Premium
[P2–422]: PREDICTING COGNITIVE TEST SCORES IN ALZHEIMER's PATIENTS USING MULTIMODAL LONGITUDINAL DATA
Author(s) -
Izquierdo Walter,
Martin Harold,
Cabrerizo Mercedes,
Barker Warren W.,
Loewenstein David A.,
Duara Ranjan,
Adjouadi Malek
Publication year - 2017
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2017.06.1078
Subject(s) - cognition , correlation , neuroimaging , alzheimer's disease neuroimaging initiative , psychology , cognitive test , medicine , cognitive psychology , audiology , cognitive impairment , mathematics , psychiatry , geometry
Alzheimer’s disease (ADAD).Methods:The search was performed on the Medline and Embase and Cochrane libraries on November 28, 2015. Critical outcomes were sensitivity and specificity and accuracy of FDG-PET in distinguishing asymptomatic ADAD from healthy controls, as assessed versus gold standard (Presenilin-1, Presenilin-2, Amyloid Precursor Protein mutation status). Minimum sample size was set a priori at 5 subjects with the target condition. Data were extracted and assessed with regard to publication bias, heterogeneity, index test (FDG-PET) imprecision, risk of bias, indirectness, and applicability. We provide assessment of the quality of evidence, of the target effect, and of strength of recommendation. Results:Out of the 13 papers obtained, only 2 papers reported the critical outcomes. These included 13 asymptomatic ADAD due to Presenilin-1 mutation and 30 non-carriers. Data provided a high level of evidence that the posterior cingulate cortex hypometabolism assessed semi-quantitatively could discriminate ADAD from controls with high sensitivity and specificity, although with large confidence intervals (Table). Strong concerns regarded the applicability of the index test, as semiquantitative methods of image analysis are increasingly but not yet routinely adopted in the clinical context, and a possible risks of bias in patients selection. Significant hypometabolism in the precuneus could be detected in mutation carriers 10 years before expected symptom onset and at the age of onset, but inconsistently across studies. Conclusions: The available data denote high quality of evidence for high diagnostic utility of semiquantitatively assessed FDG-PET in detecting early metabolic change in presymptomatic ADAD, carriers of mutation in PSEN1. However, the very low number of studies and subjects, the large confidence intervals of results, and the need of replication support only a weak recommendation for the use of FDG-PET to detect early neurodegeneration in asymptomatic subjects with familial forms of AD, and further research is warranted on this topic.