Objective To explore the clinical efficacy of radical resection with individualized surgical approach for borderline resectable pancreatic head carcinoma. Methods The retrospective descriptive study was conducted. The clinicopathological data of 54 patients with borderline resectable pancreatic head carcinoma who underwent radical resection with individualized surgical approach in the West China Hospital of Sichuan University from January 2015 to January 2018 were collected. There were 37 males and 17 females, aged from 37 to 73 years, with a median age of 59 years. For venous type borderline resectable pancreatic head carcinoma, surgery for pancreatic head carcinoma and (or) pancreatic head and neck carcinoma was performed via inferior mesenteric vein, and surgery for pancreatic uncinate process carcinoma was performed via inferior colon artery. For arterial type borderline resectable pancreatic head carcinoma, surgery for pancreatic head carcinoma and (or) pancreatic head and neck carcinoma was performed via medial uncinate artery, and surgery for pancreatic uncinate process carcinoma was performed via left posterior artery. Observation indicators: (1) surgical situations; (2) postoperative complications; (3) postoperative pathological examination; (4) follow-up. Patients were followed up by outpatient examination or telephone interview once every 3 months to detect survival up to March 2019. Measurement data with normal distribution were represented by Mean±SD. Measurement data with skewed distribution were represented by M (range), and count data were represented by absolute numbers or percentage. Kaplan-meier method was used to draw the survival curve and calculate the survival rate. Results (1) Surgical situations: all the 54 patients underwent expanded pancreatoduodenectomy combined with superior mesenteric vein/portal vein (SMV/PV) resection, including 15 via inferior mesenteric vein, 20 via inferior colon artery, 12 via medial uncinate artery, and 7 via left posterior artery. The operation time was (320±83)minutes, and the volume of intraoperative blood loss was (865±512)mL. (2) Postoperative complications: of 54 cases, 28 had postoperative complications, including 13 with grade 1 Clavien-Dindo complications, 12 with grade 2 Clavien-Dindo complications, 3 with grade 3 or above Clavien-Dindo complications. One of the 28 patients with postoperative complications died and 27 were improved after symptomatic and supportive treatment. (3) Postoperative pathological examination: of 54 patients, 31 had R0 resection and 23 had R1 resection. In the 23 patients with R1 resection, 5 underwent surgery via the inferior mesenteric vein (4 with involvement of pancreatic anterior surface, 1 with involvement of both pancreatic anterior and posterior surface), 9 underwent surgery via the inferior colon artery (2 with involvement of both pancreatic anterior and posterior surface, 2 with involvement of superior mesenteric artery margin, 2 with involvement of pancreatic posterior surface, 2 with involvement of pancreatic anterior surface, 1 with involvement of superior mesenteric artery margin and pancreatic posterior surface), 5 underwent surgery via the medial uncinate process artery (2 with involvement of superior mesenteric artery margin, 2 with involvement of both pancreatic anterior and posterior surface, 1 with involvement of pancreatic neck transected margin), and 4 underwent surgery via the left posterior artery (3 with involvement of superior mesenteric artery margin, 1 with involvement of both pancreatic anterior and posterior surface). Of 54 patients, 16 had no positive lymph nodes, 26 had 1-3 positive lymph nodes, and 12 had 4 or more positive lymph nodes. The tumor diameter was (3.20±0.14)cm. There were 48 of 54 patients with nerve infiltration, 41 with superior mesenteric vein and/or portal vein infiltration, and 11 with vascular thrombus. There were 17 of 54 patients with high differentiation and medium differentiation, and 37 with low differentiation and undifferentiation. (4) Follow-up: 54 patients were followed up for 1-42 months, with a median time of 19 months. The 1-, 3-year overall survival rate was 78.0%, 11.4%. Conclusion As for the borderline resectable pancreatic head cancer, individualized and customized surgical approach according to the location of tumor and the relationship with blood vessels is helpful to standardize the radical resection and avoid R2 resection. Key words: Pancreatic neoplasms; Pancreatic cancer; Borderline resectable; pancreaticoduodenectomy; Surgical approach; R2 resection; Efficacy