Neurology. 2026 Jul 28;107(2):e218274. doi: 10.1212/WNL.0000000000218274. Epub 2026 Jul 1.
ABSTRACT
BACKGROUND AND OBJECTIVES: Cerebral amyloid angiopathy (CAA) is a small vessel disease characterized by hemorrhagic and nonhemorrhagic brain injuries. While cortical cerebral microinfarcts (CMI) have been associated with cortical superficial siderosis (cSS), the mechanisms linking these lesions remain unknown. This study aims to determine the spatiotemporal relationship between cortical CMI and cSS in CAA, with the hypothesis that cSS can promote local ischemic injury.
METHODS: Patients with probable CAA (Boston Criteria v2.0) without previous symptomatic intracerebral hemorrhage or CAA-related inflammation were recruited from a prospective memory clinic-based cohort at Massachusetts General Hospital. CAA-related radiologic manifestations-including cortical CMI and cSS-were quantified at baseline and at the 24-month follow-up on 3-Tesla MRI. Multivariable logistic regression and generalized linear models were applied to assess associations between baseline radiologic features and radiologic progression. We also compared the local density of cortical CMIs in regions with and without overlaid cSS.
RESULTS: Among 74 patients with probable CAA (age: 73.0 [67.9-78.2] years; 38% female), 26 (35%) had cortical CMIs (n = 135, 39% underlying cSS). Compared with CMI-negative CAA, those with CMI had a higher cSS prevalence (73% vs 40%; p = 0.006), cSS volume (1.3 [0.1-8.3] vs 0 [0-0.5] mL; p < 0.001), and lobar lacunes count (50% vs 19%; p = 0.005). At baseline, cortical CMI count demonstrated a nonlinear association with cSS volume (Exp(B) = 1.12 [95% CI 1.07-1.18]; p < 0.001) and with lobar lacunes count (Exp(B) = 1.65 [95% CI 1.52-1.80]; p < 0.001). The CMI density was markedly higher in cSS-covered vs non-cSS regions (median: 20.84 [0.00-56.95] vs 0.49 [0.11-0.89]n° cortical CMI/1,000 cm2; p = 0.010). Follow-up scans (n = 36) revealed 23 incident cortical CMIs (43% underlying cSS). Baseline cSS volume was independently associated with incident CMI (OR:1.41 [95% CI 1.02-1.95]; p = 0.036), whereas baseline CMI burden was not associated with cSS progression.
DISCUSSION: Our findings demonstrate a close interplay between cortical CMI and cSS, with the spatiotemporal relationship supporting the hypothesis that convexity subarachnoid hemorrhage, and potentially resultant cSS, may trigger local ischemic injury. Given the sample size of the longitudinal cohort, we consider these data hypothesis-generating. Further studies should confirm this association and identify the driving pathophysiology, which may represent a new therapeutic target for preventing cSS-associated cortical injury.
PMID:42385122 | DOI:10.1212/WNL.0000000000218274