The aim of this study was to assess the usefulness of fluorine-18 fluoro-2-deoxy- D-glucose positron emission tomography (FDG PET) in the post-therapy surveillance of endometrial carcinomas. Forty-one fully corrected whole-body PET studies were performed in 34 women with previously treated endometrial cancers as a part of their follow-up programme. In 28 studies, FDG PET was indicated to localise a recurrence suspected at the control visits on the basis of clinical examination and/or radiological abnormalities (chest X-ray, CT or MRI) and/or elevated tumour marker levels (CA125, CEA). Another 13 studies were performed as a simple surveillance procedure. Overall, in 26 studies PET detected recurrent disease, which was confirmed either by histology ( n=7) or by clinical and radiological outcomes ( n=19). In 88% of the cases, the PET findings confirmed recurrence suggested by routine follow-up. In the remaining 12% of cases, PET detected asymptomatic recurrences that were unsuspected at the control visits. Whole-body PET accurately localised the site of confirmed recurrences as being above and below the diaphragm in 50%, only below the diaphragm in 35% and only above the diaphragm in 15%. In one patient, however, PET missed microscopic lung metastases shown on thoracic CT, and in three studies, metabolic imaging results were not confirmed. In 11 of 12 negative PET studies, no subsequent clinical or radiological recurrences were observed with a median follow-up of 10 months. Overall, the results of PET agreed well with the final diagnosis (Cohen's kappa coefficient =0.78). In 9/26 patients (35%) with confirmed recurrences, the PET findings significantly altered the treatment choice by detecting either clinically or radiologically unsuspected distant metastases. The sensitivity, specificity, diagnostic accuracy and positive and negative predictive values of FDG PET imaging in the post-therapy surveillance of endometrial carcinomas were 96%, 78%, 90%, 89% and 91%, respectively. Indeed, the high likelihood ratio for a positive test result (4.5) and the low likelihood ratio for a negative test result (0.05) demonstrated the clinical utility of metabolic imaging in "ruling in" disease as well as "ruling out" recurrence. In conclusion, whole-body FDG PET appears useful in the post-therapy surveillance of endometrial cancers, both for the accurate localisation of suspected recurrences and for the detection of asymptomatic recurrent disease.