Pre-operative image-based segmentation of the cranial nerves and blood vessels in microvascular decompression: Can we prevent unnecessary explorations?

• Published in: Clinical neurology and neurosurgery Vol. 139; pp. 159 - 165
• Main Authors: Dolati, Parviz, Golby, Alexandra, Eichberg, Daniel, Abolfotoh, Mohamad, Dunn, Ian F., Mukundan, Srinivasan, Hulou, Mohamed M., Al-Mefty, Ossama
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.12.2015; Elsevier Limited
• Subjects: Accuracy; Aged; Compression therapy; Cranial nerve segmentation; Cranial Nerves - pathology; Craniotomy; Endoscopic and microscopic view; Endoscopy; Female; Hemifacial spasm; Hemifacial Spasm - diagnosis; Hemifacial Spasm - surgery; Humans; Imaging, Three-Dimensional - methods; Magnetic Resonance Imaging - methods; Male; Microvascular Decompression Surgery - methods; Middle Aged; Neuroendoscopy - methods; Neurology; Neurosurgery; Pain; Patients; Sinuses; Trigeminal neuralgia; Trigeminal Neuralgia - diagnosis; Trigeminal Neuralgia - surgery; Veins & arteries; Hemifacial spasm; Trigeminal neuralgia; Cranial nerve segmentation; Endoscopic and microscopic view
• ISSN: 0303-8467; 1872-6968; 1872-6968
• DOI: 10.1016/j.clineuro.2015.10.006

•We validated the accuracy of image-based segmentation in neurovascular conflict cases.•Measurement accuracy was 0.45±0.21 mm (mean±SD).•Image-based segmentation is highly informative preoperatively and intraoperatively.•Larger studies are needed to prove segmentation prevents unnecessary craniotomies. This study was conducted to validate the accuracy of image-based pre-operative segmentation using the gold standard endoscopic and microscopic findings for localization and pre-operative diagnosis of the offensive vessel. Fourteen TN and 6 HS cases were randomly selected. All patients had 3T MRI, which included thin-sectioned 3D space T2, 3D Time of Flight and MPRAGE Sequences. Imaging sequences were loaded in BrainLab iPlanNet and fused. Individual segmentation of the affected cranial nerves and the compressing vascular structure was performed by a neurosurgeon, and the results were compared with the microscopic and endoscopic findings by two blinded neurosurgeons. For each case, at least three neurovascular landmarks were targeted. Each segmented neurovascular element was validated by manual placement of the navigation probe over each target, and errors of localization were measured in mm. All patients underwent retro-sigmoid craniotomy and MVD using both microscope and endoscope. Based on image segmentation, the compressing vessel was identified in all cases except one, which was also negative intraoperatively. Perfect correspondence was found between image-based segmentation and endoscopic and microscopic images and videos (Dice coefficient of 1). Measurement accuracy was 0.45±0.21mm (mean±SD). Image-based segmentation is a promising method for pre-operative identification and localization of offending blood vessels causing HFS and TN. Using this method may prevent some unnecessary explorations on especially atypical cases with no vascular contacts. However, negative pre-operative image segmentation may not preclude one from exploration in classic cases of TN or HFS. A multicenter study with larger number of cases is recommended.