A quantitative multiparametric 18 F-FPIA PET/MRI study for the characterization of primary brain gliomas

10 patients with primary brain gliomas were recruited to this study. Static and dynamic 18 F-FPIA PET, together with perfusion/diffusion MRI data were post-processed for the extraction of 3D parametric maps for each subject. Correlations among parameters were evaluated with Spearman test. Tumour grade prediction was assessed with a machine learning model. A strong correlation was found between uptake and influx rate constant of FPIA and MRI perfusion parameters. The PET/MRI methodology provided 100% accuracy in differentiating low from high grade tumours.