Advertisement

Multisite hippocampal subfields reproducibility: A european 3T study

      Background

      Hippocampal subfields are differentially affected in Alzheimer’s Disease (Hanseeuw BJ, 2011; Lim HK, 2012). However, their volume reproducibility has been poorly investigated. We report the evaluation of the across-session test-retest reproducibility of the hippocampal subfields segmentations derived from Freesurfer in 65 healthy elderly subjects. We compared the subfields reproducibility i) averaging or not two within session T1 images and ii) relative to the whole hippocampus.

      Methods

      Five healthy local volunteers (55-90 ys) were enrolled in 13 3T MRI sites (Siemens, GE, Philips) across Europe and were scanned in two sessions at least a week apart. All analyses were performed using the longitudinal pipeline of Freesurfer v5.1.0 (Reuter M, 2012; Jovicich J, 2013) on the neuGRID platform (https://neugrid4you.eu/). The whole hippocampal volume was extracted from the “aseg.stat” file of Fresurfer. For each site and ROI, volumes reliability was computed evaluating test-retest absolute differences relative to the mean (absolute error) and test-retest spatial reproducibility (DICE).

      Results

      Subfields analysis was focused on: Cornu Ammonis (CA) 1, CA2-3, CA4-dentate gyrus (DG), subiculum, presubiculum, fimbria and hippocampal fissure. Within session averaging of two T1 images gave a significant improvement in the mean test-retest reproducibility of all hippocampal subfields. Absolute errors across MRI sites was comparable to that found for whole hippocampus (about 2%) for CA2-3, CA4-DG and subiculum, about 5% for CA1 and presubiculum, around 15% for fimbria and hippocampal fissure. The DICE results were in line with the absolute error analysis (excellent for CA2-3, CA4-DG, subiculum, good for CA1 and presubiculum, poor for fimbria and hippocampal fissure).

      Conclusions

      Averaging the within session T1 images allowed to improve the Freesurfer subfields reproducibility at each site. Despite the differences of the 13 MRI scanner configurations we found good and consistent hippocampal subfields reproducibility for CA2-3, CA4-DG, subiculum, CA1 and presubiculum. Pharmacog is funded by the EU-FP7 for the Innovative Medicine Initiative (grant 115009).