Background: Stereotactic body radiotherapy (SBRT) is the standard of care in medically inoperable early-stage non-small cell lung cancer (NSCLC). Assessment of FDG-PET/CT before and after SBRT may stratify risk of disease recurrence and survival outcomes. Methods: Patients with T1-2N0M0 NSCLC who underwent PET/CT prior to SBRT (50-60 Gy over 3-5 fractions) between 2012 and 2019 were retrospectively identified. Pre-SBRT SUVmax and change in SUVmax at 3 and 6 months after SBRT were assessed as predictors of local control (LC), progression-free survival (PFS), and overall survival (OS). Optimal cutoff points for comparison were determined by receiver operator characteristic (ROC) analysis. Survival analyses were performed with Kaplan-Meier estimates with log rank testing, and Cox proportional hazards models including age, sex, T stage, histology, and performance status. Results: Out of 163 patients identified, 71 (43.6%) underwent repeat PET/CT within 6 months of SBRT completion. Median follow-up was 19 months (range 1 – 94 months). For the whole cohort, 1-year and 2-year LC, PFS, and OS were 95.0% and 80.3%, 75.9% and 47.7%, and 87.1% and 67.0%, respectively. Pre-SBRT SUVmax greater than 12.3 had an aHR of 2.80 (95% CI 1.3 – 6.2, p = 0.011) for PFS. A cutpoint of 12.6 for pre-SBRT SUVmax had an aHR of 3.00 (95% CI 1.6 – 5.8, p = 0.003) for OS. Pre-SBRT SUVmax did not significantly predict LC. A 3-month SUVmax decrease of at least 45% was associated with improved LC (aHR = 0.15, 95% CI 0.02 – 0.91, p = 0.018). At 6 months following SBRT, a cutoff point of a 53% decrease in SUVmax was associated with better LC (p = 0.038). Change in SUVmax was not significantly associated with PFS or OS at either time point. Performance status significantly predicted PFS and OS in all models. No other factors were significant. Conclusions: Pre-treatment SUVmax cutoffs can predict PFS and OS in early-stage NSCLC. At both the 3- and 6-month time points following SBRT, cutoff values for change in SUVmax can stratify risk of local recurrence.