• Kyung K Peck, Michelle Bradbury, Nicole Petrovich, Bob L Hou, Nicole Ishill, Cameron Brennan, Viviane Tabar, and Andrei I Holodny.
• Department of Radiology and Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, New York, USA.
• Neurosurgery. 2009 Apr 1;64(4):644-52; discussion 652-3.

ObjectiveFunctional magnetic resonance imaging (fMRI) is used to assess language laterality in preoperative brain tumor patients. In postsurgical patients, susceptibility artifacts can potentially alter ipsilateral fMRI activation volumes and the assessment of language laterality. The purpose of this study was to investigate the ability of fMRI to correctly measure language dominance in brain tumor patients with previous surgery because this patient cohort is vulnerable to type II statistical errors and subsequent misjudgment of laterality.MethodsTwenty-six right-handed patients with left-hemisphere gliomas (16 with and 10 without previous surgery) underwent preoperative language fMRI. Language laterality was measured using hemispheric and Broca's area regions of interest (ROIs). Hemisphere dominance, as established by laterality measurements, was compared with that determined by intraoperative electrocorticography and behavioral assessments.ResultsLocalization of primary language cortices was achieved in 24 of 26 patients studied. The hemisphere dominance evaluated by fMRI was verified by intraoperative corticography in only 14 patients (10 with and 4 without previous surgery), and only 12 of them had complete neuropsychological testing. Complete concordance of the laterality with intraoperative electrocorticography and behavioral assessments was found in patients without previous surgery. In patients with previous surgery, concordance was 75% using Broca's area ROI and 88% using hemispheric ROI, notwithstanding susceptibility artifacts. Differences in laterality between pre- and postsurgical patients, based on either hemispheric (P = 0.81) or Broca's area (P = 0.19) ROI measurements were not statistically significant. However, hemispheric ROI analyses were found to be less affected by postsurgical artifacts and may be more suitable for establishing hemisphere dominance.ConclusionfMRI mapping of eloquent language cortices in brain tumor patients after surgery is feasible and can serve as a useful baseline evaluation for preoperative neurosurgical planning. However, findings should be interpreted with caution in the presence of postsurgical artifacts.

28 keywords...

hide…