Introduction
Materials and Methods
Study Setting
CTL | MCI-s | MCI-pa
| AD | |
---|---|---|---|---|
Count | 340 | 360 | 85 | 297 |
Age, years | 75.0 ± 5.7 | 75.0 ± 6.9 | 74.3 ± 6.5 | 75.7 ± 7.0 |
Gender, Female/Male | 172/168 | 141/219 | 35/50 | 165/132 |
Education, years | 14.3 ± 4.3 | 13.9 ± 4.7 | 13.8 ± 4.2 | 12.0 ± 4.9 |
MMSE score | 29.1 ± 1.1 | 27.1 ± 1.7 | 26.5 ± 1.8 | 22.2 ± 3.7 |
CDR | 0 | 0.5 | 0.5 | 0.92 ± 0.43 |
ApoE-e4, N/P | 242/92 | 182/158 | 31/52 | 110/178 |
Cohort, ADNI/ANM | 227/113 | 260/100 | 62/23 | 175/122 |
Inclusion and Diagnostic Criteria
Imaging
MRI Data Acquisition
Regional Subcortical Volume Segmentation and Cortical Thickness Parcellation
Data Analysis
Multi and Univariate Data Analysis
Age Correction Methods
Implementation
Results
Model | Q2
| R2
| Accuracy % | Sensitivity % | Specificity % | |
---|---|---|---|---|---|---|
AD versus CTL classification (CV) | Uncorrected | 0.567 | 0.568 | 86.7 | 81.8 | 90.9 |
Covariate | 0.580 | 0.586 | 87.3 | 81.5 | 92.4 | |
Detrended | 0.582 | 0.583 | 88.2 | 82.8 | 92.9 | |
MCI prediction | Uncorrected | – | – | 62.7 | 70.2 | 60.9 |
Covariate | – | – | 62.9 | 71.4 | 60.9 | |
Detrended | – | – | 65.0 | 75.0 | 62.6 |
Time point | Model | Accuracy % | Sensitivity % | Specificity % |
---|---|---|---|---|
Month-12
MCI-p = 62, MCI-s = 260 | Uncorrected | 60.9 | 72.1 | 58.2 |
Covariate | 60.2 | 70.5 | 57.9 | |
Detrended | 63.0 | 75.4 | 60.2 | |
Month-18
MCI-p = 90, MCI-s = 232 | Uncorrected | 64.9 | 72.2 | 62.1 |
Covariate | 66.1 | 74.4 | 62.9 | |
Detrended | 67.7 | 75.6 | 64.7 | |
Month-24
MCI-p = 126, MCI-s = 196 | Uncorrected | 65.5 | 66.7 | 64.8 |
Covariate | 67.4 | 69.0 | 66.3 | |
Detrended | 68.9 | 69.8 | 68.4 | |
Month-36
MCI-p = 150, MCI-s = 172 | Uncorrected | 66.8 | 65.3 | 68.0 |
Covariate | 69.3 | 68.0 | 70.3 | |
Detrended | 70.8 | 68.7 | 72.7 |
AD versus CTL Classification | CTL subjects | AD patients | Incorrect-classified AD versus CTL | ||||
---|---|---|---|---|---|---|---|
Correct-classified | Incorrect-classified |
p value | Correct-classified | Incorrect-classified |
p value |
p value | |
Uncorrected model
| |||||||
Count | 309 | 31 | 243 | 54 | |||
Age, years | 74.5 ± 5.6 | 79.1 ± 5.2 |
<0.001
a
| 76.2 ± 6.7 | 73.3 ± 7.9 |
0.006
a
|
<0.001
a
|
Education, years | 14.3 ± 4.3 | 14.1 ± 5.5 | 0.839a
| 11.9 ± 4.8 | 12.8 ± 5.1 | 0.197a
| 0.294a
|
MMSE score | 29.1 ± 1.1 | 28.9 ± 1.2 | 0.326a
| 21.8 ± 3.7 | 24.2 ± 2.8 |
<0.001
a
|
<0.001
a
|
Gender, Male/Female | 152/157 | 16/15 | 0.797b
| 107/136 | 25/29 | 0.762b
| 0.637b
|
ApoE-e4, N/P | 217/87 | 26/5 | 0.138b
| 86/149 | 24/29 | 0.240b
|
0.001
b
|
Cohort, ADNI/ANM | 207/102 | 20/11 | 0.780b
| 141/102 | 34/20 | 0.505b
| 0.886b
|
Covariate model
| |||||||
Count | 314 | 26 | 242 | 55 | |||
Age, years | 74.8 ± 5.7 | 77.0 ± 4.3 | 0.058a
| 75.7 ± 6.8 | 75.4 ± 7.7 | 0.776a
| 0.237a
|
Education, years | 14.2 ± 4.2 | 14.0 ± 5.6 | 0.781a
| 11.9 ± 4.7 | 12.5 ± 5.3 | 0.412a
| 0.269a
|
MMSE score | 29.1 ± 1.1 | 28.9 ± 1.1 | 0.326a
| 21.7 ± 3.7 | 24.4 ± 2.7 |
<0.001
a
|
<0.001
a
|
Gender, Male/Female | 153/161 | 15/11 | 0.380b
| 103/139 | 29/26 | 0.171b
| 0.675b
|
ApoE-e4, N/P | 223/84 | 20/5 | 0.385b
| 83/151 | 27/27 |
0.048
b
|
0.012
b
|
Cohort, ADNI/ANM | 210/104 | 17/9 | 0.876b
| 143/99 | 32/23 | 0.902b
| 0.536b
|
Detrended model
| |||||||
Count | 316 | 24 | 246 | 51 | |||
Age, years | 75.0 ± 5.8 | 75.3 ± 4.6 | 0.756a
| 75.5 ± 7.0 | 76.3 ± 7.1 | 0.493a
| 0.553a
|
Education, years | 14.4 ± 4.3 | 12.8 ± 5.6 | 0.205a
| 12.0 ± 4.8 | 12.4 ± 5.0 | 0.530a
| 0.756a
|
MMSE score | 29.2 ± 1.1 | 28.8 ± 1.4 | 0.168a
| 21.8 ± 3.7 | 24.4 ± 2.7 |
<0.001
a
|
<0.001
a
|
Gender, Male/Female | 155/161 | 13/11 | 0.629b
| 105/141 | 27/24 | 0.180b
| 0.921b
|
ApoE-e4, N/P | 225/87 | 18/5 | 0.524b
| 86/151 | 24/27 | 0.156b
|
0.012
b
|
Cohort, ADNI/ANM | 213/103 | 14/10 | 0.363b
| 144/102 | 31/20 | 0.767b
| 0.840b
|
MCI prediction | MCI-s subjects | MCI-p subjects | Incorrect-predicted MCI-s versus MCI-p | |||||
---|---|---|---|---|---|---|---|---|
Correct-predicted | Incorrect-predicted |
p value | Correct-predicted | Incorrect-predicted |
p value |
p value | ||
Uncorrected model | Count | 219 | 141 | 59 | 26 | |||
Age, years | 73.6 ± 7.1 | 77.0 ± 6.1 |
<0.001
a
| 75.0 ± 6.9 | 73.1 ± 5.3 | 0.220a
|
0.003
a
| |
Education, years | 13.8 ± 4.6 | 14.1 ± 4.7 | 0.582a
| 13.8 ± 4.4 | 13.7 ± 3.8 | 0.820a
| 0.651a
| |
MMSE score | 27.4 ± 1.7 | 26.8 ± 1.6 |
0.001
a
| 26.5 ± 1.9 | 26.6 ± 1.6 | 0.802a
| 0.682a
| |
Gender, Male/Female | 144/75 | 75/66 |
0.017
b
| 35/24 | 15/11 | 0.888
b
| 0.672
b
| |
ApoE-e4, N/P | 118/87 | 65/71 | 0.077
b
| 22/36 | 8/16 | 0.694
b
| 0.190
b
| |
Cohort, ADNI/ANM | 151/68 | 109/32 | 0.084
b
| 44/15 | 18/8 | 0.609
b
| 0.375
b
| |
Covariate model | Count | 219 | 141 | 60 | 25 | |||
Age, years | 74.7 ± 7.1 | 75.3 ± 6.6 | 0.419a
| 74.3 ± 7.0 | 74.6 ± 5.0 | 0.852a
| 0.635a
| |
Education, years | 13.8 ± 4.6 | 14.2 ± 4.8 | 0.394a
| 13.9 ± 4.4 | 13.7 ± 3.9 | 0.853a
| 0.614a
| |
MMSE score | 27.4 ± 1.7 | 26.8 ± 1.7 |
0.001
a
| 26.6 ± 1.9 | 26.4 ± 1.6 | 0.740a
| 0.383a
| |
Gender, Male/Female | 149/70 | 70/71 |
<0.001
b
| 37/23 | 13/12 | 0.409
b
| 0.828
b
| |
ApoE-e4, N/P | 122/83 | 61/75 |
0.008
b
| 20/39 | 10/13 | 0.418
b
| 0.902
b
| |
Cohort, ADNI/ANM | 150/69 | 110/31 | 0.050
b
| 43/17 | 19/6 | 0.682
b
| 0.824
b
| |
Detrended model | Count | 225 | 135 | 63 | 22 | |||
Age, years | 75.2 ± 7.1 | 74.4 ± 6.6 | 0.273a
| 74.1 ± 6.9 | 75.4 ± 4.9 | 0.408a
| 0.492a
| |
Education, years | 13.8 ± 4.5 | 14.2 ± 4.8 | 0.478a
| 14.0 ± 4.1 | 13.2 ± 4.5 | 0.418T
| 0.380a
| |
MMSE score | 27.4 ± 1.7 | 26.7 ± 1.7 |
0.001
a
| 26.5 ± 1.9 | 26.6 ± 1.6 | 0.775a
| 0.803a
| |
Gender, Male/Female | 154/71 | 65/70 |
<0.001
b
| 39/24 | 11/11 | 0.329
b
| 0.872
b
| |
ApoE-e4, N/P | 127/83 | 56/75 |
0.001
b
| 40/22 | 12/8 | 0.715
b
| 0.817
b
| |
Cohort, ADNI/ANM | 156/69 | 104/31 | 0.114
b
| 46/17 | 16/6 | 0.979
b
| 0.659
b
|