Abstract
2071
Objectives Quantitative tumour perfusion values can be obtained using dynamic 15O-H2O PET studies. Up to now most data have been collected using standalone PET systems. These scanners were usually operated in 2D mode in case of high count rate 15O studies. Today, PET/CT scanners are produced exclusively and are operated in 3D mode only. The aim of this study is to compare tumour perfusion measured with a Gemini TF PET/CT system with those obtained using an ECAT Exact HR+ scanner.
Methods Dynamic 15O-H2O perfusion studies were performed in 4 patients with stage III/IV non-small cell lung cancer. For each subject 15O-H2O perfusion studies were acquired using a HR+ scanner in 2D mode and a PET/CT system in 3D mode, both prior as well as after initiation of therapy with gamma secretase inhibitors. Thus eight paired 15O-H2O perfusion studies were available. For the HR+ PET data tumour volumes of interests (VOI) were defined on a dynamic FDG study, collected directly after the perfusion study, while for the PET/CT data the CT images were used. On the CT images tumour borders were defined using a threshold set on -200 HU. Input curves were defined by drawing a VOI over the ascending aorta in a first-pass image. All VOI’s were then projected onto the dynamic 15O-H2O study to extract time activity curves (TAC). Perfusion values were obtained by fitting a single tissue compartment model to the measured data.
Results There is a significant correlation in perfusion values between most (n=7) paired studies showing a slope of 1.005 and an R2 of 0.8(intercept fixed to zero). For one subject a difference of more than 45% in perfusion was observed. This was likely due to differences in FDG versus CT based tumour delineation, mainly as a result of tracer uptake heterogeneity. Therefore this subject was excluded.
Conclusions 3D 15O-H2O tumour perfusion studies performed on the Gemini TF PET/CT system are quantitatively equivalent with those performed on an ECAT Exact HR+ system operated in 2D mode