Artikel
First Intraoperative Experiences with a Hyperspectral Imaging Setup for Cortical Brain Tumor Identification
Suche in Medline nach
Autoren
Veröffentlicht: | 9. Juni 2017 |
---|
Gliederung
Text
Objective: Besides accurate preoperative surgical planning based on MRI images, intraoperative information about tumor localization and brain tissue functionality is essential for the successful resection of brain tumors. Intraoperative Optical Imaging (IOI) is a technique that identifies areas of increased metabolism caused by stimulation of specific brain functions. IOI uses a special hardware setup that analyses selected wavelengths or broad band spectra (RGB) of the reflected light from the exposed cortex. Results after performing a stimulation protocol and complex image processing algorithms are two-dimensional navigated activity maps. IOI is already applicable in clinical routine use during neurosurgical interventions. Seeking for more information within the complete continuous light spectrum between 500 ... 1000 nm, a new Hyperspectral Imaging (HSI) setup was tested
Methods: An intraoperative usable hardware setup, based on a standard surgical microscope, a halogen light source and a hyperspectral imaging unit was built. The hyperspectral imaging system was attached to the surgical microscope using a standard beam splitter. Custom built illumination was coupled into the optical path of the surgical microscope via a fibre-optic. The image of a common RGB camera, likewise attached to the surgical microscope via beam splitter, was used to adjust focus and zoom. White and dark references were acquired before measurements on patients, that underwent resection of cortical tumors, were performed. A hyperspectral datacube, that includes the spectral information in its third dimension, was acquired and spectral characteristics of different tissue types of the cortical surface were compared.
Results: The evaluation of the hardware setup reveals that hyperspectral imaging can be integrated in a clinical setup without any remarkable changes. The additional hardware setup was easy to manage for the neurosurgeon. It was able to acquire robust hyperspectral data of the cortical surface during the neurosurgical procedure. Differences within the spectral characteristics of different tissue types were visible after postoperative evaluation of the acquired imaging data.
Conclusion: First results indicate the feasibility of the method: tissue differentiation with hyperspectral imaging is possible. Intraoperative hyperspectral imaging is able to extend optical imaging. Further work will focus on light source characteristics and on data classification algorithms.