External validation of 2 prognostic indices for patients with advanced pancreatic cancer treated with first-line therapy.

ABSTRACT

OBJECTIVES: The objective of this study was to perform an external validation of 2 Asian prognostic indices for patients with advanced pancreatic cancer. METHODS: A score was calculated in patients treated with frontline therapy derived from the factors Eastern Cooperative Oncology Group Performance Status, localization of primary tumor, and C-reactive protein level according to Sawaki and from the factors Eastern Cooperative Oncology Group Performance Status, pretreatment carcinoembryonic antigen, and presence/absence of distant metastasis following Ishii. For analysis, the Kaplan-Meier method and the log-rank test were used. An analysis of the Brier score was performed to determine how the prediction error was reduced by the introduction of prognostic factors. RESULTS: For the Sawaki and Ishii score, 112 and 105 complete cases were available, respectively. Based on the 3 prognostic categories according to the Sawaki score, median overall survival was 12.3, 9.9, and 5.9 months, respectively (not statistically significant). By adapting the Ishii score to our population, 93% of the patients were allocated to the subgroup with "good" and only 7% to the subgroup with "intermediate" prognosis. Corresponding median OS was 10.5 and 3.8 months, respectively (P = 0.0035). CONCLUSIONS: Both Asian indices may not be suitable for defining different prognostic subgroups for a white population with advanced pancreatic cancer.