Diagnostic utility of a conventional MRI-based analysis and texture analysis for discriminating between ovarian thecoma-fibroma groups and ovarian granulosa cell tumors

Table 6 Performance of each individual feature in OTFG and OGCT classifications

	Accuracy	Sensitivity	Specificity	F-measure	AUC
MBF
ADC	0.792	0.644	0.888	0.710	0.823
PC	0.679	0.952	0.500	0.702	0.640
VCE	0.792	0.524	0.969	0.667	0.843
Representative MBFs
ADC + PC + VCE	0.905	0.857	0.937	0.878	0.938
TF
10th pct	0.759	0.632	0.843	0.675	0.809
Diff.var	0.774	0.524	0.938	0.647	0.691
MCC	0.588	0.015	0.964	0.027	0.491
Representative TFs
10th pct + Diff.var + MCC	0.791	0.618	0.904	0.700	0.817
MBFs with TFs
VCE + Diff.var + ADC	0.901	0.893	0.906	0.877	0.941

All data are mean values. AUC area under the curve, OTFG ovarian thecoma-fibroma group, OGCT ovarian granulosa cell tumor, MBF MRI-based feature, TF texture feature. Feature name codes are as follows: 10th pct 10th percentile, ADC apparent diffusion coefficient, Diff.var difference variance, MCC maximal correlation coefficient, PC peripheral cystic area, VCE venous phase contrast enhancement

ISSN: 1757-2215