Clear guidelines for the investigation of occult malignancy after unprovoked venous thromboembolism are not yet available. (18)F-fluorodeoxyglucose ((18)F-FDG) PET/CT could serve as a comprehensive screening strategy for occult malignancy in this context. We aimed to compare a screening strategy based on (18)F-FDG PET/CT with a limited screening strategy for detection of malignant disease in patients with unprovoked venous thromboembolism. In an open-label, multicentre, randomised study we enrolled patients from four French university hospitals. Patients aged 18 years or older, diagnosed with unprovoked venous thromboembolism (not provoked by a major inherited or acquired risk factor) were invited to participate. Patients were randomly assigned in a 1:1 ratio to a limited screening strategy (physical examination, usual laboratory tests, and basic radiographs) or a screening strategy consisting of the limited strategy plus an (18)F-FDG PET/CT scan. Randomisation was done with a dedicated central web-based randomisation system, in block sizes of six, stratified by centre, and concealed from the investigators. Patients and investigators were not masked to study group assignment. Patients were followed up for 2 years. The primary outcome was the proportion of patients with a cancer diagnosis in each group after the initial screening assessment. Analyses were conducted in modified intention-to-test and per-protocol populations. This trial is completed and registered with ClinicalTrials.gov, number NCT00964275. Between March 3, 2009, and Aug 18, 2012, we enrolled and randomly assigned 399 patients; five withdrew consent, leaving 197 in each group for the modified intention-to-test analysis. After initial screening assessment, cancer was diagnosed in 11 (5·6%) patients in the (18)F-FDG PET/CT group and four (2·0%) patients in the limited screening group (absolute risk difference 3·6%, 95% CI -0·4 to 7·9; p=0·07). At the initial screening assessment, seven (64%) of the 11 cancers diagnosed in the (18)F-FDG PET/CT group were early-stage compared with two of four cancers diagnosed in the limited screening group (p=1·00). One (0·5%) occult malignancy was detected in 186 patients who had negative initial screening in the (18)F-FDG PET/CT group, compared with nine (4·7%) in 193 patients in the limited screening group (absolute risk difference 4·1%, 95% CI 0·8 to 8·4, p=0·01). Overall, five (42%) of the 12 cancers diagnosed in the (18)F-FDG PET/CT group were advanced stage, compared with seven (54%) of the 13 cancers diagnosed in the limited screening group (p=0·70). 16 patients died during follow-up, eight (4·1%) in each group. Two (1·0%) patients in the (18)F-FDG PET/CT group and five (2·5%) in the limited screening group had cancer-related deaths. A strategy including limited screening and a (18)F-FDG PET/CT was not associated with a significantly higher rate of cancer diagnosis after unprovoked venous thromboembolism. The risk of subsequent cancer diagnosis was, however, lower in patients who had negative initial screening that included (18)F-FDG PET/CT than in patients who had negative initial limited screening. Whether or not (18)F-FDG PET/CT might be useful in a more selected population of patients with a high risk of cancer remains to be determined. Programme Hospitalier de Recherche Clinique (French Department of Health).