Decision Criteria for Large Vessel Occlusion Using Transcranial Doppler Waveform Morphology.

Area of Science:

• Neurology
• Medical Imaging
• Biomedical Engineering

Background:

• Prehospital identification of Large Vessel Occlusion (LVO) is crucial for efficient stroke patient triage and timely treatment.
• Current tools for prehospital LVO detection are insufficient, hindering optimal patient care.
• Objective Transcranial Doppler (TCD) metrics could equip first responders with vital decision-making tools.

Purpose of the Study:

• To compare the diagnostic efficacy of two TCD metrics: Velocity Asymmetry Index (VAI) and Velocity Curvature Index (VCI).
• To evaluate a decision tree combining VAI and VCI for LVO detection.
• To assess the potential of TCD for improving prehospital stroke assessment.

Main Methods:

• Retrospective comparison of VAI and VCI accuracy, sensitivity, and specificity against CT-Angiography (CTA).
• Analysis of 66 stroke patients (33 with confirmed LVO) using TCD.
• Leave-one-out cross-validation to determine optimal thresholds and assess combined metric performance.

Main Results:

• Individual VCI and VAI metrics showed strong performance with ROC-AUC of 94% and 88%, respectively.
• VCI achieved 88% accuracy (88% sensitivity), outperforming VAI's 79% accuracy (76% sensitivity).
• A decision tree combining VCI and VAI reached 91% accuracy (94% sensitivity).

Conclusions:

• The Velocity Curvature Index (VCI) demonstrates superior diagnostic performance for LVO detection compared to VAI.
• A simple decision tree incorporating both VCI and VAI can further enhance diagnostic accuracy.
• Machine learning-based TCD analysis holds promise for robust prehospital LVO identification.