Intraoperative MRI And Diffusion Tensor Tractography Guided Surgical Resection Of Cerebral Lesions Involving Arcuate Fasciculus

Objective: To evaluate the impact of intraoperative magnetic resonance imaging(iMRI) and multimodal navigation on extent of resection, gross total resectionand preservation of language function. To assess prediction of aphasia bydiffusion tensor imaging in lesions involving arcuate fasciculus and impact oflanguage function on health related quality of life.Methods:65patients with cerebral lesions involving arcuate fasciculus (AF)were prospectively enrolled. Impact of iMRI on surgical strategies, extent ofresection and rate of gross total resection (GTR) were analyzed. To evaluate thevalidation of prediction of postoperative aphasia, a Receiver OperatingCharacteristic (ROC) curve was used to analyze relationship between AFvolume and language function scores. Influencing factors for preservation of AFwere also analyzed. Short Form36Health Survey (SF-36) was used to assesshealth related quality of life.Results: Extent of resection was overestimated in12cases (18.5%) by theneurosurgeons, compared with the first iMRI. The1stiMRI confirmed totalresection of non-glioma lesions. In54glioma cases, first iMRI demonstratedgross total resection in21patients and incomplete resection in33cases. In8outof the33cases, further resection was abandoned for the risk of postoperativelanguage dysfunction, because the residuals were close to AF. Further resectionwas continued with updated imaging in25patients. Final iMRI confirmed grosstotal resection in17cases and subtotal or partial in8, increasing the rate of grosstotal resection in glioma cases from38.9%to70.4%. Additional tumor resection in25cases increased extent of resection from87.3%to97.4%, resulting inimprovement of EOR in glioma cases from92.1%to96.8%. Compared withpreoperative language function status,15cases (23.1%) got new or worsenedlanguage dysfunction2weeks after surgery. At6-month follow-up, only2patients had a permanent postoperative language deficit. Non-glioma patientsgot no new or worsened aphasia. The mean distance between lesions and AFwas4.6mm in54glioma cases. Two weeks after surgery,14out of33patientswith distance less than5mm got new or worsened language deficit, while1hadworsened deficit in16cases with diatance between5and10mm. Distance andlesions were influencing factors for preservation of AF. The Spearmancorrelation analysis showed significant correlations of language deficits with AFinjury fraction (p<0.001). For ROC curve analysis, using variation ratio oflanguage function grade5as true positive, the area under the curve of AF was0.93. Using variation ratio of language function grade1as true negative, thearea under the curve of AF was0.90. Compared with1-month follow-up, theMental Component Summary (MCS) scores of the Short Form36Health Surveyimproved in7patients, while MCS deteriorated in2patients (1patient less than5points) at6-month follow-up. The Physical Component Summary scoresimproved in8patients with only1worsened.Conclusion: iMRI and DTI tractography help maximal resection of cerebrallesions involving AF while minimum injury to language function, improvinghealth related quality of life. DTI AF tractography is accurate in prediction ofpostoperative language deficits.