1H MR spectroscopy in common dementias.

磁共振波谱

磁共振波谱

Kantarci et al.Page 2

modalities, because the signals from several different metabolites are measured within a single

measurement period. Each metabolite in turn is sensitive to a different aspect of in- vivo

pathologic processes at the molecular or cellular level. In principle, it may be possible to

generate a disease-specific spectroscopic profile with multiple contributing components. The

metabolite N-acetylaspartate (NAA) is consistently found to be lower and the metabolite myo-

inositol (mI) higher in 1H MRS studies of patients with AD than cognitively normal elderly

(3-8). There are conflicting reports about choline (Cho) levels in AD. Some studies identified

elevated Cho levels in people with AD, and some did not (9). NAA levels are reduced in patients

with vascular dementia (VaD), and white matter NAA is lower in patients with VaD than in

patients with AD (10-12). Gray matter MI /Cr levels on the other hand are normal in patients

with VaD (13). 1H MRS metabolite changes in FTLD are similar to AD, with lower NAA /Cr

and higher mI /Cr levels than normal (14,15).

1H MRS studies in common dementias are limited to comparing the 1H MRS findings in ADNIH-PA Author Manuscript

MethodsNIH-PA Author Manuscript

NIH-PA Author Manuscriptto other dementias (14), FTLD (15) and VaD (11-13,16). We are not aware of any study thathas compared the 1H MRS metabolite profiles among the most common dementias. Theobjectives of this study were to identify the 1H MRS metabolite changes in each of the commondementias with respect to normal, and to identify 1H MRS metabolite differences among thecommon dementias.Recruitment of SubjectsAll subjects of this project were recruited through the Alzheimer's Disease Research Center(ADRC) and Alzheimer's Disease Patient Registry (ADPR) at the Mayo Clinic, Rochester(17), as part of an ongoing IRB approved MRI /1H MRS project. All individuals participatingin the ADRC/ADPR are evaluated by a behavioral neurologist and a neuropsychologist. Theyundergo neurological examination, neuropsychological testing, and basic laboratory testing.

At the completion of the evaluation, a consensus committee meeting is held involving the

behavioral neurologists, neuropsychologists, nurses and the geriatrician who evaluate the

subjects. All of the subjects undergo structural brain MRI. Subjects with structural

abnormalities other than vascular disease that could produce dementia, such as tumors and

subdural hematoma, as well as those who had concurrent illnesses or treatments interfering

with cognitive function other than AD, VaD, DLB and FTLD were excluded from this study.

We only studied patients > 40 years of age.

A diagnosis of dementia is based on DSM-IIIR criteria (18). The clinical diagnoses of AD,

VaD, DLB, and FTLD are made using published consensus criteria (19). Only patients meeting

the criteria for one of the dementias (AD, VaD, DLB, and FTLD) were included in this study.

Patients who met the criteria for two or more of the different dementias were excluded.

Normals are individuals who; 1) are independently functioning community dwellers 2) do not

have active neurological or psychiatric conditions, 3) have no cognitive complaints, 4) have a

normal neurological exam, 5) are not taking any psychoactive medications in doses that would

impact cognition.

Single Voxel 1H MRS

1H MRS studies were performed with the LX system automated single voxel 1H MRS package:

PROBE /SV (General Electric Medical Systems, Milwaukee, WI) (20). An eight cm3 (2×2×2

cm) voxel was placed on a mid-sagittal T1-weighted image covering right and left posterior

cingulate gyri and inferior precunei (Figure 1). Point resolved spectroscopy (PRESS) pulse

sequence with TR = 2000 ms, TE=30 ms, 2048 data points and 128 excitations were used for

Neurology. Author manuscript; available in PMC 2009 October 26.

磁共振波谱

Kantarci et al.Page 3

the examinations. The prescan algorithm of PROBE automatically adjusts the transmitter and

receiver gains and center frequency. The local magnetic field homogeneity is optimized with

the three-plane auto-shim procedure, and the flip angle of the third water suppression pulse is

adjusted for chemical-shift-water suppression (CHESS) prior to PRESS acquisition. The single

voxel 1H MRS scan time is about 6 min. Metabolite intensity ratios are automatically calculated

at the end of each PROBE /SV acquisition. 1H MRS variables that were analyzed for this study

are NAA /Cr, Cho /Cr, and mI /Cr ratios.

Statistical Analysis

Gender differences among the clinical groups (normal, AD, FTLD, DLB, and VaD) were tested

with the chi squared test. ANOVA was used to compare clinical groups with respect to age,

Clinical Dementia Rating (CDR), Dementia Rating Scale (DRS), Mini-Mental State Exam

(MMSE), and the three metabolite ratios. In order to control the experiment-wise error rate,

we first tested for differences in each metabolite ratio among the five groups (ie. including

normals) using a one-way ANOVA. If the test was significant at the 0.05 level, we tested which

groups differed from normals using Dunnett's test. We then performed a one-way ANOVA

excluding the normals. If this dementia-group ANOVA was significant at the 0.05 level, we

used Fisher's Least Significant Difference (LSD) procedure to make pairwise comparisons of

metabolite ratios among the groups. All tests were two-sided. The effects of age on metabolite

ratios were tested by linear regression analysis and the gender effects were tested with student's

t-tests in normal subjects only, as the data in normals should be free of the confounding effects

of disease.NIH-PA Author ManuscriptNIH-PA Author Manuscript

NIH-PA Author ManuscriptResultsDemographic AspectsWe consecutively recruited 206 cognitively normal elderly volunteers, 121 patients with AD,41 patients with FTLD, 20 patients with DLB, and 8 patients with VaD, from October 2001

until January 2003. We excluded 36 patients from this study who fulfilled the diagnostic criteria

for more than one dementia, because they constitute a clinically and pathologically

heterogeneous group, and there are no accepted research criteria available to classify the

different mixed dementia subtypes. Demographic aspects of the study groups are listed in table

1. There were no gender differences between the clinical groups (chi square p=0.06). Mean

ages of clinical groups were different (ANOVA p<0.0001). Patients with VaD were older than

patients with DLB (p=0.04). Patients with FTLD were younger than all other clinical groups

(normals, AD, VaD, DLB) (p<0.01 for all comparisons). We did not find any association

between age and meatbolite ratios in the normals, probably due to the older age range of our

subject group (Supplemental data). Because there was no association between age or gender

(p>0.05) and metabolite ratios in normals, metabolite measurements were not adjusted for age

or gender in normals or patients. The DRS and MMSE scores of patients with different

dementias were lower than the normals (p<0.0001 for all comparisons). There were no

differences between the CDR, DRS, and MMSE scores of patients with different dementias

on ANOVA, indicating that clinical dementia severity was similar between the dementia

groups. Ninety one out of 121 patients with AD, 16 /41 patients with FTLD, and 17 /20 patients

with DLB were being treated with cholinesterase inhibitors.

All patients with vascular dementia had leukoaraiosis on MRI. In addition, four had cortical

and four had subcortical infarctions. None of the lesions involved the posterior cingulate gyri

and inferior precunei. The 1H MRS voxel included only the normal appearing brain tissue on

MRI in all subjects.

Neurology. Author manuscript; available in PMC 2009 October 26.

磁共振波谱

Kantarci et al.

1H MRS differences between different dementias and normal elderlyPage 4

The 1H MRS metabolite differences from normal in patients with AD, FTLD, DLB, and VaD,are listed in table 2. NAA /Cr ratios were lower than normal in patients with AD (p<0.001),VaD (p=0.08), and FTLD (p<0.001), but not different from normal in patients with DLB. Cho /Cr ratios were higher than normal in patients with AD (p=0.05), DLB (p=0.001), and FTLD(p=0.02), but not different from normal in patients with VaD. MI /Cr ratios were higher thannormal in patients with AD and FTLD (p<0.001 for both comparisons), but not different fromnormal in patients with DLB and VaD. Examples of 1H MRS from each of the clinical groupsare shown in Figure 2.

1H MRS differences between the dementiasNIH-PA Author Manuscript

NIH-PA Author Manuscript

NIH-PA Author ManuscriptThe 1H MRS metabolite ratios, NAA /Cr (p= 0.009) and mI /Cr (p= 0.008) were differentbetween the dementia groups on ANOVA. Cho /Cr was not different between dementiagroups. 1H MRS metabolites that differed between dementias are summarized in Table 3.NAA /Cr was higher in patients with DLB than AD and FTLD (p<0.01 for both comparissons).MI /Cr was higher in patients with AD than VaD (p=0.02), and higher in patients with FTLDthan DLB (p=0.02) and VaD (p=0.003). There were no differences between patients with ADand FTLD, nor between VaD and DLB.DiscussionThe objective of this study was to identify 1H MRS metabolite signatures of commondementias, in order to determine if 1H MRS may be useful in differentiating them. While theliterature on neuroimaging findings in people with AD is extensive, the neuroimaging literatureon other common dementias is limited. Cross-sectional studies like this one are required as thefirst step towards identifying neuroimaging markers that might be useful bio-markers fordementing disorders.Our data indicates that posterior cingulate NAA /Cr levels were decreased in all dementias

studied except in patients with DLB. Normal NAA /Cr levels were previously reported in thisregion in patients with DLB (21). NAA is a neuronal metabolite and NAA levels correlate withneuronal and axonal density (22,23). Decreased NAA levels normalize with the recovery ofcognitive performance after head trauma (24), or termination of seizures after epilepsy surgery

(25), therefore NAA is also regarded as a marker for neuronal function. Patients with DLBhave preserved neuronal numbers at autopsy (26), and have normal limbic cortical volumes onvoxel based morphometry, which distinguishes them from patients with AD (27). The normalNAA /Cr levels observed in our study suggest integrity of neurons in the posterior cingulategyrus of patients with DLB, and are consistent with other studies in this respect.

Patients with VaD had a trend toward lower NAA /Cr levels than normal. All of the patientswith VaD in this study had subcortical vascular disease, and none of them had large hemisphericstrokes. The posterior cingulate voxel contained mostly gray matter with a small amount ofsubcortical white matter, which did not have elevated intensity on T2-weighted MRI. ThusMRI changes related to vascular pathology were not included in the 1H MRS voxel in patientswith VaD. We attribute this trend of depressed NAA /Cr in VaD to retrograde Walleriandegeneration or neuronal dysfunction in the posterior cingulate gyri from vascular diseaselocated elsewhere in the brain. One limitation of this data is that there were only eight patientswith dementia who fulfilled the NINDS-AIREN criteria for VaD during the 15 month periodof patient recruitment. Because of the small sample size, our findings in patients with VaDshould be regarded as preliminary, and should be confirmed in a larger VaD sample.

Neurology. Author manuscript; available in PMC 2009 October 26.

磁共振波谱

Kantarci et al.

NIH-PA Author Manuscript

NIH-PA Author Manuscript

NIH-PA Author ManuscriptPage 5Donepezil HCL, a cholinesterase inhibitor has been reported to increase NAA /Cr levels inpeople with AD (28). This effect, if present in our data, would only decrease the NAA /Crdifferences between AD, FTLD and normal subjects. This effect would have negligible impacton our data concerning differentiating the AD and DLB from each other, because most of thepatients with AD and DLB were taking cholinesterase inhibitors. However it is possible thatthe NAA/Cr differences between AD and FTLD were obscured because the proportion of ADsubjects taking cholinesterase inhibitors was greater than FTLD subjects.Cho /Cr ratios were elevated in patients with degenerative dementias compared to normals,but not in patients with VaD. Normal Cho /Cr levels in the cortical gray matter of patients withVaD is in agreement with previous studies (11-13). The largest amount of choline in the brainis in the choline bound membrane phospholipids. Among the cytosolic choline-containingspecies, free choline and acetylcholine is very low in abundance in brain tissue. The Cho peakin the 1H MRS spectra of the brain is thought to be composed of cytosolicglycerolphosphocholine and phosphocholine, which are the products of membranephosphotidylcholine breakdown, and precursors of choline and acetylcholine synthesis (29).One possible explanation for the elevation of Cho in AD and FTLD is increased membraneturnover due to dying back of the neuropil. It has also been postulated that the elevation of Chopeak is the consequence of membrane phosphotidylcholine catabolism in order to provide freecholine for the chronically deficient acetylcholine production in AD (11,30).The only measurement that was different from normal in patients with DLB was the Cho /Crratio. The activity of choline acetyltransferase, the enzyme responsible for acetylcholinesynthesis from free choline, is reduced earlier and more severely in the disease course of DLBthan AD (31). Furthermore, patients with DLB who were treated with cholinesterase inhibitorshave shown substantial cognitive improvement (32), revealing the functional significance ofacetylcholine deficiency in DLB. Although the basis for the elevation of Cho /Cr in AD andDLB is not completely understood, down regulation of choline acetyltransferase activity may

be responsible for this change in both pathological processes. The fact that Cho /Cr levelsdecrease with cholinergic agonist treatment in AD (33), raises the possibility that Cho /Cr levelscould be a biomarker of therapeutic efficacy in AD and DLB drug trials. This would howeverrequire further validation with longitudinal studies.

MI /Cr levels were elevated in patients with AD and FTLD, and were not different from normalin patients with DLB and VaD. MI is primarily isolated from glial cells (34). MI levels correlatewith glial proliferation in inflammatory CNS demyelination (23). Our findings of elevated mIlevels in AD and FTLD are in agreement with others (4,6,8,15), and support the hypothesisthat mI may be a marker for glial cell activity in these neurodegenerative diseases. Since thereis no evidence that glial proliferation is a characteristic of DLB pathology, normal mI /Cr levelswould be expected in DLB. Our observation of normal posterior cingulate mI /Cr in patientswith VaD is not surprising because the portion of the brain with vascular pathology evident onstructural MRI was not included in the 1H MRS voxel. Our data do not address whether 1HMRS acquired directly from the ischemic brain areas would reveal elevated mI /Cr levels.The metabolite measurements in this study were obtained from a single region in the brain;right and left hemispheric posterior cingulate gyri and inferior percunei. It would seem intuitiveat first that 1H MRS imaging (1H MRSI), in which 1H MRS data is simultaneously acquiredfrom multiple voxels, would be a more appropriate way to comprehensively study differentdementias. We chose to use single voxel 1H MRS for several technical reasons related to dataquality. Achieving uniform water suppression across the entire volume and removal of signalcontributions from subcutaneous lipids is difficult with short echo time 1

1H MRSI (35,http://wendang.chazidian.coming short echo time (30ms) in H MRS acquisitions was critical to this study, because mI

has a relatively short transverse relaxation time and can only be quantified at short echo times.

Neurology. Author manuscript; available in PMC 2009 October 26.