Objectives: To determine and compare the accuracy of the Boston criteria for the diagnosis of cerebral amyloid angiopathy (CAA) applied to: 1) a large hospital-based cohort; 2) a cohort of community-dwelling individuals. Methods: Among patients seen at a single academic medical center and participants enrolled in the Framingham Heart Study (FHS) we identified all individuals having had brain MRI (including hemosiderin-specific sequences) and pathological assessment of CAA at age ≥55. CAA was defined as Vonsattel degree ≥2, except in non-autopsy studies, where any vascular amloyid deposition was considered diagnostic of CAA. We excluded cases with: 1) non-lobar ICH; 2) alternative clinical diagnosis for the lobar hemorrhage/es; 3) CAA-negative studies based on small brain biopsy samples (greater diameter <1 cm) or clot evacuations with none or rare vessels identified; 4) autopsy studies not grading CAA. We determined sensitivity and specificity of “possible CAA” (1 lobar hemorrhage) and “probable CAA” (>1 strictly lobar hemorrhage) in both cohorts. Results: The study included 143 cases: 96 from the hospital cohort (mean age at time of MRI 74.1±8.2 years, 47.9% women, 47.9% autopsy studies, 54.2% cases with any lobar microbleed, 44.8% with any lobar ICH) and 47 from FHS (mean age at time of MRI 83.4±10.9 years, 48.9% women, 100% autopsy studies, 17% cases with any lobar microbleed, no lobar ICH cases). Hospital cohort: CAA prevalence was 72.6%. Specificity/sensitivity of “possible CAA” and “probable CAA” were 78.2%/33.3% and 92%/57.7%. Community cohort: CAA prevalence was 47%. Specificity/sensitivity of “probable CAA” were 88%/3%. Discussion: Specificity of detection of multiple lobar hemorrhages on MRI (“probable CAA”) for moderate-severe CAA is similarly high in both hospital and community cohorts. This increases the opportunity to identify individuals with CAA for observational studies and emerging anti-amyloid therapeutic trials. In view of its poor sensitivity, “probable CAA” diagnosis represents an inadequate screening tool for exclusion of CAA in both populations.