Evaluation of automatic multimodality fusion technique of PET and MRI/CT images for computer assisted brain tumor surgery

Research output: Contribution to journalResearch articleContributedpeer-review

Contributors

  • S. B. Sobottka - , Department of Neurosurgery, TUD Dresden University of Technology (Author)
  • R. Steinmeier - , TUD Dresden University of Technology (Author)
  • B. Beuthien-Baumann - , TUD Dresden University of Technology (Author)
  • D. Mucha - , TUD Dresden University of Technology (Author)
  • G. Schackert - , TUD Dresden University of Technology (Author)

Abstract

Positron emission tomography (PET) can provide spatial information on metabolic activity in patients with cerebral glioma and functional data in patients with lesions closely related to eloquent brain areas. Because of the limited image resolution of PET, coregistration with anatomic images such as MRI or CT is required to apply PET data for image guided surgery. The accuracy and clinical value of a novel image-fusion technique were evaluated for various PET, MRI and CT imaging modalities using an anatomic brain phantom and clinical patient images. For metabolic studies 2-[18F]-2-desoxy-D-glucose (FDG) and synthetic amino acid PET were obtained in four glioma patients. For mapping cortical functions, such as motor or speech function, specific activation tasks were used during PET acquisition in eight patients with lesions in eloquent brain areas. The PET images were automatically matched and merged with T1-, T2-weighted MRI or CT scans by rigid transformation using an automatic image fusion algorithm (BrainLAB, @Target 1.18). The intensity-based image-fusion technique is robust and capable of registering and fusing various medical image sets regardless of imaging modality and scan orientation. The brain phantom study demonstrated an excellent accuracy for rotate-translate registrations between PET versus T1-, T2-weighted MRI and CT. Excellent spatial correlation between PET and MRI was achieved in all patients during image-guided surgery using a frameless navigation system. By exactly transferring the PET data to the operative site, metabolic data can be correlated with corresponding tumor specimens and functional imaging data with intraoperative brain mapping findings.

Details

Original languageEnglish
Pages (from-to)261-267
Number of pages7
JournalInternational Congress Series
Volume1230
Issue numberC
Publication statusPublished - 1 Jun 2001
Peer-reviewedYes

Keywords

ASJC Scopus subject areas

Keywords

  • Functional brain imaging, Functional brain mapping, Glioma resection, Image fusion, Image guided surgery, Speech monitoring