In order to determine the efficacy and proper timing of routine PET scans for surveillance of recurrent head and neck squamous cell carcinoma (HNSCC), we evaluated the diagnostic performance of routine PET scans in relation to time interval from completion of treatment. Amongst 206 retrospectively evaluated post-treatment PET scans of 159 patients with HNSCC, 156 were performed for routine surveillance in subclinical cases. Diagnostic performance of PET scan and follow-up outcome were evaluated in relation to the time interval (2-6months, 6-12months, 12-24months, and >24months) of PET scan from the completion of treatment. Overall sensitivity and NPV of these PET scans for recurrence were 92.5% and 94.8%, compared with 55.0% and 76.9% for conventional evaluation methods. In the 156 routine scans, the diagnostic sensitivity, specificity, and NPV for locoregional recurrence were 90%, 91% and 97%, respectively, and the values for distant metastases and second primary cancers were 100%, 97% and 100%, respectively. The diagnostic accuracy of routine PET scans was not significantly altered by the time interval. Most (97%) of true negative cases on routine PET scans had no recurrence during a median 14months follow-up. PET scan may be a useful tool in routine surveillance for detection of recurrence in subclinical patients. For routine surveillance, the initial PET scan should be performed within 6months after completion of treatment and the proper timing of next routine PET scan for subclinical patient with initial negative PET result might be 1year after initial PET scan.