Table 1 Average ± standard deviation test characteristics for inferring glioblastoma genetic biomarkers.

From: A fully automated artificial intelligence method for non-invasive, imaging-based identification of genetic alterations in glioblastomas

BiomarkerPrevSensSpecPrecRecallF1MCCAUC
ATRX19/1900.94 ± 0.070.92 ± 0.040.60 ± 0.120.94 ± 0.070.73 ± 0.080.71 ± 0.080.97 ± 0.02
IDH18/1950.93 ± 0.080.88 ± 0.070.50 ± 0.240.93 ± 0.080.62 ± 0.160.62 ± 0.160.95 ± 0.03
7/10 aneuploidy47/670.90 ± 0.090.88 ± 0.080.95 ± 0.040.90 ± 0.090.92 ± 0.060.75 ± 0.180.93 ± 0.06
CDKN241/690.76 ± 0.060.86 ± 0.090.90 ± 0.060.76 ± 0.060.82 ± 0.040.62 ± 0.080.85 ± 0.04
EGFR80/1890.66 ± 0.060.68 ± 0.100.61 ± 0.070.66 ± 0.060.63 ± 0.050.34 ± 0.100.70 ± 0.06
TERT64/830.77 ± 0.150.59 ± 0.140.86 ± 0.030.77 ± 0.150.81 ± 0.090.36 ± 0.150.65 ± 0.08
PTEN106/1910.63 ± 0.090.66 ± 0.090.70 ± 0.040.63 ± 0.090.65 ± 0.040.28 ± 0.050.64 ± 0.03
TP53112/1860.57 ± 0.140.59 ± 0.130.44 ± 0.050.57 ± 0.140.49 ± 0.060.16 ± 0.080.57 ± 0.05
MGMT140/1900.56 ± 0.070.56 ± 0.110.52 ± 0.060.56 ± 0.070.53 ± 0.040.12 ± 0.100.55 ± 0.07
  1. Prevalence (Prev.) refers to the genetic biomarker prevalence in the study cohort. Test characteristics for the best point on the ROC curve (i.e. closest to [0,1]) include sensitivity (Sens.), specificity (Spec.), precision (Prec.), recall, F1 statistic (F1), Matthew’s correlation coefficient (MCC), and the ROC AUC (AUC).