Abstract
190
Objectives We previously described an FDG-PET based dementia prognostic index that quantifies cerebral metabolism associated with autopsy-confirmed AD (JAMA 2001; 286:2120-7) and predicts cognitive decline in less impaired subjects. Clark et al described a similar index applicable to florbetapir-PET that predicts autopsy-confirmed frequency of cerebral amyloid plaques in dementia (JAMA 2011; 305:275-83). Both indices are automatically generated by FDA-cleared software. Their relative prognostic values among clinical subpopulations remain to be defined.
Methods Baseline FDG and florbetapir PET scans of ADNI subjects with MCI and ≥2 years of subsequent follow-up were examined (n=131). Scans were analyzed by standardized volume of interest (sVOI) methods, and indices reflect ratios of summed activities in sVOI’s known to be affected by AD to summed activities in unaffected sVOI’s. Rate of cognitive decline was calculated by linear regression of Mini-Mental State Exam (MMSE) scores over 2yrs. Subjects were stratified by baseline MMSE into minimally (MMSE 30, n=36) and mildly (23-29, n=95) impaired.
Results Across all subjects, rate of decline was significantly predicted by the indices for FDG (r=0.37, p=0.00001) and florbetapir (r=-0.19, p=0.03). Cognitive decline in both the minimally impaired and mildly impaired subpopulations were significantly predicted by the FDG index (r=0.36, r=0.43, respectively), while predictive value of the florbetapir index was significant for the minimally impaired (r= -0.55) but not the mildly impaired (r=-0.16) subpopulations. Predictive value of the indices for slightly impaired subjects (MMSE 28-29, n=61) was also significant for FDG (r=0.41) but not for florbetapir (r=-0.09).
Conclusions Prognostic indices automatically derived from PET scans using clinically available software may be used for predicting with FDG the cognitive changes occurring across the spectrum of severity of subjects with amnestic MCI, and for predicting with florbetapir the changes in the most minimally impaired MCI subjects.