TCD Waveform Morphology for LVO Detection

Transcranial Doppler sonography (TCD) is not only a tool for measuring flow velocity — waveform morphology itself carries diagnostic information. A 2018 study compared two TCD-derived metrics for detecting large vessel occlusion (LVO) and tested whether they could form the basis of simple decision rules, potentially useful even in prehospital stroke triage.

Background

Rapid detection of LVO is critical for timely transfer to thrombectomy centers. While CT angiography (CTA) remains the gold standard, it is not available in the prehospital setting. TCD is portable and directly measures cerebral blood flow velocity, but its interpretation has often relied on expert assessment of waveform patterns. The challenge is to translate these morphological features into objective, quantitative criteria.

Study Aim 

The authors compared two quantitative TCD metrics:

• Velocity Asymmetry Index (VAI): measures interhemispheric velocity differences.
• Velocity Curvature Index (VCI): a newly defined biomarker that mathematically captures waveform curvature.

In addition, they tested a simple decision tree combining both indices.

Methods

The retrospective study included 66 consecutive patients presenting with stroke symptoms — 33 with CTA-confirmed LVO. TCD metrics were compared against CTA findings as reference. Receiver operating characteristic (ROC) curves were generated, and diagnostic accuracy was validated using leave-one-out cross-validation.

Key Results

• VCI achieved an ROC-AUC of 94%, outperforming VAI (88%).
• With optimal thresholds, cross-validation showed:
  • VCI: 88% accuracy / 88% sensitivity
  • VAI: 79% accuracy / 76% sensitivity
• A simple decision tree combining both metrics reached 91% accuracy with 94% sensitivity.

This indicates that incorporating waveform morphology (VCI) enhances diagnostic performance beyond asymmetry measures alone.

Clinical Relevance

• Objective criteria: VCI offers a reproducible, quantitative way to describe pathological waveforms, reducing reliance on subjective expert interpretation.
• Prehospital potential: Automated metrics could provide EMS teams with a reliable tool for stroke triage, shortening time to thrombectomy.
• Next steps: Larger, prospective studies are needed to validate robustness across operators, patients, and measurement conditions.

Conclusion

This study shows that TCD waveform morphology — particularly the Velocity Curvature Index — improves LVO detection. Combining morphology and asymmetry measures into simple decision rules yielded high sensitivity and accuracy. While larger validations are required before clinical implementation, the findings present a clear step toward more objective and practical TCD-based stroke triage.