Background and objectives: Cerebral amyloid angiopathy (CAA) is a major cause of lobar intracerebral hemorrhage (ICH) in older patients and an important contributor to cognitive decline and recurrent hemorrhagic stroke. Diagnosis of CAA is commonly based on MRI findings interpreted according to established criteria. Recently, the simplified Edinburgh criteria were proposed as a CT-based alternative to detect CAA in patients presenting with lobar ICH, potentially enabling faster diagnosis in emergency settings. The aim of this study was to evaluate the diagnostic accuracy of the simplified Edinburgh criteria in patients undergoing neurosurgical hematoma evacuation using histopathologically proven CAA as reference standard.
Methods: We conducted a retrospective case-control study including all patients admitted to Hannover Medical School between February 2013 and December 2023 with lobar ICH who underwent hematoma evacuation. During surgery, brain tissue samples were collected and processed by β-amyloid (Aβ) immunohistochemistry and Congo red staining. Samples lacking intact vessel wall architecture were excluded. CAA severity was graded according to the Vonsattel rating. Preoperative cranial CT scans were analyzed by a board-certified neuroradiologist blinded to clinical and histologic information, applying the simplified Edinburgh criteria. Interrater reliability between a neuroradiologist, neurologist, and medical student was also assessed.
Results: Overall, 84 patients were included, of whom 58 had biopsy-proven CAA. The median age was 76 years (interquartile range 72-81) in the CAA-positive group and 69 years (interquartile range 54-76.5) in the CAA-negative group; women accounted for 57% and 35%, respectively. Substantial interrater agreement was observed when applying the simplified Edinburgh criteria. However, these criteria showed limited discrimination between CAA-positive and CAA-negative patients (area under the curve 0.617; 95% CI 0.486-0.749; sensitivity 64%; specificity 58%). Logistic regression adjusted for age significantly improved discrimination (area under the curve 0.784; 95% CI 0.662-0.905). Congo red staining alone demonstrated a sensitivity of only 65% compared with immunohistochemistry.
Discussion: Validation of the simplified Edinburgh criteria in patients with surgically treated ICH revealed limited diagnostic accuracy, emphasizing the need for histopathologic confirmation when available because imaging alone may not be sufficient for reliable CAA diagnosis in this population. Limitations include the retrospective design, single-center setting, and restriction to patients with surgically treated ICH.
Classification of evidence: This study provides Class II evidence that the CT scan features of the simplified Edinburgh criteria have limited diagnostic accuracy to detect CAA in patients with lobar hematoma.