Table 4. Comparison of Segmentation Performance and Volumetric Measurement Results of Deep Learning Algorithm between Internal CT Data and External CT Data in Test Dataset-2.

Performance Statistics	External CT Data	Internal CT Data	P*
Dice similarity score†
Liver	0.982 ± 0.011 (0.932–0.999)	0.983 ± 0.007 (0.954–0.998)	0.28
Spleen	0.969 ± 0.011 (0.930–0.994)	0.968 ± 0.010 (0.936–0.993)	0.41
95% LOA of volumetric indices‡
Liver volume	-0.34 ± 2.67	-0.03 ± 3.47	NA
Spleen volume	-0.63 ± 4.34	-0.33 ± 3.70	NA
Liver/spleen volume ratio	0.30 ± 5.22	0.50 ± 4.02	NA

Data are expressed as mean ± standard deviation; data in parentheses are range. ^*p values for comparison of dice similarity score between internal CT data and external CT data using Wilcoxon test, ^†Data are expressed as mean ± standard deviation; data in parentheses are range, ^‡Data are Bland-Altman 95% LOA expressed in percentage as mean difference ± 1.96 × standard deviation of difference. NA = not applicable