Postoperative results, learning curve, and outcomes of pancreatectomy with arterial resection: a single-center retrospective cohort study on 236 procedures.

BACKGROUND:Newer chemotherapy regimens are reviving the role of pancreatectomy with arterial resection (PAR) in locally advanced pancreatic cancer. However, concerns about the early outcomes and learning curve of PAR remain. This study aimed to define the postoperative results and learning curve of PAR and provide preliminary data on oncologic outcomes. MATERIALS AND METHODS:A single center's experiences (1993-2023) were retrospectively analyzed to define the postoperative outcomes and learning curve of PAR. Oncologic results were also reported. RESULTS:During the study period 236 patients underwent PAR. Eighty PAR (33.9%) were performed until 2012, and 156 were performed thereafter (66.1%). Pancreatic cancer was diagnosed histologically in 183 patients (77.5%). Induction therapy was delivered to 18 of these patients (31.0%) in the early experience and to 101 patients (80.8%) in the last decade (P<0.0001). The superior mesenteric artery (PAR-SMA), celiac trunk/hepatic artery (PAR-CT/HA), superior mesenteric/portal vein, and inferior vena cava were resected in 95 (40.7%), 138 (59.2%), 189 (80.1%), and 9 (3.8%) patients, respectively. Total gastrectomy was performed in 35 (18.5%) patients. The thirty-day mortality rate was 7.2% and ninety-day mortality rate was 9.7%. The learning curve for mortality was 106 PAR (16.0% vs. 4.6%; odds ratio, OR=0.25 [0.10-0.67], P=0.0055). Comparison between the PAR-SMA and PAR-CT/HA groups showed no differences in severe postoperative complications (25.3% vs. 20.6%), 90-day mortality (12.6% vs. 7.8%), and median overall survival. Vascular invasion was confirmed in 123 patients (67.2%). The median number (interquartile range) of examined lymph nodes was 60.5 (41.3-83) and rate of R0 resection was 66.1% (121/183). Median overall survival for PAR was 20.9 (12.5-42.8) months, for PAR-SMA was 20.2 (14.4-44) months, and for PAR-CT/HA was 20.2 (11.4-42.7). Long-term prognosis improved by study decade (1993-2002: 12.0 [5.4-25.9] months, 2003-2012: 15.1 [9.8-23.4] months, and 2013-present: 26.2 [14.3-51.5] months; P<0.0001). CONCLUSIONS:In recent times, PAR is associated with improved outcomes despite a steep learning curve. Pancreatic surgeons should be prepared to face the technical challenge posed by PAR.