Prediction of survival after left-sided pancreatic resection for adenocarcinoma: Introduction of a new prognostic score

Abstract

Background: Due to the clinically unapparent course the entity of left-sided pancreatic adenocarcinoma is often diagnosed at advanced stages, resulting in small numbers of patients qualifying for pancreatectomy. This study strives to develop a prognostic model for survival after left-sided pancreatic resection.

Methods: A total of 54 patients were analyzed. Pre- and intra-operative predictive factors for 18-month mortality were identified with multivariable binary logistic regression analysis and compiled into a prognostic model. The applicability was evaluated by assessment of the area under the receiver operating characteristic curve (AUROC). The model was internally validated applying a randomized backwards bootstrapping analysis.

Results: The 18-month mortality rate was 74.1% (n = 40). Mean survival was 19.1 months. A prognostic model for 18-month mortality after left sided-pancreatectomy showed an AUROC >0.800: 18-month mortality risk in% = Exp(Y) / (1 + Exp(Y)) with y= -0.927 + (1.724, if CA 19-9 elevated, otherwise 0) + (1.212 × number of intra-operative transfused packed red blood cells) + (2.771, if prior abdominal surgery, otherwise 0) - (3.612, if gastric resection, otherwise 0) This model was internally validated in 40 randomized backwards bootstrapping steps with AUROCs ranging from 0.757 to 0.971.

Conclusions: The 18-month mortality risk for patients after left-sided pancreatectomy for adenocarcinoma of the pancreatic body can be assessed with the number of intra-operatively transfused packed red blood cells, elevated CA 19-9 levels, additional gastric resection and prior abdominal surgeries in the patient's history.

Keywords: Mortality; Pancreatic cancer; Risk factors.