379 Navigated Intraoperative Ultrasound With Rigid Image Fusion Decreases Neuronavigation Inaccuracies Following Brain Shift

View at publisher Summarise with VIRAT
Authors: Timothy R. West, Mercy H. Mazurek, Nicole A Perez, Ethan A. Wetzel, Alexander Wang, Zsombor Tamas Gal, Bryan Choi, Brian Vala Nahed
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

INTRODUCTION: Neuronavigation is a fundamental tool in the neurosurgeon’s armamentarium to intraoperatively identify structural anatomy and delineate tumor borders, but it suffers from accuracy loss following tissue deformation and brain shift. Historically, neurosurgeons have mentally corrected for calibration inaccuracies. Navigated intraoperative ultrasound (NiUS) offers an efficient, real-time imaging modality to co-register live anatomy on the neuronavigation system. We applied a novel Rigid Image Fusion (RIF) software to re-calibrate neuronavigation by aligning anatomical landmarks on the pre-operative MRI with NiUS. METHODS: This is a retrospective analysis on prospectively collected NiUS images obtained in patients undergoing intra-axial tumor resection. NiUS was performed pre- and post-resection in all patients, and mid-resection when clinically indicated. Mismatch between equivalent anatomical landmarks identified on co-registered pre-operative MRI and NiUS was measured before and after RIF (SNAP-to-MRI, Brainlab AG, Munich). Target Registration Error (TRE) was calculated as difference in anatomical mismatches. TRE difference percentage was calculated as TRE/initial error and represents percent improvement. RESULTS: Seventy-six NiUS scans (n = 32 pre-resection; n = 12 mid-resection; n = 32 post-resection) were acquired from 31 patients. RIF was successfully applied to 71 scans with 5 failures on poor quality NiUS scans (93% RIF-success rate). Anatomical mismatch improved in 69 scans (97.2% RIF-improvement rate) and decreased from 5.58 ± 1.75 mm (pre-RIF) to 1.85 ± 1.75 mm (post-RIF) (p = 1.56e-12). The mean TRE difference was a 3.73 ± 3.76 mm correction. RIF worsened the registration in two mid-resection scans. When stratified by tumor location, anatomical mismatch significantly improved across all locations. CONCLUSIONS: RIF offers a reliable method to correct for neuronavigation inaccuracies following brain shift in intra-axial tumor resection. Co-registration of the pre-operative MRI to the real-time NiUS image may expand the utility of neuronavigation by preserving its accuracy throughout tumor resection.

Fields of Study & Subjects

Medical Image Segmentation Techniques Anatomy and Medical Technology Glioma Diagnosis and Treatment

Publication details

Journal / source
Neurosurgery
Volume
72
Issue
Supplement_1
Pages
104–104
Publication date
26 March 2026
Language
EN