Influence of Tumor Location and Other Variables on Predictive Value of Intraoperative myogenic Motor-Evoked Potentials in Spinal Cord Tumor Surgery

Abstract

To study the influence of tumor location (cervical vs. thoracic; extramedullary vs. intramedullary) on predictive value of intraoperative myogenic motor-evoked potentials (iMEP) changes in patients undergoing surgery for spinal cord tumors. Three hundred patients retrospective data (91 intramedullary) and 209 (intradural extramedullary) with successful iMEP recordings were analyzed. Responses to transcranial electrical stimulation were recorded from the lower limb muscles. Preoperative clinical variables, iMEPs changes, and postoperative neurologic deficits were noted. Associations between categorical variables and outcome were analyzed with the Fisher exact test. Of the 300 patients 28 (9.3%) had significant intraoperative worsening of iMEPs. New postoperative deficits occurred in 23 of these 28 patients. False-positive decreases in iMEPs were observed in 5 patients. There was a significant association between changes in iMEP and postoperative new motor deficits (P ≤ 0.0001). Multivariate analysis showed that patients with changes in iMEP undergoing surgery for thoracic segment tumors, with longer duration of symptoms (>12 months) and older age (≥21.5 years) were more likely to suffer postoperative neurological decline (odds ratio 4.1, P ≤ 0.001 and odds ratio 5.4 P ≤ 0.0001, respectively). The sensitivity of iMEPs was 100% and specificity 98.2%. The positive and negative predictive values were 82% and 100%; however, the sensitivity and specificity is similar in thoracic intramedullary (TIM) (n= 53) and cervical intramedullary tumors (n= 38) (both were 100% and 97%). The positive predictive value was significantly greater for TIM tumors (93% vs. 50%). A strong association was observed between worsening of iMEPs and postoperative new neurological deficits in patients with TIM tumor.