Cox Model Analysis Of Prognostic Factors Impacting On Survival Of Carcinoma Of The Prostate

Objective The aim of the study is to explore the prognostic factors of carcinoma of the prostate(CaP) and establish Cox proportional hazards model for predicting prognosis of CaP, providing proof for the treatment and prognosis prediction of CaP.Methods Data of 145 patients who were diagnosed with CaP in Peiking University Shenzhen Hospital from January 2003 to August 2009 were collected. All case were out-patient and telephone follow-uped until February 2010. 12 cases were eliminated because of incomplete clinic data, failure of follow-up early or dying of other diseases, and the other 133 cases were inclued in the study. Kaplan-Meier method and Log-rank test were used for survival analysis, and 29 factors possiblly impacting on prognosis of CaP were univariate and multivarate analysed base on the Cox proportional hazards regression model. Then, independent prognosis factors of CaP were selected and predictive modal was established. PI(Prognostic index) threshold was identified and verified.Results The disease specific survival rates of the CaP at 1, 3, 5 year were 90.9%, 78.4%, 65.4%. Cox mutivarate analysis indicated that the most improtant independent prognostic factors impacting on CaP were X4bone pain(RR=4.707, P=0.03), X9HGB(RR=8.050, P=0.005), X20T stage(RR= 10.245, P=0.001), and X29treatments (RR=10.964, P=0.001). Therefor, the predictive modal was created as the following: PI=-2.869+1.687X4+1.756X9+2.326X20-1.779X29. Chi-square test showed X2=71.026 (P<0.001), which illustrated the equation reasonable. Each PI of the patient was calculated. Then all the patients were divided into high risk, median risk and low risk group, and Log-rank test indicated that the survival rates among these group were significant difference(P<0.001).Conclusion T stage, HGB, bone pain and treatments were all independent prognostic factors of CaP. Treatments was protective factor and the others were danger factors of CaP prognosis. The predictive modal including these four factors was able to predict the survival of CaP effectively. In clinic work with the prostate cancer patients, we can easilly know the patients prognosis base on the PI modal, just needing a simple calculation. It was indeed provide a useful tool for clinic treatment and predicting prognosis of CaP, with high practical value.