Table 1.
Characteristics of the longitudinal test sample.
Table 2.
Baseline characteristics of the MCI sample used for prediction of AD conversion.
Fig 1.
Depiction of the BrainAGE concept.
(A) The model of healthy brain aging is trained with the chronological age and preprocessed structural MRI data of a training sample (left, with an exemplary illustration of the most important voxel locations that were used by the age regression model). Subsequently, the individual brain ages of previously unseen test subjects are estimated, based on their MRI data (blue, picture modified from [56]). (B): The difference between the estimated and chronological age results in the BrainAGE score, indicating abnormal brain aging. [Image reproduced from [17], with permission from Hogrefe Publishing, Bern]
Table 3.
Mean BrainAGE scores at baseline and last follow-up for all particular allelic isoforms within the diagnostic groups of the longitudinal sample.
Fig 2.
BrainAGE scores at (A) baseline and (B) the last visit for non-carriers and carriers of APOE ε4. Shown are boxplots, presenting the distribution of the BrainAGE scores for the 4 diagnostic groups NO, sMCI, pMCI and AD. BrainAGE scores differed significantly between diagnostic groups at baseline (F = 18.9, p < 0.001) and at follow-up scans (F = 30.6, p < 0.001). Post-hoc tests showed significant differences between BrainAGE scores in NO as well as sMCI from BrainAGE scores in pMCI as well as AD at baseline and last visit (p < 0.05). The boxes include values between the 25th and 75th percentiles and the median (red line). Lines extending the boxes below and above include data within 1.5 times the interquartile range. All outliers are symbolized with a red”+”. Width of the boxes symbolizes group size.
Fig 3.
Longitudinal changes in BrainAGE score for (A) NO, (B) sMCI, (C) pMCI and (D) AD patients. Thin lines represent individual trajectories of BrainAGE score over time of follow-up. Thick lines represent the estimated average regression lines for APOE ε4 non-carriers (red) and carriers (blue).
Fig 4.
Estimated longitudinal changes in BrainAGE scores for the 4 diagnostic groups: NO (light blue), sMCI (green), pMCI (red) and AD (blue), subdivided into APOE ε4 carriers and non-carriers. Post-hoc t-tests resulted in significant differences for ε4 carriers and non-carriers as well as for NO / sMCI vs. pMCI / AD (p < 0.05).
Table 4.
Correlation coefficients between BrainAGE scores and cognitive functioning (ADAS scores) as well as disease severity (MMSE & CDR-SB scores) for each diagnostic group and the whole test sample, separately for APOE ε4 carriers and non-carriers.
Fig 5.
Mean days to conversion from MCI to AD subdivided into all allelic combinations of APOE.
Presented are the mean ± SD days to conversion within the given allelic combinations of APOE (F = 3.14; p = 0.08): ε2/ε3 (n = 2): 758 ± 356, ε3/ε3 (n = 32): 448 ± 210, ε2/ε4 (n = 5): 756 ± 201, ε3/ε4 (n = 52): 534 ± 292, ε4/ε4 (n = 21): 591 ± 246. The boxes include values between the 25th and 75th percentiles and the median (red line). Lines extending the boxes below and above include data within 1.5 times the interquartile range. All outliers are symbolized with a red”+”. Width of the boxes symbolizes group size.
Fig 6.
Cumulative probability for MCI patients of remaining AD-free, divided into patients with the score of interest below the median (light lines) and above it (dark lines). Non-carriers of the APOE ε4 gene are painted in blue, carriers in red. Shown are Kaplan-Meier survival curves based on Cox regression, comparing the cumulative incidence of AD in ε4 carriers and non-carriers in (A) BrainAGE, (B) MMSE, (C) CDR-SB, and (D) ADAS scores. The follow-up duration is truncated at 1250 days.
Table 5.
Cox Regression values for cumulative AD incidence in APOE ε4 carriers and non-carriers in the BrainAGE, MMSE, CDR-SB, and ADAS scores alone and in combination with the APOE ε4 carrier status, based on a median split.
Fig 7.
ROC curves for pMCI_early subjects, analyzing individual subject classification based on baseline BrainAGE score in either remaining MCI or converting to AD, subdivided into APOE ε4 carriers and non-carriers.
Achieved accuracies (sensitivity / specificity) for predicting conversion from MCI to AD for APOE ε4 carriers: 85% (0.79 / 0.92), non-carriers: 78% (0.71 / 0.83).
Fig 8.
ROC curves of APOE ε4 carriers vs. non-carriers based on baseline BrainAGE scores in all MCI patients.
Achieved accuracies (sensitivity / specificity) for predicting conversion from MCI to AD during follow-up for APOE ε4 carriers: 75% (0.70 / 0.92) and for non-carriers: 74% (0.65 / 0.83).
Table 6.
Results for predicting conversion to AD in MCI subjects (APOE ε4 carriers).
Table 7.
Results for predicting conversion to AD in MCI subjects (APOE ε4 non-carriers).