagnostic performance of cardiac CT for the assessment of LV function.