Intraoperative electrostimulation for awake brain mapping: how many positive interference responses are required for reliability?

OBJECTIVE The purpose of this study was to characterize the reproducibility of language trials within and between brain mapping sessions. METHODS Brain mapping and baseline testing data from 200 adult patients who underwent resection of left-hemisphere tumors were evaluated. Data from 11 additional patients who underwent a second resection for recurrence were analyzed separately to investigate reproducibility over time. In all cases, a specific protocol of electrostimulation brain mapping with a controlled naming task was used to detect language areas, and the results were statistically compared with preoperative and intraoperative baseline naming error rates. All patients had normal preoperative error rates, controlled for educational level and age (mean 8.92%, range 0%-16.25%). Intraoperative baseline error rates within the normal range were highly correlated with preoperative ones (r = 0.74, p < 10−10), although intraoperative rates were usually higher (mean 13.30%, range 0%-26.67%). Initially, 3 electrostimulation trials were performed in each cortical area. If 2 of 3 trials showed language interference, 1 or 2 additional trials were performed (depending on results). RESULTS In the main group of 200 patients, there were 82 single interferences (i.e., positive results in 1 of 3 trials), 227 double interferences (2/3), and 312 full interferences (3/3). Binomial statistics revealed that full interferences were statistically significant (vs intraoperative baseline) in 92.7% of patients, while double interferences were significant only in 38.5% of patients, those with the lowest error rates. On further testing, one-third of the 2/3 trials became 2/4 trials, which was significant in only one-quarter of patients. Double interference could be considered significant for most patients (> 90%) when confirmed by 2 subsequent positive trials (4/5). In the 11 patients who were operated on twice, only 26% of areas that tested positive in the initial operation tested positive in the second and showed the same type of interference and the same current threshold (i.e., met all 3 criteria). CONCLUSIONS Electrostimulation trials in awake brain mapping produced graded patterns of positive reproducibility levels, and their significance varied with the baseline error rates. The results suggest that caution is warranted when 2 of 3 trials are positive, although the need for additional trials depends on the individual patients' baseline error rates. Reproducibility issues should be considered in the interpretation of data from awake brain mapping.