Neoadjuvant therapy for pancreatic cancer: an intention-to-treat analysis.

PURPOSE: This study aimed to reassess the duration of neoadjuvant therapy (NAT) for patients with borderline resectable pancreatic cancer (BRPC). METHODS: The medical records of patients with BRPC who received NAT before intended curative resection were retrospectively reviewed. Patient demographics, clinicopathological factors, and prognostic factors for overall survival were analyzed. The serum carbohydrate antigen (CA) 19-9 level was examined monthly during NAT. RESULTS: A total of 118 patients with BRPC were enrolled. The median survival time and 5-year overall survival were 28.0 months and 31%, respectively. Three months after NAT, the CA19-9 levels were normal in 57% of the patients, and 92% underwent resection. Multivariate analysis showed that radiological partial response (hazard ratio (HR), 0.53; 95% confidence interval (CI), 0.26-0.99; p = 0.047); a normal CA19-9 level after NAT (HR, 0.30; 95% CI, 0.22-0.66; p = 0.006); and tumor resection (HR, 0.29; 95% CI, 0.13-0.67; p = 0.005) were independent predictors of better survival. The median CA19-9 level and the rate of normal CA19-9 levels before and after NAT were 256 (interquartile range (IQR), 23-1197) U/mL and 33%, and 27 (IQR, 7-176) U/mL and 57%, respectively. CONCLUSION: A normal CA19-9 level after NAT was an independent predictor of better survival in patients with BRPC. A longer NAT duration might contribute to improved prognosis of patients with elevated CA19-9 levels.