Abstract
In Alzheimer’s disease (AD), association of cognition with disease biomarkers is important for accurate diagnosis, prognosis and staging. We studied the value of cerebrospinal fluid (CSF), cortical thickness (CT), and volume (V) measures (in isolation & combination) to predict cognitive ability, as well as cognitive status, in a cross-sectional study of data from Alzheimer’s Disease Neuroimaging Initiative (ADNI). Cognitive status was measured by Mini Mental State Exam (MMSE). Results show that combining biomarkers allows more accurate AD diagnosis. Specifically, combining CSF, CT, and V measures seems promising. In isolation, CSF levels show significant but weak associations with MMSE. Aβ 1– 42 is associated with Mild Cognitive Impairment. CT and V measures of hippocampus and entorhinal cortex are strongly associated with Alzheimer ‘s Disease diagnosis and with ruling out the disease.
Keywords: Alzheimer’s Disease, Biomarkers, Cortical Thickness, Volumetric, CSF
1. Introduction
Brain changes related to Alzheimer’s disease (AD) exist years before manifestation of any cognitive decline1. Mild cognitive impairment (MCI) is impairment in cognition with preserved independence and functional abilities. MCI is considered a risk factor for dementia, however, not all MCI conditions are due to AD pathology. While ultimate diagnosis of AD is made at autopsy2, more accurate in vivo diagnosis can be achieved by combining biomarkers in combination with clinical tests and history3. Understanding how combinations of biomarkers improve diagnosis and assessment of the impact of novel therapeutics is important.
Proteins such as Tau, Abeta1–42, PTau181P, which are aggregated as neurofibrillary tangles and amyloid plaques and are also present in the cerebrospinal fluid (CSF), are considered as AD biomarkers 4. Decreased levels of CSF Abeta42 and increased levels of Tau and phosphorylated Tau (p-Tau) are considered biomarkers indicating underlying AD pathology 5. Moreover, AD is associated with changes in brain structure, including changes to cortical thickness and cortical and subcortical volumes6. Specifically, reduced volume of hippocampus is accepted as an AD biomarker5.
While CSF and imaging biomarkers are frequently analyzed in isolation, the current study uses a well characterized dataset to test the utility of biomarkers both in isolation and in combination, in a large population, and a unique statistical approach. The aims are to 1) differentiate groups including AD, MCI, and healthy participants with normal cognitive status (NCS), and 2) assess relationship of imaging (CT & V) and fluid biomarker (CSF) with cognition.
2. Methods
2.1 Participants and Material
Here we present a cross sectional study on 822 individuals (M=77, SD=6.9) including healthy controls with NCS and patients diagnosed with MCI or AD. Data used in the preparation of this article were ADNI 1, obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). Diagnoses were made following ADNI protocol (adni.loni.usc.edu). NCS Participants showed no sign of depression, mild cognitive impairment or dementia. MCI participants reported a subjective memory concern despite maintaining daily activities, and their performance on Wechsler Memory Scale Logical Memory II. Participants would be labeled as AD if they met the NINCDS/ADRDA criteria for probable AD7.
We used the Mini Mental State Exam (MMSE) data to assess overall cognition. CSF levels of total Tau, Abeta1–42, and PTau181P proteins were analyzed and structural measures of brain regions of interest (ROI) were computed using freesurfer version 4.3, downloaded on April 18th, 2016 (https://surfer.nmr.mgh.harvard.edu/). Regions were selected to include currently accepted biomarkers and regions known to be affected early in the AD process8. These measures included CT of 38 regions (ROIs) as well as total gray matter. ROIs were the left and the right hippocampus, ACC, insula, inferior parietal cortex, middle frontal cortex, inferior temporal cortex, pre-central cortex, post-central cortex, caudate-middle frontal cortex, and frontal poles, inferior frontal gyrus, superior frontal gyrus, inferior parietal gyrus, inferior temporal gyrus, middle temporal gyrus, post-central cortex, caudate-middle frontal cortex, entorhinal cortex, precuneus, and fusiform. Also, the volume measures of (V) of 20 ROIs were the right and the left inferior parietal, entorhinal cortex, middle temporal gyrus, temporal lobe, thalamus, hippocampus, middle temporal gyrus, putamen, caudate, and amygdala.
2.2 Data Analysis
Pearson’s correlations were calculated to measure the association between CSF protein levels, CT of brain ROIs and Vs of brain ROIs with MMSE scores. To correct for the probability of accepting null hypotheses due to multiple comparisons, Family-Wise Error Rate Bonferroni correction was used to adjust p values. For the whole population, regression models including every individual predictor were used to determine the variables explaining the highest variance of outcome of MMSE. All variables that explained more than 10% of variance were entered in separate stepwise regression models: Specifically, a linear regression model involving the CSF variables, another model involving CT variables and a third model including volumetric variables. Using Variable Inflation Factor (VIF), collinearity was checked among variables. Predictors that showed collinearity were not included in the regression model. Predictors that could significantly contribute in explaining the variance of outcome of MMSE were put in a single stepwise linear regression model. A forward stepwise logistic binary analysis was performed with the variables explaining the variance of outcome of cognitive impairment, to estimate prediction of correct diagnosis (AD, MCI, and NCS).
Table I. Predictors of the MMSE scores including CT and V of ROIs, as well as CSF proteins. Results of individual regression models, and selection of ROIs above the cut-off.
Table I.
Association of cerebrospinal fluid (CSF) proteins, region of interest (ROI)’s volume and cortical thickness with Mini Mental State Exam (MMSE).
| MMSE | Predictors | Variance of outcome |
|---|---|---|
| CSF | Tau | R2=0.07, β= .28, F(1,407)=35.012, p<.000 |
| Abeta1–42 | R2 =0.11, β= .33, F(1,412)=51.4, p<.000 | |
| PTau181P | R2=0.07, β= −.27, F(1,413)=33.51, p<.000 | |
| ROIs’ volume | L hippocampus | R2= 0.205, β= .453, F(1,818)=210.7, p<.000 |
| L entorhinal cortex | R2= 0.194, β= .382, F(1,110)=18.804, p<.000 | |
| R entorhinal cortex | R2= 0.194, β= .403, F(1,818)=159.05, p<.000 | |
| ROIs’ CT | R middle temporal gyrus | R2= 0.17, β= .417, F(1,818)=172.12, p<.000 |
| L entorhinal cortex | R2=0.21, β= .160, F(1,818)=21.58, p<.000 | |
| L inferior parietal gyrus | R2=0.21, F(1,818)=21.58, β= .160, p<.000 | |
| All predictors | Abeta1–42 | R2=0.5, F(5, 408)=106.96, p<.000 |
| L hippocampus | ||
| L entorhinal cortex | ||
| R entorhinal cortex | ||
| R middle temporal gyrus | ||
| L entorhinal cortex | ||
| L inferior parietal gyrus |
3. Results
Individual regressions showed that CSF Tau, Aβ1–42 and pTau181P were significantly associated with MMSE and explained 7%–11% of its variance. Regression models combining CT and V variables, showed that V of the left hippocampus and both entorhinal cortices and CT of the right middle temporal gyrus, as well as the left entorhinal cortex and the inferior parietal gyrus explained about 20% of variance in MMSE. Stepwise regression of a model including CSF Aβ1–42, as well as the above regions explained 50% of variance in MMSE. Table II summarizes the association of CSF proteins, ROI’s V and CT with MMSE.
Table II.
Association of cerebrospinal fluid (CSF) proteins, region of interest (ROI)’s volume and cortical thickness with Cognitive Status (CS).
| CS | Predictors | % Correctly predicted | Chi square, df Cox & Snell R square, Nagelkerke R square B, SE, ExpB |
|---|---|---|---|
| MCI | Aβ1–42, | 55.7 %, | X2=5.64, df= 1 |
| Constant | p< 0.01 | Cox & Snell R2= 0.014, Nagelkerke R2= 0.018 | |
| B=−0.004, SE=0.002, ExpB=0.996 | |||
| B=0.617, SE=0.317, ExpB=1.85 | |||
|
| |||
| AD | Aβ1–42, | 80.4%, | X2=112.20, df= 4 |
| V of R entorhinal cortex | p<0.0001 | Cox & Snell R2= 0.240, Nagelkerke R2= 0.357 | |
| CT of L inferior parietal gyrus, | B=−0.01, SE=0.003, ExpB=0.990 | ||
| CT of L entorhinal cortex | B=−1.207, SE=0.319, ExpB=0.299 | ||
| Constant | B=−2.562, SE=0.690, ExpB=0.077 | ||
| B=−0.001, SE=0.0001, ExpB=0.999 | |||
| B=11.132, SE=47.730, ExpB=68308.224 | |||
|
| |||
| NCS | Tau | 79.5%, | X2=150.75, df= 5 |
| Aβ1–42, | p<0.001 | Cox & Snell R2= 0.308, Nagelkerke R2= 0.444 | |
| V of L entorhinal cortex | B=−0.01, SE=0.004, ExpB=0.990 | ||
| V of L hippocampus | B=−0.009, SE=0.003, ExpB=1.009 | ||
| CT of R middle temporal gyrus | B=0.977, SE=0.360, ExpB=2.655 | ||
| Constant | B=−0.001, SE=0.0001, ExpB=1.001 | ||
| B=1.871, SE=0.788, ExpB=6.497 | |||
| B=−13.320, SE=2.266, ExpB=0.0001 | |||
CSF Aβ1–42 was also associated with MCI diagnosis. Aβ1–42, volume of the right entorhinal cortex and the CT of the left inferior parietal gyrus and the entorhinal cortex were associated with AD diagnosis. Normal levels of CSF Aβ1–42 Tau combined with no structural changes to the temporal lobes confirmed normal status.
Table II summarizes the association of CSF proteins, ROI’s V and CT with cognitive impairment.
4. Discussion
Combining imaging and CSF results created a more sensitive biomarker associate of cognitive function, validating previous reports4. The importance of biomarkers for diagnosis was also confirmed. Higher level of CSF Aβ 1–42 showed association with MCI (Table II). It has been argued that a positive CSF Aβ1–42 without any evidence of neuroanatomical injury has an intermediate likelihood that MCI is due to AD5. This is while lower levels of CSF Aβ 1–42 are used to differentiate AD from other dementia3, 5. Consistent with the literature4, combinations of CSF and MRI biomarkers offer stronger support for accurate diagnosis, as compared to individual biomarkers. CSF Aβ1–42 levels, combined with CT and V measures of hippocampus and entorhinal cortices were strongly associated with AD diagnosis (Table II). These results are consistent with previous studies on diagnostic biomarkers studies3: Normal levels of CSF Aβ1–42 and Tau and normal volume of the left entorhinal and the hippocampus as well as normal thickness of the right middle temporal gyrus significantly were associated with normal cognitive status (Table II).
In line with previous studies9, in isolation, various CSF levels showed significant but weak associations with MMSE scores, Aβ1–42 showing the strongest association (See Table I). Imaging biomarkers also related to cognition: temporal lobe measures, including the hippocampus and the entorhinal cortex, as well as the parietal gyrus were associated with MMSE (Table I). Changes to these regions are known to be the early signs of cognitive impairment10. Given the early course of Alzheimer’s disease in these regions impacts new memory formation, these results were expected.
This study showed that combined CSF, CT, and V measures offer promise for early AD diagnosis and tracing of cognitive decline in well-defined cohorts. However, their ability to predict outcome needs to be tested prospectively in typical clinical population and with autopsy confirmed AD pathology. Further, CSF measures are invasive. More research is required to reach noninvasive but rapid, inexpensive and reliable biomarkers.
Acknowledgments
Sources of support that require acknowledgment:
This work was supported by grant funding from the National Institute of General Medical Sciences (grant: P20GM109025)
Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). For up-to-date information, see www.adni-info.org.
Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing 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.
References
- 1.Jack Clifford R, et al. Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers. The Lancet Neurology. 2013;12(2):207–216. doi: 10.1016/S1474-4422(12)70291-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Guillozet Angela L, et al. Neurofibrillary Tangles, Amyloid, and Memory in Aging and Mild Cognitive Impairment. Archives of neurology. 2003;60(5):729–736. doi: 10.1001/archneur.60.5.729. [DOI] [PubMed] [Google Scholar]
- 3.Lista Simone, et al. Csf Aβ1-42 Combined with Neuroimaging Biomarkers in the Early Detection, Diagnosis and Prediction of Alzheimer’s Disease. Alzheimer’s & Dementia. 2014;10(3):381–392. doi: 10.1016/j.jalz.2013.04.506. [DOI] [PubMed] [Google Scholar]
- 4.Westman E, Muehlboeck JS, Simmons A. Combining MRI and CSF measures for classification of Alzheimer’s disease and prediction of mild cognitive impairment conversion. Neuroimage. 2012;62(1):229–238. doi: 10.1016/j.neuroimage.2012.04.056. [DOI] [PubMed] [Google Scholar]
- 5.Albert Marilyn S, et al. 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. Alzheimer’s & Dementia. 2011;7(3):270–279. doi: 10.1016/j.jalz.2011.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Weiner Michael W, et al. The Alzheimer’s Disease Neuroimaging Initiative: A Review of Papers Published since Its Inception. Alzheimer’s & Dementia. 2013;9(5):e111–e194. doi: 10.1016/j.jalz.2013.05.1769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology. 1984;34(7):939–44. doi: 10.1212/wnl.34.7.939. [DOI] [PubMed] [Google Scholar]
- 8.Pini Lorenzo, et al. Brain atrophy in Alzheimer’s disease and aging. Ageing research reviews. 2016;30:25–48. doi: 10.1016/j.arr.2016.01.002. [DOI] [PubMed] [Google Scholar]
- 9.Engelborghs Sebastiaan, et al. Neuropsychological and Behavioural Correlates of Csf Biomarkers in Dementia. Neurochemistry international. 2006;48(4):286–295. doi: 10.1016/j.neuint.2005.11.002. [DOI] [PubMed] [Google Scholar]
- 10.Frisoni Giovanni B, et al. The Clinical Use of Structural Mri in Alzheimer Disease. Nature Reviews Neurology. 2010;6(2):67–77. doi: 10.1038/nrneurol.2009.215. [DOI] [PMC free article] [PubMed] [Google Scholar]