A Comparison of Post-Operative MRI Signal Changes between Endoscopic Endonasal Approaches and Open Approaches for Olfactory Groove Meningiomas: A Match Paired Analysis from Two Institutions

Background: Olfactory groove meningiomas may be associated with significant brain edema and surgical removal of these tumors may be associated with further injury to the frontal lobes with potential loss of neurocognitive function. Both endoscopic and open transcranial approaches have been used to access these tumors. Endoscopic access avoids manipulation of the frontal lobes, but little objective benefit has been demonstrated.

Methods: A retrospective review was performed at the University of Toronto (UT) and University of Pittsburgh Medical Center (UPMC) from 2003 to present to identify patients who had undergone surgery (either endoscopic or open approach) for resection of olfactory groove meningiomas. Patients were eligible if they had completely resected tumors, had both pre and post-op MRI imaging, and had tumors that were accessible either by an open or endoscopic approach. Matched pairs were created by (a) matching the tumor volume (within 10%) of patients who underwent an endoscopic approach with one who underwent an open approach, and (b) matching the closest time for post-operative imaging for each approach if more than one scan was performed (minimum of 6 weeks post-operatively). Volumes of tumors were measured using gadolinium-enhanced T1-weighted images; pre-operative edema was measured on pre-operative T2-weighted imaging and post-operative frontal lobe injury was measured on FLAIR sequence. The porencephalic cave was defined as the radiographic dead space created after tumor resection on post-operative FLAIR imaging. Tumor volume, edema, FLAIR change, and porencephalic caves were volumetrically quantified based on serial free-hand outlines of tumors, edema, and FLAIR change on axial imaging using the OSIRIX (at UT) and Vitrea software platforms (at UPMC). Differences in post-operative FLAIR change and porencephalic cave were compared in univariable and multivariable analysis.

Results: Ten matched pairs (20 patients) were identified according to matching criteria. The open approach was associated with more post-operative FLAIR change on MRI compared with the endoscopic approach although not statistically significant (13.3 cm3 (SD = 12.0) versus 6.9 cm3 (SD = 10.0), p = 0.17). The endoscopic approach was also associated with smaller porencephalic cave volumes (1.7 cm3 (SD = 2.8) vs 6.9 cm3 (SD = 10.0), p = 0.058). In a multivariable model, the endoscopic approach was associated with less post-operative FLAIR change (p = 0.016) and smaller porencephalic cave volumes (p = 0.028) after adjusting for the volume of pre-operative edema and the timing of post-operative imaging.

Conclusions: This study provides preliminary evidence that the endoscopic endonasal approach is associated with quantifiable improvements in post-operative brain imaging. Further studies with inclusion of measures of neurocognitive function are needed to confirm the significance of these findings.