Adaptive Maximal Blood Flow Velocity Estimation From Transcranial Doppler Echos

Area of Science:

• Neurosonology
• Biomedical Engineering
• Signal Processing

Background:

• Transcranial Doppler (TCD) ultrasonography is crucial for assessing cerebral blood flow.
• Accurate maximal flow velocity (MFV) measurement is essential for novel TCD applications like intracranial pressure (ICP) estimation.
• Low signal-to-noise ratios in TCD spectrograms challenge precise MFV determination.

Purpose of the Study:

• To develop a calibration-free algorithm for estimating MFVs from TCD spectrograms.
• To introduce a beat-by-beat signal quality index for TCD waveform assessment.
• To enhance the reliability of TCD data for advanced clinical uses.

Main Methods:

• The algorithm employs multiple binary segmentations of TCD spectrograms.
• Envelopes, representing MFVs, are extracted using an edge-following technique with physiological constraints.
• A signal quality index guides the selection of the optimal MFV waveform.

Main Results:

• The algorithm was validated on 32 TCD recordings from middle cerebral and internal carotid arteries.
• Evaluation involved healthy individuals and neurocritical care patients.
• Relative errors were -1.5% for the whole waveform and -3.3% for peak systolic velocity compared to manual tracings.

Conclusions:

• A robust algorithm for MFV estimation was developed, leveraging a feedback loop between signal quality and segmentation.
• The algorithm has been integrated into an ICP estimation pipeline.
• Publicly sharing the code and data aims to foster new TCD applications requiring high-quality flow velocity waveforms.