Comparison of a 1-MHz and a 2-MHz probe for microembolus detection using transcranial Doppler ultrasound

Clinically silent microembolic signals (MES) can be detected by transcranial Doppler sonography (TCD). There is theoretical evidence that lower ultrasound emission frequencies may lead to a higher signal intensity and thus sensitivity to detect MES. We compared a 1-MHz probe with a 2-MHz probe regarding sensitivity in the detection of MES. Moreover, embolus detection by transcranial Doppler ultrasound is very time consuming and semi-automated detection is mandatory. Therefore, we studied an on-line algorithm using the bi-gate technique and the two transmission frequencies.

After defining detection thresholds of > or = 12 dB (1 MHz) and > or = 10 dB (2 MHz) with eight normal subjects as MES-negative controls, taking into account natural fluctuations of the Doppler spectrum, we studied 36 patients with ischaemic events and five asymptomatic patients with incidental embolic sources. All patients subsequently underwent a 1-hour unilateral embolus detection from the middle cerebral artery (MCA) or the posterior cerebral artery (PCA), respectively, using 1 and 2 MHz for 30 minutes each in a randomized order. The software algorithm was compared with a blinded off-line analysis by an experienced observer as a gold standard.

The investigator detected 198 MES (range 0-41 MES) in the recordings of 29 patients out of the 41 patients using the 1-MHz probe and 101 MES (range 0-32 MES) in the recordings of 14 patients using the 2-MHz probe (p = 0.0007). Sensitivity of the software to detect MES confirmed by the investigator was 31% using 1 MHz and 41% using 2 MHz. The positive predictive value was 6 and 30%, respectively.

The sensitivity and positive predictive values of the automated algorithm to detect MES were unacceptably low for clinical practice with both frequencies. The use of 1 MHz instead of 2 MHz may, however, be useful when evaluating the recordings off-line by an experienced blinded observer.