“…23,25 Use of ROC curves has been extended to evaluation of the performance of multivariate models for diagnosis, prognosis, machine learning (e.g., image or voice recognition), and data mining. Recently, Amato et al 26 conducted a cross-sectional study using clinical information and a bank of images from 110 patients who had undergone angiotomography of the aorta to predict identification of the artery of Adamkiewicz. Using multivariate analysis, they constructed a predictive model and evaluated its discriminatory properties using a ROC curve, showing that it enabled correct identification in 61% of the patients using a combination of nine covariates.…”