Relation between center volumes for pancreatic and esophageal cancer surgeries and outcome in Belgium: a plea for centralization

Date 22 October 2018
Event ESMO 2018 Congress
Session Proffered paper session - Public health policy
Topics Oesophageal Cancer
Bioethics, Legal, and Economic Issues
Surgical Oncology
Pancreatic Cancer
Presenter Liesbet Van Eycken
Citation Annals of Oncology (2018) 29 (suppl_8): viii562-viii575. 10.1093/annonc/mdy297
Authors H. De Schutter1, G. Silversmit1, K. Haustermans2
  • 1Research Department, Belgian Cancer Registry, 1210 - Brussels/BE
  • 2Department Of Radiation oncology, University Hospitals Leuven, 3000 - Leuven/BE



Known for its accessible health care, Belgium currently counts over 100 hospitals providing oncological care, resulting in shattered care which may negatively influence outcomes. In the context of recent centralization initiatives, this study aimed to evaluate relations between complex surgeries and outcomes for (peri)pancreatic and esophageal cancers at the Belgian population level.


All patients with (peri)pancreatic (ICD)10: C25, C17.0, C24.0-1) or esophageal (C15-C16.0) cancer between 2007 and 2014 were extracted from the Belgian Cancer Registry and linked with surgeries (Sx) from reimbursement data. Concordant with previous reports, three yearly volume categories of Sx per center were defined ((peri)pancreatic: <6, 6-14 and ≥15 Sx/yr; esophageal: <6, 6-19 and ≥20 Sx/yr). Relations between surgical volumes and 30-days postoperative mortality as well as 5-year overall survival (OS) were analyzed with multivariable regression models adjusting for case-mix (including age, stage, sex, comorbidities).


16,471 (peri)pancreatic and 12,241 esophageal cancers were retrieved, corresponding to 4,081 (peri)pancreatic and 3,387 esophageal cancer surgeries performed by 96 hospitals in total (in 2014: 68 hospitals for (peri)pancreatic and 54 for esophageal cancer Sx). Surgical volumes were significantly related with 30-days postoperative mortality for (peri)pancreatic and esophageal cancer (p = 0.005 and p < 0.0001 respectively), with 52% and 81% mortality reduction for high vs low volume hospitals, respectively. The volume effect was also seen for OS: for both cancer types, high volume hospitals had a better OS compared to low volume which remained significant after case mix adjustment. For (peri)pancreatic cancer, 1-yr and 5-yr OS for high versus low volume centers was 75% vs 69% and 34% vs 31%, respectively (HR 0.65 [0.55, 0.78], p < 0.0001). For esophageal cancer, 1-yr and 5-yr OS for high versus low volume centers was 79% vs 71% and 44% vs 38%, respectively (HR 0.88 [0.79;0.99], p = 0.04).


High surgical volume centers showed better results for postoperative mortality and survival for (peri)pancreatic and esophageal cancers, supporting centralization initiatives in Belgium.

Clinical trial identification

Legal entity responsible for the study

Foundation Belgian Cancer Registry.


Federal and Regional Authorities.

Editorial Acknowledgement


All authors have declared no conflicts of interest.