The combination of 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) and computed tomography (CT) has been reported to be more accurate than CT or PET alone in a preoperative setting. We compared the diagnostic utility of preoperative PET/CT, PET and CT/MRI in 167 patients with newly diagnosed head and neck squamous cell carcinoma (HNSCC), of whom 104 underwent FDG PET and 63 underwent combined PET/CT with all receiving CT/MRI. These preoperative PET, PET/CT, and CT/MRI results were reviewed and their accuracies were compared in patients in whom diagnosis was confirmed histopathologically. Age, sex, primary sites and stage, and nodal involvement were comparable between two groups. The accuracy of PET and PET/CT for detecting primary tumors and cervical metastases was comparable, but significantly higher than that of CT/MRI (98%-97% vs. 86-88% for primary; 92%-93% vs. 85%-86% for neck on a level-by-level basis; P<.05). PET and PET/CT gave false negative results: in 2 (2%) and 2 (3%) patients for primary tumors; in 6 (6%) and 3 (5%) patients for neck metastases, respectively. PET and PET/CT also gave false-positive results for cervical metastases in 5 (5%) and 4 (6%) patients, respectively. Compared with PET alone, preoperative FDG PET/CT may not yield significantly improved diagnostic accuracy in patients with HNSCC. Moreover, despite their high accuracy, PET and PET/CT may not abrogate the need for conventional imaging and pathologic staging based on primary resection and neck dissection.