Imaging Feature Analysis of Abnormal Uptake at Nasopharynx on 18F-FDG PET/CT from Two PET Centers in Sichuan and Guangdong Province

Purpose: The purpose of this study was to analyze the 18F-FDG PET/CT images and to compare the parameters between benign and malignant lesions of nasopharynx.

Methods: Our study retrospectively analyzed 263 patients with pathologically proven lesions from October 2010 to December 2017 from two PET centers in Sichuan and Guangdong province. Patients were divided into four groups: nasopharyngeal carcinoma, lymphoma of nasopharynx, nonspecific inflammation of nasopharynx and the other diseases. The group of the other diseases includes nasopharyngeal tuberculosis, Rosai-dorfman disease, nasopharyngeal melanoma and rhabdomyosarcoma. The nasopharynx’s SUVmax, SUVmin, SUVmean and metabolic tumor volume and the cervical lymph nodes’ SUVmax, SUVmin and SUVmean were measured. The SUVmax/SUVmean and SUVmax-SUVmin were used to evaluate the heterogeneity of lesions. In addition, we assessed the morphology of nasopharynx, the symmetry of the Waldeyer’s ring and cervical lymph nodes. Use one-way ANOVA test, Chi-Square test, Fisher’ exact test and Kruskal-Wallis Test to compare the differences in PET/CT parameters. Receiver-operating-characteristic curves were applied to evaluate the diagnostic efficacy.Results: The nasopharynx’s SUVmax and SUVmean of nasopharyngeal carcinoma were both higher than those of lymphoma of nasopharynx (P=0.007 and P=0.005) and nonspecific inflammation of nasopharynx (P<0.001 for both). Metabolic tumor volume showed no significant difference between nasopharyngeal carcinoma and lymphoma of nasopharynx. The heterogeneity of nasopharyngeal carcinoma was higher than lymphoma and inflammation of nasopharynx(P<0.001 for using SUVmax/SUVmean and SUVmax-SUVmin). Nasopharyngeal carcinoma showed symmetry of nasopharynx and Waldeyer’s ring less frequently than lymphoma of nasopharynx (P<0.001 and P=0.004, respectively). Skull base invasions in nasopharyngeal carcinoma were more than lymphoma of nasopharynx and nonspecific inflammation of nasopharynx (P<0.001 for both). SUVmax of nasopharynx in nasopharyngeal carcinoma were higher than SUVmax of cervical lymph nodes while the relations were inverse in lymphoma of nasopharynx and nonspecific inflammation of nasopharynx (P=0.003). Classified nasopharyngeal carcinoma, lymphoma of nasopharynx and nasopharyngeal melanoma as malignancy and the others were divided into benignancy. The SUVmax, SUVmean and MTV of malignancy were all higher than the benignancy (t=5.311,P<0.001; t=4.725,P<0.001; t=4.243,P<0.001, respectively). When SUVmax of 8.34 was used as cut off for differentiation of malignant from benign uptake in nasopharynx, the area under the ROC curve was 0.844±0.032 (95% confidence interval, 0.75-0.91), with a sensitivity of 73.3% and a specificity of 79.5%. Conclusions: The morphology of nasopharynx, the heterogeneity of nasopharyngeal lesions, the involvement of skull base, the Waldeyer’s ring, and the intensity of 18F-FDG uptake can provide reliable reference value to differentiate benign from malignant nasopharyngeal abnormal uptake.