Abstract Number: Esoc2026a1240 Hemodynamic Phantom for Testing Microemboli Detection by Transcranial Doppler

Abstract Background and aims Transcranial Doppler (TCD)–detected microembolic signals (MES) are strong predictors of future cerebrovascular events. However, the accuracy, differentiation and comparability of MES detection across different TCD systems remain insufficiently validated. We aimed to develop and test a novel hemodynamic phantom allowing precise and reproducible delivery of predefined number of gaseous and solid microemboli to assess the accuracy of TCD-based MES detection. Methods A hemodynamic system was constructed using dual pumps to generate controlled pulsatile flow and a 3D-printed model simulating cranial bone and intracranial arteries, with porcine blood as the circulating medium. Controlled injections of air bubbles and solid microemboli (thrombotic and polymer-based) were performed. Three commercially available TCD systems (DWL T1, DWL T2, and Delica) were evaluated for their ability to detect and differentiate MES using automatic embolus detection algorithms. Results were compared with expert sonographer assessment based on spectral analysis. Results All TCD systems reliably detected MES showers. However, automatic embolus detection failed to differentiate air from thrombotic MES, and polymer-based MES were undetectable by all systems. Expert spectral analysis identified significantly more gaseous (mean 14.7±8.3 vs. 1.4±2.3 MES/s; p 0.05) and solid thrombotic MES (mean 7.9±6.3 vs. 0.3±0.4 MES/s; p 0.05) than automatic detection. Conclusions This novel hemodynamic phantom enables standardized testing of TCD MES detection accuracy. Expert spectral analysis markedly outperforms automated detection algorithms, highlighting important limitations of current TCD systems with direct implications for clinical risk stratification and device validation. Supported by the Czech Health Research Council grant No. NU22-04-00389. Conflict of interest David Skoloudik: nothing to disclose. Roman Matejka: nothing to disclose. David Pakizer: nothing to disclose. Istvan Szegedi: nothing to disclose. Adam Skoloudik@nothing to disclose. David Netuka: nothing to disclose.