Background: Many surgeons advocate the use of neoadjuvant treatment for resectable pancreatic cancer, however little is known about variation in the utilization of neoadjuvant therapy (NAT) at the hospital level.
Methods: The National Cancer Data Base was used to identify patients undergoing resection for pancreatic cancer between 2006 and 2014 at high-volume centers. Hospitals were grouped by NAT utilization using standard deviations (SD) from the mean as follows: high neoadjuvant utilizers (> 2 SDs above the mean, > 40% of patients receiving NAT); medium-high (1-2 SDs, 27-40%), medium (0-1 SD, 14-26%); or low (- 1.1 to 0 SDs, < 14%). Overall survival (OS) was compared across NAT utilization groups.
Results: Among 107 high-volume centers, 20,119 patients underwent resection. The proportion of patients receiving NAT varied widely among hospitals, ranging from 0 to 74%, with only five centers using NAT in > 40% of patients. These five hospitals had the longest median OS at 28.9 months, compared with 21.1 months for low neoadjuvant utilizers (p < 0.001). On multivariable analysis, high and medium-high NAT utilization predicted improved OS, with a hazard ratio (HR) of 0.68 (95% confidence interval [CI] 0.56-0.83, p < 0.001) and 0.80 (95% CI 0.68-0.95, p = 0.010), respectively, compared with low utilizers. After excluding patients who underwent NAT, there remained an association of improved OS with high NAT utilization (HR 0.74, 95% CI 0.60-0.93, p = 0.009).
Conclusion: High-volume hospitals that more commonly utilize NAT demonstrated longer survival for all patients treated at those centers. In addition to altering patient selection for surgery, high neoadjuvant utilization may be a marker of institutional factors that contribute to improved outcomes.