QyScore® MRI markers' diagnostic accuracy in the clinical spectrum of Alzheimer's disease

Abstract

AbstractBackgroundDiagnostic criteria for Alzheimer’s disease (AD) acknowledge a key role of imaging markers such as hippocampus and amygdala volumes, for early diagnosis. Here we aim at assessing the diagnostic accuracy of QyScore® medial temporal lobe atrophy markers in distinguishing AD dementia and Mild Cognitive Impairment (MCI) individuals from healthy controls (HC) in US and EU cohorts.Method3DT1 images were selected from different cohorts (ADNI, MEMORA, NACC, OASIS3) and analyzed using the QyScore® automatic segmentation pipeline. Brain volumes of the hippocampus and the amygdala were normalized to the total intracranial volume. Analyses of variance were performed to compare volumetric measurements among clinical groups. Areas under the curve were calculated to identify the diagnostic accuracy of QyScore® markers in discriminating AD and MCI from HC.ResultThe sample was composed of 425 AD, 909 MCI and 993 HC with a higher prevalence of men and Apoe Ɛ4 carriers in each group. Analysis of variance showed a significant difference in hippocampal and amygdala volumes among groups. Preliminary results on the area under the curve (AUC) showed a good diagnostic accuracy in discriminating AD vs HC and MCI versus HC both for the hippocampal volumes (ADvsHC: AUC=0.86, 95%CI: 0.832‐0.877; MCIvsHC: AUC =0.75, 95%Ci: 0.726‐0.770), and the amygdala volumes (ADvsHC: AUC =0.77, 95%CI: 0.737‐0.791; MCIvsHC: AUC =0.70,95%CI 95%: 0.675‐0.720).ConclusionOur results highlight the good diagnostic performance of QyScore® markers, and support the relevance of implementing these markers in the diagnostic workflow of AD patients.