Abstract 729
Background
Inflammation, nutrition and immunity of patients are known to relate the prognosis of malignant tumors. Several prognostic scores were reported, such as modified Glasgow Prognostic Score (mGPS), and the prognostic nutritional index (PNI). We developed new Osaka Prognostic Score (OPS) and it was independent significant factor in overall survival (OS). Herein, we reported a novel model for predicting cancer prognosis based on patient’s factors (OPS) and cancer stage.
Methods
This retrospective study investigated 569 patients with colorectal cancer (CRC) from 2007 to 2013; 431 in a training set (TS) and 138 in a validation set (VS). The Osaka Prognostic Score (OPS) used comprised one point each for (CRP) > 1.0 mg/dL, albumin (< 3.5 g/dL), and lymphocyte count < 1500. Patients were classified as OPS 1 (0–1 point), OPS 2 (2 points), or OPS 3 (3 points). A classification and regression tree (CART) and a nomogram for predicting overall survival were constructed from OPS and cancer stage, developed by TS, and validated by VS.
Results
In the TS, high CRP, low albumin, and low lymphocyte were significant predictors of OS. There were 372 patients with OPS 1, 46 with OPS 2, and 13 with OPS 3 and 82 patients with Stage I disease, 135 with Stage II, 167 with Stage III, and 47 with Stage IV (UICC classification). The C-index of CART was 0.817 in the TS and 0.732 in the VS and of the nomogram was 0.818 in the TS and 0.730 in the VS.
Conclusions
Integrating TNM stage and OPS may more accurately predict prognosis of patients with CRC.
Editorial acknowledgement
Clinical trial identification
Legal entity responsible for the study
Osaka University.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.