Abstract: Objective: To investigate the possibility of predicting facial nerve involvement in the patients with parotid tumors using facial electroneurography (ENoG). Methods: In the study, 53 patients with primary parotid tumors were included in the study, 28 were benign tumors and 25 were malignant. There was no significant difference of tumor locations and sizes between the two groups. House-Brackmann facial nerve function evaluation was gradeⅠin all the patients who received examination of facial electroneurography, including stimulation strength, amplitude, and latent time bilaterally. The facial electroneurography results in the affected side were compared with the results of contralateral normal side, intraoperative appearance and postoperative histopathological diagnosis. The facial electroneurography results were analyzed by Wilcoxon signed rank test and receiver operator characteristic (ROC) curve. Results: During the facial electroneurography examination, the mean stimulation strength in the patients with benign parotid tumor was 20.0 mA. There was significant decrease in the amplitude at the affected side compared with the normal side upon posterior auricular stimulation (P<0.05). But there was no significant difference in the latent time between the bilateral sides (P>0.05). However, in the patients with malignant parotid tumor, the mean stimulation strength was higher at 24.5 mA. There was significant decrease in the amplitude or even no response at the affected side compared to the normal side upon posterior auricular stimulation (P<0.05). No response was detected in the multiple branches of facial nerve of affected sides in 36.0% patients upon posterior auricular stimulation. The amplitude of branches Ⅲ and Ⅳ was significantly lower at the affected side than that at the normal side upon anterior auricular stimulation (P<0.05). The area under the ROC curve for ENoG was 0.884. Conclusion: When the facial nerve was involved by the parotid tumors, the stimulation strength in the electroneurography was larger. There was significant difference in the amplitude and the latent time of the facial nerve between the affected side and the normal side upon the posterior auricular stimulation. The rate of absence of reaction wave was higher in the patients with malignant tumors. It was feasible to predict the facial nerve involvement by ENoG for the parotid gland tumor patients without clinical appearances of facial paralysis.

Key words: Parotid neoplasms, Facial nerve, Electroneurography