Premium
IC‐P‐027: VALUE OF AMYLOID IMAGING FOR PREDICTING CONVERSION TO DEMENTIA IN MCI SUBJECTS WITH INITIALLY INDETERMINATE FDG‐PET SCANS
Author(s) -
Silverman Daniel,
Torosyan Nare,
Manne Manogna,
Durcanova Beata,
Dahlbom Magnus,
Apostolova Liana
Publication year - 2014
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.2014.05.031
Subject(s) - dementia , neuroimaging , positron emission tomography , medicine , pet imaging , cognitive impairment , nuclear medicine , cognition , alzheimer's disease , amyloid (mycology) , clinical dementia rating , psychology , oncology , disease , pathology , psychiatry
scans. These classifications were used to initialize an iterative, voxelwise, regularized discriminant analysis to optimize the selection of a set of regions-of-interest (ROIs) and a threshold for automated classification of Ab+ and Absubjects in each cohort. The resulting ROIs and thresholds were cross-validated between cohorts. Results: The automated classifier demonstrated high accuracy, specificity, and sensitivity. The optimal ROI included areas of the default mode network (DMN), such as the precuneus and medial frontal cortex. Conclusions: We have validated an automated approach for classification of subjects as Ab+ or Abusing three different subject cohorts. This strategy represents an improvement over current methods that rely on either subjective, visual evaluation of amyloid PET images or quantitative measures based on pre-defined anatomical ROIs. Ultimately, fully-automated subject classification may complement or supplant visual reads in clinical practice and for eligibility assessment in AD clinical trials.