Abstract 21400: Morphological Analysis of Transcranial Doppler Waveforms for Large Vessel Occlusion Diagnosis

Abstract

Listen

Translate article

Introduction: Underlying cardiac conditions such as atrial fibrillation (AF) and cardiac procedures such as transcatheter aortic valve replacement (TAVR) and coronary artery bypass grafting (CABG) have been shown to lead to large vessel occlusions (LVO) causing significant mortality and morbidity and improved pre, peri, post procedural methods for diagnosis are needed. Transcranial Doppler (TCD) is a rapid, inexpensive and portable tool for assessment of cerebral blood flow velocity (CBFV) that is used clinically and has been shown to have potential as a diagnostic aid for LVO. The utilization of the morphological information from the TCD waveform has shown to provide additional diagnostic capabilities in many neurological conditions and may improve LVO detection over traditional TCD metrics of mean velocity and pulsatility index. However, traditional TCD techniques suffer from a number of limitations including operator skill level and variability due to depth and location selection. Hypothesis: The morphology of CBFV waveforms measured with TCD enables detection of LVO while relaxing the need for exact insonation depth relative to occlusion location. Methods: Bilateral TCD scans of 17 pts (66.8 ± 15.2 yrs) with confirmed MCA occlusion were performed immediately following CT/CTA. TCD was also performed on 28 healthy out of hospital controls (55.3 ± 9.5 yrs). Waveform morphology was compared between the two groups. Results: Morphological analysis wherein MCA data for each subject was randomly sampled across depths showed median AUC of 96.3% for LVO detection, with median SEN of 88.9% and SPE of 96.3%. Fig. 1 shows an example of TCD waveforms for control and LVO subjects with similar traditional TCD metrics, but different subtle morphological profiles. Conclusion: Morphological analysis demonstrated high diagnostic accuracy additionally without the need for optimal depth selection, which may reduce the requirement of a trained sonographer.