Skip to main content

Table 3 Objective measurement of different reconstruction algorithms

From: Application of a deep learning image reconstruction (DLIR) algorithm in head CT imaging for children to improve image quality and lesion detection

  FBP 50% ASIR-V 100% ASIR-V DL-H F P
5 mm images
 GM CT 32.58 ± 2.19 32.74 ± 2.36 32.77 ± 2.33 32.46 ± 2.31 0.23 0.88
 GM SD 4.51 ± 1.15 3.35 ± 0.84 2.28 ± 0.62* 2.26 ± 0.50* 93.83  < 0.001
 WM CT 25.75 ± 2.66 25.49 ± 2.82 25.52 ± 2.84 25.56 ± 2.73 0.10 0.96
 WM SD 4.28 ± 0.84 3.16 ± 0.60 2.21 ± 0.57 2.00 ± 0.34 159.18  < 0.001
 GM SNR 7.67 ± 1.83 10.35 ± 2.44 15.40 ± 4.29* 15.03 ± 3.13* 82.28  < 0.001
 WM SNR 6.20 ± 1.12 8.28 ± 1.46 12.05 ± 2.42 13.07 ± 2.24 159.09  < 0.001
 CNR 1.61 ± 0.60 2.30 ± 0.79 3.33 ± 1.10* 3.30 ± 1.06* 46.08  < 0.001
0.625 mm images
 GM CT 32.57 ± 2.62 32.71 ± 2.68 32.76 ± 2.57 32.64 ± 2.52 0.06 0.98
 GM SD 8.23 ± 1.68 5.74 ± 1.18 3.43 ± 0.80* 3.35 ± 0.68* 221.42  < 0.001
 WM CT 25.36 ± 3.00 25.37 ± 3.06 25.42 ± 3.00 25.47 ± 2.85 0.02 1.00
 WM SD 8.07 ± 1.57 5.60 ± 1.13 3.38 ± 0.81* 3.11 ± 0.58* 247.66  < 0.001
 GM SNR 4.11 ± 0.86 5.94 ± 1.31 10.09 ± 2.53* 10.09 ± 1.96* 157.54  < 0.001
 WM SNR 3.24 ± 0.67 4.68 ± 0.97 7.86 ± 1.80 8.40 ± 1.45 201.80  < 0.001
 CNR 0.91 ± 0.41 1.33 ± 0.55 2.21 ± 0.85* 2.26 ± 0.77* 54.59  < 0.001
  1. SG: clarity level of sulci/cisterns boundaries; BM: boundaries between the white and gray matters; WQ: whole image quality
  2. *Without significant statistical differences for the numbers with * between the paired comparison