Posterior-to-anterior periventricular white matter hyperintensity area ratio: a supportive MRI feature for differentiating cerebral amyloid angiopathy in lobar intracerebral hemorrhage

Diagnosing cerebral amyloid angiopathy (CAA) in primary lobar intracerebral hemorrhage (ICH) remains challenging using routine MRI markers. We investigated a quantitative posterior-dominant pattern of periventricular white matter hyperintensity (PWMH) as a potential supportive imaging feature for CAA within the Boston v2.0 framework. We retrospectively analyzed 229 patients aged ≥ 50 years with primary lobar ICH, classified as CAA (n = 123) or CAA-negative (n = 106) according to the Boston criteria v2.0. On FLAIR, the largest anterior and posterior periventricular WMH areas were measured bilaterally, and the posterior-to-anterior PWMH area ratio (PA-AR) was derived. Multivariable logistic regression models were developed, and diagnostic performance was evaluated using ROC analysis. Patients with CAA showed a greater posterior PWMH burden and higher PA-AR than CAA-negative controls (median 1.69 vs. 0.79; p < 0.001). Adding PA-AR to the covariate-adjusted model improved discrimination for CAA (AUC 0.804; bootstrap mean AUC 0.813, 95% CI 0.759–0.867). In the predefined diagnostic “gray zone” subgroup, PA-AR remained informative. The AUC was 0.887 in univariable analysis and 0.888 after age adjustment. PA-AR was also higher in patients with posterior lobar CMBs than in those with no CMBs or mixed-location hemorrhage. A posterior-dominant periventricular WMH pattern, quantified by PA-AR, is associated with CAA in lobar ICH. This pattern is also related to posterior-predominant lobar CMB topography. PA-AR may serve as a supportive MRI feature that complements Boston criteria v2.0 markers for identifying CAA within a real-world mixed lobar ICH population, particularly in diagnostically challenging patients.