Dopamine transporter single-photon emission computed tomography-derived radiomics signature for detecting Parkinson’s disease

EJNMMI Research

Table 6 Classification performance of various classification models using radiomics signature and SUR_putamen for test dataset 2

Classifier	Features AUC (95% CI)			P value (SUR_putamen vs. combination)
Classifier	Radiomics signature	SUR_putamen	Combination	P value (SUR_putamen vs. combination)
SVM	0.977 (0.934–0.994)	0.921 (0.847–0.963)	0.940 (0.746–0.995)	0.755
KNN	0.896 (0.768–0.998)	0.857 (0.729–0.941)	0.997 (0.986–1.000)	0.006
LDA	0.986 (0.947–0.998)	0.929 (0.858–0.967)	0.996 (0.961–1.000)	0.009
Decision tree	0.973 (0.929–0.990)	0.928 (0.858–0.966)	0.952 (0.909–0.985)	0.300

SVM support vector machine, KNN k-nearest neighbour, LDA linear discriminant analysis, SUR striatum uptake ratio, AUC area under the curve, CI confidence interval