We built and validated a deep learning algorithm predicting the individual diagnosis of Alzheimer's disease (AD) and mild cognitive impairment who will convert to AD (c-MCI) based on a single cross-sectional brain structural MRI scan. Convolutional neural networks (CNNs) were applied on 3D T1-weighted images from ADNI and subjects recruited at our Institute (407 healthy controls [HC], 418 AD, 280 c-MCI, 533 stable MCI [s-MCI]). CNN performance was tested in distinguishing AD, c-MCI and s-MCI. High levels of accuracy were achieved in all the classifications, with the highest rates achieved in the AD vs HC classification tests using both the ADNI dataset only (99%) and the combined ADNI + non-ADNI dataset (98%). CNNs discriminated c-MCI from s-MCI patients with an accuracy up to 75% and no difference between ADNI and non-ADNI images. CNNs provide a powerful tool for the automatic individual patient diagnosis along the AD continuum. Our method performed well without any prior feature engineering and regardless the variability of imaging protocols and scanners, demonstrating that it is exploitable by not-trained operators and likely to be generalizable to unseen patient data. CNNs may accelerate the adoption of structural MRI in routine practice to help assessment and management of patients.

Alzheimer's disease (AD) is characterized by extracellular amyloid-b (Ab) deposition, which activates microglia, induces neuroinflammation and drives neurodegeneration. Recent evidence indicates that soluble pre-fibrillar Ab species, rather than insoluble fibrils, are the most toxic forms of Ab. Preventing soluble Ab formation represents, therefore, a major goal in AD. We investigated whether microvesicles (MVs) released extracellularly by reactive microglia may contribute to AD degeneration. We found that production of myeloid MVs, likely of microglial origin, is strikingly high in AD patients and in subjects with mild cognitive impairment and that AD MVs are toxic for cultured neurons. The mechanism responsible for MV neurotoxicity was defined in vitro using MVs produced by primary microglia. We demonstrated that neurotoxicity of MVs results from (i) the capability of MV lipids to promote formation of soluble Ab species from extracellular insoluble aggregates and (ii) from the presence of neurotoxic Ab forms trafficked to MVs after Ab internalization into microglia. MV neurotoxicity was neutralized by the Ab-interacting protein PrP and anti-Ab antibodies, which prevented binding to neurons of neurotoxic soluble Ab species. This study identifies microglia-derived MVs as a novel mechanism by which microglia participate in AD degeneration, and suggest new therapeutic strategies for the treatment of the disease. Cell Death and Differentiation (2014) 21, 582-593; doi:10.1038/cdd.2013.180; published online 13 December 2013Alzheimer's disease (AD) is the major cause of dementia in humans. Neuronal loss and cognitive decline occurring in AD patients are traditionally linked to the accumulation in the brain of extracellular plaques consisting of short amyloid-b (Ab) peptides of 39-42 amino acids, generated by amyloidogenic cleavage of the amyloid precursor protein. 1 Among Ab peptides, Ab 1-42 and pyroglutamate-modified Ab very rapidly aggregate and initiate the complex multistep process that leads to mature fibrils and plaque. 2,3 Although association of amyloid plaques with AD has long been assumed, Ab load does not correlate with neuronal loss 4,5 and high plaque burden does not necessarily lead to dementia in humans. 6,7 Accordingly, recent evidence clearly showed that the amyloid load reaches a plateau early after the onset of clinical symptoms in AD patients 8 and does not substantially increase in size during clinical progression. 9 These observations agree with the current view that small, soluble pre-fibrillar Ab species, rather than plaques formed by insoluble Ab fibrils, are the most toxic forms of Ab. 10 These cause synaptic dysfunction and spine loss, and correlate most closely with the severity of human AD. 5,8,11 Recent biochemical studies indicated that natural sphingolipids and gangliosides, whose metabolism has been shown to be altered in AD patients, 12 destabilize and rapidly resolubilize long Ab fibrils to neurotoxic species. 13 These studies also showed that phospholipids stabilize toxic oligomers...

Diagnostic accuracy in FDG-PET imaging highly depends on the operating procedures. In this clinical study on dementia, we compared the diagnostic accuracy at a single-subject level of a) Clinical Scenarios, b) Standard FDG Images and c) Statistical Parametrical (SPM) Maps generated via a new optimized SPM procedure. We evaluated the added value of FDG-PET, either Standard FDG Images or SPM Maps, to Clinical Scenarios. In 88 patients with neurodegenerative diseases (Alzheimer's Disease—AD, Frontotemporal Lobar Degeneration—FTLD, Dementia with Lewy bodies—DLB and Mild Cognitive Impairment—MCI), 9 neuroimaging experts made a forced diagnostic decision on the basis of the evaluation of the three types of information. There was also the possibility of a decision of normality on the FDG-PET images. The clinical diagnosis confirmed at a long-term follow-up was used as the gold standard. SPM Maps showed higher sensitivity and specificity (96% and 84%), and better diagnostic positive (6.8) and negative (0.05) likelihood ratios compared to Clinical Scenarios and Standard FDG Images. SPM Maps increased diagnostic accuracy for differential diagnosis (AD vs. FTD; beta 1.414, p = 0.019). The AUC of the ROC curve was 0.67 for SPM Maps, 0.57 for Clinical Scenarios and 0.50 for Standard FDG Images. In the MCI group, SPM Maps showed the highest predictive prognostic value (mean LOC = 2.46), by identifying either normal brain metabolism (exclusionary role) or hypometabolic patterns typical of different neurodegenerative conditions.

White matter (WM) tract damage was assessed in patients with the behavioral variant frontotemporal dementia (bvFTD) and the 3 primary progressive aphasia (PPA) variants and compared with the corresponding brain atrophy patterns. Thirteen bvFTD and 20 PPA patients were studied. Tract-based spatial statistics and voxel-based morphometry were used. Patients with bvFTD showed widespread diffusion tensor magnetic resonance imaging (DT MRI) abnormalities affecting most of the WM bilaterally. In PPA patients, WM damage was more focal and varied across the 3 syndromes: left frontotemporoparietal in nonfluent, left frontotemporal in semantic, and left frontoparietal in logopenic patients. In each syndrome, DT MRI changes extended beyond the topography of gray matter loss. Left uncinate damage was the best predictor of frontotemporal lobar degeneration diagnosis versus controls. DT MRI measures of the anterior corpus callosum and left superior longitudinal fasciculus differentiated bvFTD from nonfluent cases. The best predictors of semantic PPA compared with both bvFTD and nonfluent cases were diffusivity abnormalities of the left uncinate and inferior longitudinal fasciculus. This study provides insights into the similarities and differences of WM damage in bvFTD and PPA variants. DT MRI metrics hold promise to serve as early markers of WM integrity loss that only at a later stage may be detectable by volumetric measures.

No abstract