Diagnostic Value of the Boston criteria v2.0 for Cerebral Amyloid Angiopathy in Individuals Without Hemorrhage: An MRI-neuropathological Validation Study
Aaron Switzer1, Antreas Charidimou2, Stuart McCarter1, Scott Przybelski1, Timothy Lesnick1, Prashanthi Vemuri1, Ross Reichard1, Alejandro Rabinstein1, David Knopman1, Ronald Petersen1, Jonathan Graff-Radford1
1Behavioral neurology, Mayo Clinic, 2Neurology, Massachusetts General Hospital
Objective:
To assess the diagnostic performance of the Boston criteria version 2.0 for cerebral amyloid angiopathy (CAA) diagnosis in a cohort of individuals without symptomatic intracerebral hemorrhage or CAA-related transient focal neurological episodes (TFNE) and who had pathological assessment for CAA.
Background:
The Boston criteria were recently updated to incorporate non-hemorrhagic MRI biomarkers of CAA and small vessel disease. The updated criteria (v2.0) demonstrated improved diagnostic yield when compared to prior versions of the criteria (v1.5) among patients presenting with symptomatic CAA-related syndromes. However, these criteria have not been validated in any cohorts without symptomatic CAA.
Design/Methods:
Fifty-four participants with a median age of 75, followed in large population-based or observational studies, were included if they had an antemortem MRI with gradient-recall echo sequences and a subsequent brain autopsy with CAA evaluation. Performance of the Boston criteria v2.0 was compared to v1.5 using pathologically verified CAA as the reference standard. Performance measures were calculated using a logistic regression. Paired AUC were compared using the pROC package in R.
Results:
Overall, 28/54 participants had pathologically confirmed CAA (i.e. moderate-to-severe CAA in at least 1 cortical region). The sensitivity and specificity of the Boston criteria v2.0 were 78.6% (59.1-91.7) and 34.6% (17.2-55.7) for any CAA diagnosis (possible + probable; AUC: 0.57) and 28.6% (13.2-48.7) and 69.2% (48.2-85.7) for probable CAA diagnosis (AUC 0.49), respectively. The v2.0 Boston criteria had superior accuracy to the prior v1.5 criteria for any CAA diagnosis (AUC 0.57 [0.45-0.69] vs. AUC 0.43 [0.30-0.56], Z= 2.10, p= 0.03).
Conclusions:
The Boston criteria v2.0 have limited accuracy for CAA diagnosis in an asymptomatic population. Additional biomarkers may need to be explored in this setting to optimize CAA diagnostic performance.