In AJR. American journal of roentgenology
Background: Radiologists generally perform clinical interpretation using thick-section images (typically 3-5 mm) because of their low image noise, and may additionally refer to thin-section images (typically 0.5-0.625 mm) for problem solving. Deep learning reconstruction (DLR) can yield thin-section images with low noise. Objective: To compare abdominopelvic CT image quality between thin-section DLR and thin- and thick-section hybrid iterative reconstruction (HIR) images. Methods: This retrospective study included 50 patients (median age, 64 years; 31 male, 19 female) who underwent abdominopelvic CT between June 15, 2020 and July 29, 2020. Images were reconstructed at 0.5-mm sections using DLR, and at 0.5-mm and 3.0-mm sections using HIR. Five radiologists independently performed pairwise comparisons (0.5-mm DLR and either 0.5-mm or 3.0-mm HIR) and recorded the preferred image for subjective image quality measures (scale from -2 to +2). Readers' pooled scores were compared with a score of 0 (no preference). Image noise was quantified using SD of ROIs on regions of homogeneous liver. Results: For comparison of 0.5-mm thick DLR and 0.5-mm thick HIR, median pooled score was +2 (definite preference for DLR) for noise and overall image quality, and +1 (slight preference for DLR) for sharpness and natural appearance. For comparison of 0.5-mm thick DLR and 3.0-mm thick HIR, median pooled score was +1 for these four measures. These assessments were all significantly different (p<.001) from 0. For artifacts, median pooled score for both comparisons was 0, which was not significant for comparison with 0.5-mm HIR (p=.03), but was significant for comparison with 0.5-mm HIR (p<.001) due to imbalance in scores of +1 (n=28) and -1 (slight preference for HIR, n=1). Noise for 0.5-mm thick DLR was lower by a mean difference of 12.8 HU compared with 0.5-mm HIR and 4.4 HU compared with 3-mm HIR (both p<.001). Conclusion: Thin-section DLR improves subjective image quality and reduces image noise compared with currently used thin- and thick-section HIR, without causing additional artefacts. Clinical Impact: Although further diagnostic performance studies are warranted, the findings suggest the possibility of replacing current use of both thin- and thick-section HIR with use of only thin-section DLR during clinical interpretations.
Oostveen Luuk J, Smit Ewoud J, Dekker Helena M, Buckens Constantinus F, Pegge Sjoert A, de Lange Frank, Sechopoulos Ioannis, Prokop Mathias