An assessment of the impact of geographic variation on resection rates and survival for early-stage pancreatic adenocarcinoma.
e15052 Background: Pancreatic resection is the only known curative option for pancreatic adenocarcinoma. Resection has been previously reported to be underutilized in patients with early stage disease. To develop a better understanding of this issue and control for treatment selection factors, we examined the relationship between geographic area resection rates and survival in patients with stage I/II pancreatic cancer. Methods: We queried Surveillance, Epidemiology, and End Results (SEER) data for patients with stage I/II cancer of the pancreatic head diagnosed from 2004-2009. We excluded patients with less than 3mo survival. Resection rates were calculated within Health Service Areas (HSAs) across all 18 SEER regions. Resection rate was defined as the number of patients who had an operation divided by the total number diagnosed with early stage pancreatic cancer. Multivariate Cox regression was used to estimate the overall survival effect of HSA rates while controlling for age, gender, marital status, poverty level, education, and AJCC stage. Results: 8,323 patients with stage I (n=1,454) and stage II (n=6,869) disease were analyzed. Pancreatectomy was performed in 476 patients (32.7%) with stage I disease and 3,846 (56.0%) with stage II disease. HSA resection rates were arranged into five groups (quintiles) which ranged from 42.7 to 65.7% (Table). Across the quintiles, median overall survival increased from 11 to 14 months, suggesting a positive association with resection rate. Multivariate analysis revealed that for every 10.00% increase in resection rate, the risk of overall death decreased by 5.26% (p<0.001). Conclusions: Patients with early stage pancreatic cancer who live in areas with higher resection rates have longer average survival times. Because geography should not influence treatment response, we conclude that efforts to raise resection rates should increase survival times in patients for whom there is uncertainty about the risk/benefits of resection. [Table: see text]