Photo Credit: Artur Plawgo

The following is a summary of “Validation of the Boston Criteria Version 2.0 for Cerebral Amyloid Angiopathy in Patients Presenting With Intracerebral Hemorrhage,” published in the March 2025 issue of Neurology by Downes et al. 

Cerebral amyloid angiopathy (CAA) causes lobar intracerebral hemorrhage (ICH) in older adults with high morbidity. The Boston criteria v2.0 add MRI biomarkers for better diagnosis. 

Researchers conducted a retrospective study to compare the diagnostic performance of Boston criteria v2.0 and v1.5 in patients with spontaneous ICH undergoing surgery and brain biopsy.  

They conducted a retrospective single-center cohort study at Mount Sinai Health System (2015–2021) on patients with spontaneous ICH undergoing surgical evacuation and brain biopsy. They included those with preoperative MRI assessing lobar hemorrhagic lesions (ICH, cerebral microbleeds [CMBs], cortical siderosis [cSS]) and nonhemorrhagic markers (severe CSO-PVS, multispot white matter hyperintensities [WMHs]). They confirmed CAA using modified Vonsattel grading and compared v2.0 with v1.5 for sensitivity, specificity, and predictive values. They used logistic regression to calculate odds ratios (ORs) and 95% CIs for MRI biomarker associations with CAA. 

The results showed that among 186 patients (median age 63; 38% female), 24% had confirmed CAA. The Boston criteria v2.0 had higher sensitivity for probable CAA (0.75 vs 0.57) with similar specificity (0.96 vs 0.99). For possible CAA, sensitivity was slightly higher (0.82 vs 0.77) with comparable specificity (0.84 vs 0.87). cSS (OR 4.14, 95% CI 1.35–13.00, P = 0.013) and lobar CMBs (OR 3.03, 95% CI 1.31–7.10, P = 0.009) were significantly associated with CAA. CSO-PVS showed strong association (OR 5.49, 95% CI 2.37–13.06, P < 0.001), while multispot WMHs did not (OR 1.10, 95% CI 0.45–2.56, P = 0.834). 

Investigators found that Boston criteria v2.0 improved sensitivity for probable CAA without compromising specificity, mainly due to nonhemorrhagic markers like CSO-PVS. They acknowledged limitations, including the retrospective design, lack of inter-rater reliability measures, and modest sample size. 

Source: neurology.org/doi/10.1212/WNL.0000000000213460