Skull Base 2007; 17 - A197
DOI: 10.1055/s-2007-984132

EMG-Monitoring of Cranial Nerves and the Posterior Fossa

J. Romstöck 1(presenter), J. Prell 1, S. Rampp 1, C. Strauss 1
  • 1Schweinfurt, Germany

Functional electromyographic (EMG) monitoring of motor cranial nerves has become an essential tool to detect imminent neurological deficits during skull base surgery. During continuous intraoperative recordings of the facial EMG, distinct waveform patterns have been described according to acoustic and visual evaluation (A-, B-, and C-trains). To improve the reliability and safety of intraoperative monitoring a system for detailed computer-aided analysis of EMG waveform patterns was developed.

In 80 patients undergoing surgery on acoustic neurinomas and other cerebellopontine or petroclival tumors, the electromyogram according to the cranial nerves V, VI, VII, IX, and X was recorded and stored continuously. Spontaneous and evoked muscle activity was classified using amplitudes and waveform patterns and was compared with postoperative clinical findings. On the basis of distinct EMG patterns found after visual offline evaluation, an online system was developed for rapid computer-aided detection of possibly harmful EMG waveforms, called A-trains.

Characteristic EMG discharges with opposite clinical relevancy were recorded: the A-train, a sinusoid high-frequency train of spikes, was found in more than 90% of the patients who revealed additional facial nerve deficits postoperatively (at least one degree in the House-Brackmann scale). The B- and C-trains showed no correlation with the surgical outcome. Single bursts served well for early detection of neural fibers during tumor dissection, but did not indicate functional deficits. The A-train was detected safely with the computer-aided online system; however, with acoustic and visual data evaluation only, large parts of data were missed in some patients.

The intraoperative detection of distinct EMG patterns provides reliable monitoring information on cranial nerve function. With automatic signal analysis, the sensitivity and specificity of more than 90% found after offline evaluation could be transferred into an online setup.