Abstract
3142
Introduction: Primary mediastinal germ cell tumor (MGCT) is a rare disease, which represents about 2%–4% of all germ cell tumors (GCT). Although MGCTs share similar histology with gonadal GCTs, the prognosis of MGCT is generally worse than that of gonadal GCT. Recently, the role of 18F-fluorodeoxyglucose positron emission tomography/CT (18F-FDG PET/CT) has been widely investigated for the diagnosis and prognostic stratification of cancer. However, 18F-FDG PET/CT-related studies on primary MGCT are rare and mainly are comprised of case reports or series. In this study we aimed to retrospectively review the distinctive visual characteristic and quantitative parameters of 18F-FDG PET/CT image for primary MGCT according to pathologic subtypes.
Methods: We retrospectively evaluated primary MGCT patients who underwent pre-operative 18F-FDG PET/CT between 2010 and 2020 at Asan medical center. MGCTs included four histologic types and were divided into two groups (benign and malignant) for analysis. Visual assessment was performed by categorizing the uptake intensity (as grade 0-3), uptake pattern (as equivocal/homogenous/heterogeneous), and contour (as round/lobulated/infiltrative) of the primary mass. 18F-FDG PET/CT quantitative parameters including maximum standardized uptake value (SUVmax), tumor-to-background ratio (TBR), metabolic tumor volume (MTV, segmented with a threshold of the relative value of more than 50% of SUVmax by VOI), total lesion glycolysis (TLG, SUV mean multiplied by MTV), and maximum diameter values were compared between benign versus malignant MGCT. Receiver operating characteristic curve (ROC) analysis was used to evaluate the diagnostic performance of PET/CT variables in differentiating malignant from benign MGCT. In addition, subgroup analysis between seminoma versus nonseminomatous germ cell tumor (NSGCT) and according to high versus low level of tumor markers was performed.
Results: A total of 35 patients with 24 mature teratomas, 4 seminomas, 5 yolk cell tumors, and 2 mixed GCTs were included. When compared with the benign MGCT group, the malignant group showed a significantly younger age distribution (p<0.001). All 11 patients in the malignant group were male. In visual analysis, all 6 cases of grade 0 among 35 patients were teratomas but none of the teratomas showed grade 3 uptake. All the malignant GCT groups showed uptake with either grade 2 (82%, 9/11) or 3 (18%, 2/11). Most of the MGCT showed a heterogeneous uptake pattern (83%, 29/35). Most of the MGCT showed either round (74% 24/35) or lobulated contour (23%, 8/35). In quantitative analysis, all the PET/CT parameters including SUVmax, TBR, MTV, and TLG showed significantly higher value in the malignant MGCT group than those in the benign MGCT group. In ROC curve analysis, SUVmax (Area Under Curve [AUC] = 0.947, p < 0.0001), TBR (AUC = 0.917, p < 0.0001), MTV (AUC =0.727, p = 0.0198), and TLG (AUC = 0.920, p < 0.0001) showed excellent diagnostic performance in discriminating between benign and malignant MGCT. Especially, SUVmax demonstrated a significantly higher diagnostic value compared to MTV and maximum diameter (p = 0.0254 and 0.0114, respectively). With an optimal cut-off value of SUVmax 4.54, sensitivity and specificity for differentiating malignant from benign MGCT were 81.8% and 100%, respectively. In the subgroup analysis of differentiating between seminoma and NSGCT among malignant MGCTs, SUVmax, TBR, and maximum diameter showed significance (p = 0.042, 0.042, and 0.012, respectively). A high level of alpha fetoprotein was correlated with a higher value of SUVmax, TBR, maximum diameter (p = 0.012, 0.034, 0.044, respectively), and no significant difference was found according to human chorionic gonadotrophin level.
Conclusions: Visual assessment of MGCT on 18F-FDG PET/CT showed discriminative findings between benign and malignant MGCT. 18F-FDG PET/CT quantitative parameters give additional information in differentiating benign versus malignant MGCT and seminoma versus NSGCTs.