Skip to main content
. 2019 Nov 14;19:1111. doi: 10.1186/s12885-019-6341-6

Table 3.

Receiver operator characteristics of textural parameters for predicting KRAS status

Feature AUC SE(AUC) 95%CI(AUC) Criterion Se(%) Sp(%)
Mean 0.754 0.040 0.674~0.833 >72.072 58.33% 83.67%
Variance 0.759 0.038 0.684~0.834 >281.700 83.33% 60.20%
Skewness 0.703 0.043 0.618~0.787 >0.554 80.00% 69.39%
Entropy 0.800 0.035 0.731~0.870 >1.893 76.67% 78.57%
RLNU 0.802 0.036 0.731~0.872 >186.350 78.33% 76.53%
GLNU 0.813 0.034 0.746~0.880 >7.846 78.33% 74.49%

Abbreviation: AUC area under the curve, SE standard error, Se sensitivity, Sp specificity, RLNU run-length nonuniformity, GLNU gray-level nonuniformity