Diagnostic Accuracies of Cerebral Amyloid Angiopathy Criteria in the First Thai Pathologically Confirmed Cohort

Abstract Backgrounds The gold standard of cerebral amyloid angiopathy (CAA) diagnosis are surgical pathology or postmortem examination, which are scarcely done in clinical practice. Thus, the current diagnostic criteria are mainly based on clinic-radiological characteristics, including modified Boston criteria v1.5 [Linn J, et al. 2010] and Boston criteria v2.0 [Charidimou A, et al. 2022], which are magnetic resonance imaging (MRI)-based, as well as simplified [Sembill JA, et al. 2022] and full [Rodrigues MA, et al. 2018] Edinburg criteria, which are computerized topography (CT)-based. These criteria had been developed based on patients in western countries. Therefore, the application of these criteria in Asian population, especially in South-East Asian, is limited. This study aimed to explore the clinico-radiological characteristics and the accuracy of the current diagnostic criteria among Thai CAA patients. Methods Congo Red staining brain histopathological specimens in King Chulalongkorn Memorial Hospital, Thailand during 2011–2021 were reviewed. Patients’ characteristics were gathered from the best available data in electronic medical records. Each clinical event was analyzed separately. Radiological biomarkers in brain CT and MRI which had been done closet to the pathological confirmed date were systematically rated by trained investigators. Level of diagnosis was assessed based on the modified Boston criteria v1.5, Boston criteria v2.0, and simplified Edinburg criteria. Results Eight patients had pathological-confirmed CAA and 11 clinical events from these patients were reviewed. The clinico-radiological characteristics are shown in Table 1. Comparisons of the level of diagnosis between the criteria are shown in Table 2 and the diagnostic accuracy of the criteria using the Boston criteria v1.5 and v2.0 as reference standards are summarized in Table 3 . The Boston criteria v2.0 revealed the better sensitivity compared with v1.5 (probable CAA, 63.6% vs 54.4%), likely due to an inclusion of non-hemorrhagic biomarkers; whereas the simplified Edinburgh criteria showed lower sensitivity compared with Boston criteria. The AUC of probable CAA in Boston criteria v1.5 using v2.0 as a reference standard was 0.93 (95%CI [0.76-1.00], sensitivity 85.7%, and specificity 100%). Conclusions We performed the application of most recent current diagnostic criteria in CAA Thai population. The larger and well-designed confirmatory study in Asian cohort is needed.