Publication Type:
Journal ArticleSource:
J Neuroimaging, Volume 24, Issue 5, p.509-14 (2014)Keywords:
Aged, Aged, 80 and over, Alzheimer Disease, Anatomic Landmarks, Europe, Guideline Adherence, Hippocampus, Humans, Imaging, Three-Dimensional, Magnetic Resonance Imaging, Male, Observer Variation, Practice Guidelines as Topic, Reproducibility of Results, Sensitivity and SpecificityAbstract:
<p><b>BACKGROUND AND PURPOSE: </b>An effort to define and validate a Harmonized Protocol for standard hippocampal segmentation is being carried out. We wished to estimate the effect of magnetic resonance image (MRI) spatial orientation on manual hippocampal segmentations to define optimal standard orientation of MRIs for hippocampal volumetry.</p><p><b>METHODS: </b>Three expert tracers segmented twice the hippocampi of 10 ADNI subjects on MRI slices oriented perpendicular to the anterior-posterior commissure (AC-PC) line and the long hippocampal axes plane, following internationally harmonized landmarks. We computed intra and interrater reliability figures for total volumes and similarity coefficients.</p><p><b>RESULTS: </b>Total volume reliability was similar for both orientations. Similarity coefficients were significantly higher for the AC-PC orientation (exact P = 0.002).</p><p><b>DISCUSSION: </b>These data show that AC-PC orientation is slightly more reliable for manual segmentations, possibly due to better visualization of the cerebrospinal fluid spaces separating hippocampal head and amygdala. A Delphi panel of experts has used these data to decide on the optimal orientation for a Harmonized Protocol for hippocampal segmentation.</p>