Focal Aβ-amyloid deposition precedes cerebral microbleeds and Superficial siderosis: a case report

1CSIRO Preventative Health National Research Flagship, Australian e-Health Research CentreBioMedIA, Brisbane, QLD, Australia 2Department of Radiology, University of Melbourne and the Royal Melbourne Hospital, Melbourne, VIC, Australia 3Department of Molecular Imaging & Therapy, Austin Health, Melbourne, VIC, Australia 4Department of Aged Care, Austin Health, Melbourne, VIC, Australia 5The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne, VIC, Australia 6Department of Medicine, Austin Health, University of Melbourne, VIC, Australia


Introduction
Cerebral amyloid angiopathy (CAA), a condition where Aβ-amyloid deposits in and around the media of small arteries and arterioles of the cerebral cortex and leptomeninges has been found to be present in most patients with Alzheimer's disease (AD) [1]. To this date, neuropathologic examination of the brain remains the only de initive method for diagnostic con irmation of CAA. However the combination of Aβ-amyloid imaging with 11 C-PiB-PET [2] and T2* susceptibility weighted (SWI) MR imaging [3], allows for the concomitant assessment of molecular and structural changes in-vivo.
There have not been any reports in the literature of longitudinal follow-up in subjects evaluating the relationship between MH, SS and Aβ-amyloid deposition. In this case-study we report the spatial and temporal relationship between Aβ-amyloid deposition and radiological markers of MH and SS.

Clinical evaluation
Studies were approved by the Austin Health Human Research Ethics Committee. Written informed consent was obtained prior to the examinations. Medical history was obtained from the participant, carer and from physical examination. Vascular risk factors were identi ied using self-report, physical examination and laboratory indings: hypertension, hypercholesterolemia, diabetes, current smoking history, atrial ibrillation, prior or current history of vascular disease (coronary or peripheral vascular disease), and dichotomised as present or absent according to published diagnostic guidelines.

Neuropsychological evaluation and Imaging data
Details of neuropsychological evaluation and imaging data are provided in the supplementary materials. Brie ly, a neuropsychological battery of assessments was conducted including MiniMental State Examination (MMSE) and Clinical Dementia Rating (CDR) as previously described [7]. 18 F-FDG and 11 C-PiB PET scans were obtained on a Phillips Allegro™ PET camera as previously described [2,8]. MRI imaging was obtained on a 3Tesla Siemens scanner with sequences including a 3D T1-Weighted Magnetization Prepared Rapid Gradient Echo (MPRAGE), a SWI and a Fluid Attenuated Inversion Recovery (FLAIR) scan.

Genetic profi le
ApoE allele subtype was determined by PCR ampli ication of genomic DNA.

Image analysis
All MRI were inspected blind to clinical and 11 C-PiB scan indings. Number and location of MH and SS were recorded for each SWI scan by an expert neuroradiologist (P.D). A difference map between the 2009 and 2007 SWI scans was computed and manually thresholded. The difference map and the manual segmentation were used to highlight evolving (change>5%), stable (change<5%) and new MH and SS. WMH volumes were computed from manually segmented FLAIR images.

Case report
A 65-year-old female with family history of AD, presented with memory complaints since the beginning of 2005. Physical examination in August 2005 revealed no focal neurological signs, normal blood tests (including B12, folate and thyroid function), normal blood pressure (140/85 mm Hg) and no cardiovascular risk factors. Participant reported no use of anticoagulant medication. Her MMSE at that stage was 28/30, CDR was 0.5 and she was classi ied as suffering mild cognitive impairment (MCI). She underwent an 18 F-FDG PET study that was suggestive of AD, with evidence of mild to moderate temporoparietal hypo metabolism, with sparing of the frontal, sensorimotor and occipital cortices ( Figure 1A). Seven months after initial presentation, she underwent her irst PiB study while her MMSE decreased to 27/30. There was progressive cognitive decline and 18 months after initial presentation she was diagnosed as probable AD and started on Galantamine (dose increased up to 16 mg/d).
She was subsequently enrolled in the AIBL study [9] and she received her second PiB PET scan and irst set of MRIs at 24 months after initial presentation. At that stage, her MMSE was 22/30 with a CDR of 1.0; there were still no focal neurological signs and blood tests were normal. In April 2009, after her second set of MRIs and before her next 11 C-PiB or neuropsychological evaluation, she was withdrawn from the study by her husband. The demographic information of the subject along with memory scores and quantitative imaging measures at different time-points are summarized in table 1.  C-PiB retention, in frontal, striatum, anterior cingulate, temporal and parietal cortices, particularly on the right side, with sparing of the sensorimotor and occipital cortices ( Figure 1B). The PiB PET study from 2007 showed similar but more extensive focal 11 C-PiB retention ( Figure 1C). This was also evidenced through Z-score maps of the subject's follow-up PiB scan compared to AD patients (Supplementary Figure 1).  Table 1).

Discussion
As previously reported MH and SS were intimately associated with Aβ-amyloid deposition [6,10]. While lobar MH are a frequent inding in AD patients or even in cognitively-normal older individuals, they are strongly associated with increasing age and Aβ-amyloid deposition [11]. This association between Aβ-amyloid and vascular lesions has crucial implications not only for the selection and risk strati ication of individuals undergoing anticoagulant therapies, but also in those enrolled in anti-Aβamyloid therapeutic trials [12]. It is thought that the typical Aβ-amyloid deposits in and around the media of small arteries and arterioles of the brain weakens the vessel walls predisposing to MH.
The focal and asymmetrical nature of the Aβ-amyloid deposition in this case is of particular interest, and in contrast to the diffuse cortical pattern usually observed in AD. Conversely, the clinical history was consistent with Alzheimer's type dementia, and without classical features of stepwise cognitive decline or progressive neurological signs to suggest a vascular aetiology. While it has been proposed that the regional pattern of Aβ-amyloid deposition may separate AD from CAA, without histopathological con irmation it is impossible to ascertain if the 11 C-PiB retention in our patient is mainly attributable to vascular deposits, plaques, or both.
Much more striking are the observations derived from the longitudinal aspect of the study, where areas with focal Aβ-amyloid deposition and no MH in the respective PiB and MRI SWI studies in 2007, matched the new MH and larger areas of SS observed in the MRI SWI study from 2009. All this imaging data points to CAA, both by the location of the MH-areas largely towards the surface of the cerebral cortex where the interstitial luid normally diffuses on its way out of the brain-as well as for the SS usually the end result of repeated small haemorrhages, most likely to happen in leptomeningeal arteries as they transverse from the subarachnoid space penetrating into the parenchyma and forming cortical arteries. Both the leptomeningeal and cortical arteries have been found to be heavily involved in CAA [12].
The distribution of WMH in this patient is in agreement with previous reports on subjects with AD, CAA and vascular dementia [18]. There were no WMH in the brainstem on either the 2007 or 2009 MRI scans; an observation that has been linked with increased likelihood of cerebrovascular disease [14]. The involvement of capillaries in CAA, more prominent in subjects with APOE e4 alleles, has been linked with the development of occlusions [15] and presumably ischemia related white matter hyperintensities.
An increase of almost 12% in Aβ-amyloid deposition was also detected in the 16 months between the 2006 and 2007 PiB studies, higher than our reported 5.7% in a large cohort of AD patients [16] but similar to the increase in Aβ-amyloid deposition reported by Rinne and colleagues in AD patients on the placebo arm of their therapeutic trial [17]. This increase in neocortical Aβ-amyloid deposition was accompanied by a decline in cognitive performance leading to the clinical diagnosis of AD.
This study demonstrates that the combination of different imaging techniques provides essential complementary information crucial to ascertain the underlying pathological disorder [18]. Further longitudinal population-based studies combining clinical information with these neuroimaging techniques will help better elucidate the relationship between Aβ-amyloid deposition and microhemorrhages and super icial siderosis.