Improving image quality with model-based iterative reconstruction algorithm for chest CT in children with reduced contrast concentration.

Abstract

To evaluate model-based iterative reconstruction (MBIR) in improving the image quality of chest CT in children with reduced concentration contrast medium (CM). Fifty-six children (median age of 4years) who received low-dose enhanced chest CT were enrolled as the study group and compared with the control group of 56 children. Both groups used the automatic tube current modulation to achieve age-based noise index values of 11-15HU. The study group used 100kVp and reduced CM concentration of 270mgI/ml, and the images in this group were reconstructed with 50% adaptive statistical iterative reconstruction (ASIR) and MBIR. The control group used 120kV and standard CM of 320mgI/ml, and the images in this group were reconstructed with ASIR only. Subjective image quality and objective image quality of the three image sets were evaluated. The subjective quality included overall image noise, enhancement degree, lesion (including mediastinum mass, pulmonary space-occupying lesions, and parenchymal infiltrative lesions) conspicuity, and beam-hardening artifacts. The objective quality included the measurement of noise in the left ventricle and back muscle to calculate signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of ventricle. There was no difference in radiation dose between the study (CTDIvol of 1.79 ± 1.45mGy) and control (1.68 ± 0.92mGy) groups (p = 0.65). However, the study group used 19.7% lower CM dose than the control group (5.84 ± 2.69 vs. 7.27 ± 3.80gI), and the enhancement in all images met the diagnostic requirements. MBIR reduced image noise by 58.6% and increased SNR and CNR by 143.6% and 165.7%, respectively, compared to ASIR images in the control group. The two ASIR image sets had similar image quality. MBIR improved the image quality of low-radiation-dose chest CT in children at 19.3% reduced CM dose.