A Pretreatment Prognostic Score to Stratify Survival in Pancreatic Cancer

Objective To develop and validate a pretreatment prognostic score in pancreatic cancer (PDAC). Background Pretreatment prognostication in PDAC is important for treatment decisions but remains challenging. Available prognostic tools are derived from selected cohorts of patients who underwent resection, excluding up to 20% of patients with exploration only, and do not adequately reflect the pretreatment scenario. Methods Patients undergoing surgery for PDAC in Heidelberg from 07/2006 to 06/2014 were identified from a prospective database. Pretreatment parameters were extracted from the database and the laboratory information system. Parameters independently associated with overall survival by uni- and multivariable analyses were used to build a prognostic score. A contemporary cohort from Verona was used for external validation. Results In 1197 patients, multiple pretreatment parameters were associated with overall survival by univariable analyses. ASA-classification, CA19-9, CEA, CRP, albumin, and platelet count were independently associated with survival and were used to create the Heidelberg Prognostic Pancreatic Cancer (HELPP)-score. The HELPP-score was closely associated with overall survival (median survival between 31.3 and 4.8 months; 5-year survival rates between 35% and 0%) and was able to stratify survival in subgroups with or without resection as well as in CA19-9 non-secretors. In the resected subgroup the HELPP-score stratified survival independently of pathological prognostic factors. The HELPP-score was externally validated and was superior to CA19-9 in both the development and validation cohorts. Conclusion The HELPP-score is a readily available prognostic tool based on pretreatment routine parameters to stratify survival in PDAC independently of resection status and pathological tumor stage.