Implementation Of Intraoperative Magnetic Resonance Imaging And Multimodal Navigation In Surgical Resection Of Gliomas Involving Language Areas And Brain Biopsy

Objective To investigate the impact of intraoperative magnetic resonance imaging(iMRI) and multimodal navigation on glioma surgery adjacent to language areasand brain biopsy. Methods The first part was designed to estimate theintraobserver and interobserver variability of reconstructed arcuate fasciculus.Perioperative volumes of arcuate fasciculus and aphasia quotients (AQ) weremeasured in40patients, who underwent glioma surgery adjacent to languageareas. Coefficients of variability of arcuate fasciculus volume and languagefunction denoted as CV_afvand CVlfwere calculated, and the relationship betweenthe two coefficients were further analyzed. In the second part,217suspectedgliomas involving language areas were divided into study group and control group.124cases in study group underwent iMRI and multimodal navigation-guidedneurosurgery, whereas93cases in control group underwent conventionalnavigation-guided neurosurgery. The differences with regard to extent of resection(EoR), rate of gross total removal (GTR), short-term and long-term languagefunction as well as progression free survival (PFS) in patients with glioblastomasbetween two groups were statistically analyzed. The third part enrolled51patients,who underwent iMRI and multimodal navigation-guided brain biopsy. Metabolicinformation was used for biopsy target selection. Intraoperative guidance helpedbiopsy trajectory avoid the eloquent structures. iMRI was performed to prove thebiopsy accuracy and to revise the incorrect biopsy. Procedure data, diagnostic rate,postoperative neurological status and surgical outcome were recorded. Results Inthe first part, intraobserver arcuate fasciculus showed excellent congruency, andminor differences of interobserver arcuate fasciculus near the cortical endpointswere noted. Positive linear correlations was demonstrated between CVlfand CV_afv(P ﹤0.001) with a r value of0.483. To estimate the sensitivity of ROC whenpredicting CVlfwith CV_afv, the area under the curve was0.96with95%CIbetween0.91and1.00(P﹤0.001). To estimate the specificity of ROC whenpredicting CVlfwith CV_afv, the area under the curve was0.91with95%CIbetween0.82and1.00(P=0.001). In the second part,198cases were available fordata analysis, including106low-grade gliomas (LGGs) and92high-gradegliomas (HGGs).112cases were in study group and86cases were in controlgroup. The initial baseline was well balanced between two groups. EoR and therate of GTR were significant higher in study group than in control group (EoR:overall,95.50%﹥89.85%, P﹤0.001; LGGs,95.27%﹥88.49%, P=0.016; HGGs,95.74%﹥91.49%, P=0.011);(GTR: overall,69.60%﹥47.7%, P=0.001; LGGs,no difference, P=0.069; HGGs,75.5%﹥48.7%, P=0.008). For short-termlanguage function, AQ of HGGs in study group was higher than in control group(76.16﹥64.94, P=0.032), whereas no statistical difference in AQ was found foroverall and LGGs between two groups (overall, P=0.056; LGGs, P=0.910). Forlong-term language function, all patients in study group had higher AQ than thosein control group (overall,87.47﹥78.73, P=0.001; LGGs,91.22﹥84.20, P=0.003;HGGs,82.62﹥70.80, P=0.015). New aphasia occurred in34.8%of the controlgroup, whereas it occurred only in2.3%of the study group (P<0.001). In addition,patients with glioblastomas had a longer PFS in study group than in control group(median PFS,12.5﹥6.6months, P=0.003) with a HR value of0.3124.Furthermore, the use of iMRI and multimodal navigation was proved to be thesingle independent beneficial factor for progression free survival (HR,0.325; 95%CI,0.157-0.677; P=0.003). In the third part, first iMRI helped revise4(7.8%)incorrect biopsy site, and finial iMRI confirmed biopsy accuracy in all cases.Postoperative diagnostic rate was96.1%(49/51) without any postoperativeneurological deficits, despite28(54.9%) lesions were adjacent to eloquent areas.Additionally, iMRI detected2(3.9%) intraoperative hematomas that were treatedimmediately. Owing to brain biopsy,38(58.8%) avoided an unnecessarycraniotomy. Conclusions Diffusion tensor imaging-based tractography is areliable and accurate technique to reconstruct and localize arcuate fasciculus,which is useful for prediction of postoperative language function. The use of iMRIand multimodal navigation in surgical resection of gliomas involving languageareas can maximize safe tumor resection and minimize long-term languagedeficits. Furthermore, it enhanced progression free survival in patients withglioblastomas involving language areas. Brian biopsy with iMRI and multimodalnavigation is a safe, novel and efficient biopsy modality. This technique may helpincrease the biopsy accuracy with low morbidity and mortality.