This study was performed to assess the potential benefit of positron emission tomography (PET) using fluorine-18 fluorodeoxyglucose (FDG) in patients with cervical metastasis of squamous cell carcinoma from an unknown primary tumour. Eighteen patients with cervical metastasis of squamous cell carcinoma from an unknown primary who were assessed by physical examination including transnasal fibre-endoscopy and radiological work-up with computed tomography were included in this prospective tertiary referral centre cohort study. The results of the search for the primary with rigid panendoscopy of the upper aerodigestive tract were compared to the evaluation with FDG PET. Panendoscopy revealed a primary tumour in 8/18 (44%) patients. PET accurately diagnosed five of these eight primary tumours, and gave one false positive and three false negative scans, resulting in a sensitivity of 63%, a specificity of 90%, an accuracy of 78%, a positive predictive value of 83% and a negative predictive value of 75%. Small primaries or primaries in areas with physiologically increased FDG uptake can be missed with PET owing to the limited resolution of the camera (approximately 5 mm). Our study in a small number of patients suggests that PET does not provide benefit in terms of detecting additional primary tumours if applied in addition to extensive clinical work-up. Considering its high specificity, PET could be of value as an initial evaluation instrument, reserving the need for extensive work-up to patients with negative scans.