S. Kulshrestha1,4, P. J. Sweigert1, C. Tonelli1,6, C. Bunn1, F. Luchette1,6, T. M. Pawlik5, M. S. Baker1,6 1Loyola University Chicago Stritch School Of Medicine,Department Of Surgery,Maywood, IL, USA 4Loyola University Chicago Stritch School Of Medicine,Burn And Shock Trauma Research Institute,Maywood, IL, USA 5Ohio State University,Department Of Surgery,Columbus, OH, USA 6Edward J. Hines VA Hospital,Department Of Surgery,Hines, IL, USA
Introduction: Advocacy groups continue to promote regionalization in pancreatic cancer care. There is little empiric evidence to suggest regionalization is resulting in improved clinical outcomes. We examine the effects of contemporary trends in regionalization on a composite metric of clinical quality for pancreaticoduodenectomy (PD), a textbook oncologic outcome (TOO).
Methods: We queried the National Cancer Database to identify patients presenting with resectable (clinical stage I and II) pancreatic adenocarcinoma undergoing PD between 2006 and 2015. TOO was defined as a margin negative resection with formal lymph node assessment, no prolonged hospitalization, readmission, or 30-day mortality, and receipt of adjuvant chemotherapy within 12 weeks of surgery. Facility volume was classified by annual number of pancreatectomies performed based on Leapfrog Group criteria: centers performing 20 pancreatectomies per year were designated as high volume; those performing 11-19 as moderate, 5-10 as low, <4 as very low volume.
Results: 16,394 patients met inclusion criteria. 9,514 (58.0%) underwent PD at high, 2,665 (16.2%) at moderate, 2,635 (16.0%) low and 1,580 (9.6%) very low volume centers. 2,643 (16.1%) patients had a TOO. The proportion of patients treated in high volume centers increased yearly from 46% in 2006 to 64% in 2015 (p<0.01). Annual rates of TOO increased each year from 12.4% in 2006 to 19.6% in 2015 (p<0.01). On univariate analysis, mean TOO rates correlated directly with surgical volume (7.4% at very low facilities vs. 20.0% at high volume facilities (p<0.01). On mixed effects multivariable regression adjusting for patient, tumor, and facility characteristics surgical volume was consistently associated with rates of TOO. Very low (OR 0.35, 95%CI [0.27,0.45]), low (OR 0.47, 95%CI [0.39,0.57]), and moderate (OR 0.74, 95%CI [0.63,0.87]) volume centers all demonstrated lower adjusted odds of achieving TOO than high volume centers (all p<0.01). Positive regional lymph nodes (OR 1.38, 95%CI [1.23,1.54) and facility location in the East North Central United States (OR 1.50, 95%CI [1.11,2.02], p<0.01) were independently associated with increased odds of TOO while age over 75 (OR 0.44, 95%CI [0.37,0.52]) and Charlson comorbidity score >3 (OR 0.52, 95%CI [0.36,0.75]) were associated with lower odds of TOO (all p<0.05). Academic affiliation was not associated with TOO.
Conclusions: Rates of TOO in PD are relatively low across all thresholds of surgical volume. There is a direct association between volume and rate of TOO. Rates of TOO have increased over time as has the percentage of cases done in high volume centers. Such trends support continued efforts to promote regionalization in pancreatic cancer care.
