Skip to main content

Table 4 Brain Regions Classifier predictive accuracy analysis in the training set

From: Brain metabolic pattern analysis using a magnetic resonance spectra classification software in experimental stroke

Training set Sensitivity Specificity PPV NPV AUC
Non-infarcted parenchyma 84% (27/32) 84% (64/76) 69% (27/39) 93% (64/69) 0.90 ± 0.06
SVZ 76% (41/54) 93% (50/54) 91% (41/45) 79% (50/63) 0.92 ± 0.04
Infarcted parenchyma 100% (22/22) 98% (84/86) 92% (22/24) 100% (84/84) 1.00 ± 0.00
  1. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and the mean ± standard error area under curve (AUC) value of the receiver operating curve (ROC) of the dichotomization of each class compared to the other two classes combined. Results are given in percentage with the number of spectra within parentheses. Total number of spectra in the training set (n = 108) corresponded to non-infarcted parenchyma (n = 32), SVZ (n = 54) and infarcted parenchyma (n = 22).