Table 4.

Confusion matrix analyses of 10 optimal radiomics features.

Feature	AUC	CI	Cutoff-value	Sensitivity	Specificity	Accuracy	PPV	NPV
Firstorder_Minimum	0.623	0.548-0.698	0.777	0.291	0.929	0.625	0.789	0.59
GLCM_SumEntropy_waveletHLL	0.587	0.511-0.663	1.528	0.330	0.814	0.583	0.618	0.571
Firstorder_Maximum_waveletLLL	0.597	0.522-0.673	3.980	0.487	0.68	0.579	0.547	0.625
GLDM_DependenceEntropy_waveletHLH	0.730	0.663-0.796	4.454	0.466	0.867	0.676	0.762	0.641
GLCM_MaximumProbability_waveletHLH	0.628	0.554-0.702	0.279	0.718	0.487	0.597	0.561	0.655
GLCM_Autocorrelation_waveletLHL	0.657	0.584-0.731	2.331	0.825	0.504	0.657	0.603	0.760
GLSZM_LargeAreaHighGrayLevelEmphasis_waveletHHH	0.748	0.683-0.812	27502194.84	0.718	0.681	0.699	0.673	0.726
GLSZM_LargeAreaLowGrayLevelEmphasis	0.628	0.554-0.702	482639205.8	0.612	0.593	0.602	0.578	0.626
Firstorder_Minimum_waveletLHH	0.668	0.596-0.739	-0.147	0.573	0.690	0.634	0.628	0.639
GLRLM_LongRunHighGrayLevelEmphasis_waveletLHH	0.771	0.709-0.833	14.646	0.563	0.867	0.722	0.795	0.685

AUC, area under the curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value.