INTRODUCTION: This study examined rates of tumor progression in treatment-naive patients with non-small-cell lung cancer (NSCLC) as determined by repeat treatment-planning fluorine-18 ((18)F) fluorodeoxyglucose positron emission tomography/computed tomography ((18)F-FDG PET/CT). METHODS AND MATERIALS: This study assessed patients who underwent PET/CT simulation for NSCLC stage II/III, radiation-naive, nonmetastatic NSCLC. It compared planning PET/CT with previous PET/CT images. Patients were analyzed for change in stage, treatment intent, or both. Progression was defined as a change in TNM status leading to upstaging, and standardized uptake value (SUV) velocity was defined as [(SUVscan2 - SUVscan1)/interscan interval in days]. RESULTS: Of 149 consecutive patients examined between April 2009 and April 2011, 47 had prior PET/CT scans and were included. The median age was 68 years. New nodal disease or metastatic disease was identified in 24 (51%) of 47 patients. Fourteen (30%) had evidence of extrathoracic metastatic disease; the remaining 10 (21%) had new nodal disease that required substantial alteration of treatment fields. At a scan interval of 20 days, the rate of upstaging was 17%. SUV velocity was analyzed in the subset of patients who had their studies on the identical PET/CT scanner (n = 14). Nonupstaged patients had a mean SUV velocity of 0.074 units per day, compared with 0.11 units per day in patients that were upstaged by their second PET/CT scan (P = .020). CONCLUSION: Radiation treatment planning with hybrid PET/CT scans repeated within 120 days of an initial staging PET/CT scan identified significant upstaging in more than half of patients. For a subset of patients who underwent both scans on the same instrument, SUV velocity predicts upstaging, and the difference between those upstaged and those not was statistically significant.