The investigation of hippocampal and hippocampal subfield volumetry, morphology and metabolites using 3T MRI

McLean, John (2012) The investigation of hippocampal and hippocampal subfield volumetry, morphology and metabolites using 3T MRI. PhD thesis, University of Glasgow.

Abstract

A detailed account of the hippocampal anatomy has been provided. This thesis will explore
and exploit the use of 3T MRI and the latest developments in image processing techniques
to measure hippocampal and hippocampal subfield volumes, hippocampal metabolites and
morphology.
In chapter two a protocol for segmenting the hippocampus was created. The protocol was
assessed in two groups of subjects with differing socioeconomic status (SES). This was a
novel, community based sample in which hippocampal volumes have yet to be assessed in
the literature.
Manual and automated hippocampal segmentation measurements were compared on the two
distinct SES groups. The mean volumes and also the variance in these measurements were
comparable between two methods. The Dice overlapping metric comparing the two methods
was 0.81.
In chapter three voxel based morphometry (VBM) was used to compare local volume differences
in grey matter volume between the two SES groups. Two approaches to VBM were
compared. DARTEL-VBM results were found to be superior to the earlier ’optimised’ VBM
method. Following a small volume correction, DARTEL-VBM results were suggesitive of
focal GM volumes reductions in both the right and left hippocampi of the lower SES group.
In chapter four an MR spectroscopy protocol was implemented to assess hippocampal metabolites
in the two differing SES groups. Interpretable spectra were obtained in 73% of the 42
subjects. The poorer socioeconomic group were considered to have been exposed to chronic
stress and therefore via inflammatory processes it was anticipated that the NAA/Cr metabolite
ratio would be reduced in this group when compared to the more affluent group. Both
NAA/Cr and Cho/Cr hippocampal metabolite ratios were not significantly different between
the two groups.
The aim of chapter 5 was to implement the protocol and methodology developed in chapter
2 to determine a normal range for hippocampal volumes at 3T MRI.

3D T1-weighted IR-FSPGR images were acquired in 39 healthy, normal volunteers in the
age range from 19 to 64. Following the automated procedure hippocampal volumes were
manually inspected and edited.
The mean and standard deviation of the left and right hippocampal volumes were determined
to be: 3421mm3 ± 399mm3 and 3487mm3 ± 431mm3 respectively. After correcting for total
ICV the volumes were: 0.22% ± 0.03% and 0.23% ± 0.03% for the left and right hippocampi
respectively.
Thus, a normative database of hippocampal volumes was established. The normative data
here will in future act as a baseline on which other methods of determining hippocampal
volumes may be compared. The utility of using the normative dataset to compare other
groups of subjects will be limited as a result of the lack of a comprehensive assessment of IQ
or education level of the normal volunteers which may affect the volume of the hippocampus.
In chapter six Incomplete hippocampal inversion (IHI) was assessed. Few studies have assessed
the normal incidence of IHI and of those studies the analysis of IHI extended only
to a radiological assessment. Here we present a comprehensive and quantitative assessment
of IHI. IHI was found on 31 of the 84 normal subjects assessed (37%). ICV corrected IHI
left-sided hippocampal volumes were compared against ICV corrected normal left-sided hippocampal
volumes (25 vs. 52 hippocampi). The IHI hippocampal volumes were determined
to be smaller than the normal hippocampal volumes (p<< 0.05). However, on further inspection
it was observed that the ICV of the IHI was significantly smaller than the ICV of
the normal group, confounding the previous result.
In chapter seven a pilot study was performed on patients with Rheumatoid Arthritis (RA).
The aim was to exploit the improved image quality offered by the 3T MRI to create a
protocol for assessing the CA4/ dentate volume and to compare the volume of this subfield
of the hippocampus before and after treatment. Two methodologies were implemented.
In the first method a protocol was produced to manually segment the CA4/dentate region
of the hippocampus from coronal T2-weighted FSE images. Given that few studies have
assessed hippocampal subfields, an assessment of study power and sample size was conducted
to inform future work.
In the second method, the data the DARTEL-VBM image processing pipeline was applied.
Statistical nonparametric mapping was applied in the final statistical interpretation of the
VBM data. Following an FDR correction, a single GM voxel in the hippocampus was deemed to be statistically significant, this was suggestive of small GM volume increase following antiinflammatory
treatment.
Finally, in chapter eight, the manual segmentation protocol for the CA4/dentate hippocampal
subfield developed in chapter seven was extended to include a complete set of hippocampal
subfields. This is one of the first attempts to segment the entire hippocampus into its subfields
using 3T MRI and as such, it was important to assess the quality of the measurement procedure.
Furthermore, given the subfield volumes and the variability in these measurements,
power and sample size calculations were also estimated to inform further work.
Seventeen healthy volunteers were scanned using 3T MRI. A detailed manual segmentation
protocol was created to guide two independent operators to measure the hippocampal subfield
volumes. Repeat measures were made by a single operator for intra-operator variability
and inter-operator variability was also assessed. The results of the intra-operator comparison
proved reasonably successful where values compared well but were typically slightly poorer
than similar attempts in the literature. This was likely to be the result of the additional
complication of trying to segment subfields in the head and tail of the hippocampus where
previous studies have focused only on the body of the hippocampus. Inter-rater agreement
measures for subfield volumes were generally poorer than would be acceptable if full exchangeability
of the data between the raters was necessary. This would indicate that further
refinements to the manual segmentation protocol are necessary. Future work should seek to
improve the methodology to reduce the variability and improve the reproducibility in these
measures.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Hippocampus, subfields, metabolites, MRI, 3T, malrotation, deprivation, socioeconomic status
• Subjects: R Medicine > R Medicine (General); R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine; T Technology > T Technology (General)
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Psychology & Neuroscience
• Supervisor's Name: Condon, Prof. Barrie
• Date of Award: 2012
• Depositing User: Mr John McLean
• Unique ID: glathesis:2012-3133
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 23 Feb 2012
• Last Modified: 10 Dec 2012 14:04
• URI: https://theses.gla.ac.uk/id/eprint/3133