NI-15 * USE OF VOLUMETRIC MR SPECTROSCOPIC IMAGING AND 5-ALA FLUORESCENCE-GUIDED SURGERY FOR RESECTION OF GLIOBLASTOMA

Abstract

RATIONALE: Glioblastoma (GBM) resection based on neuronavigation with contrast-enhanced MRI (CE-MRI) results in a high rate of local recurrence as infiltrating tumor cells extend beyond areas of contrast enhancement. Integrating metabolic maps from magnetic resonance spectroscopic imaging (MRSI) into neuronavigation may identify high-risk tumor infiltration zones outside of CE-MRI for surgery. Coupling MRSI with fluorescence-guided surgery (FGS) using 5-aminolevulinic acid (5-ALA) may provide a means to further enhance the degree of tumor resection. METHODS: In a new Phase II trial for new and recurrent GBM patients, 3D-echo planar spectroscopic imaging was performed and processed to give full-brain metabolite volumes for import into a neuronavigation station. Patients were administered 5-ALA orally 3-5 h before surgery. To minimize error due to shift, needle biopsies were collected from tumor regions with elevated choline/NAA values within T1 CE and/or T2/FLAIR hyperintense regions before tumor resection. Fluorescence intensity of fresh tissue was quantified ex vivo using a hand-held spectroscopic device. Samples were then fixed/stained with H&E and analyzed for tumor cell density and tissue infiltration using automated image analysis techniques. Quantitative tumor segmentation was used to evaluate extent of tumor resection. RESULTS: Choline/NAA values show a positive linear trend with the number of tumor nuclei per high power field and the degree of tumor infiltration. Real-time intraoperative tumor fluorescence was also confirmed ex vivo and shows a similar positive linear trend with choline/NAA ratios. CONCLUSIONS: The linear trend that Cho/NAA values exhibit with histopathology and quantitative intraoperative fluorescence of tumor tissue outside of CE-MRI regions supports its use for identifying regions of tumor infiltration. This is the first time that 5-ALA induced tumor fluorescence has been shown to correlate with MRSI-derived metabolic markers in GBM tumors. We believe the continued combination of MRSI-neuronavigation with 5-ALA FGS in this trial will result in more complete GBM resections.