P4‐045: EFFICIENCY OF THE RATIO OF CEREBROSPINAL FLUID Aβ42/Aβ40 CONCENTRATIONS IN DETECTING ALZHEIMER'S DISEASE: A STEP FORWARDS
| Autor: | Tamara Eisele, Panagiotis Alexopoulos |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Oncology
Apolipoprotein E medicine.medical_specialty Epidemiology business.industry Health Policy Memory clinic Disease medicine.disease Regression Psychiatry and Mental health Cellular and Molecular Neuroscience Cerebrospinal fluid Developmental Neuroscience Internal medicine Cohort medicine Biomarker (medicine) Dementia Neurology (clinical) Geriatrics and Gerontology business |
| Zdroj: | Alzheimer's & Dementia. 10 |
| ISSN: | 1552-5279 1552-5260 |
| DOI: | 10.1016/j.jalz.2014.05.1559 |
| Popis: | classify patients with only one biomarker value beyond the cut-off. Alternative classification strategies must be developed. The aim of the current study was to apply a classification and decision tree model for better classification of AD patients from controls with subjective memory complaints (SMC).Methods:We included patients from the memory clinic based Amsterdam Dementia Cohort with defined diagnosis of AD (n1⁄4631), and SMC (n1⁄4251). Classification and regression treefit using R package ’tree’ was applied for optimal discrimination of AD vs SMC, including the variables amyloid beta 1-42 (abeta) total tau (tTau), ptau181, age and ApoE status. Sensitivity and specificity of these models were compared to single and combination cut offs of an independent cohort consisting of 432 AD patients and 287 controls. Results: The optimal decision treemodel for discrimination ofADpatient fromSMC included as first node Abeta(1-42) with 669.5 pg/mL as cut-point. The second and third nodes included tTau >322.5 pg/mL or Abeta(1-42) 322.5 pg/mL classifying 555 as AD including 26 SMC/ 529 AD; 3) Abeta (1-42) levels< 487 pg/mL classifying 64 as AD including 13 SMC/51 AD. Sensitivity of the training set was 88% at a specificity of 84%which is comparablewith the sensitivity and specificity of the independent test set 89% and 83% respectively. Conclusions: Classification and regression tree fit is a powerful tool to identify specific AD subgroups with unique biomarker characteristics. Depending on Abeta level patients (also SMC) can be classified as AD in several ways, either with normal CSF tTau or even with relatively high CSF abeta(1-42) level. The unique phenotypes enable application of personalized treatment of patients. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|