Table 3.
Conventional DWI | DL DWI | P Value | ||
---|---|---|---|---|
Overall image quality | 4.16 ± 0.47 | 4.17 ± 0.50 | 0.77 | |
GW differentiation | 4.86 ± 0.42 | 4.83 ± 0.43 | 0.21 | |
DG differentiation | 4.78 ± 0.44 | 4.77 ± 0.44 | 0.70 | |
Artifact |
Motion Others |
5.9% (n = 37) 43.2% (n = 16) 56.8% (n = 21) |
21.5% (n = 136) 11.0%(n = 15) 89.0%(n = 121) |
< 0.0001 |
Lesion conspicuity | ≤ 5 mm | 4.14 ± 0.79 | 4.58 ± 0.72 | 0.03 |
> 5 mm | 4.46 ± 0.76 | 4.63 ± 0.69 | 0.02 | |
Lesion conspicuity | 4.43 ± 0.77 | 4.62 ± 0.69 | 0.005 |
Data are mean ± standard deviation and averaged scores of two observers.
DL DWI indicates diffusion weighted imaging with deep learning reconstruction.
GW: gray-white matter.
DG: deep gray matter.