6505 Background: The incidence of stage I and II lung cancer has been increasing due to the reported benefits of CT screening in high risk individuals. Guidelines suggest continued surveillance following diagnosis, however the extent to which this is done is unclear. We examined the patterns and predictors of surveillance imaging in patients following a diagnosis of stage I-II non-small cell lung cancer (NSCLC). Methods: The SEER-Medicare database was used to identify patients with stage I-II, NSCLC who underwent curative intent surgical resection from 2000-2006.Three surveillance periods (6-18, 18-30, 30-42 months) after diagnosis were identified. Use of imaging was quantified for each period and logistic regression models developed to examine predictors of thoracic imaging during all 3 surveillance periods. Results: Between 2000-2006 5269 patients were identified. During the first year of surveillance use of chest CT (48% in 2000 to 78% in 2006) and PET (2% to 23%) increased while use of bone scan (12% to 6%) and chest radiography (87% to 71%) decreased over time. Similar trends were noted for surveillance periods 2 and 3. During the study, surveillance imaging was performed in all 3 surveillance periods in 38% of the cohort. Yearly surveillance imaging increased with time; 21% of patients diagnosed in 2000 vs. 78% diagnosed in 2006 underwent testing in the first 3 years of follow-up (P<0.0001). In a multivariable model, in addition to year of diagnosis, patients with high socioeconomic status (OR=1.96; 95% CI, 1.51-2.55 compared to low SES) were more likely to undergo yearly surveillance imaging, while Hispanics (OR=0.42; 95% CI, 0.21-0.85) and older patients (OR=0.52; 95% CI, 0.43-0.63) were less likely to undergo testing. Conclusions: The use of costly diagnostic imaging of uncertain value is increasing rapidly for patients with localized NSCLC. Non-clinical factors including high socioeconomic status and white race are among the strongest predictors of use of high intensity surveillance.