Abstract Number: Esoc2026ys10 the Decryptoring Score for Occult Atrial Fibrillation Detection in Cryptogenic Stroke: A Prospective Validation Study

Abstract Background and aims In up to one-third of ischemic stroke cases, no definite cause is identified, classifying them as cryptogenic strokes, often falling under the concept of embolic stroke of undetermined source (ESUS). Occult atrial fibrillation (AF) is frequently encountered and an often undetected source, owing to its paroxysmal and asymptomatic nature. Current monitoring strategies, including Holter and implantable loop recorders, show variable detection rates. The Decryptoring Score (PMID: 34689053) integrates clinical, echocardiographic, and biomarker parameters to predict AF in cryptogenic stroke and to identify who may benefit from prolonged rhythm monitoring. This study aims to prospectively validate its diagnostic performance regarding sensitivity, specificity, predictive values, and AUC. Methods Consecutive patients aged ≥50 years with ischemic stroke or transient ischemic attack (TIA) are prospectively enrolled. Based on an expected atrial fibrillation (AF) prevalence of approximately 15%, the study aims to recruit at least 100 patients to ensure adequate statistical power for validating the diagnostic performance of the Decryptoring Score. All participants undergo continuous in-hospital ECG monitoring using a telemetric AF detection system (SRA+, Apoplex Medical Technologies), followed by 7–15 days of ambulatory Holter monitoring and analysis by an independent core laboratory. Echocardiography includes standardized left atrial strain (LAS) assessment with a temporal resolution of ≥50 Hz. NT-proBNP, troponin (hs-TnT), comorbidities, and relevant laboratory parameters are systematically recorded. Results Patient recruitment is nearing completion, with more than 90% of the planned cohort enrolled. Preliminary evaluations indicate high data quality and strong feasibility of this comprehensive diagnostic approach. Conflict of interest nothing to disclose