Effect of adaptive statistical iterative reconstruction to decrease low-dose chest CT image noise

Objective To evaluate the effect of adaptive statistical iterative reconstruction(ASIR) to decrease CT image noise in the low-dose chest screening.Methods Totally 180 patients who underwent chest screening scan using HDCT with automatic tube current(mA) modulation were categorized randomly into 3 groups.The noise index(NI) for group 1,2 and 3 was set to 20,30 and 40,respectively.Images ASIR in each group were reconstructed 0,30%,50% and 80%.CT dose index volumes(CTDIvol) and dose length product(DLP) were recorded,while effective dose(ED) was estimated.Image quality was assessed by two radiologists on a 5-points scale.The degree of interobservers concordance was evaluated using Kappa test.Results The images of each group had acceptable image quality,and there was interobserver agreement in diagnosis acceptability(Kappa value was 0.485).The average CTDIvol for Group 1,2 and 3 was(3.41±1.28),(1.40±0.58) and(0.76±0.46)mGy,respectively.ED in Group 3 decreased to(0.35±0.20)mSv.There was no statistical difference of image noise between group 2 at ASIR50 and group 3 at ASIR80.Conclusion ASIR helps to reduce low-dose chest CT image noise,and acceptable image quality can be obtained with ED reduced to 0.35 mSv.