Our aim was to determine the diagnostic limitations of low-dose, unenhanced CT scans performed for anatomic reference and attenuation correction during PET/CT.The Radiology Information System at our oncologic hospital was queried during the 9-mo period from July 2002 to April 2003 for patients with PET/CT scans and diagnostic enhanced CT within 2 wk of each other. One radiologist interpreted the CT portion of the PET/CT (CT(p)) unaware of the PET results and the associated enhanced diagnostic CT (CT(d)). A medical student compared this interpretation with the official report of the CT(d) and listed all discrepancies between reports. A separate radiologist compared CT(p) and CT(d) images and classified true discrepant findings as due to lack of intravenous contrast, arm-position artifact, lack of enteric contrast, low milliamperage (mA), and quality of lung images.Among 100 patients, the most common malignancies were lymphoma (n = 37), cancer of the colorectum (n = 31), and esophageal cancer (n = 15). Among 194 true discrepancies in which findings were missed at CT(p), causes were as follows: (a) lack of intravenous contrast (128/194, 66%), (b) arm-down artifact (17/194, 9%), (c) quality of lung images (26/194, 13%), (d) lack of enteric contrast (15/194, 8%), and (e) low mA (8/194, 4%). Discrepancies were seen most commonly in detecting lymphadenopathy and visceral metastases.Most missed findings on the unenhanced CT portion of the PET/CT scans were due to technical factors that could be altered. Discrepant findings would have led to altered management in only 2 patients, suggesting a role for limited repeat imaging to reduce radiation and use of valuable resources.