18F-FDG-PET/CT-based machine learning model evaluates indeterminate adrenal nodules in patients with extra-adrenal malignancies

Table 3 Individual variables obtained from the ROC analysis for the differentiation of AMs from ABNs

Variables	Cutoff	AUC	Sensitivity	Specificity	Precision	Accuracy
Gender	–	0.656	0.659	0.711	0.600	0.670
CTU	32.5	0.788	0.742	0.711	0.776	0.784
SUV/liver	1.493	0.851	0.830	0.722	0.953	0.946
Clinical stage of primary cancer	–	0.688	0.703	0.918	0.459	0.659

ROC receiver operating characteristic curve, AMs adrenal metastases, ABNs adrenal benign nodules, CTU the value on unenhanced CT, AUC area under the receiver operating characteristic curve, SUV/liver the ratio of the adrenal, SUVmax to the mean liver SUV

ISSN: 1477-7819