¹⁸ F-FD

ObjectiveThe diagnosis and therapeutic monitoring is challenging for central nerve system germ cell tumors. PET/CT provides functional images of targeted organs or lesions and has been widely used in diagnosis and evaluation of most tumors. However, the application of PET/CT in central nerve system germ cell tumors has not been well studied. This study was prospectively designed to evaluate18F-FDG PET/CT and68Ga-TATE PET/CT in detecting and monitoring central nerve system germ cell tumors.MethodsThirty-eight patients underwent baseline18F-FDG PET/CT. Among which,21patients repeated the scans4weeks after chemotherapy and/or radiotherapy. The lesion detectability was compared with that of MRI. The lesion SUVmax (L-SUVmax), white matter SUVmean (WM-SUVmean) and cortex SUVmean (C-SUVmean) were measured and lesion to white matter ratio (SUVL/WM) and lesion to cortex ratio (SUVL/C) were calculated. Pearson correlation coefficient was used to compare the correlations between Ki67index and the SUV parameters (L-SUVmax, SUVL/WM, SUVL/C). Paired t-test was performed to compare the baseline and post-therapy SUV parameters. The expression of glucose transporter1(GLUT1) was determined by immunohistochemical staining and mean integrated optical density (IOD). Pearson correlation coefficient was used to compare the correlation between SUV parameters and mean IOD. Nineteen patients underwent68Ga-TATE PET/CT, and8of them repeated the scans4weeks after therapy. L-SUVmax was measured and paired t-test was performed to compare baseline L-SUVmax and follow-up L-SUVmax. The expression of somatostatin receptor type2(SSTR2) was determined by immunohistochemical staining and mean IOD was correlated with SUVmax using Pearson correlation coefficient.Results1.18F-FDG PET/CT detected67lesions in total, located in suprasellar region (57.8%), pineal region (7.9%), basal ganglia (7.9%) and multiple location (23.7%), including bifocal lesions (18.4%). Compared with MRI,18F-FDG PET/CT detected11more lesions, most of which were implantation metastasis in brain ventricles, while1brain ventricle lesion and1pituitary stalk lesion were missed. Baseline L-SUVmax, SUVL/WM and SUVL/C were5.5±3.6,2.92±1.16, and0.93±0.55, respectively, which was strongly correlated with Ki67index (P<0.001in all; γpearson was0.660,0.687and0.703, respectively). 2. The post-treatment scans showed significant reductions in all SUV parameters, with the L-SUVmax, SUVL/WM and SUVL/c reduced to3.2±1.8,1.47±0.65and0.50±0.21, respectively (P<0.001in all).3. Nineteen patients underwent68Ga-TATE PET/CT and8patients underwent post-therapy scans.68Ga-TATE PET/CT revealed71.4%lesions compared with18F-FDG PET/CT but the lesion images were clearer and more definite. The initial L-SUVmax was2.5±1.8and the post-therapy values significantly reduced to0.3±0.44weeks after the therapy (P=0.019).4. The over expression of GLUT1and SSTR2was observed in biopsy tissues but there was no obvious correlation between mean IOD and SUV parameters (P>0.05).Conclusions1.18F-FDG and68Ga-TATE PET/CT are useful tools in diagnosis of CNS germ cell tumors.18F-FDG PET/CT is more senstive of implantation metastasis in brain ventricles and reflects tumor viability, while68Ga-TATE PET/CT defines the lesions with better clarity and with no interference of backgroud uptake.2. The changes of L-SUVmax, SUVuwm and SUVL/C in18F-FDG PET/CT may be effective markers for treatment evaluation, and thereby provides quantified standards compared with MRI.3. The L-SUVmax, SUVL/WM and SUVL/C are strongly correlated with the Ki67index, and thereby reflect the proliferation activity of CNS germ cell tumors, which can be indications for prognosis.4. GLUT1and SSTR2are overexpressed in CNS germ cells but show no correlation with SUV parameters. This result may be explained by the bias of lesion size and the mixed tumor components.