Table 2

Microbleed, plaque and miscellaneous papers: 7T MRI in vivo studies including subjects with AD

PaperImaging methodsParticipantsQuestionResults
Microbleed and microinfarct papers:
Brundel et al 34 7T MRI: dual-echo 3D T2*-weighted acquisition; voxel size 0.5×0.5×0.7 mm
3T MRI: 3D T2*-weighted acquisition; voxel size 0.99×0.99×3 mm
Visual identification of microbleeds (MARS criteria)68
What is the true prevalence of microbleeds in AD/MCI?At 7T MRI≥1 microbleeds in 78% of AD/MCI vs 44% NC (sig).
At 3T MRI≥1 microbleeds in 33% of AD/MCI vs 17% NC (non-sig).
Higher quantity of microbleeds in AD/MCI (max 80 in one subject) than NC (max 5 in one subject) at 7 T (sig).
van Veluw et al 60 3D FLAIR acquisition; voxel size 0.8×0.8×0.8 mm
3D T2-weighted acquisition; voxel size 0.7×0.7×0.7
3D T1-weighted acquisition; voxel size 1.0×1.0×1.0 mm
3D dual-echo gradient echo acquisition; voxel size 0.5×0.5×0.7 mm
Semiautomated identification of microbleeds69
Are cerebral microinfarcts (CMI) increased in aMCI/AD?CMIs found in 45% NC and 55% aMCI/AD (non-sig).
No sig diff between CMIs in aMCI and AD.
In all subjects, CMIs uniformly distributed throughout cortex.
No relationship between MTL atrophy or MMSE and number of CMIs.
Amyloid plaque papers:
Nakada et al (2008)30 Parietal association cortex SWI: T2*-weighted 2D GRE acquisition; voxel size 0.156×0.156×3 mm; scan time 3:48 min
Visual identification of plaques described in Ref. 70
Can senile plaques be visualised using 7T MRI in vivo in AD?Hypointense ‘black dots’ (‘senile-plaque-like-pathology’) seen throughout parietal cortex in 10/10 AD, in 2/10 ONC, in 0/20 YNC.
Sig diff between YNC and ONC and between AD and ONC (Ryan multiple comparison).
van Bergen et al 33 Structural: T1-weighted MP2-RAGE acquisition; voxel size 0.6×0.6×0.6 mm; scan time 7:50 min
QSM: 3D GRE acquisition with 3 echoes; voxel size 0.5×0.5×0.5 mm; scan time 13:48 min
rs-fMRI: 3D T2-prep GRE acquisition; voxel size 1.5×1.5×1.5 mm; scan time 7:03 min
Also: C[11]-PiB-PET
QSM maps created in multiple steps. Laplacian-based phase unwrapping converted to frequency shift images (in hertz), inverse dipole calculation used to obtain susceptibility maps, values reported relative to reference CSF71
rs-fMRI images created using SPM12 (
NC=22 (7 APOE e4 positive)
MCI†=15 (6 APOE e4 positive)
What is the relationship between cerebral iron (using QSM) MCI and APOE-e4 status?
What are the relationships between cerebral iron burden, Aβ -plaque density and MPFC-coupling (using rs-fMRI)?
APOE e4 associated with increased cortical iron and higher Aβ-plaque-load in MCI (sig); not in NC.
High iron burden in MCI associated with increased MPFC-coupling in ROIs, including cingulate, paracingulate, frontal regions (sig).
In areas with increased MPFC coupling iron-burden and Aβ-plaque-load correlated (sig).
van Rooden et al 31 T2*-weighted 2D GRE acquisition; 20 slices; voxel size 0.24×0.24×1 mm (FOV included frontal and parietal regions); scan time approximately 10 min
Hippocampus: 2D T2*-weighted GRE acquisition; 32 slices; voxel size 0.5×0.5×3 mm; scan time approximately 6 min
Visual identification of plaques
Phase shift values for four different ROIs (right and left TMP, frontal and parietal) averaged and phase shift with subcortical WM calculated
Can amyloid plaques be visualised in the cerebral cortex using 7T MRI in vivo in AD?No focal hypointensities found.
Increased cortical phase shift in left TMP (AD=0.90, NC=79), right TMP (AD=0.97, NC=0.85), frontal region (AD=0.70, NC=0.62), parietal region(AD=0.87, NC=0.74) in AD (all sig).
Phase shift difference between groups in hippo (left hippo AD=0.09, NC=0.07; right hippo AD=0.10, NC=0.08) not sig diff.
Association between whole brain phase shift and MMSE scores r=−0.54 (sig).
van Rooden et al 32 T2*-weighted 2D GRE acquisition; 20 slices; voxel size 0.24×0.24×1 mm; scan time approximately 10 min (FOV included frontal and parietal regions)
Phase shift values for four different ROIs (right and left TMP, frontal and parietal) averaged and phase shift with subcortical WM calculated
EOAD*=12 (onset before 65 years)
LOAD*=17 (onset after 65 years)
Does amyloid deposition (measured as per van Rooden et al)31 differ between EOAD and LOAD?Increased cortical phase shift in LOAD/ EOAD vs NC in right TMP (LOAD=1.23, EOAD=1.31, NC=0.98), left TMP (LOAD=1.18, EOAD=1.25, NC=0.96), frontal (LOAD=1.04, EOAD=1.15, NC=0.88), parietal (LOAD=1.16, EOAD=1.27, NC=0.98) regions and whole brain (LOAD=1.15, EOAD=1.25, NC=0.95).
Phase shift in all regions and whole brain sig diff between NC and LOAD, between NC and EOAD, between LOAD and EOAD groups.
Miscellaneous papers:
Cai et al 72 T2-weighted 3D TSE SPACE acquisition; 224 slices; voxel size 0.4×0.4×0.4 mm; scan time approximately 7.5 min (interpolated reconstruction from 0.42×0.42×1 mm acquired voxels)
PVS automatically segmented using MATLAB
Can PVS be imaged and quantified using 7T MRI in vivo?Increase in PVS density in AD vs NC (AD=8.0, NC=4.9) (sig).
Versluis et al 44 T2*-weighted acquisition with additional navigator echo technique application; 20 slices; voxel size 0.24×0.24×1 mm; scan time approximately 10 minNC=5
What causes increased  imaging artefacts in AD? Does application of an additional navigator echo technique correct for increased artefact production?Increase in AD group artefact production due to increase in F0 variations (physiological, eg, due to chest volume changes with breathing) and large jumps (overt movement).
Application of navigator echo technique corrected for increased F0 variations, improved image quality, reduced ghosting in 9/10 AD scans.
  • AD diagnostic criteria used: *,2.

  • MCI diagnostic criteria used:†24.

  • AD diagnostic criteria used: ‡criteria not specified.

  • 3D, three dimensional; Aβ, amyloid beta; AD, Alzheimer’s disease; B0, static field strength; C[11] -PiB, Carbon 11 Pittsburgh B compound (binds to amyloid); CA, cornu ammonis; CSF, cerebrospinal fluid; DG, dentate gyrus; EOAD, early-onset AD; ERC, entorhinal cortex; FA, fractional anisotropy; FSE/TSE, fast/turbo spin Echo (these terms are synonymous); GM, grey matter; GRE, gradient echo sequence; hippo, hippocampus; ICV, intracranial volume; LOAD, late-onset AD; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; MP2-RAGE, magnetisation-prepared 2 rapid acquisition gradient echo; MTL, medial temporal lobe; NC, cognitively normal controls; ONC, older cognitively normal controls; PHC, parahippocampal cingulum; PVS, perivascular space; QSM, quantitative susceptibility mapping; rs-fMRI, resting state functional MRI; RF, radio frequency; ROI, region of interest; sig, significant (p<0.05); SP, stratum pyramidale; SPACE, sampling perfection with application optimised contrasts by using different flip angle evolutions; SRLM, strata radiatum, lacunosum and moleculare; SUB, subiculum; SWI, susceptibility-weighted imaging; TE, echo time; THV, total hippocampal volume; TMP, temporoparietal region; WM, white matter; YOC, younger cognitively normal controls.