Correlation of factors predicting intraoperative brain shift with successful resection of malignant brain tumors using image-guided techniques

Background: Intraoperative brain shift may cause inaccuracy of stereotactic image guidance on the basis of preoperatively acquired imaging data. The purpose of our study was to determine whether factors predicting brain shift affect the success of image-guided resection of malignant brain tumors. Methods: We retrospectively studied 54 patients who underwent image-guided resections of histopathologically confirmed malignant brain tumors (9 metastases, 45 high-grade gliomas). Precautions were taken during surgery to minimize brain shift, but intraoperative imaging was not performed. The following factors predictive of intraoperative brain shift were assessed: tumor size, periventricular location, patient age, prior surgery or radiation therapy, patient positioning, use of mannitol, and length of operative time. Postoperative magnetic resonance imaging was obtained in all cases within 48 hours of surgery to assess extent of resection. Results: Perioperative mortality was 0% in our series; perioperative morbidity was 3 of 54 patients (5.5%); 1 patient required reoperation for a hematoma, and 2 had transient neurological deficits. Successful resection was accomplished in 93% of tumors less than 30 cm³ compared with 63.6% of tumors greater than 30 cm³ (P = .026, Fisher exact test). This difference was more pronounced for patients with malignant gliomas. However, other factors predictive of intraoperative brain shift were not associated with unsuccessful resection. Conclusions: Intraoperative brain shift does not significantly affect the likelihood of successful resection of malignant brain tumors smaller than 30 cm³. Larger tumors are less likely to be successfully resected, although factors other than brain shift can contribute to unsuccessful resection.