Skip to main content

Table 5 ROC curve analysis for binary classification of hormonal cancer subtype

From: Clustering subtypes of breast cancer by combining immunohistochemistry profiles and metabolism characteristics measured using FDG PET/CT

Variable

Criteria

Sensitivity (%)

Specificity (%)

P

AUC

Nuclear Grade

2

99.1

100

< 0.001

0.995

Ki-67

9 %

52.8

85.5

< 0.001

0.733

p53

60 %

12.3

100

0.082

0.572

SUV

2.310

65.1

65.8

< 0.001

0.677

EGFR

positive

2.8

98.7

0.468

0.508

ER

6

11.3

98.7

0.372

0.523