A practical computerized decision support system for predicting the severity of Alzheimer’s disease of an individual

Abstract

1Computerized clinical decision support systems can help to provide objective, standardized, 2 and timely dementia diagnosis. However, current computerized systems are mainly based 3 on the group analysis, discrete classification of disease stages, or expensive and not readily 4 accessible biomarkers, while current clinical practice relies relatively heavily on cognitive and 5 functional assessments (CFA). In this study, we developed a computational framework using 6 a suite of machine learning tools for identifying key markers in predicting the severity of 7 Alzheimer's disease (AD) from a large set of biological and clinical measures. Six machine 8 learning approaches, namely Kernel Ridge Regression (KRR), Support Vector Regression 9 (SVR), and k-Nearest Neighbor (kNN reg ) for regression and Support Vector Machine (SVM), 10Random Forest (RF), and k-Nearest Neighbor (kNN class ) for classification, were used for the 11 development of predictive models. We demonstrated high predictive power of CFA. 12Predictive performance of models incorporating CFA was shown to be consistently higher 13 accuracy than those based solely on biomarker modalities. We found that KRR and SVM 14 were the best performing regression and classification methods respectively. The optimal 15 SVM performance was observed for a set of four CFA test scores (FAQ, ADAS13, MoCA, 16MMSE) with multi-class classification accuracy of 83.0%, 95%CI = (72.1%, 93.8%) while the 17 best performance of the KRR model was reported with combined CFA and MRI 18 neuroimaging data, i.e., R 2 = 0.874, 95%CI = (0.827, 0.922). Given the high predictive power 19 of CFA and their widespread use in clinical practice, we then designed a data-driven and 20 self-adaptive computerized clinical decision support system (CDSS) prototype for evaluating 21 the severity of AD of an individual on a continuous spectrum. The system implemented an 22 automated computational approach for data pre-processing, modelling, and validation and 23 used exclusively the scores of selected cognitive measures as data entries. Taken together, 24we have developed an objective and practical CDSS to aid AD diagnosis. 25Keywords: dementia; Alzheimer's disease; decision support system; machine learning; 26 diagnosis support; cognitive impairment 27 97 CDSS for evaluating the severity of AD of an individual on a continuous spectrum. In order to 98 achieve this, we first utilize a suite of machine learning techniques to extract useful 99 information from large volumes of patient data and provide a disease outcome prediction for 100 different types and combinations of AD markers. We demonstrate that CFA can reliably and 101 accurately provide prediction of AD severity. Next, we design a CDSS that incorporates an 102 automated computational approach for data pre-processing, modelling, and validation and 103 uses selected CFA scores as data input. Since our system was designed to utilize 104 information from readily available and cost-effective CFA markers, it can be easily 105 implemented in general cli...

A metabolite‐based machine learning approach to diagnose Alzheimer‐type dementia in blood: Results from the European Medical Information Framework for Alzheimer disease biomarker discovery cohort

“…In a study using the Alzheimer's Disease Neuroimaging Initiative (ADNI) data for AD prediction, XGBoost demonstrated superior results (AUC = 0.97 (0.01) when including imaging parameters (MRI and PET) as predictors and when compared to RF, Support Vector Machines, Gaussian Processes and Stochastic Gradient Boosting [16]. In another study where cognition and MRI were used as predictors, Kernel Ridge Regression was performed to R 2 = 0.87 (0.025) when cognition and MRI predictors were included [17].…”

Practical Strategies for Extreme Missing Data Imputation in Dementia Diagnosis

“…In our previous work, we have shown that CFAs are among the most predictive features for classifying AD severity [40], [41]. For the current clinical dataset, the CFAs included Addenbrooke’s Cognitive Examination (ACE-III) and the Mini-ACE [42], the Bristol Activities of Daily Living Scale [43], the Geriatric Depression Scale [44], the NPI-Q behavioral, distress and severity measurements [45], and the Zarit Caregiver Burden [46].…”

Association of the use of hearing aids with the conversion from mild cognitive impairment to dementia and progression of dementia: a longitudinal retrospective study

“…CDRSB has a total score range from 0 to 18 points, with higher scores indicating greater cognitive impairment. CDRSB has been commonly used as a reliable tool for assessing dementia severity [27,28].…”