Prediction of Cognitive Decline from White Matter Hyperintensity and Single-Photon Emission Computed Tomography in Alzheimer’s Disease

Background While several studies support an association of white matter hyperintensity (WMH) volume and regional cerebral blood flow (rCBF) with cognitive decline in Alzheimer’s disease (AD), no reports have simultaneously considered the effects of both factors on cognitive decline. Objective The purpose of the present study was to compare WMH volume and rCBF in relation to cognitive function by developing a new software program to fuse magnetic resonance imaging (MRI) and single-photon emission computed tomography (SPECT) data. Method We used MRI, SPECT, and neuropsychological data from 182 serial outpatients treated at the memory clinic of our hospital. Results Twenty-nine AD patients fulfilled the inclusion criteria (18 females, mean age: 73.1 ± 7.9 years, mean Mini-Mental State Examination: 23.1 ± 3.0). Analysis of variance revealed that posterior deep WMH (DWMH) volume was significantly larger than both anterior periventricular hyperintensity (PVH) and DWMH, and posterior PVH volumes. Multivariate regression analysis showed that increased volumes of the anterior PVH and the posterior DWMH and decreased rCBF of the parietal cortex negatively affected cognitive function. The other areas had no significant negative effects on cognitive function. Conclusion Our findings show that the volume of the posterior WMH was significantly larger than that of other areas, and the increased posterior WMH volume and decreased rCBF of the parietal cortex negatively affected cognitive function. Therefore, the posterior WMH volume and the parietal rCBF are key parameters of cognitive decline in AD patients.

While several studies support an association between the WMH volume or rCBF and cognitive decline in AD, no reports have simultaneously considered the effects of both WMH volume and rCBF on cognitive decline. Several methods have been proposed for performing quantitative white matter lesion load measurements on MRI (33)(34)(35)(36)(37)(38)(39)(40). However, there is currently no software that allows fusion of MRI and SPECT data. The purpose of the present study was to compare WMH volume and rCBF in relation to cognitive function by using a software to fuse MRI and SPECT data. We hypothesized that WMH volume and rCBF have specific additive and independent effects on cognitive decline. We developed a novel software program to examine the relationship between WMH volume and cognitive function for different cognitive domains and used technetium-99m-ethyl cysteinate diethylester ( 99m Tc-ECD) SPECT to evaluate whether the relationship between WMH volume and cognitive function is independent of rCBF.

Participants
In accordance with the principles of the Declaration of Helsinki, we prospectively registered 182 serial patients who consulted the memory clinic of the Mie University Hospital. All procedures followed the clinical study guidelines of the ethics committee of the Mie University hospital and were approved by the internal review board. All procedures were described to the patients, and informed consent was obtained from them or their caregivers in the written form. Neurologists with sufficient experience in examining patients with dementia comprehensively examined each patient in our study. We collected data from the patients who fulfilled the following inclusion criteria: (1) patients who consulted the Memory Clinic of the Mie University Hospital from October 2013 to August 2016 and were diagnosed with AD based on preestablished criteria, fulfilling the criteria for probable AD of the National Institute of Neurologic Disorders and Stroke/Alzheimer Disease and Related Disorders Association (NINCDS-ADRDA) (41); (2) patients who received a neuroimaging examination on a 3T-MRI; (3) patients who received a 99m Tc-ECD SPECT examination; and (4) patients who underwent neuropsychological assessments. The exclusion criteria were as follows: (1) patients who were not examined using MRI and 99m Tc-ECD SPECT; (2) patients who were not administered neuropsychological assessments; (3) patients who were diagnosed with dementia other than AD; and (4) patients who had normal cognitive function.
Tissue quantification was performed using a novel in-house software (FUsed Software for Imaging Of Nervous system: FUSION) that yielded an individualized volumetric profile of brain tissue. The obtained T1-weighted and FLAIR images were imported from DICOM format files for processing. To increase the accuracy of segmentation, we used the Lesion Segmentation Tool for lesion filling (42). Lesion filling was applied to T1-weighted images that were in alignment with the lesion probability map. For the preprocessing level, T1-weighted images were coregistered to FLAIR images. Next, to separate WM, segmentation was performed by using the T1-weighted images and a mask of cerebral ventricles. The preprocessing function was based on SPM 8 (Wellcome Trust Centre for Neuroimaging, UCL). Second-level tissue segmentation was performed to separate WMHs from WM, using a semiautomated operation that extracted the pixels falling within predetermined value as WMHs. The WMH volume, which appeared as hyperintense areas on FLAIR images, was quantified for each area. The brain tissue was classified into four areas based on the division of the longitudinal fissure of the cerebrum and central sulcus. WMHs were automatically classified as periventricular hyperintensity (PVH) or deep WMH (DWMH), and their corrected volumes were calculated in cubic centimeters (cc).
sPecT Protocol and evaluation of rcBF Intravenous radionuclide angiography was performed by bolus injection of the reconstituted 99mTc-ethyl cysteinate dimer (ECD) (600 MBq). Passage of the tracer from the aortic arch to the brain was monitored in a 128 × 128 format for 120 s at 1-s intervals using a three-head gamma camera system (GCA-9300A/DI, Toshiba, Tokyo, Japan) equipped with low-energy high-resolution fanbeam collimators. SPECT images were reconstructed by filtered back-projection using a ramp filter follower and postprocessing with a Butterworth filter. The tripleenergy window technique was employed for scatter correction. ROIs were placed manually over the aortic arch and bilateral cerebral hemispheres. Time activity curves of these two ROIs  were plotted and the brain perfusion index (BPI) was determined as described previously. BPI was then converted to the mean CBF (mCBF) value. The rCBF values were obtained by the conversion of total counts in brain SPECT into mCBF, using Lassen's correction (43). A three-dimensional stereotactic region of interest (ROI) template (3DSRT) program (FUJIFILM RI Pharma Co., Ltd.) was applied to assess the regional quantitative value (44)(45)(46)(47)(48). 3DSRT is a fully automated rCBF quantification program that can be used to examine a total of 636 ROIs. These ROIs are categorized into six brain segments on the 3DSRT template: bilateral parietal, temporal, and posterior cingulate cortices determined by previous studies (24)(25)(26)(27)(28)(29). The blood flow to each ROI was quantified in mL/100 g/min.
Finally, the results of evaluation of the SPECT images were fused with WMHs on MRI scans (Figure 1).

neuropsychological assessments
The Mini-Mental State Examination (MMSE) (49) and Raven's Colored Progressive Matrices (RCPM) (50) were used to quantify intellectual function. Memory was evaluated using the Rivermead Behavioural Memory Test (RBMT) (51). Assessment of visuospatial constructional ability was based on the method described by Strub and Black (52). A simple cube and a Necker cube were shown to the participants, and they were asked to draw them one by one. Each drawing was scored by assigning one of four possible grades (0: poor, 1: fair, 2: good, and 3: excellent). Frontal lobe function was assessed using two tasks, word fluency (WF) and the trailmaking test A and B (TMT-A/-B) (53). The WF test consisted of category and letter domains. For the categorical WF, participants were asked to name as many animals as possible in 1 min. For the letter WF task, participants were asked to say the name of objects that begin with each of four phonemes, ka, sa, ta, and te (54).

statistical analyses
Statistical analyses were performed with the Statistical Package for the Social Sciences, Version 20 (IBM Corp., Armonk, New York, NY, USA). Statistical analyses were conducted using oneway analysis of variance for continuous variables and the Ryan method for pairwise comparisons. We performed multivariate regression analysis using a generalized linear model. As predictor variable, we used age, and the PVH, DWMH, and CBFs of temporal, posterior cingulate, and parietal regions. As a response variable, we used MMSE, RCPM, RBMT, TMT-A/B, WF, and visuospatial constructional ability. A p-value of <0.05 was regarded as statistically significant.

resUlTs clinical Data
A total of 182 patients were registered for the present study, and 29 patients fulfilled the inclusion criteria. Their clinical characteristics are shown in Table 1. The mean (±SD) age was ( Table 3). The analysis showed that increased age correlated negatively with intellectual function (p = 0.008). Multivariate regression analysis also showed that WMHs and rCBF were significantly associated with memory and intellectual, frontal, and visuospatial functions, as follows. An increased anterior PVH volume negatively affected visuospatial constructional ability (p = 0.016) and increased posterior DWMH volume negatively affected WF (category), whereas increased volumes in other areas did not have negative impacts on cognitive function. In addition, decreased rCBF in the parietal areas negatively affected RCPM (time) (p = 0.033), TMT-A (p < 0.001), and visuospatial constructional ability (p = 0.012), whereas decreased rCBF in other areas did not.
In particular, the anterior PVH, the posterior DWMH, and the parietal regions were key areas that negatively affected cognitive function. We present a representative patient with increased posterior WMH volume and decreased rCBF of the parietal region as a typical example in Figure 1.

summarize the results
We compared MRI and SPECT findings with neuropsychological data from outpatients of the memory clinic of our hospital. An analysis of variance revealed that the posterior DWMH volume was significantly larger than the anterior PVH, anterior DWMH, and posterior PVH volumes. Multivariate regression analysis showed that increased anterior PVH and left posterior DWMH volumes and decreased rCBF of the parietal area correlated negatively with cognitive function, whereas other areas did not have negative effects on cognitive function.

DiscUssiOn
We developed a novel software program for quantitative volume analysis of WMHs and SPECT using fused imaging data. Our software has some advantages over other programs and showed reasonable results in accordance with previous studies. Although several software programs have been developed for WMH analysis (39,40), none are available for discrimination of PVH and DWMH on the anatomical basis of blood flow cliff or for the combined evaluation of rCBF. Measuring white matter volume is critical for the diagnosis of dementia, and fusion images of WMHs and rCBF are also useful in clinical settings. In addition to these regional analyses, we examined brain tissue after classification into four areas based on division of the longitudinal fissure of the cerebrum and central sulcus. Other anatomical ROIs have also been used (55); nonetheless, the regions we investigated may be useful in practical settings.
Previous studies have reported that the WMH posterior volume was larger than the anterior volume (56)(57)(58) in patients with cerebral amyloid angiopathy (CAA) as well as AD (59). CAA induces cerebral hypoperfusion in the white matter as a result of amyloid β deposition in the microvessels of the cerebral cortices and is found in more than 80% of all patients with AD (60). It is, therefore, likely that most patients in our cohort had comorbid CAA. Our results showed that the posterior WMH volume, especially that of the DWMH, was larger than the anterior and posterior PVH volumes. The results of the present study are in 73.1 ± 7.9 years, and 18 of the participants were female. The mean (±SD) MMSE was 23.1 ± 3.0.

association of WMh, rcBF, and cognitive Decline
The effect of regional WMHs and rCBF on neuropsychological test results was evaluated with multivariate regression analysis The present study showed that WMHs were significantly associated with cognitive functions, which is in accordance with previous findings (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23). Additionally, our results indicated that increased volumes of the anterior PVH and posterior DWMH negatively affected frontal and visuospatial function, both of which decline in AD (33). Given the larger posterior WMH volume, it is reasonable to hypothesize that posterior WMH negatively affected cognitive function in AD. On the other hand, SPECT results indicated that the parietal region more negatively affected cognitive function than did other regions. The area also negatively affected intellectual, frontal, and visuospatial function. A previous study showed that distinct cognitive profiles are associated with anterior and posterior WMH progression (61), When simultaneously considering the effects of both WMHs and rCBF factors (as shown in Figure 1), modulation of posterior regions may underscore neurodegeneration in the posterior association cortex and cognitive decline in AD. While increased WMH volume and decreased rCBF negatively affected cognitive function, some areas were associated with improvement in certain cognitive domains.
The present study has several limitations, including the small number of patients, investigating only patients with AD, and a lack of gray matter volume measurements. The patient cohort was small due to the prohibitive cost of SPECT imaging and standardized data collection form ECD as a radiolabeled ligand. Most patients in our cohort had AD, preventing us from establishing a significant effect of pathological background. Besides patients with AD, information is lacking in other types of dementias. Future studies that investigate other types of dementias, as well as gray and white matter volumes, could advance WMH and rCBF analysis to predict cognitive decline more accurately.

cOnclUsiOn
Our results agreed with those of previous studies, indicating that our software is a reliable tool. Collectively, the existing evidence suggests that posterior WMH volume and parietal cortex rCBF may predict cognitive decline in AD.

eThics sTaTeMenT
In accordance with the principles of the Declaration of Helsinki, we prospectively registered 182 serial patients who consulted the memory clinic of the Mie University Hospital. All procedures followed the clinical study guidelines of the ethics committee of the Mie University hospital and were approved by the internal review board. All procedures were described to the patients, and informed consent was obtained from them or their caregivers.