2-[18F]FELP, a novel LAT1-specific PET tracer, for the discrimination between glioblastoma, radiation necrosis and inflammation

Abstract Introduction Considering the need for rapid change of treatment in recurrent glioblastoma (GB), it is of utmost importance to characterize PET radiopharmaceuticals that allow early discrimination of tumor from therapy-related effects. In this study, we examined the value of 2-[18F]FELP as a LAT1 tumor-specific PET tracer in comparison with [18F]FDG and [18F]FET in a combined orthotopic rat radiation necrosis and glioblastoma model. A second experiment compared 2-[18F]FELP to [18F]FDG in a mouse glioblastoma - inflammation model. Methods Using the small animal radiation research platform (SARRP), radiation necrosis (RN) was induced in the left frontal lobe of the rat brain. When radiation-induced changes were visible on MRI, F98 rat glioblastoma cells were stereotactically inoculated in the contralateral right frontal lobe. When tumor growth was confirmed on MRI, 2-[18F]FELP, [18F]FET and [18F]FDG PET scans were acquired on three consecutive days. In an inflammation experiment, mice were inoculated in the left thigh with U87 human glioblastoma cells. After heterotopic tumor growth was confirmed macroscopically, inflammation was induced by injection of turpentine subcutaneously in the right thigh. Subsequently, 2-[18F]FELP and [18F]FDG scans were acquired on two consecutive days. Results The in vivo PET images demonstrated that 2-[18F]FELP could differentiate glioblastoma and radiation necrosis using SUVmean (p = 0.0016) and LNRmean (p = 0.009), while [18F]FET was only able to differentiate both lesions by means of the SUVmean. (p = 0.047) Delayed [18F]FDGlate PET (4 h postinjection) was also able to distinguish glioblastoma from radiation necrosis, but smaller lesion-to-normal brain ratios were observed (SUVmean: p = 0.009; LNRmean: p = 0.028). In the inflammation study, 2-[18F]FELP showed no significant uptake in the inflammation lesion when compared to the control group (SUVmean: p = 0.149; LNRmean: p = 0.083). In contrast, both conventional and delayed [18F]FDG displayed significant uptake in the turpentine-invoked lesion (SUVmean: p = 0.021; LNRmean: p = 0.021). Conclusion This study suggests that the 2-[18F]FELP PET is able to differentiate glioblastoma from radiation necrosis and that the 2-[18F]FELP uptake is less likely to be contaminated by the presence of inflammation than the [18F]FDG signal. Advances in knowledge These results are clinically relevant for the differential diagnosis between tumor and radiation necrosis because radiation necrosis always contains a certain amount of inflammatory cells. Hence, 2-[18F]FELP is preferred to discriminate tumor from radiation necrosis.