Simplified Edinburgh CT Criteria for Identification of Lobar Intracerebral Hemorrhage Associated With Cerebral Amyloid Angiopathy

Abstract

Background and objectives: In patients with lobar intracerebral hemorrhage (ICH), etiologic characterization represents a tradeoff between feasibility, resource allocation, and diagnostic certainty. This study investigated the accuracy and clinical utility of the simplified Edinburgh CT criteria to identify underlying cerebral amyloid angiopathy (CAA).

Methods: This external validation analyzed 210 consecutive patients with lobar ICH and available CT and MRI studies from a prospective single-center observational cohort study (2006-2015, Longitudinal Cohort Study on ICH Care [UKER-ICH,] NCT03183167). We investigated the interrater variability and diagnostic accuracy of the simplified Edinburgh CT-based criteria for identification of ICH associated with probable CAA according to MRI-based modified Boston criteria as a reference standard. We evaluated the utility of the simplified Edinburgh criteria by decision curve analysis, comparing the theoretical clinical net benefit (weighted benefit-harm at varying threshold probabilities) of the high-risk category (finger-like projections and subarachnoid hemorrhage) for ruling in and the low-risk category (neither finger-like projections nor subarachnoid hemorrhage) for ruling out with the assumptions of no or all patients having CAA (default strategies).

Results: Of 210 patients, 70 (33.3%) had high risk, 67 (31.9%) had medium risk, and 73 (34.8%) had low risk for CAA-associated ICH according to simplified Edinburgh CT criteria, showing moderate interrater variability. Discrimination was good (area under the receiver operating characteristics curve 0.74, 95% CI 0.67-0.81) without evidence of poor calibration (Hosmer-Lemeshow, p = 0.54) for validation of MRI-based diagnosis of probable CAA (n = 94 of 210, 44.8%). The rule-in criteria (high risk), had 87.1% (79.3%-92.3%) specificity, and the rule-out criteria (low risk), had 80.9% (71.1%-88.0%) sensitivity. Decision curve analysis suggested a theoretical clinical net benefit for ruling in but not for ruling out probable CAA compared to default strategies.

Discussion: Applying the simplified Edinburgh CT criteria during diagnostic workup seems clinically useful and may accurately identify CAA in patients with lobar ICH.

Trial registration information: ClinicalTrials.gov Identifier: NCT03183167.

Classification of evidence: This study provides Class II evidence that in patients with lobar hemorrhages, the simplified Edinburgh criteria accurately identify those at high risk of CAA.