Amyloid-PET in sporadic cerebral amyloid angiopathy A diagnostic accuracy meta-analysis

Objective: To perform a meta-analysis synthesizing evidence of the value and accuracy of amyloid-PET in diagnosing patients with sporadic cerebral amyloid angiopathy (CAA). Methods: In a PubMed systematic literature search, we identified all case-control studies with extractable data relevant for the sensitivity and specificity of amyloid-PET positivity in symptomatic patients with CAA (cases) vs healthy participants or patients with spontaneous deep intracerebral hemorrhage (ICH) (control groups). Using a hierarchical (multilevel) logistic regression model, we calculated pooled diagnostic test accuracy. Results: Seven studies, including 106 patients with CAA (>90% with probable CAA) and 151 controls, were eligible and included in the meta-analysis. The studies were of moderate to high quality and varied in several methodological aspects, including definition of PET-positive and PET-negative cases and relevant cutoffs. The sensitivity of amyloid-PET for CAA diagnosis ranged from 60% to 91% and the specificity from 56% to 90%. The overall pooled sensitivity was 79% (95% confidence interval [CI] 62-89) and specificity was 78% (95% CI 67-86) for CAA diagnosis. A predefined subgroup analysis of studies restricted to symptomatic patients presenting with lobar ICH CAA (n = 58 vs 86 controls) resulted in 79% sensitivity (95% CI 61-90%) and 84% specificity (95% CI 65-93%). In prespecified bivariate diagnostic accuracy meta-analysis of 2 studies using F-18-florbetapir-PET, the sensitivity for CAA-ICH diagnosis was 90% (95% CI 76-100%) and specificity was 88% (95% CI 74-100%). Conclusions: Amyloid-PET appears to have moderate to good diagnostic accuracy in differentiating patients with probable CAA from cognitively normal healthy controls or patients with deep ICH. Given that amyloid-PET labels both cerebrovascular and parenchymal amyloid, a negative scan might be useful to rule out CAA in the appropriate clinical setting.