%0 Journal Article %J J Alzheimers Dis %D 2018 %T A Clinically-Translatable Machine Learning Algorithm for the Prediction of Alzheimer's Disease Conversion in Individuals with Mild and Premild Cognitive Impairment. %A Grassi, Massimiliano %A Perna, Giampaolo %A Caldirola, Daniela %A Schruers, Koen %A Duara, Ranjan %A Loewenstein, David A %K Aged %K Aged, 80 and over %K Algorithms %K Alzheimer Disease %K Area Under Curve %K Atrophy %K Brain %K Cognitive Dysfunction %K Disease Progression %K Female %K Humans %K Logistic Models %K Machine Learning %K Magnetic Resonance Imaging %K Male %K Neuropsychological Tests %K Predictive Value of Tests %K Prognosis %K Regression Analysis %K Sensitivity and Specificity %K Support Vector Machine %X

BACKGROUND: Available therapies for Alzheimer's disease (AD) can only alleviate and delay the advance of symptoms, with the greatest impact eventually achieved when provided at an early stage. Thus, early identification of which subjects at high risk, e.g., with MCI, will later develop AD is of key importance. Currently available machine learning algorithms achieve only limited predictive accuracy or they are based on expensive and hard-to-collect information.

OBJECTIVE: The current study aims to develop an algorithm for a 3-year prediction of conversion to AD in MCI and PreMCI subjects based only on non-invasively and effectively collectable predictors.

METHODS: A dataset of 123 MCI/PreMCI subjects was used to train different machine learning techniques. Baseline information regarding sociodemographic characteristics, clinical and neuropsychological test scores, cardiovascular risk indexes, and a visual rating scale for brain atrophy was used to extract 36 predictors. Leave-pair-out-cross-validation was employed as validation strategy and a recursive feature elimination procedure was applied to identify a relevant subset of predictors.

RESULTS: 16 predictors were selected from all domains excluding sociodemographic information. The best model resulted a support vector machine with radial-basis function kernel (whole sample: AUC = 0.962, best balanced accuracy = 0.913; MCI sub-group alone: AUC = 0.914, best balanced accuracy = 0.874).

CONCLUSIONS: Our algorithm shows very high cross-validated performances that outperform the vast majority of the currently available algorithms, and all those which use only non-invasive and effectively assessable predictors. Further testing and optimization in independent samples will warrant its application in both clinical practice and clinical trials.

%B J Alzheimers Dis %V 61 %P 1555-1573 %8 2018 %G eng %N 4 %1 http://www.ncbi.nlm.nih.gov/pubmed/29355115?dopt=Abstract %R 10.3233/JAD-170547 %0 Journal Article %J J Alzheimers Dis %D 2018 %T Comparison between FCSRT and LASSI-L to Detect Early Stage Alzheimer's Disease. %A Matías-Guiu, Jordi A %A Cabrera-Martín, María Nieves %A Curiel, Rosie E %A Valles-Salgado, María %A Rognoni, Teresa %A Moreno-Ramos, Teresa %A Carreras, José Luis %A Loewenstein, David A %A Matías-Guiu, Jorge %K Aged %K Aged, 80 and over %K Alzheimer Disease %K Cues %K Female %K Fluorodeoxyglucose F18 %K Humans %K Male %K Memory Disorders %K Mental Recall %K Middle Aged %K Neuropsychological Tests %K Positron-Emission Tomography %K Psychiatric Status Rating Scales %K ROC Curve %K Semantics %X

BACKGROUND: The Free and Cued Selective Reminding Test (FCSRT) is the most accurate test for the diagnosis of prodromal Alzheimer's disease (AD). Recently, a novel cognitive test, the Loewenstein-Acevedo Scale for Semantic Interference and Learning (LASSI-L), has been developed in order to provide an early diagnosis.

OBJECTIVE: To compare the diagnostic accuracy of the FCSRT and the LASSI-L for the diagnosis of AD in its preclinical and prodromal stages using 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) as a reference.

METHODS: Fifty patients consulting for subjective memory complaints without functional impairment and at risk for AD were enrolled and evaluated using FCSRT, LASSI-L, and FDG-PET. Participants were evaluated using a comprehensive neurological and neuropsychological protocol and were assessed with the FCSRT and LASSI-L. FDG-PET was acquired concomitantly and used for classification of patients as AD or non-AD according to brain metabolism using both visual and semi-quantitative methods.

RESULTS: LASSI-L scores allowed a better classification of patients as AD/non-AD in comparison to FCSRT. Logistic regression analysis showed delayed recall and failure to recovery from proactive semantic interference from LASSI-L as independent statistically significant predictors, obtaining an area under the curve of 0.894. This area under the curve provided a better discrimination than the best FCSRT score (total delayed recall, area under the curve 0.708, p = 0.029).

CONCLUSIONS: The LASSI-L, a cognitive stress test, was superior to FCSRT in the prediction of AD features on FDG-PET. This emphasizes the possibility to advance toward an earlier diagnosis of AD from a clinical perspective.

%B J Alzheimers Dis %V 61 %P 103-111 %8 2018 %G eng %N 1 %1 http://www.ncbi.nlm.nih.gov/pubmed/29125488?dopt=Abstract %R 10.3233/JAD-170604