Microvascular morphology alteration using relaxation rate change with gadolinium-based magnetic resonance imaging contrast agent in patients with Alzheimer’s disease

Introduction

Alzheimer’s disease (AD) is the most common progressive and complex age-related neurodegenerative disorder that can lead to cognitive decline (1) and memory loss (2). Concurrent cerebrovascular disease is common in patients with AD (3). This microvascular pathology is an important factor contributing to the worsening of AD through deteriorating microvascular hemodynamics leading to hypoxia, which in turn precipitates amyloid retention (4). Magnetic resonance imaging (MRI) is an important method if it can non-invasively map microvascular pathology in the AD brain.

The microvasculature consists of three types of small vessels: arterioles, capillaries, and venues. It maintains the blood-brain barrier (BBB). Information of microvessel densities and sizes are mostly provided by pathology after biopsy or samples extracted from clinics. Quantitative analysis of microvessel structures is an important method to demonstrate characteristics of brain diseases related to cerebrovascular disorders (5). MRI has been used to obtain information of microvessel structures. For example, previous studies have measured cerebral blood flow and blood volume to evaluate microvascular changes in AD (6,7). However, those values could not directly map microvascular changes, although those are valuable to evaluate alterations of blood flow and volume in AD brains. A more directly method based on measurement of relaxation rates has been introduced to map microvessel structures. Measuring transverse relaxation rate differences between before and after injection of a contrast agent, ∆R2 and ∆R2*, has been applied to map microvessel structures of those with brain tumors (8), brain ischemic stroke (9), AD (10,11), and vascular dementia (12). In addition, researchers have further developed microvascular indices based on ∆R2 and ∆R2*, including vessel size index (VSI) (13), mean vessel diameter (mVD) (14), microvascular density (Q) (15), blood volume fraction (BVf) (16), and microvessel-weighted imaging (MvWI) (17). Previous AD-model animal studies have shown that the microvascular indices of mVD, VSI, and Q have the characteristics of dynamically changes with age and disease progression (18). The other animal studies showed mVD and VSI increased in AD mice and Q has no changes (11).

Except our recent published paper which used gadolinium (Gd)-based contrast agent (12), most of previous studies were performed with using intravascular super-paramagnetic (SPION) and Manganese (Mn) contrast agents. These contrast agents are not immediately removed from the circulation by the reticuloendothelial system and have a longer plasma half-life (T1/2>2.5 h) (19), thus provide higher signal-to-noise (SNR) ratio, but have not recommended for human, yet. The characteristic of Gd-based contrast agent is double excretion of biliary system and renal system, excreting the contrast medium out of the body as soon as possible (20). It reduces the side effects caused by contrast medium residue in the body and is more suitable for clinical research. Relaxation rate differences as well as microvascular indices can provide useful quantitative metrics of in vivo vascular morphology (14), using a clinically approved contrast agent should be beneficial to directly apply it in clinics. To date, there are few studies on the use of Gd-based contrast agent in 3T MRI to research microvascular injuries of AD. Therefore, it is necessary to evaluate the value of microvascular mapping in clinical practice.

Despite its potential in the assessment of microvascular pathology, microvascular imaging has not been addressed in the diagnosis of AD in humanity to the best of our knowledge. Therefore, the objective of this study was to assess cerebral microvascular alterations in patients with AD relative to a cognitively normal (CN) population using in vivo microvascular index maps with Gd-based contrast agent in a 3T MRI system. We present the following article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-22-524/rc).

Methods

Patient population

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). Institutional review board of Kyung Hee University Hospital at Gangdong (2018-01-22–2020-01-21; No. KHNMC 2017-11-028) approved this cross-sectional prospective study. Informed consent was obtained from participants. Participants provided a detailed medical history. They underwent a neurologic examination, standard neuropsychological testing, and MRI scan. Cognitive function was assessed using the Seoul Neuropsychological Screening Battery (SNSB) (21) that covers five cognitive subsets: attention, memory, language, visuospatial function, and frontal/executive function as well as which was included in the Korean version of the Mini-Mental State Examination (MMSE) for global cognitive ability. Based on results of the SNSB examination and MRI findings, patients with probable AD were included in this study for the AD patient group. They had Clinical Dementia Rating (CDR) score of 0.5, 1, or 2 according to the criteria of the National Institute of Neurological and Communicative Disorders and the Stroke-Alzheimer Disease and Related Disorders Association (NINCDS–ADRDA) (21,22). Patients were excluded if they had evidence of intracranial diseases or if they had any other medical disorders severe enough to account for memory and related symptoms according to NINCDS–ADRDA criteria (22). Figure 1 shows the screening process of participants: 35 participants were willing to participate in our study; 34 participants were enrolled in the study and completed the neurologic examination, standard neuropsychological testing, and MRI scan. Finally, we included 22 participants according to results of standard neuropsychological testing and MRI scan. We excluded 12 participants who had any other neurodegenerative disorders or intracranial diseases. A total of 22 participants, including 11 CN (CDR 0–0.5) and 11 AD (CDR 0.5–1) subjects recruited in the neurological center of our institution of Kyung Hee University Hospital at Gangdong were included in this study. There is no missing data in this study. A neurologist with 12 years of experience (HYR) and a psychiatrist with 16 years of experience (ARC) evaluated the SNSB examination. Two neuroradiologists with 10 and 16 years of experience (SP and CWR, respectively) evaluated brain MRI of each participant to detect any evidence of prior cortical infarctions or other space-occupying lesions. Table 1 summarizes demographic characteristics and results of their neuropsychological tests.

Figure 1 Participant recruitment and image screening flow diagram. MRI, magnetic resonance imaging; WM, white matter; AD, Alzheimer’s disease; ICA, intracranial carotid artery; CN, cognitively normal elderly.

MRI acquisition

Figure 2A shows the MRI scan protocol. MRI data were acquired using a 3T scanner (Ingenia, Philips Healthcare, The Netherlands) for all participants. First, to register microvascular index maps to a standard brain template, sagittal structural 3D T1-weighted (3DT1W) images were acquired before and after contrast injection with a turbo fast field echo sequence, similar to the magnetization-prepared rapid acquisition of the gradient echo (MPRAGE) sequence with the following parameters: repetition time (TR) =8.1 ms, echo time (TE) =3.7 ms, flip angle (FA) =8°, field-of-view (FOV) =236×236 mm², and voxel size =1×1×1 mm³. This scan was repeated after admitting a contrast agent to examine any abnormality of the brain. Second, fluid-attenuated inversion recovery (FLAIR) image was acquired to calculate lesion probability map and examine any abnormality of the brain before admitting a contrast agent. Third, 3D time-of-flight (TOF) image was obtained to allow visual assessment of large vessel integrity before admitting a contrast agent. Fourth, a diffusion-weighted imaging (DWI) sequence based on spin-echo echo-planar imaging (EPI) was acquired with b=0 and 1,000 s/mm² to calculate the apparent diffusion coefficient (ADC) before admitting a contrast agent. ADC value was used to calculate VSI map. Finally, to map microvascular indices, a multi-echo (ME) turbo-spin-echo (TSE) sequence was scanned to calculate R2 relaxation rates before and after contrast injection. Imaging parameters for TSE imaging were as follows: TR =11,072 ms, TEs =16 ms +32 ms*n where n is the number of echoes, number of echoes =7, FOV =210×210 mm², acquisition voxel size =1.1×1.1×3 mm³, reconstruction voxel size =0.4×0.4×3 mm³, reconstruction matrix size =512×512, number of slices =45, TSE factor =14, and scan time =7 min 45 s. Furthermore, a multi-echo gradient-echo (GRE) sequence was also run to calculate R2* relaxation rates before and after contrast injection. Imaging parameters for GRE imaging were the same as those for TSE imaging except the following: TR =2,882 ms, TEs =2 ms +10 ms*n, n=7, FOV =210×210 mm², acquisition voxel size =1.1×1.1×3 mm³, reconstruction voxel size =0.4×0.4×3 mm³, reconstruction matrix size =512×512, number of slices =45, FA =60°, and scan time =4 min 43 s. A dose of 0.13 mL Gadobutrol (Gadovist, Bayer Schering Pharma AG, Berlin, Germany) per kg body weight was injected at a speed of 5 mL/s with a time delay of 18 s.

Figure 2 Scan protocol and representative images and maps. (A) Scan protocol with imaging acquisition sequences. R2 and R2* and ADC maps were calculated to map BVf, VSI, mVD, mean vessel density (Q), and MvWI. 3DT1W, 3D T1-weighted imaging; FLAIR, fluid attenuation inversion recovery to obtain white matter legion volume; 3D TOF, 3D Time-of-flight MR angiography to evaluate any large vessel integrity; DWI, DWI to calculate ADC; ME preSE, ME SE imaging before injection of contrast agent to map preR2 (relaxation rate); ME preGE, multi-echo GE imaging before injection of contrast agent to map preR2*; pre, imaging before injection of contrast agent; post, imaging after injection of contrast agent; CE using Gd-based contrast agent (Gadovist, Bayer Schering Pharma AG, Berlin, Germany). (B) Representative images and maps of the microvascular indices obtained from one CN participant (75-year-old female) and one AD patient (78-year-old female). The window level for each representative map was the same for both the CN and AD participant. ADC, apparent diffusion coefficient; BVf, blood volume fraction; VSI, vessel size index; mVD, mean vessel diameter; MvWI, microvessel-weighted imaging; DWI, diffusion-weighted imaging; ME, multi-echo; SE, spin-echo; GE, gradient-echo; CE, contrast enhancement; Gd, Gadolinium; CN, cognitively normal; AD, Alzheimer’s disease.

Mapping for microvascular imaging indices

Multi-echo TSE and GRE images before and after contrast injection were co-registered and resliced using the Statistical Parametric Mapping Version 12 (SPM12) software (http://www.fil.ion.ucl.ac.uk/spm/software/spm12/) to minimize motion during scans. The voxel-based R2* relaxation rates before (preR2*) and after (postR2*) contrast agent were calculated by fitting exponential decay curves from multi-echo GRE images. Then ΔR2* (i.e., postR2*- preR2*) was calculated by the difference in the relaxation rate between before and after contrast injection. We did not use the subtraction method of R2 relaxation rate to calculate ΔR2 because this method provided a poor resulting map. ΔR2 was calculated by using the signal ratio before and after contrast injection obtained from multi-echo TSE images proposed by Tropres et al. (23):

ΔR2=1/TE⋅ln(Spost/Spre)[1]

Where S_pre and S_post were signal intensities obtained with TE =80 ms before and after injection of the contrast agent, respectively.

The following five microvascular indices were calculated based on ΔR2, ΔR2*, and ADC maps. First, the total blood volume fraction (BVf, %) in a voxel was calculated as follows, assuming that a voxel had thousands of microvessels and that it occupied a small blood fraction of the voxel volume (16):

BVf=3/(4π⋅γ⋅Δχ⋅Β0)⋅ΔR2*⋅100(%)[2]

where γ was the gyromagnetic ratio of hydrogen proton (Hz/T), B₀ was the strength of the main magnetic field (i.e., 3T), and ∆χ (in cgs units) was the change in intravascular magnetic susceptibility due to the administration of the paramagnetic contrast agent. BVf was proportional to changes of ΔR2*. Note that Δχ depends on the type and concentration of the contrast agent. We used 0.2 ppm as the Δχ value. Second, the mVD (no unit) value in a voxel was calculated as follows (14):

mVD=ΔR2*ΔR2[3]

mVD provides an estimate of microvascular changes. It is sensitive to the average vessel size in the voxel. It is valid for small coefficients of variation in vessel size distribution. Third, the VSI (expressed in µm) in a voxel was obtained from the following equation with the assumption that the microvasculature had a uniform distribution of randomly oriented straight cylinders as well as a small BVf (<<1) and a large echo time (TE≫1/(2π⋅γ⋅Δχ⋅B0) (13):

VSI=0.424⋅(ADCγ⋅Δχ⋅B0)12⋅(ΔR2*ΔR2)32[4]

VSI is a characteristic of the distribution of the vessel radii. It considers water diffusion rate (ADC: µm²/s) and contrast agent concentration. Fourth, the mean vessel density (Q, expressed in s^-1/3) in a voxel was calculated as follows (15):

Q=ΔR2(ΔR2*)23[5]

Q was introduced to minimize the dependence of the contrast agent concentration. It is sensitive to intrinsic tissue properties. Therefore, Q is dependent on intrinsic tissue properties only. Finally, a MvWI (expressed in s^-2) in a voxel was calculated by using the following equation (17):

MvWI=ΔR2×ΔR2*[6]

MvWI was proposed to sensitively map small blood vessels.

Post-processing of microvascular index maps

We performed the following processing steps using SPM12 software. First, the 3DT1W image was segmented into different brain tissues using a Computational Anatomy Toolbox (CAT12) (http://www.neuro.uni-jena.de/cat/). It was spatially normalized to the dementia template generated by our lab (24). Second, co-registration was performed between 3DT1W image and the first-echo image of multi-echo GRE, all microvascular index maps, and FLAIR image. Third, all microvascular index maps and FLAIR image were spatially normalized into the dementia standard template using deformation field information of the 3DT1W image. Fourth, Gaussian smoothing was performed using the full-width at half maximum (FWHM) of 10×10×10 mm³ for voxel-based statistical analyses of all microvascular index maps.

Statistical analyses

Demographic data and descriptive clinical scores

Demographic data and descriptive clinical scores of all continuous variables such as age and MMSE scores were compared between the CN and AD groups using Mann-Whitney U-test. Sex was tested by the chi-squared test.

Voxel-based analyses

First, a voxel-based two-sample t-test was used to evaluate the difference of each microvascular index map between the two participant groups with age as a covariate. Second, voxel-based regression analysis was performed to evaluate the relationship between each microvascular index map and age without separating the two participant groups. Furthermore, voxel-based regression analysis was performed to evaluate the relationship between each microvascular index map and MMSE without separating the two participant groups with age as a covariate. Both analyses were performed to define atlas-based region-of-interests (ROIs). For both analyses, a significance level of α=0.001 was applied for clusters with at least 50 contiguous voxels without correction for multiple comparisons.

ROI-based analyses

ROI areas were defined based on both results of the voxel-based analysis and AD signature areas from previous reports as the hippocampus, parahippocampal gyrus, precuneus, caudate, corpus callosum, globus pallidus, lateral ventricle, putamen, and thalamus using wfu_pickAtlas software (https://www.nitrc.org/projects/wfu_pickatlas/). ROI values for each microvascular index map were extracted with a Marsbar toolbox (http://marsbar.sourceforge.net/). In addition, threshold-based ROIs were generated as more than 70% probability that a voxel is gray matter (GMV70) and more than 70% probability that a voxel is white matter (WMV70) to represent the dominant gray matter volume and the dominant white matter volume based on the brain tissue volume of 3DT1W image for each participant, respectively. In addition, areas at the white matter hyperintensity (WMHI) on FLAIR image were defined as ROIs by automatic segmentation with the lesion growth algorithm (25) as implemented in the LST toolbox version 3.0.0 (www.statisticalmodelling.de/lst.html) for SPM. ROI values for each microvascular index map were extracted from each ROI.

For ROI values, we performed the following analyses. First, Mann-Whitney U-test was used to explore differences of each microvascular index in each ROI between the two participant groups. Furthermore, to evaluate the influence in each microvascular index by the disease group and age, we performed the multiple regression analysis with each microvascular index as dependent variable and both the participant group and age as independent variables: each microvascular index = group + age. Second, rank correlation analysis was performed to evaluate the relationship between each microvascular index and subjects’ age. Furthermore, partial correlation analysis was performed to evaluate the degree of relationship between each microvascular index and subjects’ MMSE scores after adjusting for age in each ROI. Finally, Wilcoxon signed-rank test was used to compare ROI values obtained from the three threshold-based ROIs of GMV70, WMV70, and WMHI in each participant group separately. For ROI analyses, the significant level was defined as α=0.05. All statistical analyses were performed using the Medcalc statistical program (MedCalc Software, Acacialaan, Ostend, Belgium).

Results

Figure 2B shows representative images and corresponding maps of FLAIR, ADC, and microvascular index maps of BVf, mVD, VSI, Q, and MvWI obtained from one CN participant (75-year-old female, CDR =0) and one AD patient (78-year-old female, CDR =1). We did not see any signals of microvascular indices at the CSF areas. By visual inspection, signals from BVf, Q and MvWI looked higher in AD than in CN. However, mVD and VSI signals looked lower in the AD than in the CN.

Demographic data and results of neuropsychological tests

Table 1 lists statistical analysis results of demographic data and results of neuropsychological tests for CN and AD groups. Age (P=0.013, z=2.495) was significantly different between the two groups. Sex (P=0.637, χ²=0.222) was not significantly different between the two groups. MMSE was significantly different between the two groups (P=0.002, z=3.088), which was expected.

Voxel-based analysis of microvascular indices

Group difference

Figure 3A shows results of voxel-based two-sample t-test of microvascular index maps for between CN and AD groups. In summary, BVf and MvWI were higher while mVD and VSI were lower in the AD group than in the CN group. Q was not significant difference between two groups. Table S1 lists locations of significantly different results of the voxel-based group comparisons for each map. Results of comparisons of GMV and WMV are also listed in Table S1.

Figure 3 Result of voxel-based analyses of each microvascular index map. (A) Comparison between participant groups. The red color indicates greater microvascular index value in the AD than in CN group and the blue color indicates vice versa. (B) Multiple regression with age. The red color indicates the positive association with age, and the blue color indicates the negative association with age. None indicates no association between VSI and age. BVf, blood volume fraction; Q, mean vessel density; mVD, mean vessel diameter; VSI, vessel size index; MvWI, microvessel-weighted imaging; AD, Alzheimer’s disease; CN, cognitive normal.

Multiple regression

Figure 3B shows results of voxel-based multiple regression analysis for microvascular index maps and age. In summary, microvascular indices of BVf, Q, and MvWI were significantly positively associated with age, but mVD was significantly negatively associated with age. VSI did not show any association with age. Table S2 lists locations of the significant relationship with age. All microvascular indices failed to show significant associations with MMSE scores.

ROI-based analysis of microvascular indices

Group comparison

Table 2 lists Mann-Whitney U-test results for each microvascular index between CN and AD groups. Compared with the CN group, BVf was significantly increased at the corpus callosum (P=0.024), thalamus (P=0.005), GMV70 (P=0.020), and WMV70 (P=0.012) in the AD group. MvWI was significantly increased at the hippocampus (P=0.020), parahippocampal gyrus (P=0.006), precuneus (P=0.008), globus pallidus (P=0.001), putamen (P=0.002), thalamus (P=0.020), GMV70 (P=0.003), WMV70 (P=0.005), and WMHI (P=0.017) in the AD group. Table 2 lists results of comparisons of each microvascular index between CN and AD groups. Table S3 lists results of comparisons of ADC, ΔR2, ΔR2*, GMV, and WMV between CN and AD groups. GMV was significantly decreased in the AD group in the hippocampus and parahippocampal gyrus.

Multiple regression analysis

Results of the multiple regression are listed in Table 2. BVf significantly associated with the participant group in thalamus (F=4.860, P=0.029). MvWI significantly associated with the participant group in precuneus (F=4.289, P=0.029), globus pallidus (F=8.316, P=0.003), and WMV70 (F=10.529, P=0.001). Age was not associated with BVf value in any ROIs, but associated with MvWI values in hippocampus (F=9.297, P=0.002), putamen (F=19.964, P<0.0001), thalamus (F=6.460, P=0.007), and WMV70 (F=10.529, P=0.001). Results of the multiple regression analyses of ADC, ΔR2, and ΔR2* are listed in Table S3.

Correlation analysis

Figure 4 illustrates correlation heat map of each microvascular index and age (a), MMSE (b) in each ROIs. BVf was positively correlated with age at GMV70 (r=0.443, P=0.039) and WMV70 (r=0.449, P=0.036). Q was positively correlated with age at lateral ventricle (r=0.452, P=0.035). mVD was negatively correlated with age at WMHI (r=−0.439, P=0.041). MvWI was positively correlated with age at all ROIs (r>0.464, P<0.030), except precuneus. VSI were not significantly correlated with age. Only BVf was significant negatively correlated with MMSE at WMV70 (r=−0.445, P=0.043). Table S4 lists results of correlation analyses between ADC, ΔR2, ΔR2*, GMV, or WMV and age or MMSE score in each ROI.

Figure 4 Correlation heat map in each brain area. (A) Between each microvascular index and age. (B) Between each microvascular index and MMSE. Color coding indicates the strength of correlation between microvascular index and age or MMSE. BVf, blood volume fraction; Q, mean vessel density; mVD, mean vessel diameter; VSI, vessel size index; MvWI, microvessel-weighted imaging, GMV70, more than 70% gray matter volume; WMV70, more than 70% white matter volume; WMHI, white matter hyperintensity; MMSE, mini-mental state examination.

Paired comparisons analysis

Figure 5 illustrates paired comparisons of microvascular indices among GMV70, WMV70, and WMHI ROIs in CN or AD participants. BVf was significantly higher in GMV70 than in WMV70 for both the CN (P=0.001) and AD (P=0.007) group. BVf also was significantly higher in WMV70 than in WMHI (P=0.001) for AD group. mVD was significantly higher in GMV70 than in WMV70 for CN (P=0.005) group, and higher in WMV70 than in WMHI for AD (P=0.042) group. VSI was significantly higher in WMHI than in WMV70 for only the CN group (P=0.010). Both Q and MvWI failed to show significant difference between GMV70 and WMV70 and between WMHI and WMV70 in both CN and AD groups. Table S5 lists results of paired comparisons of ADC, ΔR2, ΔR2* among GMV70, WMV70, and WMHI ROIs in CN or AD participants.

Figure 5 The paired comparisons of microvascular indices among GMV70, WMV70, and WMHI ROIs. (A) CN group. (B) AD group. The dark bar indicates the GMV70, the grey bar indicates the WMV70, and the white bar indicates the WMHI. The bar shows as median (95% CI for the median). *, P<0.05, **, P<0.01. CN, cognitively normal; AD, Alzheimer’s disease; GMV70, brain area of more than 70% of gray matter volume; WMV70, brain area of more than 70% of white matter volume; WMHI, white matter hyperintensity on FLAIR; BVf, blood volume fraction; Q, mean vessel density; mVD, mean vessel diameter; VSI, vessel size index; MvWI, microvessel-weighted imaging; ROI, region of interest.

Discussion

Measurements of microvascular index changes were generally studied in tumor angiogenesis and microvascular remodeling of ischemic stroke. The morphological and structural changes of cerebral microvessels in patients with AD were rarely reported before. Cerebrovascular morphology and structure are crucial for the maintenance of cognitive functions during aging and in AD, because cerebrovascular alterations may alter vascular integrity and blood supply. We investigated alterations of microvascular morphology and structure in AD and CN participants using relaxation rate difference obtained before and after injection of a clinically available Gd-based contrast agent with a clinical 3T MRI system. We found the following results. First, ROI-based comparison showed that BVf and MvWI were increased in AD patients compared with CN participants, but, Q, mVD, and VSI were not increased or decreased in AD patients depending on brain area. Second, voxel-based analysis showed mVD, and VSI were decreased in the prefrontal lobe. Finally, BVf and mVD in the WMHI area was significantly decreased compared to that in the WMV70 area in AD participants.

Microvessel quantity increased in AD patients

BVf is the portion of tissue occupied by blood, characterizing angiogenesis (26,27). BVf was commonly used to monitor angiogenesis in tumor or bone regeneration studies (28,29). In this study, BVf index was used to detect alteration of cerebral microvascular. Increased BVf may indicate more neovascularization in the AD condition. The proliferation of microvessels may be related to the deposition of amyloid protein in the patient’s brain. Aβ is a modulator of vessel density and vascular remodeling through angiogenic mechanisms. Previous studies showed that microvessel amount in the Tg4510 (15-mo-old) was increased compared with that in normal mice using two-photon microscopy (30), and MRI reported a microvascular density increased in the APP23 (14-mo-old) (18). However, vascular density in AD patients was decreased compared with normal elderly based on postmortem analysis (31). Such discrepancies between the previous postmortem study and this study might be explained by variation in disease stage. Microvessel density could change depending on the disease stage. It could show a transient increase or a decrease depending on AD progression. Microvessel density values are generally significantly reduced in the late stage of Aβ pathology (32). Our study reported that the directly related index Q with density was no significant difference in the AD condition. However, the index BVf and MvWI, which can indirectly reflect the microvessel quantity was increased. It is unclear if the increased vessel density in human brain and animal will lead to abnormal or loss of physiologic function. Thus, further studies are required to assess which vascular index is more sensitive to density and the effect of AD progression on microvascular density.

This study showed that MvWI increased in AD patients, especially in the hippocampus, parahippocampal gyrus. MvWI was introduced to improve the sensitivity to map intercortical penetrating arterioles and deep venules within the brain while simultaneously reducing the vessel size overestimation by multiplication of ∆R2 and ∆R2*. It is unclear which one of ∆R2 and/or ∆R2* dominantly contributes to the increase of MvWI value in patients with AD. ∆R2 is widely used to detect cerebral arterioles and venules because it is less susceptible to air tissue interface artifacts (17). ∆R2* is sensitively mapped to relatively large vessels.

Microvessel diameter changes

mVD is a dimensionless parameter, qualitatively measures the mean vessel diameter, and represents the distribution of a vascular size. Unfortunately, this quantitative ROI-based analysis did not find a significantly increase or decrease in the mVD in Alzheimer’s susceptible areas such as the parahippocampal gyrus and hippocampus. However, voxel-based analysis suggested that both mVD and VSI in the prefrontal regions decreased significantly. A review indicates that executive abilities depend on prefrontal areas (33). Therefore, the decrease of prefrontal vessel diameter may be the reason for the slow metabolism and implicated in executive dysfunction in AD. An animal micro-optical sectioning also concluded that the microvessel diameter of AD was decreased (34). In addition, tau overexpressing changes blood vessel morphology, including the abnormal structure and spiraling morphologies (30). Another previously animal study also showed that the vascular diameter of APP23 mice underwent a morphology variation from slight constriction in the early stage to obvious dilation in the late stage of AD (18). In the early phase of AD, the deposition of Aβ peptides can lead to hypoxia and hypoperfusion, resulting in the vasoconstriction by the regulation of vascular contractile endothelin-1 and angiotensin II (35,36). As the disease continues to advance, the serious Aβ pathological deposition contributes to the vascular elasticity loss and degeneration of the vascular smooth muscle cell. It enables the inner surface of the vascular lumen to be rugged, eventually enabling vascular to dilate (34,35). A significant dilation generally exists in the late stage of AD (18). VSI is also usually used to monitor the contraction and dilation of microvessels. It is more quantitative than mVD due to considering the contrast agent concentration water apparent diffusion rate and magnetic field intensity. The variation of VSI values was generally consistent with mVD.

Difference of microvascular index between GMV and WMV or between WMV and WMHI

BVf, and mVD value were higher in dominant gray matter areas than in dominant white matter areas of brains for CN participants, indicating that the blood supply or vascular distribution and vessel structure in the brain was not uniform and the microvascular distribution within the white matter was relatively less. Thus, white matter is more vulnerable to chronic hypoperfusion and hypoxic-ischemic injury than gray matter (37). In the AD group, this difference still exists. The difference of BVf between GMV70 and WMV70 was 0.670 for the CN group and 0.100 for the AD group, indicating the microvascular structure in the white matter area was destructed in AD patients.

WMHI on FLAIR image reflects neuropathological substrates or tissue damage with varying severity. Compared to the normal white matter (WMV70) area, VSI was increased in the WMHI area of CN participants. However, BVf was not changed, indicating that the alteration of microvascular structure was a normal function in the WMHI area and a compensative vasodilation to maintain normal blood supply. However, in the AD group, BVf in the WMHI area was significantly decreased compared to that in the WMV70 area (P=0.001), and mVD also significantly decreased. Therefore, it can be considered that the pathological mechanism WMHI in the brain of patients with AD is completely different from that of normal population. WMHI area in the normal elderly is associated with damage caused by demyelinate, axonal loss, and enlarged perivascular spaces (38), while WMHI in AD is often associated with damage caused by neuronal loss and microvascular injury evolved by microglia activation (39). Therefore, when the BVf in the WMHI area decreased accompany with vasoconstriction, it might be considered that it is caused by AD pathology.

Limitation of this study

Our study has several limitations. First, microvascular alterations were not confirmed with autopsy or pathological diagnosis. Second, the sample size for each group was too small and age was younger in the CN group. During the voxel-based analysis, we always used age as a covariate. Results of the voxel-based two-sample t-test showed that both BVf and MvWI are still significantly different between the two groups, indicating that both indices are demonstrated the microstructure changes in AD patients. In a future study, it should be better to match ages for both groups. Third, in the AD group, we only included mild AD patients (CDR 0.5–1), no severe AD include, that might have further changes of microvascular morphology caused by plaque-related inflammation and anti-inflammatory cytokine interleukin expression (40). Fourth, since this was a cross-sectional study, values of microvascular indices for predicting progression from CN to AD were not assessed. Furthermore, this study performed by 3T clinical MRI, the SNR ratio was lower than the high magnetic field (7T, 9.3T) used in previous animal experiments. Finally, compared with contrast agents used in animal experiments Mn and USPIO, the half-life of contrast agent Gd-based is shorter and can be excreted from the body quickly.

Conclusions

This is the first study to evaluate cerebral microvascular alterations in patients with AD using microvascular indices obtained with the relaxation rate difference before and after injection of a Gd-based contrast agent in 3T clinical MRI. We found that BVf and MvWI were increased in AD patients, mVD and VSI were decreased in the prefrontal lobe in AD patients. In addition, the BVf and mVD in the WMHI area was significantly decreased in AD patients. We can conclude that some microvascular indices might be useful as imaging biomarkers to evaluate the alteration of brain microstructure in AD. However, additional studies with a relatively large sample size are needed to validate our findings.

Acknowledgments

The authors appreciate Miss Seon Hwa Lee (Clinical Research Institute, Kyung Hee University Hospital at Gangdong, Seoul, Korea) for providing advice on the statistical analyses.

Funding: This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (No. 2016R1D1A1B03930720 to GH Jahng) and by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (No. 2020R1A2C100474912 to GH Jahng).

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-22-524/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-22-524/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and approved by the Institutional Review Board of Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea (2018-01-22–2020-01-21; No. KHNMC 2017-11-028). Informed consent was obtained from participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Albertini V, Benussi L, Paterlini A, Glionna M, Prestia A, Bocchio-Chiavetto L, Amicucci G, Galluzzi S, Adorni A, Geroldi C, Binetti G, Frisoni GB, Ghidoni R. Distinct cerebrospinal fluid amyloid-beta peptide signatures in cognitive decline associated with Alzheimer's disease and schizophrenia. Electrophoresis 2012;33:3738-44. [Crossref] [PubMed]
2. Jahn H. Memory loss in Alzheimer's disease. Dialogues Clin Neurosci 2013;15:445-54. [Crossref] [PubMed]
3. Toledo JB, Arnold SE, Raible K, Brettschneider J, Xie SX, Grossman M, Monsell SE, Kukull WA, Trojanowski JQ. Contribution of cerebrovascular disease in autopsy confirmed neurodegenerative disease cases in the National Alzheimer's Coordinating Centre. Brain 2013;136:2697-706. [Crossref] [PubMed]
4. Ewers M, Mielke MM, Hampel H. Blood-based biomarkers of microvascular pathology in Alzheimer's disease. Exp Gerontol 2010;45:75-9. [Crossref] [PubMed]
5. Hunter JM, Kwan J, Malek-Ahmadi M, Maarouf CL, Kokjohn TA, Belden C, Sabbagh MN, Beach TG, Roher AE. Morphological and pathological evolution of the brain microcirculation in aging and Alzheimer's disease. PLoS One 2012;7:e36893. [Crossref] [PubMed]
6. Weidensteiner C, Metzger F, Bruns A, Bohrmann B, Kuennecke B, von Kienlin M. Cortical hypoperfusion in the B6.PS2APP mouse model for Alzheimer's disease: comprehensive phenotyping of vascular and tissular parameters by MRI. Magn Reson Med 2009;62:35-45. [Crossref] [PubMed]
7. Nielsen RB, Parbo P, Ismail R, Dalby R, Tietze A, Brændgaard H, Gottrup H, Brooks DJ, Østergaard L, Eskildsen SF. Impaired perfusion and capillary dysfunction in prodromal Alzheimer's disease. Alzheimers Dement (Amst) 2020;12:e12032. [Crossref] [PubMed]
8. Farrar CT, Kamoun WS, Ley CD, Kim YR, Kwon SJ, Dai G, Rosen BR, di Tomaso E, Jain RK, Sorensen AG. In vivo validation of MRI vessel caliber index measurement methods with intravital optical microscopy in a U87 mouse brain tumor model. Neuro Oncol 2010;12:341-50. [Crossref] [PubMed]
9. Kang M, Jin S, Lee D, Cho H. MRI Visualization of Whole Brain Macro- and Microvascular Remodeling in a Rat Model of Ischemic Stroke: A Pilot Study. Sci Rep 2020;10:4989. [Crossref] [PubMed]
10. Zerbi V, Jansen D, Dederen PJ, Veltien A, Hamans B, Liu Y, Heerschap A, Kiliaan AJ. Microvascular cerebral blood volume changes in aging APP(swe)/PS1(dE9) AD mouse model: a voxel-wise approach. Brain Struct Funct 2013;218:1085-98. [Crossref] [PubMed]
11. Chang SK, Kim J, Lee D, Yoo CH, Jin S, Rhee HY, Ryu CW, Lee JK, Cho H, Jahng GH. Mapping of microvascular architecture in the brain of an Alzheimer's disease mouse model using MRI. NMR Biomed 2021;34:e4481. [Crossref] [PubMed]
12. Choi HI, Ryu CW, Kim S, Rhee HY, Jahng GH. Changes in Microvascular Morphology in Subcortical Vascular Dementia: A Study of Vessel Size Magnetic Resonance Imaging. Front Neurol 2020;11:545450. [Crossref] [PubMed]
13. Troprès I, Grimault S, Vaeth A, Grillon E, Julien C, Payen JF, Lamalle L, Décorps M. Vessel size imaging. Magn Reson Med 2001;45:397-408. [Crossref] [PubMed]
14. Dennie J, Mandeville JB, Boxerman JL, Packard SD, Rosen BR, Weisskoff RM. NMR imaging of changes in vascular morphology due to tumor angiogenesis. Magn Reson Med 1998;40:793-9. [Crossref] [PubMed]
15. Jensen JH, Chandra R. Strong field behavior of the NMR signal from magnetically heterogeneous tissues. Magn Reson Med 2000;43:226-36. [Crossref] [PubMed]
16. Yablonskiy DA, Haacke EM. Theory of NMR signal behavior in magnetically inhomogeneous tissues: the static dephasing regime. Magn Reson Med 1994;32:749-63. [Crossref] [PubMed]
17. Jung HS, Jin SH, Cho JH, Han SH, Lee DK, Cho H. UTE-ΔR2 -ΔR2 * combined MR whole-brain angiogram using dual-contrast superparamagnetic iron oxide nanoparticles. NMR Biomed 2016;29:690-701. [Crossref] [PubMed]
18. Xu X, Meng T, Wen Q, Tao M, Wang P, Zhong K, Shen Y. Dynamic changes in vascular size and density in transgenic mice with Alzheimer's disease. Aging (Albany NY) 2020; Epub ahead of print. [Crossref] [PubMed]
19. Ullrich RT, Jikeli JF, Diedenhofen M, Böhm-Sturm P, Unruh M, Vollmar S, Hoehn M. In-vivo visualization of tumor microvessel density and response to anti-angiogenic treatment by high resolution MRI in mice. PLoS One 2011;6:e19592. [Crossref] [PubMed]
20. Do C, DeAguero J, Brearley A, Trejo X, Howard T, Escobar GP, Wagner B. Gadolinium-Based Contrast Agent Use, Their Safety, and Practice Evolution. Kidney360 2020;1:561-8. [Crossref] [PubMed]
21. Ahn HJ, Chin J, Park A, Lee BH, Suh MK, Seo SW, Na DL. Seoul Neuropsychological Screening Battery-dementia version (SNSB-D): a useful tool for assessing and monitoring cognitive impairments in dementia patients. J Korean Med Sci 2010;25:1071-6. [Crossref] [PubMed]
22. Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D, Gauthier S, Jicha G, Meguro K, O'brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ, Scheltens P. Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria. Lancet Neurol 2007;6:734-46. [Crossref] [PubMed]
23. Troprès I, Pannetier N, Grand S, Lemasson B, Moisan A, Péoc'h M, Rémy C, Barbier EL. Imaging the microvessel caliber and density: Principles and applications of microvascular MRI. Magn Reson Med 2015;73:325-41. [Crossref] [PubMed]
24. Guo XY, Chang Y, Kim Y, Rhee HY, Cho AR, Park S, Ryu CW, San Lee J, Lee KM, Shin W, Park KC, Kim EJ, Jahng GH. Development and evaluation of a T1 standard brain template for Alzheimer disease. Quant Imaging Med Surg 2021;11:2224-44. [Crossref] [PubMed]
25. Schmidt P, Gaser C, Arsic M, Buck D, Förschler A, Berthele A, Hoshi M, Ilg R, Schmid VJ, Zimmer C, Hemmer B, Mühlau M. An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis. Neuroimage 2012;59:3774-83. [Crossref] [PubMed]
26. Seevinck PR, Deddens LH, Dijkhuizen RM. Magnetic resonance imaging of brain angiogenesis after stroke. Angiogenesis 2010;13:101-11. [Crossref] [PubMed]
27. Ungersma SE, Pacheco G, Ho C, Yee SF, Ross J, van Bruggen N, Peale FV Jr, Ross S, Carano RA. Erratum to: Ungersma SE, Pacheco G, Ho C, Yee SF, Ross J, van Bruggen N, Peale FV Jr, Ross S, Carano RA. Vessel imaging with viable tumor analysis for quantification of tumor angiogenesis. Magn Reson Med 2010;63:1637–1647. Magn Reson Med 2011;65:889-99. [Crossref] [PubMed]
28. Righesso LAR, Terekhov M, Götz H, Ackermann M, Emrich T, Schreiber LM, Müller WEG, Jung J, Rojas JP, Al-Nawas B. Dynamic contrast-enhanced magnetic resonance imaging for monitoring neovascularization during bone regeneration-a randomized in vivo study in rabbits. Clin Oral Investig 2021;25:5843-54. [Crossref] [PubMed]
29. Qi XL, Burns P, Hong J, Stainsby J, Wright G. Characterizing blood volume fraction (BVF) in a VX2 tumor. Magn Reson Imaging 2008;26:206-14. [Crossref] [PubMed]
30. Bennett RE, Robbins AB, Hu M, Cao X, Betensky RA, Clark T, Das S, Hyman BT. Tau induces blood vessel abnormalities and angiogenesis-related gene expression in P301L transgenic mice and human Alzheimer's disease. Proc Natl Acad Sci U S A 2018;115:E1289-98. [Crossref] [PubMed]
31. Fischer VW, Siddiqi A, Yusufaly Y. Altered angioarchitecture in selected areas of brains with Alzheimer's disease. Acta Neuropathol 1990;79:672-9. [Crossref] [PubMed]
32. Ielacqua GD, Schlegel F, Füchtemeier M, Xandry J, Rudin M, Klohs J. Magnetic Resonance Q Mapping Reveals a Decrease in Microvessel Density in the arcAβ Mouse Model of Cerebral Amyloidosis. Front Aging Neurosci 2016;7:241. [PubMed]
33. Collette F, Hogge M, Salmon E, Van der Linden M. Exploration of the neural substrates of executive functioning by functional neuroimaging. Neuroscience 2006;139:209-21. [Crossref] [PubMed]
34. Zhang X, Yin X, Zhang J, Li A, Gong H, Luo Q, Zhang H, Gao Z, Jiang H. High-resolution mapping of brain vasculature and its impairment in the hippocampus of Alzheimer's disease mice. Natl Sci Rev 2019;6:1223-38. [Crossref] [PubMed]
35. Weller RO, Boche D, Nicoll JA. Microvasculature changes and cerebral amyloid angiopathy in Alzheimer's disease and their potential impact on therapy. Acta Neuropathol 2009;118:87-102. [Crossref] [PubMed]
36. Merlini M, Wanner D, Nitsch RM. Tau pathology-dependent remodelling of cerebral arteries precedes Alzheimer's disease-related microvascular cerebral amyloid angiopathy. Acta Neuropathol 2016;131:737-52. [Crossref] [PubMed]
37. van Gijn J. Leukoaraiosis and vascular dementia. Neurology 1998;51:S3-8. [Crossref] [PubMed]
38. Haller S, Kövari E, Herrmann FR, Cuvinciuc V, Tomm AM, Zulian GB, Lovblad KO, Giannakopoulos P, Bouras C. Do brain T2/FLAIR white matter hyperintensities correspond to myelin loss in normal aging? A radiologic-neuropathologic correlation study. Acta Neuropathol Commun 2013;1:14. [Crossref] [PubMed]
39. Block ML, Zecca L, Hong JS. Microglia-mediated neurotoxicity: uncovering the molecular mechanisms. Nat Rev Neurosci 2007;8:57-69. [Crossref] [PubMed]
40. Cuello AC. Early and Late CNS Inflammation in Alzheimer's Disease: Two Extremes of a Continuum? Trends Pharmacol Sci 2017;38:956-66. [Crossref] [PubMed]