Comparing Western and Eastern criteria for sarcopenia and their association with survival in patients with pancreatic cancer.

BACKGROUND & AIMS: Sarcopenia and cachexia are associated with pancreatic cancer and serve as important adverse prognostic factors. Body composition can be analyzed by routine computed tomography (CT) for cancer staging and has been used to study many types of cancer. The CT measurements are robust, but the diagnostic criteria for sarcopenia vary among different studies. Age, sex and race are important factors that affect muscle and fat masses. This study aimed to analyze the effect of different sarcopenia diagnostic criteria on the prognosis of patients with pancreatic cancer.

METHODS: Patients with newly diagnosed pancreatic cancer at National Taiwan University Hospital between October 2013 and October 2016 were retrospectively reviewed in this study. Body composition was assessed using cross-sectional CT images to calculate the total skeletal muscle (TSM) index. The concordance and interobserver variability of the TSM measurements were evaluated using both the Western criteria and the Eastern criteria. Kaplan-Meier analyses and the Cox proportional hazard ratio with two different diagnostic criteria for sarcopenia were used to compare the effect on overall survival (OS).

RESULTS: A total of 146 patients with pancreatic cancer were enrolled. The TSM index measured by the Western institute was highly correlated with that measured by the Eastern institute (r = 0.953, p < 0.001). The prevalence of sarcopenia in the patient group at baseline was 66.4% (97/146) by the Western criteria and 11.0% (16/146) by the Eastern criteria, and only low agreement was found between the Western and Eastern criteria (Kappa value = 0.028, p = 0.149). Patients who were sarcopenic by the Western criteria showed no significant difference in OS versus those who were not sarcopenic (p = 0.807). However, patients who were sarcopenic by the Eastern criteria showed a significant difference in OS versus those who were not sarcopenic in a univariate analysis (p = 0.008) and multivariate analysis after adjustment for AJCC stage (p = 0.014).

CONCLUSIONS: Our study demonstrates that different diagnostic criteria may result in different diagnoses and that sarcopenia is an important poor prognostic factor for pancreatic cancer when appropriate diagnostic criteria are selected.