Differentiation of Alzheimer’s Disease from Frontotemporal Dementia and Mild Cognitive Impairment Based on Arterial Spin Labeling Magnetic Resonance Imaging: A Pilot Cross-Sectional Study from PUMCH Dementia Cohort

Abstract

Background:

Arterial spin labeling (ASL) is helpful in early diagnosis and differential diagnosis of Alzheimer’s disease (AD), with advantages including no exposure to radioactivity, no injection of a contrast agent, more accessible, and relatively less expensive.

Objective:

To establish the perfusion pattern of different dementia in Chinese population and evaluate the effectiveness of ASL in differentiating AD from cognitive unimpaired (CU), mild cognitive impairment (MCI), and frontotemporal dementia (FTD).

Methods:

Four groups of participants were enrolled, including AD, FTD, MCI, and CU based on clinical diagnosis from PUMCH dementia cohort. ASL image was collected using 3D spiral fast spin echo–based pseudo-continuous ASL pulse sequence with background suppression and a high resolution T1-weighted scan covering the whole brain. Data processing was performed using Dr. Brain Platform to get cerebral blood flow (ml/100g/min) in every region of interest cortices.

Results:

Participants included 66 AD, 26 FTD, 21 MCI, and 21 CU. Statistically, widespread hypoperfusion neocortices, most significantly in temporal-parietal-occipital cortices, but not hippocampus and subcortical nucleus were found in AD. Hypoperfusion in parietal lobe was most significantly associated with cognitive decline in AD. Hypoperfusion in parietal lobe was found in MCI and extended to adjacent temporal, occipital and posterior cingulate cortices in AD. Significant reduced perfusion in frontal and temporal cortices, including subcortical nucleus and anterior cingulate cortex were found in FTD. Hypoperfusion regions were relatively symmetrical in AD and left predominant especially in FTD.

Conclusion:

Specific patterns of ASL hypoperfusion were helpful in differentiating AD from CU, MCI, and FTD.

Keywords

Alzheimer’s disease arterial spin labeling frontotemporal dementia hypoperfusion mild cognitive impairment

INTRODUCTION

Alzheimer’s disease (AD) is the most common neurodegenerative dementia causing progressive cognitive deterioration and brain atrophy. The most common presentation of AD is insidious, progressive problems of episodic memory in elderly. As the condition progresses, cognitive impairment become more profound and widespread so as to interfere with activities of daily living [1]. Frontotemporal dementia (FTD) is an insidious neurodegenerative clinical syndrome characterized by progressive deficits in behavior, executive function, and language, which is classified into three clinical variants: behavioral-variant FTD, non-fluent variant primary progressive aphasia, and semantic-variant primary progressive aphasia [2]. There are some atypical clinical syndromes of AD, include posterior cortical atrophy, logopenic aphasia, and the frontal variant. The predominant clinical presentations of language problem and/or behavioral symptoms make it difficult in the differential diagnosis between AD and FTD [1, 3].

Biomarkers promote the precise and early diagnosis of AD. ATN research framework proposed in 2018 is the most significant biomarkers profile. A represents Aβ pathology, including cerebrospinal fluid (CSF) Aβ₄₂, Aβ₄₂/Aβ₄₀, and Aβ-positron emission tomography-computed tomography (PET/CT); T represents tau pathology, including CSF p-tau and tau-PET/CT; N represents neurodegeneration, including CSF t-tau, FDG-PET/CT, and structure magnetic resonance imaging (MRI) [4]. However, due to the invasive property of CSF test and potential radiation risk of PET/CT, the widely use of them in clinical settings is restricted. MRI is the most cost-effective evaluation method in clinical practice. Bilateral atrophy of temporal and/or parietal lobe supports diagnosis of AD, while predominant frontal or temporal atrophy, and atrophy in the frontoinsular region is especially indicative of frontotemporal dementia [5]. Functional MRI including arterial spin labeling (ASL) MRI is proved to be helpful in early and prodromal AD diagnosis [6].

ASL is a non-invasive MRI technique to measure cerebral blood flow (CBF). It uses the water in arterial blood as an endogenous, freely diffusible contrast medium and provides quantitative parametric images of tissue perfusion [7]. CBF is tightly correlated to brain metabolism revealed by FDG-PET, which is proved to be a neurodegenerative marker, and it is considered as an alternative for FDG-PET [8, 9]. ASL MRI has advantages including no exposure to radioactivity or injection of a contrast agent, accompanying with the structural MRI to assess structure-function relationships, more accessible, and relatively less expensive [6].

Declines in regional CBF, including cingulate, precuneus, parietal lobes, and inferior frontal regions in participants with AD were demonstrated in several studies, and were found to be related to global cognition measured with Mini-Mental State Examination (MMSE) [10]. Other studies found decreased precuneus and/or posterior cingulate blood flow in AD. Posterior cingulate and especially parietal effects appeared moderately related to disease severity [11]. However, researchers found hypoperfusion in frontal regions in patients with FTD which was correlated with deficits in judgment and problem solving [5]. Research found progressive decline of CBF across the continuum from cognitively normal older adults to mild AD, suggesting ASL as a promising functional biomarker for early diagnosis and disease tracking in AD [12]. However, increased CBF was found in the left hippocampus, right amygdala, and rostral head of the right caudate nucleus and ventral putamen and globus pallidus in mild cognitive impairment (MCI) participants compared with normal control as reported by Dai et al. [13]. The inconsistent results of hyperperfusion and hypoperfusion might be caused by methodological differences, patient demographics or heterogeneity of capillary blood flow patterns in the preclinical stage of AD, called early compensatory hyperperfusion [14].

Although it is proved that ASL is helpful in early diagnosis and differential diagnosis of AD in many cohorts, the heterogeneity of demographics, different parameters of imaging acquisition, the various influencing factors in clinical settings prevent the widely use of ASL in AD clinical practice and incorporating ASL in clinical guidelines. The aim of this article is to establish the perfusion pattern of different dementia and evaluate the effectiveness of ASL in differentiating AD from CU, MCI, and FTD in Chinese dementia population, which may add to the world-wide validation of ASL clinical use.

MATERIALS AND METHODS

Participants inclusion

Participants in this study were recruited from the dementia clinic database, department of neurology in the Peking Union Medical College Hospital (PUMCH dementia cohort). The study was conducted according to the declaration of Helsinki in 1964 and the subsequent revisions. All participants and/or their caregivers were informed regarding the purpose of the study and written consent was obtained. The ethics committee of Peking Union Medical College Hospital had approved the study (No. JS2998).

66 AD participants were enrolled. The inclusion criteria were the criteria for probable Alzheimer’s disease in 2011 NIA-AA recommendation [15]. 26 FTD participants were enrolled. The inclusion criteria was based on report of the Work Group on Frontotemporal Dementia and Pick’s Disease proposed by McKhann et al. in 2001 [16]. 21 MCI participants were enrolled. The inclusion criteria were based on recommendation of NIA-AA in 2011 [17]. Besides, 21 cognitive unimpaired (CU) control participants were included. Participants with stenosis of carotid artery detected by ultrasound were excluded for influence of brain perfusion. Detailed demographic and clinical information including age, education, symptoms onset and progression were provided by reliable informants. All participants underwent a systemic physical examination, laboratory blood tests (including counts of blood cells, hepatic function, renal function, homocysteine, thyroid function, folic acid, vitamin B₁₂, erythrocyte sedimentation rate, blood ammonia, anti-HIV antibody, and rapid plasma reagin test) and neuroimaging that were necessary for diagnosis and differential diagnosis. Besides, we analyzed the genotype of apolipoprotein E (APOE).

All participants finished detailed cognition battery examination including screening tests: MMSE, Montreal Cognitive Assessment-PUMCH edition (MoCA-PUMCH edition) [18], Activities of Daily Living (ADLs), Clinical Dementia Rating (CDR), and Hospital Anxiety and Depression Scale (HAD). After screening, detail domain specific battery was performed, including story memory, category fluency, Trail Making A and B, digit symbol, digit span, auditory verbal learning test, block design test, clock drawing test, language evaluation, similarities test and modified Rey figure, among others. All these chosen scales had been validated in Mandarin language with available normative data for Chinese population.

MRI data acquisition

Head MRI examinations were performed using a 3-T MRI scanner (Discovery MR750 3 T, GE Healthcare, Milwaukee, WI, USA). Thirty-two channel head coils were used. Axial T1-weighted, T2-weighted, Fluid attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI), and susceptibility weighted imaging (SWI) were performed. ASL image was collected using 3D spiral fast spin echo–based pseudo continuous ASL (pcASL) pulse sequence with background suppression and a high resolution T1-weighted scan covering the whole brain. The following parameters were used: Label/Control pair = 1 pair; TR/TE = 4865/10.5 ms; labeling duration = 1500 ms; post labeling delay time = 2025 ms; field of view (FOV) = 240 mm×240 mm; slice thickness = 4.0 mm; slice gap = 0 mm; in plane matrix = 128×128; number of excitations (NEX) = 2; spiral arms = 8; points per arm = 512; and TA = 3 min 24 s [19].

Image data processing and CBF calculation

Data processing and analysis were performed using Dr. Brain Platform (YiWei Medical Technology Co. Ltd., Shenzhen, China), shown in Fig. 1. ASL used electromagnetic pulses to inverse the spin labeling of hydrogen atoms in the carotid artery inflow, and after a certain period of time, the labeled blood flows into the brain tissue, and the signal was collected to obtain two images, a label and a control. The difference image of the two eliminated the background signal and yielded a perfusion weighted brain image. Firstly, head motion correction could reduce the noise of head movement if necessary. Then calculate the difference image between the label and control image. The signal of the difference image was proportional to CBF. The MRI signal was usually an arbitrary number without units, so the signal from the differential image needed to be standardized with a reference signal. An equilibrium proton density image was used as the reference signal to standardize signals from the differential images. A specific formula for quantifying CBF can be obtained based on the Buxton hemodynamic model [20]. Then quantify the CBF diagram by simply bringing the differential image, the reference image, and some scan parameters such as labeling duration, post-labeling delay, and some constants such as blood T1 and brain tissue T1 into the formula. Compared to images with arbitrary signal values such as T1, T2, etc., the CBF obtained from ASL is a quantitative value that represented how many milliliters of blood per unit of brain tissue were perfused per unit of time. So, the unit of CBF was ml/100 grams of brain tissue per minute. Based on different brain segmentation templates designed by the platform, which are listed in Supplementary Table 1, the module could generate CBF averages of CBF for the corresponding segmented regions (region of interest, ROI), as well as left and right asymmetric index. Based on the templates, CBF analysis was restricted to regions of gray matter. The brain segmentation templates we used included templates that were divided by anatomical structures and blood supply territories.

Fig. 1

Data processing procedure of ASL imaging.

Table 1

Demographic and cognitive status of participants

	AD	FTD	MCI	CU	p	p ^*
No.	66	26	21	21	–	–
Sex (Male: Female)	21:45	15:11	11:10	6:15	0.052	–
Age ( $\bar{x} \pm s$ )	65.8±10.2	61.6±9.1	66.5±7.5	64.2±6.5	0.188	0.068
Education (years, $\bar{x} \pm s$ )	9.7±4.6	9.9±3.9	12.0±3.4	13.4±4.0	0.007	0.898
APOE (ɛ4 carrier: ɛ4 noncarrier)	38:28	6:20	12:9	–	0.009	0.003
MMSE ( $\bar{x} \pm s$ )	15.8±7.5	15.8±7.3	27.7±1.7	28.9±0.8	<0.0001	0.970
MoCA ( $\bar{x} \pm s$ )	13.9±7.2	12.5±6.0	23.5±2.1	28.0±1.5	<0.0001	0.509
ADL ( $\bar{x} \pm s$ )	37.4±11.9	37.8±13.1	23.4±2.4	20.6±1.2	<0.0001	0.885
CDR ( $\bar{x} \pm s$ )	1.6±0.7	1.4±0.7	0.5±0	0±0	<0.0001	0.393

^*p, p value between AD and FTD group.

Statistical analysis

Statistical analyses were performed using SPSS V.25.0 software. Measurement data was described as average±standard deviation ( $\bar{x} \pm s$ ), and numeration data was described as ratio. Measurement data including CBF difference between AD and other group were statistically analyzed with two sample t test, while numeration data was analyzed with χ² test. ROC curve analysis was performed. The level of statistical significance was set to p < 0.05.

RESULTS

General characteristics of the participants

66 AD, 26 FTD, 21 MCI, and 21 CU participants were enrolled. The demographic and cognitive status was summarized in Table 1. The sex and age distribution in the four groups were comparable. Education years were higher in MCI and CU groups, but comparable between AD and FTD groups. APOE ɛ4 carrier was significantly less in FTD group. The cognitive status including scores of MMSE, MoCA, ADL, and CDR was comparable between AD and FTD groups.

Regions of CBF difference between AD and CU

CBF of different ROI cortices (see Supplementary Table 1) in AD and CU groups were compared and statistically analyzed with t test. Lower CBF in AD group was found in majority of neocortices especially temporal-parietal-occipital area (listed in order based on p value in Table 2). Regions of cortices with lower CBF in AD than CU were also presented as anatomical view in Fig. 2 (sectional view, A1–A4) and Fig. 3 (superficial view, A1-A2). Relative symmetrical distribution of low CBF cortices was found in AD. However, CBF of bilateral hippocampus and subcortical nucleus were not significantly lower in AD than CU.

Table 2

Regions of CBF lower in AD than CU (in order based on p value)

Regions	CBF in AD	CBF in CU	p
	(ml/100g/min)	(ml/100g/min)
Right Inferior Parietal Lobule	17.2±7.5	24.8±8.8	0.0002
Right Angular Gyrus	18.6±8.2	26.9±9.3	0.0002
Right Middle Occipital Gyrus	17.4±7.4	24.4±8.3	0.0004
Left Angular Gyrus	19.9±8.5	27.3±9.3	0.001
Right Middle Temporal Gyrus	20.2±8.1	27.1±9.4	0.001
Left Inferior Parietal Lobule	18.4±7.2	24.8±8.1	0.001
Left Posterior Cingulate Gyrus	22.8±9.2	30.5±10.4	0.002
Right Supramarginal Gyrus	20.9±8.3	29.8±9.2	0.002
Left Superior Parietal Lobule	15.2±6.3	20.1±6.5	0.003
Left Middle Occipital Gyrus	18.5±7.9	24.6±8.4	0.003
Right Inferior Temporal Gyrus	17.8±6.5	22.8±7.0	0.004
Right superior Occipital Gyrus	17.3±7.6	22.9±7.9	0.004
Right Inferior Occipital Gyrus	16.5±6.0	21.2±6.8	0.004
Left Precuneus	20.5±8.1	26.5±8.4	0.005
Right Superior Parietal Lobule	14.2±5.5	18.0±5.7	0.007
Left Supramarginal Gyrus	21.8±8.0	27.6±9.5	0.007
Left Superior Occipital Gyrus	16.7±7.5	21.9±7.4	0.007
Left Inferior Occipital Gyrus	17.4±6.2	21.9±7.5	0.007
Left Inferior Temporal Gyrus	17.3±6.4	21.7±6.5	0.008
Right Posterior Cingulate Gyrus	23.0±9.2	29.2±10.0	0.010
Left Cuneus	19.3±7.7	24.5±8.3	0.010
Left Middle Temporal Gyrus	20.1±8.3	25.7±8.2	0.012
Right Superior Temporal Gyrus	22.8±9.2	28.8±9.9	0.013
Left Dorsolateral Superior Frontal Gyrus	19.3±7.1	23.6±7.2	0.016
Right Dorsolateral Superior Frontal Gyrus	19.3±7.1	23.6±7.2	0.016
Right Middle Frontal Gyrus	20.4±7.8	25.2±8.0	0.018
Left Middle Frontal Gyrus	20.5±7.9	25.1±7.5	0.023
Right Precuneus	22.0±8.7	27.2±9.3	0.024
Left Superior Temporal Gyrus	22.2±8.3	27.0±8.8	0.025
Right Middle Cingulate Gyrus	24.5±8.9	29.4±9.4	0.032
Left Occipital Fusiform Gyrus	19.8±6.6	23.4±7.2	0.035
Left Middle Cingulate Gyrus	24.0±8.5	28.6±9.1	0.035
Left Entorhinal Area	23.6±8.0	28.2±10.4	0.037
Right Postcentral Gyrus	18.3±6.5	21.6±6.5	0.042
Left Parahippocampus Gyrus	19.7±5.8	22.9±6.9	0.043
Left Insula	23.0±8.1	27.3±9.2	0.046
Left Temporal Pole: Superior Temporal Gyrus	15.4±5.0	18.0±6.0	0.046
Left Anterior Cingulate Gyrus	26.7±9.6	31.8±11.5	0.046
Left Opercular part of Inferior Frontal Gyrus	20.8±7.4	24.8±9.0	0.048
Left Temporal Pole: Middle Temporal Gyrus	14.0±5.4	16.8±6.1	0.049

Fig. 3

A1, A2) Regions of cortices had lower CBF in AD than CU in blue. (A1: left view; A2: right view.) B1, B2) Regions of cortices had lower CBF in FTD than AD in blue. (B1: left view; B2: right view.).

On the other hand, we performed ROC curve analysis of ten ROI in differentiating AD from CU. Ten ROI cortices all located in temporal-parietal-occipital area, and ROC curves were shown in Fig. 4 (left). ROC curve of combination of the ten ROI increased the sensitivity of differentiating AD from CU (AUC = 0.823, p = 0.000) (Fig. 4, right). The combined sensitivity and specificity was 0.810 and 0.742, respectively.

Fig. 4

Left: ROC curve of ten ROI cortices: right inferior parietal lobule (AUC = 0.757, p = 0.000), right angular gyrus (AUC = 0.755, p = 0.000), right middle occipital gyrus (AUC = 0.741, p = 0.001), left angular gyrus (AUC = 0.715, p = 0.003), right middle temporal gyrus (AUC = 0.724, p = 0.002), left inferior parietal lobule (AUC = 0.726, p = 0.002), left posterior cingulate gyrus (AUC = 0.720, p = 0.002), right supramarginal gyrus (AUC = 0.725, p = 0.002), left superior parietal lobule (AUC = 0.713, p = 0.003) and left middle occipital gyrus (AUC = 0.700, p = 0.006). Right: the combined ROC curve in differentiating AD from CU (AUC = 0.823, p = 0.000).

Regions of CBF difference between AD and FTD

CBF of different ROI cortices in AD and FTD groups were compared and statistically analyzed with t test. Lower CBF in FTD than AD was found in the following regions (listed in order based on p value in Table 3). Regions of cortices with lower CBF in FTD than AD were also presented as anatomical view in Fig. 2 (sectional view, B1–B6) and Fig. 3 (superficial view, B1-B2). Obvious asymmetrical distribution of low CBF cortices was found in FTD, which was left predominant.

Table 3

Regions of CBF lower in FTD than AD (in order based on p value)

Regions	CBF in AD (ml/100g/min)	CBF in FTD (ml/100g/min)	p
Left Entorhinal Area	23.6±8.0	18.9±5.1	0.007
Left Temporal Pole: Superior Temporal Gyrus	15.4±5.0	12.4±4.0	0.008
Left Parahippocampus Gyrus	19.7±5.8	16.5±3.7	0.010
Left Anterior Cingulate Gyrus	26.7±9.6	21.5±5.8	0.011
Left Amygdala	19.0±5.8	16.0±4.6	0.020
Right Anterior Cingulate Gyrus	24.5±9.2	19.9±5.5	0.021
Left Medial Superior Frontal Gyrus	22.5±8.0	18.4±5.8	0.021
Left Insula	23.0±8.1	19.0±5.1	0.021
Left Opercular part of Inferior Frontal Gyrus	20.8±7.4	17.1±5.7	0.022
Left Hippocampus	21.1±6.4	17.9±4.0	0.022
Left Dorsolateral Superior Frontal Gyrus	19.3±7.1	15.8±4.8	0.022
Left Gyrus Rectus	24.3±8.3	20.1±6.4	0.022
Left Triangular part of Inferior Frontal Gyrus	20.3±7.2	16.6±6.3	0.024
Right Entorhinal Area	22.7±8.3	18.7±5.1	0.024
Right Parahippocampus Gyrus	20.7±6.9	17.4±4.4	0.026
Left Caudate	17.6±6.3	14.5±4.7	0.027
Left Medial Orbital part of Superior Frontal Gyrus	26.4±9.4	21.8±7.0	0.028
Left Middle Frontal Gyrus	20.5±7.9	16.8±5.5	0.028
Right Gyrus Rectus	23.5±8.4	19.6±6.3	0.035
Right Medial Orbital part of Superior Frontal Gyrus	25.7±9.6	21.4±6.3	0.036
Right Medial Superior Frontal Gyrus	21.2±7.7	17.7±5.5	0.038
Right Hippocampus	21.0±7.2	17.8±4.5	0.038
Left Temporal Pole: Middle Temporal Gyrus	14.0±5.4	11.6±3.7	0.039
Right Triangular part of Inferior Frontal Gyrus	22.0±8.7	18.1±6.7	0.044
Left Orbital part of Superior Frontal Gyrus	20.9±8.1	17.3±6.3	0.046
Left Middle Cingulate Gyrus	24.0±8.5	20.3±5.5	0.046

Fig. 2

A1–A4) Regions of cortices had lower CBF in AD than CU in blue. (A1: lateral ventricle plane; A2: basal ganglia plane; A3: midsagittal; A4: coronal plane with hippocampus). B1–B6) Regions of cortices had lower CBF in FTD than AD in blue. (B1: basal ganglia plane; B2: lateral ventricle plane; B3: midsagittal; B4: left lateral sagittal plane; B5: coronal plane with hippocampus; B6: coronal plane with anterior frontal lobe.).

Also, we performed ROC curve analysis of ten ROI in differentiating AD from FTD. Ten ROI cortices all located in frontal-temporal area and ROC curves were shown in Fig. 5 (left). ROC curve of combination of the ten ROI increased the sensitivity of differentiating AD from FTD (AUC = 0.707, p = 0.002) (Fig. 5, right). The combined sensitivity and specificity were 0.808 and 0.545, respectively.

Fig. 5

Left: ROC curve of ten ROI cortices: left entorhinal area (AUC = 0.686, p = 0.006), left temporal pole: superior temporal gyrus (AUC = 0.672, p = 0.010), left parahippocampus gyrus (AUC = 0.661, p = 0.017), left anterior cingulate gyrus (AUC = 0.647, p = 0.029), left amygdala (AUC = 0.651, p = 0.025), right anterior cingulate gyrus (AUC = 0.641, p = 0.036), left medial superior frontal gyrus (AUC = 0.656, p = 0.020), left insula (AUC = 0.647, p = 0.029), left opercular part of inferior frontal gyrus (AUC = 0.642, p = 0.034) and left hippocampus (AUC = 0.644, p = 0.032). Right: the combined ROC curve in differentiating AD from FTD (AUC = 0.707, p = 0.002).

Regions of CBF difference between MCI and AD/CU

CBF of different ROI brain regions in MCI and AD groups were compared and statistically analyzed with t test. Lower CBF in AD than MCI was found in the following regions: left angular gyrus (19.9±8.5 versus 24.4±7.0 ml/100 g/min, p = 0.031), left posterior cingulate gyrus (22.8±9.2 versus 27.4±8.3 ml/100 g/min, p = 0.043), left inferior occipital gyrus (17.4±6.2 versus 20.5±5.7 ml/100 g/min, p = 0.044), right angular gyrus (18.6±8.2 versus 22.9±6.1 ml/100 g/min, p = 0.032), right inferior occipital gyrus (16.5±6.0 versus 19.6±5.9 ml/100 g/min, p = 0.044), right Temporal Pole: Middle Temporal Gyrus (15.7±6.7 versus 19.4±7.9 ml/100 g/min, p = 0.037) and right middle occipital gyrus (17.4±7.4 versus 21.3±6.4 ml/100 g/min, p = 0.035).

CBF of different ROI brain regions in MCI and CU groups were compared and statistically analyzed with t test. Lower CBF in MCI than CU was found in the following regions: left superior parietal lobule (16.5±4.5 versus 20.1±6.5 ml/100 g/min, p = 0.048) and right inferior parietal lobule (19.9±5.4 versus 24.8±8.8 ml/100 g/min, p = 0.034).

Regions associated with cognitive decline in AD

Linear regression was used to evaluate regions with CBF decline associated with cognitive impairment in AD. General cognition was represented with MMSE scores. Ten regions with most decreased CBF along with MMSE score decline were as follows: left angular gyrus (R² = 0.2491, p < 0.0001), right posterior cingulate gyrus (R² = 0.1840, p = 0.0003), left posterior cingulate gyrus (R² = 0.1781, p = 0.0004), left middle temporal gyrus (R² = 0.2159, p < 0.0001), right precuneus (R² = 0.1706, p = 0.0006), right angular gyrus (R² = 0.1931, p = 0.0002), right middle temporal gyrus (R² = 0.1836, p = 0.0003), left supramarginal gyrus (R² = 0.1879, p = 0.0003), left middle occipital gyrus (R² = 0.1869, p = 0.0003), and left precuneus (R² = 0.1617, p = 0.0008).

DISCUSSION

The present study established the perfusion pattern of AD dementia and evaluated the effectiveness of ASL in differentiating AD from CU and FTD in Chinese dementia population. Compared with CU, the hypoperfusion region of AD was found to be widespread in frontal, temporal, parietal, and occipital neocortices, most significantly in temporal-parietal-occipital cortices. However, CBF in bilateral hippocampus and subcortical nucleus were not significantly decreased. Also, the most significant perfusion decrease associated with cognitive decline in AD was located in the parietal lobe, adjacent temporal, and occipital lobe. These findings further proved posterior brain changes in early onset AD. Hypoperfusion in parietal lobe was found in MCI and extended to adjacent temporal, occipital, and posterior cingulate cortices in AD. However, compared with AD, significant reduced perfusion in frontal and temporal lobe, including subcortical nucleus and anterior cingulate cortex suggested FTD. We also found hypoperfusion regions were relative symmetrical in AD and mostly left predominant especially in FTD.

Specific regional CBF decrease mode was related to the disease and severity. In our results, we found regions with most significant CBF decrease along with cognitive decline were located in posterior brain, such as posterior parietal and posterior cingulate gyrus, CBF measures the amount of blood supply to the brain. The decrease of CBF reflects neural dysfunction or reduced neural activity because of neurovascular coupling. In comparison to structural properties, CBF may be more sensitive in predicting cognitive changes. De Vis et al. revealed whole brain CBF could predict general cognitive ability and specific cognitive domains at follow up in elderly normal individuals [21]. Alsop et al demonstrated significant CBF decreases in AD relative to control subjects in temporal, parietal, frontal, and posterior cingulate cortices. Increased severity of disease correlated with posterior parietal and posterior cingulate decreases [22]. Kaneta et al. also found significant positive correlations in CBF with cognition were in the posterior cingulate cortices and temporo-parietal association cortices [23]. Other research also proved ASL-CBF had potential as a measure for disease progression in AD and lower CBF was associated with more rapid cognitive decline in AD [24, 25]. Posterior cortical CBF reduction was detectable in PCA showed that ASL could be a useful diagnostic adjunct in atypical AD [26]. Besides, research using amyloid-β load measured by florbetapir PET found higher amyloid-β load was related to lower CBF in several regions, suggesting CBF decrease in AD was related to pathogenesis [27].

Hypoperfusion regions of AD revealed by ASL in our study included frontal, temporal, parietal, and occipital neocortices, most significantly in temporal-parietal-occipital cortices, but not hippocampus and subcortical nucleus. It was reported that the areas of most severe hypoperfusion detected with ASL were located in the parietal lobes, the precuneus, the right posterior cingulate cortex, and the frontal lobes in early onset AD, which was similar to the distribution of hypometabolism detected by FDG-PET [28]. This was also in concordance with cortical atrophy distribution in early onset AD [29]. Verclytte et al. revealed more severe hypoperfusion in non-amnestic early onset AD patients in the bilateral temporo-parietal neocortex, the precuneus, the posterior cingulate cortices, and frontal lobes [30]. More hypoperfusion was found in the left parietal lobe, posterior cingulate cortices, and precuneus in early onset AD than late onset AD [30]. The hypoperfusion area distribution in this study was similar to literature, as participants enrolled were mostly early onset AD.

ASL was also shown to detect early hypoperfusion in prodromal AD or MCI. Dolui et al. found patterns of ASL-hypoperfusion in MCI particularly in midline parietal regions versus cognitively normal adults [31]. Johnson et al also revealed significant regional hypoperfusion in the inferior right parietal lobe in MCI and extended into the bilateral posterior cingulate gyri, frontal gyri, and left inferior parietal lobe in AD [32]. However, there were results in controversy. In the research by Riederer et al., FDG-PET demonstrated hypometabolism in the precuneus in MCI, but no hypoperfusion was detected by ASL, probably due to the limited signal-to-noise ratio of pulsed ASL [33]. Also, Ding et al. reported decreased CBF in the left occipital lobe, bilateral inferior temporal cortex, and right middle temporal cortex, but significantly increased CBF in bilateral frontal lobes and right temporal subgyral regions in amnestic MCI [34]. Therefore, inconsistency was found in the CBF distribution pattern of MCI in different studies, affected by the methodology of imaging, heterogeneity of MCI, and other factors needing further clarification. However, longitudinal evolution of CBF measures with ASL was found in the AD continuum, and baseline right inferior parietal and right middle frontal perfusion were associated with conversion to dementia [35, 36]. These studies suggested that hypoperfusion in MCI could predict subsequent cognitive decline and dementia progression in the future. In our results, we also found hypoperfusion in parietal lobe in MCI and extended to adjacent temporal, occipital, and posterior cingulate cortices in AD, similar to FDG-PET evaluation results in literature.

We also found cerebral perfusion pattern was helpful in differentiation between AD and FTD. As the most common causes of degenerative dementia in especially early onset patients, the accurate diagnosis of early-stage AD and FTD based on clinical presentation and structure MRI was difficult. It was reported that significant inferior, medial, and dorsolateral frontal hypoperfusion in FTD, and significant hypoperfusion in bilateral lateral temporal-parietal and medial parietal/precuneus regions in AD, suggesting ASL was a potentially useful biomarker for distinguishing the two diseases [37]. Other research also reported regional hypoperfusion in anterior cingulate cortex and right frontal lobe in FTD, which helped with distinguishing from AD [5, 38]. However, although concurrent atrophy and hypoperfusion were found in the right prefrontal cortex and bilaterally medial frontal lobe, there were regions discordance between brain perfusion and atrophy in FTD, suggesting the perfusion change in FTD was determined by more complex mechanism than AD [39]. Thus, combining structure MRI and advanced techniques including ASL could improve the accurate diagnosis of dementia [40]. In our results, significant reduced perfusion was found in frontal and temporal lobe, including subcortical nucleus and anterior cingulate cortex, predominant left sided. Regional perfusion in FTD was lower than AD, probably due to the difference of global and regional cortical involvement between AD and FTD. The general cognitive function was comparable between AD and FTD.

However, there were limitations in our study and future research is needed. Firstly, the diagnosis of AD and FTD was clinical based, without pathological and biomarker validation, which might cause misdiagnosis, especially when multiple pathologies existed; Secondly, we did not include atypical AD and divide FTD into the three clinical subtypes due to the sample size, which might have distinct atrophy and perfusion patterns on MRI. Further studies of larger sample size and more subgroups were needed. Thirdly, combination of multimodal imaging including structure MRI, FDG-PET was needed in further studies to achieve more accurate and sensitive tools for dementia diagnosis. Also, relationship analysis between ASL and FDG-PET would help clarifying the usefulness of ASL as an alternative for FDG-PET. Fourthly, different cortical regions were associated with different cognitive domains. Further incorporating detailed neuropsychological analysis and atypical AD participants would help understanding the relationship between perfusion and clinical cognitive impairment.

Footnotes

ACKNOWLEDGMENTS

The authors sincerely thank the participants for their joining the research.

FUNDING

This work was supported by National Key Research and Development Program of China [No. 2020YFA0804500, 2020YFA0804501]; CAMS Innovation fund for medical sciences (CIFMS) [No. 2020-I2M-C&T-B-010, 2021-I2M-1-020]; National High Level Hospital Clinical Research Funding [2022-PUMCH-A-254]; STI 2030-Major Project (2021ZD0201106); National Natural Science Foundation of China [No.81550021, 30470618].

CONFLICT OF INTEREST

Dr. Jing Gao is an Editorial Board Member of this journal but was not involved in the peer-review process nor had access to any information regarding its peer-review.

DATA AVAILABILITY

The data supporting the findings of this study are available within the article and/or its supplementary material.

The supplementary material is available in the electronic version of this article: .

References

Lane

, Hardy

, Schott

(2018) Alzheimer’s disease, Eur J Neurol 25, 59–70.

Bang

, Spina

, Miller

(2015) Frontotemporal dementia, Lancet 386, 1672–1682.

Graff-Radford

, Yong

KXX

, Apostolova

, Bouwman

, Carrillo

, Dickerson

, Rabinovici

, Schott

, Jones

, Murray

(2021) New insights into atypical Alzheimer’s disease in the era of biomarkers, Lancet Neurol 20, 222–234.

Jack

Jr. , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

(2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease, Alzheimers Dement 14, 535–562.

, Jahng

, Hayasaka

, Kramer

, Rosen

, Gorno-Tempini

, Rankin

, Miller

, Weiner

, Schuff

(2006) Hypoperfusion in frontotemporal dementia and Alzheimer disease by arterial spin labeling MRI, Neurology 67, 1215–1220.

Wolk

, Detre

(2012) Arterial spin labeling MRI: An emerging biomarker for Alzheimer’s disease and other neurodegenerative conditions, Curr Opin Neurol 25, 421–428.

Grade

, Hernandez Tamames

, Pizzini

, Achten

, Golay

, Smits

(2015) A neuroradiologist’s guide to arterial spin labeling MRI in clinical practice, Neuroradiology 57, 1181–1202.

Weyts

, Vernooij

, Steketee

, Valkema

, Smits

(2017) Qualitative agreement and diagnostic performance of arterial spin labelling MRI and FDG PET-CT in suspected early-stage dementia: Comparison of arterial spin labelling MRI and FDG PET-CT in suspected dementia, Clin Imaging 45, 1–7.

Haller

, Zaharchuk

, Thomas

, Lovblad

, Barkhof

, Golay

(2016) Arterial spin labeling perfusion of the 569 brain: Emerging clinical applications. Radiology 281, 337–356.

10.

Zhang

, Gordon

, Goldberg

(2017) Cerebral blood flow measured by arterial spin labeling MRI at resting state in normal aging and Alzheimer’s disease, Neurosci Biobehav Rev 72, 168–175.

11.

Alsop

, Dai

, Grossman

, Detre

(2010) Arterial spin labeling blood flow MRI: Its role in the early characterization of Alzheimer’s disease, J Alzheimers Dis 20, 871–880.

12.

Wang

, Das

, Xie

, Arnold

, Detre

, Wolk

(2013) Arterial spin labeled MRI in prodromal Alzheimer’s disease: A multi-site study, Neuroimage Clin 2, 630–636.

13.

Dai

, Lopez

, Carmichael

, Becker

, Kuller

, Gach

(2009) Mild cognitive impairment and alzheimer disease: Patterns of altered cerebral blood flow at MR imaging, Radiology 250, 856–866.

14.

Wierenga

, Hays

, Zlatar

(2014) Cerebral blood flow measuredby arterial spin labeling MRI as a preclinical marker of Alzheimer’sdisease, J Alzheimers Dis 42(Suppl 4), S411–S419.

15.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr. , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease, Alzheimers Dement 7, 263–269.

16.

McKhann

, Albert

, Grossman

, Miller

, Dickson

, Trojanowski

(2001) Clinical and pathological diagnosis of frontotemporal dementia: Report of the Work Group on Frontotemporal Dementia and Pick’s Disease, Arch Neurol 58, 1803–1809.

17.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease, Alzheimers Dement 7, 270–279.

18.

Tan

, Li

, Gao

, Wang

, Zhao

, Yu

, Du

, Zhang

, Cui

, Wang

, Li

, Yang

, Yu

, Xia

, Zhou

(2015) Optimal cutoff scores for dementia and mild cognitive impairment of the Montreal Cognitive Assessment among elderly and oldest-old Chinese population, J Alzheimers Dis 43, 1403–1412.

19.

Zhang

, Dou

, Zuo

, You

, Lv

, Hou

, Shi

, Feng

(2021) Brain volume and perfusion asymmetry in temporal lobe epilepsy with and without hippocampal sclerosis, Neurol Res 43, 299–306.

20.

Buxton

, Uludağ

, Dubowitz

, Liu

(2004) Modeling the hemodynamic response to brain activation, Neuroimage 23 Suppl 1, S220–233.

21.

De Vis

, Peng

, Chen

, Li

, Liu

, Sur

, Rodrigue

, Park

, Lu

(2018) Arterial-spin-labeling (ASL) perfusion MRI predicts cognitive function in elderly individuals: A 4-year longitudinal study, J Magn Reson Imaging 48, 449–458.

22.

Alsop

, Detre

, Grossman

(2000) Assessment of cerebral blood flow in Alzheimer’s disease by spin-labeled magnetic resonance imaging, Ann Neurol 47, 93–100.

23.

Kaneta

, Katsuse

, Hirano

, Ogawa

, Shihikura-Hino

, Yoshida

, Odawara

, Hirayasu

, Inoue

(2017) Voxel-wise correlations between cognition and cerebral blood flow using arterial spin-labeled perfusion MRI in patients with Alzheimer’s disease: A cross-sectional study, BMC Neurol 17, 91.

24.

Binnewijzend

, Benedictus

, Kuijer

, van der Flier

, Teunissen

, Prins

, Wattjes

, van Berckel

, Scheltens

, Barkhof

(2016) Cerebral perfusion in the predementia stages of Alzheimer’s disease, Eur Radiol 26, 506–514.

25.

Benedictus

, Leeuwis

, Binnewijzend

, Kuijer

, Scheltens

, Barkhof

, van der Flier

, Prins

(2017) Lower cerebral blood flow is associated with faster cognitive decline in Alzheimer’s disease, Eur Radiol 27, 1169–1175.

26.

Lehmann

, Melbourne

, Dickson

, Ahmed

, Modat

, Cardoso

, Thomas

, De Vita

, Crutch

, Warren

, Mahoney

, Bomanji

, Hutton

, Fox

, Golay

, Ourselin

, Schott

(2016) A novel use of arterial spin labelling MRI to demonstrate focal hypoperfusion in individuals with posterior cortical atrophy: A multimodal imaging study, J Neurol Neurosurg Psychiatry 87, 1032–1034.

27.

Mattsson

, Tosun

, Insel

, Simonson

, Jack

Jr. , Beckett

, Donohue

, Jagust

, Schuff

, Weiner

(2014) Association of brain amyloid-β with cerebral perfusion and structure in Alzheimer’s disease and mild cognitive impairment, Brain 137, 1550–1561.

28.

Verclytte

, Lopes

, Lenfant

, Rollin

, Semah

, Leclerc

, Pasquier

, Delmaire

(2016) Cerebral hypoperfusion and hypometabolism detected by arterial spin labeling MRI and FDG-PET in early-onset Alzheimer’s disease, J Neuroimaging 26, 207–212.

29.

Mao

, Li

, Huang

, Dong

, Liu

, Peng

, Cui

, Gao

(2021) White matter hyperintensities and patterns of atrophy in early onset Alzheimer’s disease with causative gene mutations, Clin Neurol Neurosurg 203, 106552.

30.

Verclytte

, Lopes

, Viard

, Rollin

, Vanhoutte

, Pasquier

, Pruvo

, Leclerc

(2020) Differences in cortical perfusion detected by arterial spin labeling in nonamnestic and amnestic subtypes of early-onset Alzheimer’s disease, J Neuroradiol 47, 284–291.

31.

Dolui

, Li

, Nasrallah

, Detre

, Wolk

(2020) Arterial spin labeling versus (18)F-FDG-PET to identify mild cognitive impairment, Neuroimage Clin 25, 102146.

32.

Johnson

, Jahng

, Weiner

, Miller

, Chui

, Jagust

, Gorno-Tempini

, Schuff

(2005) Pattern of cerebral hypoperfusion in Alzheimer disease and mild cognitive impairment measured with arterial spin-labeling MR imaging: Initial experience, Radiology 234, 851–859.

33.

Riederer

, Bohn

, Preibisch

, Wiedemann

, Zimmer

, Alexopoulos

, Förster

(2018) Alzheimer disease and mild cognitive impairment: Integrated pulsed arterial spin-labeling MRI and (18)F-FDG PET, Radiology 288, 198–206.

34.

Ding

, Ling

, Zhang

, Huang

, Zhang

, Wang

, Yan

(2014) Pattern of cerebral hyperperfusion in Alzheimer’s disease and amnestic mild cognitive impairment using voxel-based analysis of 3D arterial spin-labeling imaging: Initial experience, Clin Interv Aging 9, 493–500.

35.

Camargo

, Wang

(2021) Longitudinal cerebral blood flow changes in normal aging and the Alzheimer’s disease continuum identified by arterial spin labeling MRI, J Alzheimers Dis 81, 1727–1735.

36.

Chao

, Buckley

, Kornak

, Schuff

, Madison

, Yaffe

, Miller

, Kramer

, Weiner

(2010) ASL perfusion MRI predicts cognitive decline and conversion from MCI to dementia, Alzheimer Dis Assoc Disord 24, 19–27.

37.

, Wang

, Lee

, Trojanowski

, Detre

, Grossman

(2010) Distinct cerebral perfusion patterns in FTLD and AD, Neurology 75, 881–888.

38.

Steketee

, Bron

, Meijboom

, Houston

, Klein

, Mutsaerts

, Mendez Orellana

, de Jong

, van Swieten

, van der Lugt

, Smits

(2016) Early-stage differentiation between presenile Alzheimer’s disease and frontotemporal dementia using arterial spin labeling MRI, Eur Radiol 26, 244–253.

39.

Shimizu

, Zhang

, Laxamana

, Miller

, Kramer

, Weiner

, Schuff

(2010) Concordance and discordance between brain perfusion and atrophy in frontotemporal dementia, Brain Imaging Behav 4, 46–54.

40.

Bron

, Smits

, Papma

, Steketee

RME

, Meijboom

, de Groot

, van Swieten

, Niessen

, Klein

(2017) Multiparametric computer-aided differential diagnosis of Alzheimer’s disease and frontotemporal dementia using structural and advanced MRI, Eur Radiol 27, 3372–3382.