The development of ultra-high field MRI guidance technology for neuronavigation

Aaron E. Rusheen, Abhinav Goyal, Robert L. Owen, Elise M. Berning, Dane T. Bothun, Rachel E. Giblon, Charles D. Blaha, Kirk M. Welker, John Huston, Kevin E. Bennet, Yoonbae Oh, Andrew J. Fagan, Kendall H Lee

Research output: Contribution to journalArticlepeer-review

Abstract

OBJECTIVE Magnetic resonance imaging at 7T offers improved image spatial and contrast resolution for visualization of small brain nuclei targeted in neuromodulation. However, greater image geometric distortion and a lack of compatible instrumentation preclude implementation. In this report, the authors detail the development of a stereotactic image localizer and accompanying imaging sequences designed to mitigate geometric distortion, enabling accurate image registration and surgical planning of basal ganglia nuclei. METHODS Magnetization-prepared rapid acquisition with gradient echo (MPRAGE), fast gray matter acquisition T1 inversion recovery (FGATIR), T2-weighted, and T2*-weighted sequences were optimized for 7T in 9 human subjects to visualize basal ganglia nuclei, minimize image distortion, and maximize target contrast-to-noise and signal-to-noise ratios. Extracranial spatial distortions were mapped to develop a skull-contoured image localizer embedded with spherical silicone fiducials for improved MR image registration and target guidance. Surgical plan accuracy testing was initially performed in a custom-developed MRI phantom (n = 5 phantom studies) and finally in a human trial. RESULTS MPRAGE and T2*-weighted sequences had the best measures among global measures of image quality (3.8/4, p < 0.0001; and 3.7/4, p = 0.0002, respectively). Among basal ganglia nuclei, FGATIR outperformed MPRAGE for globus pallidus externus (GPe) visualization (2.67/4 vs 1.78/4, p = 0.008), and FGATIR, T2-weighted imaging, and T2*-weighted imaging outperformed MPRAGE for substantia nigra visualization (1.44/4 vs 2.56/4, p = 0.04; vs 2.56/4, p = 0.04; vs 2.67/4, p = 0.003). Extracranial distortion was lower in the head's midregion compared with the base and apex ((equation presented)1.17-1.33 mm; MPRAGE and FGATIR, p < 0.0001; T2-weighted imaging, p > 0.05; and T2*-weighted imaging, p = 0.013). Fiducial placement on the localizer in low distortion areas improved image registration (fiducial registration error, (equation presented)0.79-1.19 mm; p < 0.0001) and targeting accuracy (target registration error, (equation presented)0.60-1.09 mm; p = 0.04). Custom surgical software and the refined image localizer enabled successful surgical planning in a human trial (fiducial registration error = 1.0 mm). CONCLUSIONS A skull-contoured image localizer that accounts for image distortion is necessary to enable high-accuracy 7T imaging-guided targeting for surgical neuromodulation. These results may enable improved clinical efficacy for the treatment of neurological disease.

Original languageEnglish (US)
Pages (from-to)1265-1277
Number of pages13
JournalJournal of neurosurgery
Volume137
Issue number5
DOIs
StatePublished - Nov 2022

Keywords

  • deep brain stimulation
  • functional neurosurgery
  • image distortion
  • image registration
  • MRI-guided therapy
  • stereotactic engineering
  • surgical technique
  • ultra-high field MRI

ASJC Scopus subject areas

  • Surgery
  • Clinical Neurology

Fingerprint

Dive into the research topics of 'The development of ultra-high field MRI guidance technology for neuronavigation'. Together they form a unique fingerprint.

Cite this