Grading of Glioma: Comparison between18F-FMISO-PET, Apparent Diffusion Coefficient and Gadolinium Enhancement

398 Purpose: Pre-operative prediction of glioma grading is always challenging, and high accuracy for prediction is always preferable. Recently, classification of glioma has been updated to use integrated phenotypic and genotypic parameters for brain tumor classification. In this study, we investigated whether hypoxic tracer, 18F-fluoromisonidazole (FMISO), add the values in prediction of glioma grading compared with apparent diffusion coefficient (ADC) and gadolinium enhancement on MRI. Materials and Methods: Fifty-six patients with histologically proven cerebral glioma grade 2, 3, and 4 were finally included. FMISO were intravenously administered to them and PET/CT scan were performed 4 hours after the injection. PET images, ADC map images of MRI, and Gd-enhanced 3D T1-weighted images were anatomically co-registered to individual pre-contrast 3D T1-weighted images respectively by using SPM12. Mean and max value, tumor-normal ratio (TNR) of mean and max value were calculated for SUV of FMISO, ADC, and enhancement ratio (ER), respectively. ROC analysis for differentiating high-grade tumor (grade 3 or 4) from low-grade tumor (grade 2) was conducted between SUV, ADC, and ER. Results: FMISOmax_TNR showed the highest ROC values (0.946) among all variables. Enhancement_mean was high (0.926) among ADC paramters, ADCmin_TNR was high among ADC parameters (0.846). Combination of FMISO, ADC and ER showed ROC values of 0.975. Conclusions: FMISO PET added more diagnostic values in differentiation between high-grade glioma and low-grade glioma.