Intraoperative language localization in multilingual patients with gliomas

OBJECTIVE: Intraoperative localization of speech is problematic in patients who are fluent in different languages. Previous studies have generated various results depending on the series of patients studied, the type of language, and the sensitivity of the tasks applied. It is not clear whether languages are mediated by multiple and separate cortical areas or shared by common areas. Globally considered, previous studies recommended performing a multiple intraoperative mapping for all the languages in which the patient is fluent. The aim of this work was to study the feasibility of performing an intraoperative multiple language mapping in a group of multilingual patients with a glioma undergoing awake craniotomy for tumor removal and to describe the intraoperative cortical and subcortical findings in the area of craniotomy, with the final goal to maximally preserve patients' functional language. METHODS: Seven late, highly proficient multilingual patients with a left frontal glioma were submitted preoperatively to a battery of tests to evaluate oral language production, comprehension, and repetition. Each language was tested serially starting from the first acquired language. Items that were correctly named during these tests were used to build personalized blocks to be used intraoperatively. Language mapping was undertaken during awake craniotomies by the use of an Ojemann cortical stimulator during counting and oral naming tasks. Subcortical stimulation by using the same current threshold was applied during tumor resection, in a back and forth fashion, and the same tests. RESULTS: Cortical sites essential for oral naming were found in 87.5% of patients, those for the first acquired language in one to four sites, those for the other languages in one to three sites. Sites for each language were distinct and separate. Number and location of sites were not predictable, being randomly and widely distributed in the cortex around or less frequently over the tumor area. Subcortical stimulations found tracts for the first acquired language in four patients and for the other languages in three patients. Three of these patients decreased their fluency immediately after surgery, affecting the first acquired language, which fully recovered in two patients and partially in one. The procedure was agile and well tolerated by the patients. CONCLUSION: These findings show that multiple cortical and subcortical language mapping during awake craniotomy for tumor removal is a feasible procedure. They support the concept that intraoperative mapping should be performed for all the languages in which the patient is fluent in to preserve functional integrity.