Increased prediction of right nonrecurrent laryngeal nerve in thyroid surgery using preoperative computed tomography...

A nonrecurrent laryngeal nerve (NRLN) is a rare but potentially serious anatomical variant. Although the incidence is reported to be 0.3% to 1.3%, it carries a much higher risk of palsy during thyroid surgery. The objective of this study is to investigate the usefulness of computed tomography (CT) for preoperative identification and intraoperative neuromonitoring identification (IONM) of NRLN in thyroid cancer patients.

The preoperative neck CT scans from 1,574 patients who needed thyroid surgery were examined. Absence of the brachiocephalic artery (BCA) and the presence of arteria lusoria were defined as positive with NRLN. Systematic intraoperative neuromonitoring (IONM) was also carried out for these 1,574 patients to localize and identify NRLN. A negative electromyography (EMG) response from lower vagal stimulation but a positive EMG response from the upper position indicated the occurrence of an NRLN.

Nine NRLN (0.57%) were intraoperatively identified out of the 1,574 patients, and no patient with a NRLN showed preoperative clinical symptoms related to NRLN. Prior to the operation, surgeons identified only seven suspected NRLN cases based on identification of arteria lusoria. But a review of CT scans revealed that all cases could be identified by vascular anomalies. All patients were successfully detected at an early stage of operation using intraoperative neuromonitoring (IONM). Postoperative vocal cord function was normal in all patients.

CT of the neck is a reliable method for predicting NRLN before thyroid cancer surgery. However, some image features can be easily missed. Neurophysiology helps the surgeon to identify the NRLNs more precisely. Combining the two evaluation methods may decrease the incidence of nerve palsy, especially in cases of NRLN. Considering that CT is expensive, requires an X-ray, and achieves less information than ultrasound (US) concerning thyroid nodules, we suggest that applying US and IONM is more reasonable.