Characterization of a glioma mouse model using multiparametric [18F]FET-PET-MRI

Ziel/Aim:

Glioblastoma (GBM) is the most aggressive malignant tumor of the brain with remarkably reduced survival. Many animal tumor models have been designed to understand glioma development. PET-MRI allows the evaluation of tumor progression and intrinsic tumor tissue characteristics in vivoand longitudinally. Understanding the dependencies between multiple imaging parameters provides useful information regarding early therapy response and the time-dependent changes in the tumor. Here, we modeled GBM in mice and longitudinally characterized the tumors using [¹⁸F]FET-PET-MRI.

Methodik/Methods:

Mice (n = 8) intracranially induced with Glioblastom were measured weekly using [¹⁸F]FET-PET-MRI (6 weeks). The acquisition protocol consisted of T1-weighted images pre- and post-administration of Gadovist, T2-weighted images, Apparent Diffusion Coefficient (ADC) maps and dynamic (45 min) [¹⁸F]FET-PET. Images were analyzed using Pmod software. Regions of interest were drawn on the tumor and overlaid on the coregistered PET/MRI.

Ergebnisse/Results:

We successfully identified tumors 4 weeks post-injection. Tumor volumes increased in the following 2 weeks. There was a consistent concentration Gadovist increment that correlated with tumor growth and time. Likewise, % ID/cc of [¹⁸F]FET in the tumor increased in time. Deposition of [¹⁸F]FET was also confirmed by autoradiography. Histology of the surviving animals showed permeability through identification of IgG inside tumor vessels.

Schlussfolgerungen/Conclusions:

We evaluated development of GBM using PET-MRI. The consistent increment in volume starting at 4 weeks, increasing deposition of gadolinium and [¹⁸F]FET and the histological results, indicate that the model is well vascularized and highly permeable. Further voxelwise analysis (including ADC maps) will likely permit further evaluation of different tissue populations in the tumor. The similarity of this model with the highly vascularized human GBM shows potential for machine learning evaluations.