A comprehensive framework for cerebral amyloid angiopathy diagnosis: Diagnostic accuracy meta-analyses of different biomarkers
Antreas Charidimou1, Gregoire Boulouis2
1Neurology Department, Boston Medical Centre, 2• Diagnostic and Interventional Neuroradiology Department, University Hospital of Tours
Objective:

We synthesized evidence on the accuracy of different diagnostic approaches for cerebral amyloid angiopathy (CAA), including the clinical-MRI based Boston criteria, amyloid-PET and core CSF biomarkers.

Background:

An accurate non-invasive diagnosis of CAA, an age-related small vessel disease, characterised by progressive deposition of amyloid β in the cerebrovascular wall, is important for clinical decision-making, as well as research in the field. Yet, the diagnostic accuracy and validity of key MRI markers and molecular markers of the disease remain (paradoxically) poorly studied.

Design/Methods:

In a systematic literature search, we identified case-control studies with data relevant for sensitivity/specificity of the Boston criteria, amyloid PET positivity and core CSF biomarkers in symptomatic CAA patients vs. different comparison groups, as applicable. Using a hierarchical (multilevel) logistic regression model we calculated pooled diagnostic test accuracy.

Results:

Five studies had data on the Boston criteria v.1.5 (n=123): the pooled sensitivity and specificity for probable CAA diagnosis was 73.1% (95% CI, 45%-90.1%) and 86% (95% CI, 41.4%-98.1%), respectively. Thirteen amyloid-PET studies (211 CAA, 181 controls) were included in the meta-analysis. The overall pooled sensitivity of amyloid-PET for CAA diagnosis was 80% (95%CI: 67%-84%) and specificity was 82% (95%CI: 71%-86%). Seven studies (153 CAA, 185 Alzheimer’s disease patients), provided data on core CSF biomarkers (Aβ40, Aβ42, t-tau, and p-tau). Aβ40 demonstrated the best overall performance for CAA diagnosis vs Alzheimer’s disease with 80% (95%CI: 68%-83%) pooled sensitivity and 68% (95%CI: 59%-74%) specificity. The core CSF pattern characteristic of CAA was low Aβ40, with intermediate t-tau/p-tau levels (higher than healthy controls, but lower than Alzheimer’s disease) (p < 0.0001 for all comparisons).

Conclusions:

Amyloid-PET and specific CSF patterns of Aβ40, t-tau, and p-tau have moderate-to-good diagnostic accuracy for CAA diagnosis and might supplement the Boston criteria in certain clinical settings (such an approach will be presented at the meeting).

10.1212/WNL.0000000000202673