ABSTRACT   To evaluate the role of 18-fluoro-deoxy-d-glucose (FDG)-positron emission tomography (PET) scan for detecting distant metastases in the preoperative assessment of patients with recurrent head and neck squamous cell carcinoma (HNSCC).   Retrospective study.   University Teaching Hospital.   Thirty-seven consecutive patients who presented, between April 2008 and April 2010, a local and/or regional recurrence of head and neck squamous cell carcinoma after treatment with radio-chemotherapy were studied. The work-up included thoraco-abdominal computed tomography (CT), fluoro-deoxy-D-glucose-positron emission tomography scan and abdominal ultrasound. The imaging results, when positive, were compared to histology or cytology (conducted during targeted examinations, for example, fiberoptic oesophago-gastro-duodenal search, colonoscopy, bronchoscopy, liver biopsy) or targeted imaging examination (Abdominal MRI, sonography) combined with clinical follow-up. All patients were followed-up for at least 6 months. Positron emission tomography performances were then analysed and compared with those of conventional imaging for detecting distant metastases.   Among the 37 patients, 9 (24%) had visceral metastases. The sensitivity, specificity, positive predictive value and negative predictive value for detecting metastasis or second primary were, respectively, 100%, 94%, 86% and 100% for CT and 92%, 87%, 74% and 97% for positron emission tomography. Computed tomography and positron emission tomography were strictly concordant in 32/37 (86%) of cases. No false-negative result was found for CT, while we found one case of false-negative positron emission tomography. The number of false-positive results was two for CT and four for positron emission tomography.   From our study, positron emission tomography does not appear to offer a first-choice technique for the detection of metastases before salvage surgery as CT detected all lesions visible on positron emission tomography.