The purpose of this study was to evaluate the role of [18F] fluorodeoxyglucose positron emission tomography (FDG-PET) in influencing salvage treatment decision making for locally persistent nasopharyngeal carcinoma (NPC). A total of 33 NPC patients with histologic persistence at nasopharynx 1 to 6 weeks after a full course of radiotherapy underwent both computed tomography (CT) and FDG-PET/CT simulation at the same treatment position. The salvage treatment decisions, with regard to the decision to offer salvage treatment and the definition of gross tumor volume (GTV), were made before knowledge of the FDG-PET findings. Subsequently the salvage treatment decisions were made again based on the FDG-PET findings and compared with the pre-FDG-PET decisions. All 33 patients were referred for salvage treatment in the pre-FDG-PET decision. After knowledge of the FDG-PET results, the decision to offer salvage treatment was withdrawn in 4 of 33 patients (12.1%), as no abnormal uptake of FDG was found at nasopharynx. Spontaneous remission was observed in repeat biopsies and no local recurrence was found in these 4 cases. For the remaining 29 patients, GTV based on FDG-PET was smaller than GTV based on CT in 24 (82.8%) cases and was greater in 5 (17.2%) cases, respectively. The target volume had to be significantly modified in 9 of 29 patients (31%), as GTV based on FDG-PET images failed to be enclosed by the treated volume in the salvage treatment plan performed based on GTV based on CT simulation images. Use of FDG-PET was found to influence the salvage treatment decision making for locally persistent NPC by identifying patients who were not likely to benefit from additional treatment and by improving accuracy of GTV definition in salvage treatment planning.