ABSTRACT
PURPOSE: To assess the correlation between profile and severity deterioration in the neuropsychological assessment and the most affected regions in amyloid PET semiquantification. The influence of vascular risk and other potential confounding factors was also evaluated. METHODS: A retrospective, observational, and multicenter study including all patients referred for amyloid PET in daily practice was conducted. Patients underwent neuropsychological assessment, and cognitive decline severity and domain(s) affected were recorded. The patients were grouped according to cognitive impairment (CI) profile and severity: (A) no CI, single-domain amnestic CI, multiple-domain amnestic CI, and nonamnestic CI; and (B) mild CI, moderate and severe dementia. An adapted Framingham Stroke Risk Profile was calculated for each individual. Depression and parkinsonism were also recorded. Standardized quantitative analysis software was used to obtain standardized uptake value ratio (SUVR) values from PET/CT images. The corresponding associations were assessed with the most appropriate statistical tests. RESULTS: One hundred twenty-nine patients were included (62 men, 67 women; 64.67 ± 7.47 years old). Significant differences in global and regional amyloid load were exclusively found in women between non-CI and moderate dementia ( P = 0.006, for total-cerebellum SUVR). Posterior and anterior cingulates and prefrontal cortex best represented CI severity ( P = 0.003, 0.006, and 0.006, respectively). No relationship between the CI profile and the regional amyloid load was shown. A significantly high positive correlation was found between age and vascular risk and between these variables and amyloid load in nearly all regions, especially in women with moderate dementia. CONCLUSION: Semiquantitative analysis of amyloid PET by SUVR values revealed a significant correlation between amyloid burden and CI severity, although only in women.
