The prognostic value of somatosensory evoked potential monitoring and tumor data in supratentorial tumor removal.

The aim of this study was to find appropriate somatosensory evoked potential (SSEP) and tumor data that would predict immediate postoperative outcome. Seventeen patients were evaluated, all with supratentorial mass lesions. Intraoperative SSEP monitoring was carried out, and central conduction time, N20 amplitude, and N20 to N13 amplitude ratio were recorded throughout the operation. The differences between the end and the start of the procedure and between the end of the operation and the most substantial changes during tumor removal were calculated for statistical evaluation. Tumor location and extent were carefully determined and calculated by computed tomographic reconstructions in the axial and sagittal planes. Several tumor parameters were measured: the distances from the midpoint of the tumor to the central sulcus, midline, and base of the skull and the longitudinal and cross-sectional diameters. The electrophysiologic and tumor data sets were correlated with the immediate postoperative neurologic deterioration. For this purpose, patients were divided into two groups: group 1 = no neurologic deterioration after operation (13 patients); group 2 = neurologic deterioration after operation (4 patients). The difference in central conduction times between the end and the start of the procedure was the only variable that exhibited a significant influence on the immediate postoperative condition (P = 0.007), as determined by analysis of variance. The other electrophysiologic data as well as the tumor data failed the significance threshold of P = 0.05. Discriminant analysis was applied to test the classification capacity of the measured variables. Involving all measured variables (electrophysiologic and tumor data), discriminant analysis allowed a correct classification of all 17 patients to their proper neurologic deterioration group. Discriminant analysis for SSEP data alone led to 15 correct classifications. Tumor data, used alone for discriminating procedures, revealed 14 correct classifications. When each variable was analyzed separately, only the difference in central conduction times between the end and the start of the procedure gave significant predictions, namely, 15 correct classifications. This was the same number as achieved by all electrophysiologic variables together. No other variable could on its own yield any valid prognosis for assessment of immediate postoperative neurologic deterioration. The data confirm the importance of central conduction time recovery before the end of the operation on the patient's immediate postoperative neurologic condition.