ObjectiveCurrently, there are no studies showing which neoadjuvant therapy modality can provide better prognosis for patients after pancreatic cancer surgery. This study explores the optimal neoadjuvant therapy model by comparing the survival differences between patients with non-metastatic pancreatic cancer (cT1-4N0-1M0) who received neoadjuvant chemotherapy (NACT) and neoadjuvant chemoradiotherapy (NARCT).MethodsWe retrospectively analyzed the clinical data of 723 patients with cT1-4N0-1M0 pancreatic cancer who received neoadjuvant therapy before surgery from the Surveillance, Epidemiology, and End Results (SEER) database. After propensity score matching (PSM), we compared the effects of NACT and NARCT on overall survival (OS) and cancer-specific survival (CSS) in patients with non-metastatic pancreatic cancer, and then performed subgroup analyze. Finally, we used univariate and multivariate Cox regression analysis to explore potential risk factors for OS and CSS in patients with non-metastatic pancreatic cancer treated with preoperative neoadjuvant therapy.ResultBefore PSM, mOS (30.0 months VS 26.0 months, P=0.122) and mCSS (30.0 months VS 26.0 months, P=0.117) were better in patients with non-metastatic pancreatic cancer treated with NACT compared with NARCT, but this was not statistically significant (P>0.05). After PSM, mOS (30.0 months VS 25.0 months, P=0.032) and mCSS (33.0 months VS 26.0 months, P=0.028) were better in patients with non-metastatic pancreatic cancer treated with NACT compared with NARCT, and this difference was statistically significant (P<0.05). Multivariate Cox regression analysis results showed that age, lymph node positivity, and NARCT were independent adverse prognostic factors for OS and CSS in patients with non-metastatic pancreatic cancer.ConclusionThe study results show that compared with NARCT, NACT is the best preoperative neoadjuvant therapy mode for patients with non-metastatic pancreatic cancer. This result still needs to be confirmed by more prospective randomized controlled trials.