Table 3. Agreement of Volumetric Indices between Deep Learning Segmentation and Ground Truth Segmentation in Test Dataset-1.
| Liver Volume | Spleen Volume | Liver/Spleen Volume Ratio | ||||
|---|---|---|---|---|---|---|
| 95% LOA* | P† | 95% LOA* | P† | 95% LOA* | P† | |
| Total | -0.17 ± 3.07 | 0.19 | -0.56 ± 3.78 | 0.001 | 0.39 ± 4.89 | 0.06 |
| Subgroups | ||||||
| Healthy liver | -0.42 ± 2.49 | 0.08 | -0.97 ± 4.15 | 0.02 | 0.55 ± 5.30 | 0.28 |
| Fatty liver disease | -0.33 ± 2.23 | 0.12 | -0.77 ± 2.84 | 0.007 | 0.44 ± 3.04 | 0.13 |
| Non-cirrhotic chronic liver disease | 0.12 ± 2.63 | 0.64 | -0.14 ± 4.32 | 0.74 | 0.25 ± 4.95 | 0.59 |
| Liver cirrhosis | -0.75 ± 3.58 | 0.03 | 0.06 ± 4.21 | 0.88 | -0.81 ± 5.15 | 0.10 |
| Post-hepatectomy | 0.55 ± 3.64 | 0.11 | -0.99 ± 2.79 | 0.001 | 1.54 ± 4.82 | 0.002 |
*Data are Bland-Altman 95% LOA expressed in percentages as mean difference ± 1.96 × standard deviation of difference, †p values for statistically significant difference in mean difference from zero. LOA = limits of agreement