Evaluation of Diffuse Reflectance Spectroscopy and Fluorescence Spectroscopy for Detection of Glioma Brain Tumors

Abstract

Imaging instruments are required for accurate tumor resection during neurosurgery, especially in the case of glioblastoma multiforme (GBM) - the most common and aggressive malignant glioma. However, current intraoperative imaging techniques for detection of glioma either suffer low sensitivity and low specificity or require a significant capital cost. Advances in diffuse reflectance spectroscopy and fluorescence spectroscopy have offered high sensitivity and high specificity in differentiating tumors from normal tissues with much lower capital cost. Whereas diffuse reflectance spectroscopy alone and fluorescence spectroscopy alone has been used in limited studies to differentiate normal brain tissues from brain tumors with moderate sensitivity and specificity, low specificity and sensitivity were usually observed when studying high grade glioma (HGG) such as GBM. Furthermore, optical properties and diffuse reflectance signal of HGG and low grade glioma (LGG) have not been observed separately, and thus a relation between optical properties and glioma progression has not been established. Intraoperative differentiation of GBM and LGG can be helpful in making treatment plan at the first surgery. This thesis focuses on characterizing a previous integrated system of diffuse reflectance spectroscopy and fluorescence spectroscopy to extract optical properties and fluorescence properties of LGG and GBM. First, tissue-simulating phantom models were developed to calibrate the integrated system. The direct method and Mie theory were used to calculate optical scattering of the phantoms while Beer-Lambert’s law was used to calculate optical absorption. Second, an experimental method was introduced to recover intrinsic fluorescence because the measured fluorescence signal is likely distorted by the presence of scatterers and absorbers in tissue (i.e. hemoglobin). Third, an experimental method was developed to recover optical properties of both GBM and LGG. In addition, the sensitivity and specificity of the integrated system was optimized.