The main aim of the study was to evaluate the use of positron emission tomography using fluoro-deoxyglucose (PET-FDG) imaging for the detection of squamous cell carcinoma of the head and neck. Fifty-four consecutive patients with malignancies involving the head and neck were studied prospectively. Thirty-one patients presented with primary disease and 23 were suspected of recurrent or residual disease. All patients underwent full clinical staging, PET-FDG scans and anatomical imaging, 37 underwent computed tomography (CT), 13 magnetic resonance (MR) and four had both CT and MR. Clinical assessment, CT/MR, PET-FDG and histological examination were all evaluated independently of each other. All 31 primary head and neck malignant tumours were detected by PET-FDG. Based on 16 patients who underwent neck dissections, the sensitivity and specificity of PET-FDG for detecting nodal disease was 67% and 100% respectively, compared with clinical assessment of 58% and 75% and CT/MR of 67% and 25%. In all 12 patients, PET-FDG correctly identified the presence of absence or recurrent or residual disease. PET-FDG staged 13 post-treatment necks with an accuracy of 100% as compared to CT/MR which was accurate in 7 of 13 and clinical assessment which was accurate in eight. Three sites of abnormal tracer uptake unrelated to malignancy were recorded as incidental findings (mandibular osteomyelitis, 1: post glossectomy site, 2). PET-FDG was more accurate than CT/MR for identifying primary and recurrent tumours as well as metastatic lesions in the neck. If these diagnostic properties of PET-FDG are confirmed in further prospective studies, it could prove a valuable adjunct for the management of head and neck cancer.