Table 3.

ROC analysis for classification of tumor characteristics by texture features using principal component analysis.

	Tumor characteristic		AUC	Sensitivity	Specificity	Accuracy
dbPET	T-category	T1 vs. T2-4	0.89	0.82	0.94	0.86
	N-category	Negative vs. Positive	0.66	0.85	0.54	0.68
	Molecular subtype	Luminal A vs. Others	0.73	0.52	0.89	0.68
	Ki67 level	20%> vs. 20% ≤	0.75	0.72	0.79	0.74
PET/CT	T-category	T1 vs. T2-4	0.94	0.93	0.88	0.91
	N-category	Negative vs. Positive	0.71	0.70	0.75	0.73
	Molecular subtype	Luminal A vs. Others	0.82	0.68	0.84	0.75
	Ki67 level	20%> vs. 20% ≤	0.86	0.76	0.86	0.79

ROC, receiver operating characteristic; dbPET, dedicated breast positron emission tomography; PET/CT, positron emission tomography/computed tomography; AUC, area under the curve.