PROJECT SUMMARY Intracerebral hemorrhage (ICH) affects over two million people per year worldwide and is associated with a 54% one-year mortality. One important risk factor for ICH is chronic kidney disease (CKD), an entity that affects over 20 million people in the United States and is associated with higher ICH incidence and worse outcomes. ICH patients with advanced CKD have a 2.3-fold larger hematoma volume and a >4-fold increase in 1-year mortality. Given a prevalence as high as 14% in the U.S. population, CKD is a prime yet understudied disease in the pathogenesis of ICH. In this mechanistic clinical investigation, we hypothesize that CKD-induced vascular injury leading to ICH can be characterized by two intermediate imaging biomarkers: (1) intracranial arterial calcification (IAC), and (2) cerebral microbleeds (CMB). Specifically, we hypothesize that arteriolar injury in CKD results in higher IAC and CMB burden, which are predictive of ICH volume and hematoma expansion. To better characterize these relationships, this study leverages a 10-year institutional database of 1,500 spontaneous ICH patients spanning July 2011 and June 2020 as well as 150 prospective ICH patients aggregated over the first study year. For each patient, deep learning imaging analysis is used to quantify ICH volume as well as IAC burden on computed tomography scans. For patients with corresponding brain magnetic resonance imaging within 2-weeks of hospital admission, deep learning analysis will also be used to quantify CMB burden. Key target outcomes include prediction of ICH volume and hematoma expansion as a function of IAC and CMB burden stratified by CKD stage. In Specific Aim 1, the relationship between CKD status and IAC burden is characterized, which is used in combination with clinical risk factors to predict ICH volume and hematoma expansion. In Specific Aim 2, the relationship between CKD status and CMB burden is characterized, which is used in combination with IAC burden and clinical risk factors to predict ICH volume and hematoma expansion. All statistical models derived from the 10-year historic cohort of 1,500 patients are validated against the prospective cohort of 150 patients. Through an improved understanding of the shared pathophysiology between these disease states, quantitative models derived from the proposed analysis can be used for early identification of patients at high-risk for ICH and associated complications, who in turn are optimal candidates for aggressive management of underlying CKD.