Premium
P2‐274: Reliability of Three Diffusion Tensor Imaging Processing Programs
Author(s) -
Yang Dong Won,
Lim Eun Ye,
Nae Seunghee,
Shim YongSoo,
Yoon Bora,
Hong Yun Jeong
Publication year - 2016
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.2016.06.1534
Subject(s) - splenium , corpus callosum , diffusion mri , fractional anisotropy , internal capsule , gyrus , psychology , medicine , neuroscience , white matter , radiology , magnetic resonance imaging
clinician diagnostic bias, reader bias, or other factors. This study systematically evaluated the factors associated with negative amyloid PET in a series of multicenter AD therapeutic trials. Methods: 5,357 F18 Florbetapir PET scans were obtained at screening in preclinical, prodromal, and early Alzheimer’s disease participants recruited in six Phase 2-3 therapeutic trials. Each scan was visually interpreted according to the florbetapir visual read method and required agreement by either two expert independent readers or a reader and an independent composite standard uptake value ratio (SUVr) depending on the trial. A series of candidate clinical, technical, and methodological parameters were identified and analyzed in a discriminate function model for determining those factors most strongly associated with a negative florbetapir PET scan. Results: The most important individual factors associated with a negative amyloid PET scan were: cohort status with at risk>prodromal> early AD, enrolling site clinician, and scan interpretation methodology (visual vs visual + quantitation). Conclusions: The clinical stage of the cohort is the most important factor predicting the number negative scans. While there is high variability between rates of negative scans amongst recruiting sites, there is no correlation between total screening scans performed and the rate of negative studies. The influence of individual readers is largely minimized by consistent training and checks and balances in the read algorithms.