Computerised tomography (CT) is a useful adjunct in the diagnosis of cervical pathology in head and neck carcinoma. However, different criteria used in CT analysis make the comparison of outcome studies difficult. This study compares two different CT criteria (two different lymph node sizes) for detection of cervical lymph node pathology for head and neck tumours in the same patient population. A prospective study was performed on 55 patients, with minimum T2 intraoral squamous cell carcinoma and high risk extraoral malignancies, who underwent 65 neck dissections. Sensitivity, specificity and accuracy of clinical palpation, CT considering lymph nodes larger than 10 mm as positive and CT considering lymph nodes larger than 15 mm as positive are compared. The results are as follows: sensitivity, specificity and accuracy of clinical palpation are 86, 84 and 85%, respectively; sensitivity, specificity and accuracy of CT with lymph nodes exceeding 10 mm are 95, 47 and 63%, respectively; sensitivity, specificity and accuracy of CT with lymph nodes exceeding 15 mm are 86, 81 and 83%, respectively. The outcomes of the study reveal that CT evaluation of cervical lymph nodes with a cutoff value of 15 mm should be preferred to a cutoff value of 10 mm in order to prevent over-treatment in cervical management of head and neck tumours. When accuracy of the results is considered, CT atone does not have an advantage over clinical palpation in this study. (c) 2007 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier Ltd. All rights reserved.