Abstract P36: Lesional Perfusion and Permeability Are Diagnostic and Prognostic Biomarkers of Cavernous Angioma With Symptomatic Hemorrhage

Abstract

Introduction: Diagnosis of cavernous angioma with symptomatic hemorrhage (CASH) requires MRI evidence of lesional bleeding associated directly with attributable symptoms. However, hemorrhagic signs of CASH may become clinically silent on conventional MRI after 3 months. As CASH is likely to rebleed for several years, accurate diagnosis of CASH that bled more than 3 months prior is needed. Hypothesis: Perfusion and permeability derivations of dynamic contrast-enhanced quantitative perfusion (DCEQP) MRI can diagnose CASH and predict bleeding/growth in CAs. Methods: CAs of 205 consecutively enrolled patients scanned with DCEQP during clinical visits were classified as CASH that bled 3 - 12 months prior (N = 55) versus non-CASH (N = 658) or CA with (N = 23) versus without (N = 721) bleeding/growth within a year after MRI. Demographics and 13 perfusion and 13 permeability derivations of DCEQP were assessed via machine learning and univariate analyses. Logistic regression models ln ( P / 1 - P ) = Σ (β i x i ) + β 0 were selected as the best diagnostic and prognostic biomarkers by minimizing the Bayesian information criterion (BIC). Results: The best diagnostic biomarker of CASH that bled 3 - 12 months prior (BIC = 321.6, Figure A) showed 80% sensitivity and 82% specificity. Permeability derivations did not add diagnostic efficacy when combined with perfusion. The best prognostic biomarker of bleeding/growth (BIC = 201.5, Figure B) showed 77% sensitivity and 72% specificity. Conclusion: Perfusion imaging may diagnose CASH even after hemorrhagic signs disappear on conventional MRI. A combination of permeability and perfusion derivations may help predict bleeding/growth in CAs.