OBJECTIVE: Embolus detection by transcranial Doppler ultrasound is very time consuming and semi-automated detection is mandatory. The device studied, a TC4040, Nicolet-EME, uses the four-gate technique and allows for audiovisual off-line verification of the recorded events. METHODS: Twenty controls, 10 patients with mechanical prosthetic heart valves and 12 patients with occlusive carotid artery disease were investigated by transcranial colour-coded duplex sonography and, subsequently, underwent a 1-h unilateral embolus detection from the middle cerebral artery using four-gate TCD. We investigated the Doppler spectrum background, microembolic signals (MES) and artefacts produced. A detection threshold of 5 dB or more was defined taking into account natural fluctuations of the Doppler spectrum. RESULTS: Sensitivity of the software was 91.9% and observer-software agreement on MES was 7.8% in the valve patients, and 77.7% and 7.5% in the carotid artery disease patients, respectively. Weaker MES were more likely not to be detected in all four channels. The artefact signal rejection rate was 62%. MES produced either positive or zero time delays in adjacent channels. Artefact signals produced either no delay, or a positive or a negative time delay. Duration of MES ranged from 1-88 ms. CONCLUSIONS: Besides refined recognition of MES using the time delay, four gates give faint MES no less than four opportunities to overcome the detection threshold. With this device's satisfying sensitivity, regions of interest in a 1-h recording can audiovisually be evaluated off-line in a few minutes by an investigator.
OBJECTIVE: Embolus detection by transcranial Doppler ultrasound is very time consuming and semi-automated detection is mandatory. The device studied, a TC4040, Nicolet-EME, uses the four-gate technique and allows for audiovisual off-line verification of the recorded events. METHODS: Twenty controls, 10 patients with mechanical prosthetic heart valves and 12 patients with occlusive carotid artery disease were investigated by transcranial colour-coded duplex sonography and, subsequently, underwent a 1-h unilateral embolus detection from the middle cerebral artery using four-gate TCD. We investigated the Doppler spectrum background, microembolic signals (MES) and artefacts produced. A detection threshold of 5 dB or more was defined taking into account natural fluctuations of the Doppler spectrum. RESULTS: Sensitivity of the software was 91.9% and observer-software agreement on MES was 7.8% in the valve patients, and 77.7% and 7.5% in the carotid artery diseasepatients, respectively. Weaker MES were more likely not to be detected in all four channels. The artefact signal rejection rate was 62%. MES produced either positive or zero time delays in adjacent channels. Artefact signals produced either no delay, or a positive or a negative time delay. Duration of MES ranged from 1-88 ms. CONCLUSIONS: Besides refined recognition of MES using the time delay, four gates give faint MES no less than four opportunities to overcome the detection threshold. With this device's satisfying sensitivity, regions of interest in a 1-h recording can audiovisually be evaluated off-line in a few minutes by an investigator.