Monitoring language networks in the asleep patient during surgery: an electrophysiological approach
Introduction: Preservation of language function is critical in neurosurgical procedures, since damage to essential cortico-subcortical networks leads to aphasia. In current surgical practice, awake surgery is the only tool for monitoring language function, however conditions of impaired abilities or poor compliance require different surgical strategies. First described almost fifteen years ago, cortico-cortical evoked potentials (CCEPs) may enable intraoperative monitoring of language networks1. Nevertheless, their adoption in clinical practice has been anecdotal and the technique remains ancillary to awake procedures2. We aim to evaluate feasibility of recording cortico-cortical evoked potentials of the major language tract, the arcuate fasciculus (AF), in the asleep patient.
Methods: We prospectively recruited seven patients undergoing surgery for lesions in the left perisylvian cortex. Inclusion criteria were a tumour in a language eloquent area and contraindication for awake surgery. Preoperative spherical deconvolution HARDI tractography was performed in all patients to identify cortical terminations of the AF. This was used to guide preoperative functional mapping for assessing language eloquent cortices performed with navigated transcranial magnetic stimulation (nTMS). Intraoperatively, strip electrodes were placed over the identified eloquent cortical regions with the aid of neuronavigation and CCEPs measured. CCEP amplitudes and latencies were evaluated offline with Matlab.
Results: AF-CCEPs were successfully evoked in all but one patient. Mean intensity used was 15 mA ±2 mA. The mean peak latency (N1) was 13.2 ms ±1.7 ms and mean amplitude was 0.36 mV ±0.13 mV. Clinically, only the patient in which AF-CCEPs were not monitorable showed post-operative aphasia. In three patients in which CCEPs were monitored prior and after the resection, persistence of CCEPs coincided with preserved language function. Of interest, a decrease below 50% corresponded postoperatively with light expressive aphasia while a decrease higher to 50% with mild language impairment.
Conclusions: We preliminary demonstrate that CCEPs can be used to monitor and preserve fibres of the arcuate fasciculus in the asleep patient. Notably, the latency of N1 is much shorter than described for the awake patient (22-36 ms) while maintenance of N1 corresponded with preserved language function, as previously reported1,3. This may unravel novel surgical perspectives for patients that are not candidates for awake surgery.