Background: The optimal multidisciplinary treatment algorithm for pancreatic ductal adenocarcinoma (PDAC) is not well established. We studied outcomes in non-metastatic PDAC treated with chemotherapy and stereotactic body radiation therapy (SBRT) with or without pancreatic resection. Methods: Between August 2011 and July 2015, 73 patients with non-metastatic PDAC were treated with chemotherapy and SBRT with or without resection. Variables considered included: ECOG, CA 19-9, clinical stage, vascular involvement, pathologic stage and margin status. Resectability status was determined by an expert abdominal radiologist on initial staging imaging. Chemotherapy included FOLFIRINOX, FOLFOX, gemcitabine monotherapy (gem) or with nab-paclitaxel (gem/nab). SBRT was delivered as 30 Gy in 3 fractions over 5 days via CyberKnife. The Kaplan-Meier method and log-rank test were used to compare median overall survival (mOS), local progression free survival (LPFS) and metastasis free survival (MFS). Results: After a median follow-up of 19.3 months (mo) the mOS was 30.0 mo (95% CI, 19.4 to 36.5). The surgical group had longer mOS (36.5 vs 19.4 mo; P < 0.001), LPFS (29.0 vs 16.3 mo; P = 0.03) and MFS (29.0 vs 15.1 mo; P < 0.001) as compared to the nonsurgical group. FOLFIRINOX or gem/nab was associated with better mOS as compared to other chemotherapy (33.2 vs 12.8 mo; P < 0.001). There was a trend towards longer mOS in pts with initial imaging deemed less resectable (36.5 vs 30.1 vs 13.0 mo in unresectable vs borderline vs resectable; P = 0.19). In a multivariate analysis significant predictors of OS were resection, ECOG and FOLFIRINOX or gem/nab chemotherapy. Conclusions: Patients who had surgery after neoadjuvant chemotherapy and SBRT had significantly longer mOS, LPFS and MFS than those without surgery. FOLFIRINOX or gem/nab and better ECOG were also associated with improved outcomes. Worse resectablity status per imaging was associated with longer OS despite less likelihood of surgery; This may reflect more intensive neoadjuvant therapy or suggest that radiologic resectability status is a poor predictor of OS. Further investigation of factors underlying this discrepancy may have implications on neoadjuvant strategy and resectability determination.