Fig. 4
From: A deep learning MRI approach outperforms other biomarkers of prodromal Alzheimer’s disease
Predicting progression to Alzheimer’s dementia. Survival analyses were performed comparing the deep learning MRI scores to other measures, and example curves illustrate that the deep learning MRI score (left panel) outperforms the CSF measure of the tau/Aβ ratio (right panel). The high risk (indicated by red) and low risk (indicated by blue) curves were fitted from 75% and 25% percentile of the measures, respectively. The shaded area indicates the 95% confidence interval. The deep learning MRI scores outperformed CSF Aβ, tau, or tau/Aβ; MRI-derived measures of hippocampal volume, entorhinal cortex volume, and entorhinal thickness; behavioral measures, Mini-Mental State Exam (MMSE), and RAVLT retention; and, when available, PET measures of amyloid using the AV45 radioligand or metabolism using fluorodeoxyglucose (FDG)