Background People with subjective cognitive decline (SCD) may be at increased risk for Alzheimer’s disease (AD). However, not all studies have observed this increased risk. This project examined whether four common methods of defining SCD yields different patterns of atrophy and future cognitive decline between cognitively normal older adults with (SCD+ ) and without SCD (SCD−). Methods Data from 273 Alzheimer’s Disease Neuroimaging Initiative cognitively normal older adults were examined. To operationalize SCD we used four common methods: Cognitive Change Index (CCI), Everyday Cognition Scale (ECog), ECog + Worry, and Worry. Voxel-based logistic regressions were applied to deformation-based morphology results to determine if regional atrophy between SCD− and SCD+ differed by SCD definition. Linear mixed-effects models were used to evaluate differences in future cognitive decline. Results Results varied between the four methods of defining SCD. Left hippocampal grading was more similar to AD in SCD+ than SCD− when using the CCI ( p = .041) and Worry ( p = .021) definitions. The right ( p= .008) and left ( p= .003) superior temporal regions had smaller volumes in SCD+ than SCD−, but only with the ECog. SCD+ was associated with greater future cognitive decline measured by Alzheimer’s Disease Assessment Scale, but only with the CCI definition. In contrast, only the ECog definition of SCD was associated with future decline on the Montreal Cognitive Assessment. Conclusion These findings suggest that the various methods used to differentiate between SCD− and SCD+ influence whether volume differences and findings of cognitive decline are observed between groups in this retrospective analysis.

Background: While both cognitive and magnetic resonance imaging (MRI) data has been used to predict progression in Alzheimer’s disease, heterogeneity between patients makes it challenging to predict the rate of cognitive and functional decline for individual subjects. Objective: To investigate prognostic power of MRI-based biomarkers of medial temporal lobe atrophy and macroscopic tissue change to predict cognitive decline in individual patients in clinical trials of early Alzheimer’s disease. Methods: Data used in this study included 312 patients with mild cognitive impairment from the ADNI dataset with baseline MRI, cerebrospinal fluid amyloid-β, cognitive test scores, and a minimum of two-year follow-up information available. We built a prognostic model using baseline cognitive scores and MRI-based features to determine which subjects remain stable and which functionally decline over 2 and 3-year follow-up periods. Results: Combining both sets of features yields 77%accuracy (81%sensitivity and 75%specificity) to predict cognitive decline at 2 years (74%accuracy at 3 years with 75%sensitivity and 73%specificity). When used to select trial participants, this tool yields a 3.8-fold decrease in the required sample size for a 2-year study (2.8-fold decrease for a 3-year study) for a hypothesized 25%treatment effect to reduce cognitive decline. Conclusion: When used in clinical trials for cohort enrichment, this tool could accelerate development of new treatments by significantly increasing statistical power to detect differences in cognitive decline between arms. In addition, detection of future decline can help clinicians improve patient management strategies that will slow or delay symptom progression.

Background: Finding an early biomarker of Alzheimer's disease (AD) is essential to develop and implement early treatments. Much research has focused on using hippocampal volume to measure neurodegeneration in aging and Alzheimer's disease (AD). However, a new method to measure hippocampal change, hippocampal grading, has shown enhanced predictive power in older adults. It is unknown whether this method can capture hippocampal changes at each progressive stage of AD better than hippocampal volume. The goal of this study was to determine if hippocampal grading is more strongly associated with group differences between normal controls (NC), early MCI (eMCI), late (lMCI), and AD than hippocampal volume. Methods: Data from 1666 Alzheimer's Disease Neuroimaging Initiative older adults with baseline MRI scans were included in the first set of analyses (513 normal controls NC, 269 eMCI, 556 lMCI, and 328 AD). Sub-analyses were also completed using only those that were amyloid positive (N=834; 179 NC, 148 eMCI, 298 lMCI, and 209 AD). We compared seven different classification techniques to classify participants into their correct cohort using 10-fold cross-validation. The following classifiers were applied: support vector machines, decision trees, k-nearest neighbors, error-correcting output codes, binary Gaussian kernel, binary linear, and random forest. These multiple classifiers allow for comparison to other research and examination of the most suitable classifier for Scoring by Nonlocal Image Patch Estimator (SNIPE) grading, volume, and Freesurfer volume. This model was then validated in the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL). Results: SNIPE grading provided the highest classification accuracy over SNIPE volume and Freesurfer volume for all classifications in both the full sample and amyloid positive sample. When classifying NC from AD, SNIPE grading provided an accuracy of 89% for the full sample and 87% for the amyloid positive group. A much lower accuracy of 65% and 46% was obtained when using Freesurfer in the full sample and amyloid positive sample, respectively. Similar accuracies were obtained in the AIBL validation cohort for SNIPE grading (NC vs AD: 90% classification accuracy) Conclusion: These findings suggest that SNIPE grading offers increased prediction accuracy compared to both SNIPE volume and Freesurfer volume. SNIPE grading offers promise as a means to classify between people with and without AD. Future research is needed to determine the predictive power of grading at detecting conversion to MCI and AD in amyloid positive cognitively normal older adults (i.e., early in the AD continuum).