Meeting Banner
Abstract #2867

Test-Retest Reliability of Brain Volume Measurements

Julian R. Maclaren1, Zhaoying Han1, Sjoerd B. Vos1, 2, Christoph Seeger1, 3, Alexander Brost1, Nancy J. Fischbein1, Roland Bammer1

1Center for Quantitative Neuroimaging, Dept. of Radiology, Stanford University, Stanford, CA, United States; 2Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands; 3Pattern Recognition Lab, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany

The monitoring of neurodegenerative disease progression may be assisted by quantitative measurement of the volume of structures in the human brain. We performed 120 scans on 3 volunteers over a period of 31 days. The resulting dataset is available online and is a valuable resource for quantifying the repeatability of brain volume measurements. Results show that lateral ventricle volume, in particular, varies significantly from day to day. This could potentially mask the effects of disease and should be taken into consideration in any analysis performed using ventricle volume.

Abstract PDF Presentation Video

Keywords

advanced affect affecting aims alcohol allow allowed analyzed apparent applicable assist assisted assurance audience automated available back balance bars biological brain break caffeine calculation caution channel clinical coil comprised computed confidence confounding consistent correlation created dale days decoupled dehydration dept determine developments deviation deviations directly disease ensured errors exceeds exercise experiment expressed fluctuations foundation future greater head hydration hypothesize importantly improved included incorrectly indicate indicates intake inter interpreting interval intervals intra involved manual mask measured medial medical month neurology noise observe often paired particularly pattern percentage perhaps phantom positioning process progression project protocol quantification quantitative radiology recent recognition recommended recorded reflects reliability removed repeatability replicate repositioning researchers retest scanned scanner scanning sciences segmentation segmented separate session sessions seven since software stability structure structures subject subjects table target technology temporal thalamus throughout tools true twice twofold uncertain underestimates validation variability variance variation variations ventricles ventricular visually volume volumes volumetric volunteers whether worse