Low-dose computed tomography of the chest using iterative reconstruction versus filtered back projection: comparison of image quality.

Objective: The purpose of this study was to compare image quality of iterative reconstruction (IR) to filtered back projection (FBP) in low-dose computed tomography of the chest.

Methods: Forty-three consecutive patients were retrospectively enrolled. Eight series of images were reconstructed using FBP and 7 levels of IR in each subject. Image noise, signal-to-noise ratio (SNR), and SNR improvement were measured. Two radiologists evaluated subjective artifact, image artificiality, and subjective overall image quality with 4- or 5-point scales.

Results: Iterative reconstruction showed significantly lower image noise (135.5 ± 36.6 vs 219.9 ± 40.9) and higher SNR (0.36 ± 0.12 vs 0.21 ± 0.05) than FBP (P < 0.001). Signal-to-noise ratio improvement was 72.4% ± 44.9%. Subjective artifact of FBP was significantly higher than IR images (P < 0.001). Image artificiality of IR was significantly higher than that of FBP (P < 0.001). Overall, subjective image quality was poor in FBP and acceptable or good in IR.

Conclusions: With the use of IR, low-dose computed tomography of the chest would achieve less image noise and better image quality compared to the FBP.