The aim of this study was to retrospectively compare the value of integrated PET/CT and separate PET plus morphological imaging studies for lesion localisation in cancer patients. Two different series of consecutive patients who had previously been treated for neoplastic disease were considered. One series consisted of 105 patients who had undergone [(18)F]fluorodeoxyglucose (FDG) PET/CT ( n=70) or [(11)C]choline PET/CT ( n=35) studies (PET/CT group). The other series comprised 105 patients who had undergone FDG PET scan ( n=70) or [(11)C]choline PET scan ( n=35) alone; in this series, PET findings were correlated with the results of morphological imaging (MI) studies, i.e. CT ( n=92) or MR imaging ( n=13) (PET+MI group). Regions of abnormal tracer uptake at PET scanning were classified as ambiguous or unambiguous depending on their precise anatomical localisation. A total of 207 and 196 lesions were found in the PET/CT and PET+MI groups, respectively. The difference in terms of number of lesions per patient detected with the two imaging protocols was not statistically significant ( P=0.718). When analysis of lesion localisation was performed, there were 7/207 (3.4%) and 30/196 (15.3%) ambiguous lesions in the PET/CT and PET+MI groups, respectively. The number of ambiguous lesions was significantly higher in the PET+MI group than in the PET/CT group (chi(2)=15.768, P<0.0001). Comparison of the effect of use of the different tracers on reporting of PET/CT versus PET+MI revealed that the improvement in the final report in [(11)C]choline PET/CT studies was similar to that observed in [(18)F]FDG studies. In cancer patients, PET/CT shows higher diagnostic accuracy for lesion localisation than PET plus morphological imaging studies performed independently. This result does not seem to be affected by the type of tracer used.