Image quality assessment of ultra low-dose chest CT using sinogram-affirmed iterative reconstruction

Objectives: To assess the image quality of ultra-low-dose computed tomography (ULDCT) using sinogram-affirmed iterative reconstruction (SAFIRE) compared to reduced dose CT (RDCT). Methods: Eighty-one consecutive patients underwent non-enhanced ULDCT using 80 kVp and 30 mAs and contrast-enhanced RDCT using automated tube potential selection and tube current modulation. CT images were reconstructed with SAFIRE. Image noise and subjective image quality of normal structures and various pulmonary lesions were assessed. Results: The mean effective doses were 0.29±0.03 and 2.88±1.11 mSv for ULDCT and RDCT, respectively. ULDCT had significantly higher noise (p<0.001). Image quality of five normal structures was diagnostic in 91.1 % of ULDCT and 100 % of RDCT. With ULDCT, the frequencies of non-diagnostic image quality were 2.0 (1/50), 4.6 (13/280), 25.5 (14/55), and 40.0 (8/20)% for BMIs of<20, 20-25, 25-30, and >30. In the assessment of pulmonary lesions, non-diagnostic image quality was observed for 11.2 % of all lesions, 60.9 % of decreased attenuation (significantly more frequent for upper lung lesions), and 23.5 % of ground-glass nodules. Conclusion: ULDCT generates diagnostic images in patients with a BMI ≤25, but is of limited use for lesions with decreased attenuation, ground-glass nodules, or those located in the upper lobe. Key Points: • Iterative reconstruction enables ultra-low-dose CT (ULDCT) with very low radiation doses. • Image quality of ULDCT depends on the patient body mass index (BMI). • Selection of kVp and mAs depends on both BMI and lesion type. • Diagnosis of pulmonary emphysema or ground-glass nodules requires higher radiation doses.