Applying arterial enhancement fraction (AEF) texture features to predict the tumor response in hepatocellular carcinoma (HCC) treated with Transarterial chemoembolization (TACE)

Table 2 Comparisons of AEF texture features between “Improved” and “Un-improved” patients

AEF texture features	Improved	Un-improved	U	p
MaxIntensity	17.50^0.85/0.76	28.26^1.54/0.86	132	0.006
Skewness	18.59^{–0.43/−0.45}	27.22^1.95/0.33	156	0.028
Energy	18.14^0.01/0.01	27.65^0.03/0.02	146	0.015
Entropy	27.00^6.98/7.12	19.17^6.30/6.38	165	0.046
InverseDifferenceMoment	18.23^0.46/0.48	27.57^0.56/0.55	148	0.017
HaralickCorrelation	29.05^{1093970.86/965593.50}	17.22^{598050.88/386531.00}	120	0.003
HighGreyLevelRunEmphasis	29.25^{3683.28/3936.19}	16.90^{2072.46/1683.97}	104	0.001
ShortRunHighGreyLevelEmphasis	29.68^{3685.10/3926.28}	16.61^{2033.09/1667.16}	106	0.001

Note: Considering there were 32 texture features overall, only the features with significant between-group distinctions were listed in this table. AEF refers to the Arterial enhancement fraction. Data was described using the format of “Rank Mean^Mean/Median”. Rank Mean was used to perform the Mann-Whitney U test, Mean and Median were listed as the references to help indicate the distribution of the values

ISSN: 1470-7330