Reducing the acquisition time of whole-body fluorine-18 fluorodeoxyglucose positron emission tomography (18F-FDG PET) (corrected for attenuation) is of major importance in clinical practice. With the introduction of lutetium oxyorthosilicate (LSO), the acquisition time can be dramatically reduced, provided that patients are injected with larger amounts of tracer and/or the system is operated in 3D mode. The aim of this study was to determine the optimal dose of 18F-FDG required in order to achieve good-to-excellent image quality when a "3-min emission, 2-min transmission/bed position" protocol is used for an LSO PET camera. A total of 218 consecutive whole-body 18F-FDG PET studies were evaluated retrospectively. After excluding patients with liver metastases, hyperglycaemia and paravenous injections, the final study population consisted of 186 subjects (112 men, 74 women, age 59 +/- 15 years). Patients were injected with an activity of 18F-FDG ranging from 2.23 to 15.21 MBq/kg. Whole-body images corrected for attenuation (3 min emission, 2 min transmission/bed position) were acquired with an LSO PET camera (Ecat Accel, Siemens) 60 min after tracer administration. Patients were positioned with their arms along the body. Image reconstruction was done iteratively and a post-reconstruction filter was applied. Image quality was scored visually by two independent observers using a five-point scoring scale (poor, reasonable, good, very good, excellent). In addition, the coefficient of variability (COV) was measured in a region of interest over the liver in order to quantify noise. Of the images obtained in 118 patients injected with > or =8 MBq/kg 18F-FDG, 92% and 90% were classified as good, very good or excellent by observer 1 and observer 2, respectively. The COV averaged 10.63% +/- 3.19% for doses > or =8 MBq/kg and 16.46% +/- 5.14% for doses <8 MBq/kg. Administration of an 18F-FDG dose of > or =8 MBq/kg results in images of good to excellent quality in the vast majority of patients when using an LSO PET camera and applying a 3-min emission, 2-min transmission/bed position acquisition protocol. At lower doses, a rapid decline in image quality and increasing noise are observed. Alternative protocols should be adopted in order to compensate for the loss in image quality when doses <8 MBq/kg are used.