Palliative prognostic tools for the terminally ill patients with gastric cancer.

Abstract

120 Background: Terminally ill patients with gastric cancer have specific gastrointestional symptoms and signs related with cancer progression. To estimate accurate survival expectancy of gastric cancer patients is important for timely decision making of their end of life issues. Methods: We reviewed the 276 patients with terminally ill gastric cancer who were treated at Yonsei Cancer Center between January 2007 and December 2011 and eventually were died. Retrospectively, we conducted the data of clinical signs, symptoms, and laboratory results at the time of cessation of the active treatment. Then, we established the palliative survival estimation model by stratification of risk group. Results: Median palliative survival time from the decision to stop further treatment to death was 42days. In the multivariate Cox regression analysis, 5 parameters were identified as prognostically significant factors: anorexia, dyspnea, hypoalbuminemia, elevated blood urea nitrogen, and elevated serum alkaline phosphatase. We scored each variables as 1-3 for symptom(1:asymptomatic, 2:symptomatic, 3:symptomatic requiring intervention) and 1-2 for lab results(1:normal, 2:abnormal) and summed up each scores. Using the total score, patients were divided into 3 risk groups: low-risk(5-7points), intermediate-risk(8-11points), and poor-risk patients(12point). As a result, median palliative survival for low-risk group(n=110) was 87.0±7.4days, intermediate-risk group(n=158) and poor-risk group(n=8) were 31.0±2.1days and 6.0±2.1days, respectively (p<0.0001). Conclusions: Using multivariate analysis and summation of each prognostic factor score, 3 risk groups were determined. After validation by prospective multicenter trial, this palliative survival time estimation tool will be helpful to inform the accurate survival for terminally ill gastric cancer patients.