Background: SBRT treatment for very elderly ( > 80 years) patients with early stage NSCLC has been reported to be well tolerated with good short term efficacy. Using a large hospital based registry, we report a comparison of patterns of practice, outcomes, and prognostic factors for very elderly patients undergoing any treatment for early-stage NSCLC. Methods: The NCDB was queried for patients with clinical Stage I-IIA NSCLC with age ≥ 80 years diagnosed from 2001-2015 treated with surgery or SBRT alone. Patients were excluded if they received chemotherapy /immunotherapy or non-standard SBRT doses (i.e. > 5 fractions of RT, < 30 Gy or > 70 Gy total dose). Survival analyses were performed with propensity-matching, Kaplan-Meier estimates, Cox proportional hazards regression, and log rank testing. Results: 26039 patients met search criteria, median age 83 (80-90) years. 17141 (65.8%) patients underwent surgery, and 8898 (34.2%) underwent SBRT. Median follow up was 31 months. Median survival was 52 and 35 months for surgery and SBRT. Of patients receiving SBRT, 2044 (23%) had a contraindication to primary surgery due to patient risk factors. Age, clinical stage, tumor size, surgery type, CDCC score, BED, bronchial involvement, and type of treatment facility were predictive of median survival. BED > 154 Gy was associated with greater median survival (p < 0.01). Lobectomy was associated with greater median survival vs sub-lobar resection/pneumonectomy (p < 0.0001). For stage I tumors, surgery was associated with better median survival (56 vs. 35 months, p < 0.0001), but for stage IIA patients both modalities had similar median survival (30 vs 29 months, p = 0.04). Conclusions: Surgery remains the predominant treatment modality for early stage NSCLC in this very elderly population, and is associated with good outcomes for patients with stage I tumors. For elderly patients who are poor surgical candidates due to medical co-morbidities SBRT is associated with reasonable median survival. With limited information on patient comorbidities, more robust studies are needed to determine the effects of patient selection on treatment outcomes in this population.