Improved Cerebrospinal Fluid-Based Discrimination between Alzheimer's Disease Patients and Controls after Correction for Ventricular Volumes
SourceJournal of Alzheimer's Disease, 56, 2, (2017), pp. 543-555
Article / Letter to editor
Display more detailsDisplay less details
Journal of Alzheimer's Disease
SubjectRadboudumc 1: Alzheimer`s disease DCMN: Donders Center for Medical Neuroscience; Radboudumc 3: Disorders of movement DCMN: Donders Center for Medical Neuroscience
Cerebrospinal fluid (CSF) biomarkers may support the diagnosis of Alzheimer's disease (AD). We studied if the diagnostic power of AD CSF biomarker concentrations, i.e., Abeta42, total tau (t-tau), and phosphorylated tau (p-tau), is affected by differences in lateral ventricular volume (VV), using CSF biomarker data and magnetic resonance imaging (MRI) scans of 730 subjects, from 13 European Memory Clinics. We developed a Matlab-algorithm for standardized automated segmentation analysis of T1 weighted MRI scans in SPM8 for determining VV, and computed its ratio with total intracranial volume (TIV) as proxy for total CSF volume. The diagnostic power of CSF biomarkers (and their combination), either corrected for VV/TIV ratio or not, was determined by ROC analysis. CSF Abeta42 levels inversely correlated to VV/TIV in the whole study population (Abeta42: r = -0.28; p < 0.0001). For CSF t-tau and p-tau, this association only reached statistical significance in the combined MCI and AD group (t-tau: r = -0.15; p-tau: r = -0.13; both p < 0.01). Correction for differences in VV/TIV improved the differentiation of AD versus controls based on CSF Abeta42 alone (AUC: 0.75 versus 0.81) or in combination with t-tau (AUC: 0.81 versus 0.91). In conclusion, differences in VV may be an important confounder in interpreting CSF Abeta42 levels.
Upload full text