Comparison of pharmacokinetic parameters in dynamic 18F-FET PET of glioma

Objectives Glioma grading with dynamic 18F-FET PET is typically performed by qualitative analysis of tumoral 18F-FET uptake over time (increasing vs. decreasing shape of time-activity curves). The aim of this work was to investigate quantitative parameters derived from pharmacokinetic and pseudo-kinetic models for the differentiation of WHO grade 2, 3 and 4 glioma in dynamic 18F-FET PET (0-40 min p.i.).

Methods For each of the 39 patients with glioma (WHO2: 11; WHO3: 17; WHO4: 11) one blood volume-of-interest (VOI; 55% isocontour) and one tumor VOI (SUV &gt 1.8 x background) were segmented. Furthermore, within each tumor VOI, sub-VOIs including voxels with similar time-to-peak (TTP) were defined. Kinetic parameters for the entire tumor VOI and the TTP-dependent sub-VOIs were derived from Patlak-plot, RE-plot, 2- and 3-compartment models (2CM, 3CM), and a linear fit to the late frames (15-40 min).

Results The resulting parameters obtained for the entire tumor VOIs are summarized in the two tables. The differentiation between low grade (WHO grade 2) and high grade (WHO grade 3 and 4) glioma (p &lt 0.01) was successful with TTP, SUV at TTP, linear fit (slope and intercept), Patlak-plot (VT), RE-plot (VT and intercept), 2CM (K1 and k2), and 3CM (k4, K1/k2 and k3/k4). No significant differences were found for the remaining parameters. None of the parameters allowed for the differentiation of grade 3 and 4 glioma (p &gt 0.1). The analysis of each tumor’s TTP-dependent sub-VOIs showed that approx. 53% of the voxels of each high grade tumor volume had a maximum activity between 10 and 20 min p.i. (3CM: K1/k2=0.6, k3/k4=1.6), whereas only 15% of the voxels exhibited a peak activity after 30 min. p.i. (3CM: K1/k2=0.7, k3/k4=3.7). In the low grade glioma, 55% of the voxels presented their maximum activity later than 30 min. p.i. (3CM: K1/k2=0.4, k3/k4=3.3). However 38% of the voxels in low grade glioma revealed a peak between 15 and 30 min. p.i. (3CM: K1/k2=0.4, k3/k4=2.7).

Conclusions We found significant differences between high and low grade gliomas for various kinetic parameters. Notably, areas with early peak uptake could be detected with a voxel-based analysis within histologically proven WHO grade 2 gliomas, supporting the identification of tumor heterogeneities. In combination with other parameters, e.g. a high uptake rate, this observation might raise suspicion for the presence of aggressive sub-areas. Whether this may have an impact on clinical procedures and patient’s prognosis has to be further investigated in future studies.