ABSTRACT NUMBER: ESOC2026A1206 PREDICTING IN-HOSPITAL OUTCOMES IN ACUTE CEREBRAL VENOUS THROMBOSIS: THE CLOT-VENUS STRATIFICATION SCORE Authors: Leonardo Cruz-Criollo, Aaron Rodriguez-Calienes, Eric Kontowicz, Anderson Brito, Nashwa Abdelhakim, Vanessa Cano Nigenda, Andres Alberto Mercado, Brian J Smith, Antonio Arauz Gongora, Santiago Ortega-Gutierrez Published: 2026 ABSTRACT AbstractBackground and aimsEarly identification of high-risk Acute Cerebral Venous Thrombosis (CVT) patients is vital for aggressive management. We developed and validated machine-learning models to predict in-hospital mortality and poor functional outcome (mRS 3-6) at discharge.MethodsThis retrospective observational cohort study of patients diagnosed with acute CVT from the international CoLlabOraTion on Cerebral VENoUs Thrombosis Study (CLOT-VENUS) registry (2004-2024), including two comprehensive stroke centers in the USA and Mexico. Most relevant admission clinical, imaging, and laboratory results were collected. We developed machine learning models using LASSO and a gradient-boosted model (GBM), to predict poor functional outcome (mRS 3-6) at discharge and in-hospital mortality. We then developed nomograms for the outcomes. Parsimonious predictors were selected via Bayesian optimization and recursive feature elimination. Models were trained on two-thirds of the available data and tested on the remaining one-third holdout dataResultsOf 432 patients, 394 met the inclusion criteria. Median age was 40 years [IQR 28-55], and 65% were female. Volume of ICH, Glasgow coma scale, mRS baseline, age, cerebral edema, anemia, leucocytes, hemoglobin, and NLR were among the important predictors identified for discharge score and in-hospital mortality. The final model achieved high predictive accuracy, with ROC-AUCs of 0.87 (95 %CI: 0.83–0.91) for discharge mRS (3-6) and 0.87 (95% CI: 0.8 – 0.94) for in-hospital mortalityConclusionsThe new scales provide a rapid and reliable tool for risk-stratifying patients with acute CVT upon admission, helping clinicians identify candidates for intensive monitoring and aggressive interventions to prevent secondary injury.Conflict of interestSantiago Ortega-Gutierrez: NIH-NINDS (RO1NS127114, RO1NS138765, RO3NS126804), Leonardo Cruz-Criollo and rest of the team: nothing to discloseFigure 1 - belongs to Conclusions