Table 3.
Results for qualitative image analyses
| Reader 1 | Reader 2 | Reader 3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| DLR | Hybrid IR | P-values | DLR | Hybrid IR | P-values | DLR | Hybrid IR | P-values | |
| Noise | 2/37/16/0/0 | 0/3/25/26/1 | < 0.001 | 0/21/30/4/0 | 0/5/38/12/0 | < 0.001 | 3/44/6/2/0 | 0/2/23/29/1 | < 0.001 |
| Sharpness | 0/0/33/22/0 | 0/0/53/2/0 | < 0.001 | 0/13/40/2/0 | 0/16/36/3/0 | 0.694 | 0/22/16/17/0 | 0/4/29/22/0 | 0.004 |
| Artifact | 0/12/33/10/0 | 0/16/25/13/1 | 0.848 | 0/13/37/5/0 | 0/6/41/8/0 | 0.008 | 1/23/26/5/0 | 0/3/39/12/1 | < 0.001 |
| Overall | 0/1/43/11/0 | 0/0/51/4/0 | 0.117 | 0/22/30/3/0 | 0/15/32/8/0 | 0.043 | 0/31/12/12/0 | 0/2/28/24/1 | < 0.001 |
The numbers of patients for each score (5/4/3/2/1) are shown
DLR deep learning reconstruction; Hybrid IR hybrid iterative reconstruction
Comparisons were performed with the Wilcoxon signed-rank test