Prognostic factors of multidisciplinary treatment for borderline resectable and locally advanced pancreatic adenocarcinoma: Results of a dual-center study.

680 Background: Multidisciplinary treatment of borderline resectable (BR)/unresectable locally advanced (UR-LA) pancreatic adenocarcinoma (PDAC) has not yet been established. The purpose of this study is to explore factors that improve prognosis in radical surgery after multidisciplinary treatment for pancreatic cancer. Methods: We evaluated the following prognostic factors in 240 PDAC patients who underwent radical resection after multidisciplinary treatment. Patients were classified into 3 groups according to NCCN guidelines (BR PDAC invading the portal vein (BR-PV), BR pancreatic cancer in contact with the major arteries such as the hepatic artery, celiac axis and superior mesenteric artery (BR-A), and UR-LA), and prognostic factors were investigated. Patients with BR PDAC were treated with chemotherapy followed by surgery, while radiation therapy was added preoperatively in most cases with arterial invasion. All patients with UR-LA underwent surgery after nab-paclitaxel plus gemcitabine (GnP) followed by chemoradiotherapy (CRT) with S-1. Results: BR-PV/BR-A/UR-LA patients were 88/111/41, respectively. Prognosis was significantly better in the NAT group than in the upfront surgery group for both BR-PV/A (P=0.004/<0.001). In univariate analysis of overall survival (OS) in 36 patients with BR-PV who underwent resection after NAT, the following factors were significantly favorable prognostic factors; tumor marker (TM) normalization (P=0.028), preoperative Glasgow prognostic score=0 (P=0.025), and preoperative prognostic nutritional index (PNI)>42.5 (P=0.022). In univariate analysis in 39 patients with BR-A, the following factors were significantly favorable prognostic factors; TM normalization (P=0.033), preoperative PNI>42.5 (P=0.013), intraoperative blood loss>830 ml (P=0.013). Multivariate analysis revealed that high preoperative PNI was an independent prognostic factor (hazard ratio 0.15 [0.02-0.85]; P=0.014) in BR-A patients. In patients with UR-LA who underwent radical resection after GnP and subsequent CRT, median duration of NAT was 8.8 months, and R0 resection was achieved in 36 patients (88%). 3-year OS was 77.4%, and 5-year OS 58.6%. Multivariate analysis revealed that CA19-9 normalization (hazard ratio 0.23 [0.02-0.88]; P=0.032) and PNI≥41.7 (HR 0.05 [0.01-0.62]; P=0.021) were independent prognostic factors. Conclusions: In both BR/UR-LA pancreatic cancer, normalization of TM and maintenance of good nutritional status during NAT until surgery may contribute to prolonged prognosis.