Table 3 Distribution of subjective image scoring for different reconstruction techniques by two pediatric radiologists
From: Image quality assessment of pediatric chest and abdomen CT by deep learning reconstruction
Parameter | Reviewer 1 | Reviewer 2 | *Agreement (κ) |
|---|---|---|---|
Overall image quality | |||
(1/2/3/4) | |||
ASIR-V 50 | 0/47/4/0 | 0/47/4/0 | 0.728 |
ASIR-V 100 | 0/4/47/0 | 0/0/51/0 | 0 |
DLR-M | 0/0/49/1 | 0/0/49/1 | -0.02 |
DLR-H | 0/0/38/13 | 0/0/35/16 | 0.568 |
Noise | |||
(1/2/3/4) | |||
ASIR-V 50 | 0/47/4/0 | 0/49/2/0 | 0.297 |
ASIR-V 100 | 0/8/43/0 | 0/0/51/0 | 0 |
DLR-M | 0/1/48/1 | 0/0/51/0 | 0 |
DLR-H | 0/1/34/16 | 0/0/33/18 | 0.62 |
Artifact | |||
(1/2/3/4) | |||
ASIR-V 50 | 0/11/40/0 | 0/12/39/0 | 0.944 |
ASIR-V 100 | 0/11/40/0 | 0/12/39/0 | 0.944 |
DLR-M | 0/11/40/0 | 0/12/39/0 | 0.94 |
DLR-H | 0/11/40/0 | 0/12/39/0 | 0.944 |