Is There an Indication for Intraoperative MRI in Subtotal Resection of Glioblastoma? A Multicenter Retrospective Comparative Analysis.

OBJECTIVE

Surgery in patients with highly eloquent glioblastoma (GB) remains challenging. The aim of this study was to evaluate influence of use of intraoperative magnetic resonance imaging (iMRI) on extent of resection (EOR), clinical outcome, and survival in patients with preoperatively intended subtotal resection of GB.

METHODS

Retrospective assessment was performed in 3 neurosurgical centers (2008-2013). All patients with primary GB, unilocular growth, and adjuvant radiochemotherapy in whom subtotal resection was intended were included. Imaging data were pseudonymized and volumetrically assessed by a central reviewer. Clinical outcome was evaluated based on complications, new permanent neurologic deficits, and survival. Based on subtraction of potential EOR (as defined by reviewer) from effective EOR, EOR change (ΔEOR) was calculated for each case. We calculated multivariate linear regressions to assess influence on effective EOR and ΔEOR, binary regressions to assess influence on complications and new permanent neurologic deficits, and Cox regression models to assess survival.

RESULTS

Assessment included 70 patients. In iMRI group, a significantly lower ΔEOR (4% vs. 22%, P = 0.008) was found. In linear regression model controlling for age, tumor volume, neurophysiologic mapping, and iMRI, only iMRI influenced ΔEOR significantly (P = 0.03, odds ratio 0.27). In binary regression model controlling for age, tumor volume, effective EOR, neurophysiologic mapping, and iMRI, none of the variables showed a significant difference concerning complications and new permanent neurologic deficits. Overall survival was significantly influenced by age (P = 0.03, odds ratio 1.02) and iMRI (P = 0.026, odds ratio 0.51).

CONCLUSIONS

Maximum safe resection is an important prognostic factor for patients with eloquent GBs. iMRI seems to be a relevant tool to achieve this goal.