Laboratory prognostic score predicting 2-week mortality in terminally ill cancer patients

Abstract

Accurate prognostic information in palliative settings is needed for patients to make decisions and set goals and priorities. A number of prognostic scoring systems for terminal cancer patients have been developed; however, most of them include subjective or categorical variables that may not account for the intensity or severity of the disease and may be a limitation for making more objective evaluations. In this study, we developed a laboratory prognostic score (LPS) to predict 2-week mortality using routine blood tests and investigated its validity. LPS is the sum of 7 blood indices: C-reactive protein level of ≥ 6.0 mg/dL, albumin level of ≤ 2.7 g/dL, blood urea nitrogen level of ≥ 26 mg/dL, white blood cell count of ≥ 10.2 × 10⁹/L, eosinophil rate of ≤ 0.4%, lymphocyte count of ≤ 0.620 × 10⁹/L, and platelet count of ≤ 183 × 10⁹/L. The optimal cut-off LPS for predicting 2-week mortality was 4, which predicted death within 2 weeks with a sensitivity of 69% and a specificity of 82%. Thus, LPS can be a useful objective prognostic tool for predicting the 2-week mortality of terminally ill cancer patients.