Image quality improvement using model-based iterative reconstruction in low dose chest CT for children with necrotizing pneumonia

BMC Medical Imaging

Research Article

Jinrong Liu¹  Yun Peng²  Xiaomin Duan²  Di Hu²  Yong liu²  Jihang Sun²  Tong Yu² 

[1] Department of respiratory, Beijing Children’s Hospital, Capital Medical University, 100045, Beijing, People’s Republic of China;Imaging Center, Beijing Children’s Hospital, Capital Medical University, No.56, Nanlishi Road, Xicheng District, 100045, Beijing, People’s Republic of China;

关键词: Computed tomography (CT);    Model-based iterative reconstruction;    Filtered back-projection;    Necrotizing pneumonia;    Child;   

DOI  :  10.1186/s12880-017-0177-9

 received in 2016-09-09, accepted in 2017-01-09,  发布年份 2017

来源: Springer

【 摘 要 】

BackgroundModel-based iterative reconstruction (MBIR) is a promising reconstruction method which could improve CT image quality with low radiation dose. The purpose of this study was to demonstrate the advantage of using MBIR for noise reduction and image quality improvement in low dose chest CT for children with necrotizing pneumonia, over the adaptive statistical iterative reconstruction (ASIR) and conventional filtered back-projection (FBP) technique.MethodsTwenty-six children with necrotizing pneumonia (aged 2 months to 11 years) who underwent standard of care low dose CT scans were included. Thinner-slice (0.625 mm) images were retrospectively reconstructed using MBIR, ASIR and conventional FBP techniques. Image noise and signal-to-noise ratio (SNR) for these thin-slice images were measured and statistically analyzed using ANOVA. Two radiologists independently analyzed the image quality for detecting necrotic lesions, and results were compared using a Friedman’s test.ResultsRadiation dose for the overall patient population was 0.59 mSv. There was a significant improvement in the high-density and low-contrast resolution of the MBIR reconstruction resulting in more detection and better identification of necrotic lesions (38 lesions in 0.625 mm MBIR images vs. 29 lesions in 0.625 mm FBP images). The subjective display scores (mean ± standard deviation) for the detection of necrotic lesions were 5.0 ± 0.0, 2.8 ± 0.4 and 2.5 ± 0.5 with MBIR, ASIR and FBP reconstruction, respectively, and the respective objective image noise was 13.9 ± 4.0HU, 24.9 ± 6.6HU and 33.8 ± 8.7HU. The image noise decreased by 58.9 and 26.3% in MBIR images as compared to FBP and ASIR images. Additionally, the SNR of MBIR images was significantly higher than FBP images and ASIR images.ConclusionsThe quality of chest CT images obtained by MBIR in children with necrotizing pneumonia was significantly improved by the MBIR technique as compared to the ASIR and FBP reconstruction, to provide a more confident and accurate diagnosis for necrotizing pneumonia.

【 授权许可】

CC BY   
© The Author(s). 2017