Enhanced Fiber Tractography in Glioblastoma Using a Novel Edema Correction Algorithm Correlated With Direct Intraoperative Electrical Stimulation and fMRI. (16th November 2020)

• Record Type: Journal Article
• Authors: Henderson, Fraser; Parker, Drew; Wolf, Ronald; Hodges, Wes; Elliott, Mark; Lucas, Timothy; McGarvey, Michael; Vijayakumari, Anupa; Harsch, Jessica; Maloney-Wilensky, Eileen; Brem, Steven; Verma, Ragini
• Is Part Of: Neurosurgery. Volume 67(2010)Supplement 1
• Publication Date: 2020-11-16
• Subjects: Nervous system -- Surgery -- Periodicals; 617.48005
• DOI: 10.1093/neuros/nyaa447_802
• Languages: English
• ISSNs: 0148-396X

Abstract

Abstract: INTRODUCTION: Fiber-tracking aids neurosurgical planning by delineating peritumoral eloquent tracts. The most commonly used FDA-approved solutions employ deterministic tracking strategies based on diffusion tensor imaging (DTI) acquisitions. This strategy can be confounded by edema. Resultantly, underlying white matter tracts may erroneously appear truncated on tractography maps which may contribute to a suboptimal surgical plan. METHODS: Edema correction was performed on clinically-acquired DTI data, by applying a two-compartment modeling algorithm (FERNET), followed by deterministic fiber-tracking using the corrected DTI dataset. Results were compared to tracking with the original data prior to correction visually and quantitatively using metrics mean track count, voxel count, and mean track length. Ten patients presenting with glioblastoma and peritumoral edema in proximity to and/or infiltrating language (5 patients) and motor (5 patients) pathways were included. Preoperative fMRI and intraoperative DES were available in each case for correlation. RESULTS: Twenty-five tracts were analyzed before and after free-water correction. Mean track count increased to 441 from 337 ( P = .017) with the application of edema invariance algorithm. Voxel count, a proxy measurement for tract volume, increased to 2266 versus 1382 ( P = .00005). Mean track length increased to 100 mm versus 91mm ( P = .0007). Standard fractional anisotropy across all tracts decreased to 0.46 from 0.50 ( P = .00004). Language patients in particular showed a clear increase in visible tracts. Overlay with functional activation maps confirms increased number of visualized fibers near functional locations. Correlation with DES also demonstrated increased number of visualized fibers resulting based on the corrected model. CONCLUSION: Edema correction in standard clinical DTI results in improved visualization of motor and language tracts in patients with glioblastoma-associated peritumoral edema, informing pre-surgical planning for optimal tumor resection.