RESUMEN
PURPOSE: In patients with metastatic nonseminomatous germ cell cancer (NSGCT), residual masses after chemotherapy (CTX) can consist of vital carcinoma, mature teratoma, or necrosis. This prospective trial has evaluated the accuracy of [(18)F]fluorodeoxyglucose positron emission tomography (FDG-PET) for the prediction of histology compared with computed tomography (CT) and serum tumor markers (STM). PATIENTS AND METHODS: A total of 121 patients with stage IIC or III NSGCT scheduled for secondary resection after cisplatin-based CTX were included. FDG-PET was performed after completion of CTX. All results were confirmed by histopathology and correlated to STM and CT. RESULTS: Prediction of tumor viability with FDG-PET was correct in 56%, which did not reach the expected clinically relevant level of 70%, and was not better than the accuracy of CT (55%) or STM (56%). Sensitivity and specificity of FDG-PET were 70% and 48%. The positive predictive values were not significantly different (55%, 61%, and 59% for CT, STM, and PET, respectively). Judging only vital carcinoma as a true malignant finding, the negative predictive value increased to 83% for FDG-PET. CONCLUSION: The presence of vital carcinoma and mature teratoma is common (55%) in residual masses in patients with NSGCT, and CT and STM cannot reliably predict absence of disease. In contrast to prior studies, this prospective trial, which is the only with histologic confirmation in all patients, demonstrated that FDG-PET is unable to give a clear additional clinical benefit to the standard diagnostic procedures, CT and STM, in the prediction of tumor viability in residual masses.