The Use of F-18 FDG PET-Based Cognitive Reserve to Evaluate Cognitive Decline in Alzheimer’s Disease, Independent of Educational Influence

Background and Objectives: The optimal assessment of cognitive function, including the impact of education, is crucial in managing Alzheimer’s disease (AD). This study aimed to evaluate the role of cognitive reserve (CR), represented by the metabolic status of regions of the cerebral cortex, to evaluate cognitive decline considering the educational attainment of patients with AD. Materials and Methods: We used data from the Alzheimer’s Disease Neuroimaging Initiative database, and selected 124 patients who underwent both baseline F-18 fluorodeoxyglucose (FDG) and F-18 florbetaben (FBB) positron emission tomography (PET) scans. Demographics, cognitive function variables (Clinical Dementia Rating—Sum of Boxes [CDR]; AD Assessment Scale 11/13 [ADAS11/13] Mini-Mental State Examination [MMSE]), and the average standardized uptake value ratio (SUVR) of cerebral cortex regions to those of the cerebellum were obtained from the data. The participants’ education level was divided into low and high education subgroups using four cut-offs of 12, 14, 16, and 18 years of educational attainment (G12, G14, G16, and G18, respectively). Demographic and cognitive function variables were compared between the two subgroups in each of the four groups, and their correlations with the SUVRs were evaluated. Results: There was no significant difference between the high and low education subgroups in each of the four groups, except for ADAS11/13 and MMSE in G14 and age in G16. The SUVRs of FDG PET (FDGSUVR) were significantly correlated with CDR, ADAS11/13, and MMSE scores. FDGSUVR showed different trajectories of neurodegeneration between the low and high education groups. Conclusions: FDGSUVR correlated moderately but significantly with neuropsychological test results, without being influenced by education level. Therefore, FDG PET may reflect CR independent of education level, and therefore could be a reliable tool to evaluate cognitive decline in AD.


Background
Alzheimer's disease (AD) is a neurodegenerative disorder characterized by the mismatched production and clearance of β-amyloid (Aβ) and tau proteins. This is followed by adverse neuro-inflammation. Cognitive decline is the most common symptom of AD, and is a key factor in the diagnosis and evaluation of AD progression and treatment response. Conventional neuropsychological tools used to measure cognitive function include the Mini-Mental State Examination (MMSE), the Clinical Dementia Rating (CDR), and the AD Assessment Scale (ADAS) [1][2][3].
The discrepancy between the severity of cognitive decline and postmortem pathological findings among participants in the low and high education groups has been termed "cognitive reserve (CR)" [4,5]. The CR signifies the influence of education in the assessment of cognitive function. The MMSE, which is a representative cognitive test, may also be influenced by the education level of individuals [6]. Therefore, an appropriate measurement tool is required to assess cognitive decline, particularly for cases in which the education level of the patient may have impacted the optimal assessment of cognitive function [4].
F-18 F-18 fluorodeoxyglucose (FDG) positron emission tomography (PET) is used as an image biomarker to assess AD, and for the differential diagnosis of other dementia types. Amyloid PET depicts the accumulation of Aβ and plays a role in the diagnosis of AD [7][8][9]. The correlation of cognitive function with FDG and amyloid PET has been extensively evaluated [10][11][12]. The standardized uptake value ratio (SUVR) of FDG PET, using the cerebellum as a reference, is used as a marker of cognitive decline [11] and may more accurately represent the CR without the interference of education level, thereby enabling the optimal assessment of cognitive function [4].
Therefore, this study aimed to evaluate the correlation of FDG PET-based SUVR with neuropsychological test results according to education level and to assess whether the measurement of SUVR by FDG PET can be a surrogate for conventional cognitive measures and a proxy for education-independent CR in AD.

Dataset from the Alzheimer's Disease Neuroimaging Initiative
Several key variables from various clinical information reports and biomarker results from the Alzheimer's Disease Neuroimaging Initiative (ADNI) protocols are merged to create the "adnimerge" table. A total of 2159 participants in ADNI-3 datasets were screened and 124 participants who had both baseline FDG PET and florbetaben (FBB) PET amyloid imaging were finally enrolled. The FDG SUVR and FBB SUVR of 124 participants were analyzed in this study. The FDG SUVR was defined as the average of the FDG PET-based SUVR values of the angular, temporal, and posterior cingulate regions. The FBB SUVR was defined as the average of the FBB PET-based SUVR values of the frontal cortex, anterior cingulate gyrus, precuneus cortex, and parietal cortex. The educational attainment in years (EDU), number of apolipoprotein E4 alleles (APOE), MMSE score, CDR, and ADAS11/13 data were also obtained from the baseline study of ADNI datasets. The APOE results were available for 99 of the 124 (79.8%) participants. Cerebrospinal fluid Aβ data were not available. The Institutional Ethics Committee of Ulsan University Hospital confirmed that ethical approval was not required for this observational study, and waived the requirement for informed consent (IRB file number: UUH2022-05-029).

Statistical Analyses
All data are described as mean and standard deviation for continuous variables and as numbers (percentages) for categorical variables. Age, sex, EDU, APOE, CDR, ADAS11/13, MMSE, FDG SUVR , and FBB SUVR were compared between mild cognitive impairment (MCI) and AD groups using an independent t-test and a Mann-Whitney U test for continuous and categorical variables, respectively. Four cut-offs of 12, 14, 16, and 18 years for EDU were used and participants were divided into low and high education subgroups (G12: ≤12 vs. >12, G14: ≤14 vs. >14, G16: ≤16 vs. >16, and G18: ≤18 vs. >18, respectively). The association of each subgroup of G12, G14, G16, and G18 with neuropsychological results was evaluated using Pearson's and Spearman's correlation analyses. For the MCI and AD groups, univariable and multivariable analyses were performed to evaluate the association of age, sex, EDU, APOE, FDG SUVR , and FBB SUVR with neuropsychological tests to assess cognitive function. In multivariable regression analyses, dummy variables were used for categorical variables, including sex and APOE.

Baseline Demographics
There were no statistically significant differences regarding age, sex, and EDU between the MCI and AD groups; however, APOE, CDR, ADAS11/13, and MMSE scores, as well as FDG SUVR and FBB SUVR , showed significant differences. One of the three healthy participants was 90 years old and had CDR, ADAS11, ADAS13, and MMSE scores of 7, 29.33, 41.33, and 20, respectively. The other two healthy participants exhibited normal neuropsychological results. Detailed demographics of all participants are presented in Table 1.

Correlation between Demographics, FDG SUVR , FBB SUVR , and Neuropsychological Tests
Correlation analyses for all participants revealed that APOE, FDG SUVR , and FBB SUVR were correlated with CDR, ADAS11/13, and MMSE scores. Age correlated with ADAS11/13; however, the correlation was not statistically significant (Table 3). In the MCI group, similar correlation patterns for FDG SUVR , FBB SUVR , and age with CDR, ADAS11/13, and MMSE were observed. The AD group showed almost no correlations with the results of the neuropsychological tests (Supplementary Tables S3 and S4). Table 3. Correlation analyses for demographics, FDG SUVR , FBB SUVR , and neuropsychological tests.

Multivariable Analyses for Correlations between Demographics, FDG SUVR , FBB SUVR , and Neuropsychological Tests
Multivariable linear regression analyses revealed that FDG SUVR was a significant independent factor correlating with CDR, ADAS11/13, and MMSE in most groups, followed by FBB SUVR and age. Sex, EDU, and APOE were not significant variables ( Table 4). The regression lines of FDG SUVR and FBB SUVR for CDR, ADAS11/13, and MMSE scores are shown in Figures 1 and 2. The FDG SUVR downward slopes for the participants with high education levels for the ADAS11/13 and MMSE scores in the G12, G14, and G16 groups were steeper than those for the participants with low education levels. The slopes for CDR in all EDU groups as well as the ADAS11/13 and MMSE scores in the G12, G14, and G16 groups were gentler for those with high than for those with low education levels ( Figure 1). The upward sloping lines of FBB SUVR were not as regular as the downward sloping lines of FDG SUVR (Figure 2).

Discussion
With the global aging of society, AD is increasingly becoming a public health issue [13]. Cognitive decline becomes more severe over time, because of the progressive neurodegenerative nature of AD. However, education has demonstrated a protective influence against AD [14,15], which led to the hypothesis that CR might cause the delayed detection

Discussion
With the global aging of society, AD is increasingly becoming a public health issue [13]. Cognitive decline becomes more severe over time, because of the progressive neurodegenerative nature of AD. However, education has demonstrated a protective influence against AD [14,15], which led to the hypothesis that CR might cause the delayed detection of AD or its progression, particularly in highly educated populations [4]. Therefore, we evaluated whether FDG SUVR correlated with the results of various neuropsychological tests used to evaluate CR, and whether FDG SUVR could reflect CR in a manner that was independent of the influence of education. We observed that FDG SUVR was significantly and moderately correlated with CDR, ADAS11/13, and MMSE scores.
In our study, the age of high-education-level participants at baseline was expected to be older than that of the low-education-level participants, as their neuropsychological results did not differ according to the concept of CR. However, we found that the age at baseline did not differ between patients with low or high education levels. In contrast, the ADAS11/13 and MMSE differed for all participants in G14, and the MMSE scores for the MCI group differed in G12 and G14 between the low-and high-education groups. Only in the MCI group for G16 did age differ significantly according to different MMSE results. None of the other groups exhibited significant differences. These results suggest that CR may not affect the early detection of AD, regardless of the education level.
Therefore, the question might arise regarding whether the concept of CR is meaningful. Since CR could delay AD diagnosis in the higher education group, FDG and amyloid PET were evaluated as tools to assess CR, regardless of the level of education [4,5,15]. Interestingly, FDG SUVR and FBB SUVR revealed a moderate correlation with neuropsychological test results across all EDU-based subgroups in our study. Regression lines of FDG SUVR and FBB SUVR (Figures 1 and 2) data at baseline showed no significant differences between the low-and high-education level groups (left side of the lines), in which neuropsychological results ranged in severity from mild to moderate. However, more distinctive differences might be observed if the lines were extended to more severe neuropsychological results on the right side. The virtual differences in FDG SUVR and FBB SUVR between the low-and high-education level groups might indicate the existence of CR, although this was not based on an analysis with real data. In addition, longitudinal correlations between FDG or FBB PET/CT and ADAS11 have been presented in other studies [5,16]. Another study using C-11 Pittsburgh Compound B (PiB) reported no relationship between education and CR in participants with lower PiB uptake, which might represent the early stage of AD pathology, similar to the left side of the regression lines in our study. In addition, the PiB study highlighted that the duration of education was correlated with the CDR and MMSE scores in participants with higher PiB uptake, which might reflect the advanced pathological changes of AD (similar to the right side of the lines in our study) [17]. FDG and amyloid PET imaging are representative biomarkers of pathology and neurodegeneration in AD [9], thereby suggesting that CR might explain the different neurodegenerative trajectories between low-and high-education-level groups. The gap in the trajectories in terms of subjective or objective cognitive decline between low-and high-education-level groups might be minimal at the time of the initial workup, but may become more pronounced as cognitive impairment progresses. Therefore, CR might play a more important role in evaluating treatment responses than in the diagnosis of AD.
Neuropsychological tests remain the mainstay for evaluating treatment responses in patients with AD. Conducting neuropsychological tests in an AD population with profound cognitive decline is challenging. Questionnaire-based neuropsychological assessments may be difficult to obtain, particularly for individuals with intellectual disabilities. Indeed, the low sensitivity of questionnaire-type cognitive assessments has been previously reported [18]. The evaluation of cognitive function using neuropsychological assessments, including the MMSE, remains controversial [19,20]. Other studies have indicated that the educational level of participants might affect the MMSE results [21,22]. Our study revealed that the regression lines of FDG SUVR and FBB SUVR were more reliable in representing CR with the current cognitive status demonstrated by neuropsychological test results, independent of educational levels. In multivariable analyses, FDG SUVR was the factor most significantly correlated with CDR, ADAS11/13, and MMSE scores, followed by FBB SUVR and age. Other studies also showed that FDG SUVR and FBB SUVR were significantly correlated with cognition [23]. In addition, those studies hypothesized that the correlation of FDG SUVR with cognitive status may be more significant than that of FBB SUVR , as our study also indicated [16,23,24]. Similarly, a previous study reported that FDG and amyloid PET/CT might be useful tools to evaluate the concept of CR in mild cases of AD [5,10]. These findings suggest that FDG PET could be useful as a surrogate for the evaluation of cognitive decline, considering the concept of CR, not only for diagnosis but also for treatment response.
Our study had some limitations. This was a cross-sectional study, in which chronological changes in cognitive decline were not monitored. The duration of subjective or objective cognitive impairment could vary and affect the time-based correlation between FDG SUVR and neuropsychological results, which were not evaluated in this study. However, neuropsychological assessments become more difficult as cognitive decline progresses. Nevertheless, our study indicated that even in advanced cases FDG SUVR could provide reliable information on CR, which showed different trajectories between low and high education level participants. Our study was further limited in that only the average FDG SUVR of the angular, temporal, and posterior cingulate regions in the brain were included. As regional changes in the brain glucose metabolism in AD have been well evaluated, more detailed correlations with each brain region based on time and neuropsychological results might provide a better understanding of CR. Further studies should be designed based on these considerations.

Conclusions
In conclusion, FDG SUVR showed significant correlations with the results of neuropsychological tests based on questionnaires. The FDG SUVR displayed different trajectories of neurodegeneration between participants with low and high education levels based on diminished cognitive function, supporting the concept of CR. Therefore, FDG PET could be a reliable and objective imaging tool to assess cognitive decline and might be a proxy of CR which is independent of educational influence in patients with AD.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/medicina59050945/s1, Table S1: Comparisons in subgroups by the educational attainment for MCI participants; Table S2: Comparisons in subgroups by the educational attainment for AD participants; Table S3: Correlation analyses in subgroups by the educational attainment for MCI participants; Table S4: Correlation analyses in subgroups by the educational attainment for AD participants.

Institutional Review Board Statement:
The institutional ethics committee of Ulsan University Hospital confirmed that no ethical approval was required for this observational study and waived the informed consent (IRB file number UUH2022-05-029).
Informed Consent Statement: Patient consent was waived because data used in this study was public data from ADNI. and Transition Therapeutics. The Canadian Institutes of Health Research provides funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer's Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. Alzheimer's Disease Neuroimaging Initiative (ADNI) data can be accessed by accepting the Data Use Agreement and submitting an online application form (https://adni.loni.usc.edu/data-samples/access-data/). All researchers who are interested and have no commercial purpose may freely access ADNI images.

Conflicts of Interest:
The authors declare no conflict of interest.