COX proportional hazards model applied to the resection of colorectal cancer Analysis of prognostic factors
【Abstract】 Objective To investigate the colorectal cancer prognostic factors after surgery. Methods 523 cases of non-
operative deaths in patients after surgical resection of colorectal cancer research. Select 11 colorectal cancer patients may have an impact on the characteristics of factors, through the computer COX proportional hazard model, the cumulative survival rate for colorectal cancer prognosis of
multi-factor analysis. Results 5-year follow-up rate was
92.5%. The whole group of 3-year survival rate was 57.2%, 5-
year survival rate was 42.7%. The results showed that the main factor affecting the prognosis lymph node metastasis, Duke's
stage, tumor depth of invasion, tumor diameter and histological type length (P <0.0001). Conclusion lymph node metastasis of colorectal cancer is surgical resection of the most important factor affecting the prognosis, the survival rate must be strengthened to improve the treatment of regional lymph nodes.
Key words Colorectal cancer survival prognosis analysis of COX Risk Analysis Model
Surgical resection is the preferred method of treatment of colorectal cancer, many factors may be the survival time after radical impact. To this end, we 523 cases of colorectal cancer following radical surgery follow-up data, through the
computer COX proportional hazard model, the cumulative survival rate after resection of colorectal cancer affecting
the main prognostic factors were statistically analyzed.
1 Data and methods
1.1 The clinical data from 1990 to 1995, a three A-level
cancer hospital with a complete record of information, non-
operative death, postoperative complications have not occurred 523 cases of colorectal cancer surgery. 342 cases were male, female 181 cases; age "40 years 140 cases, 40 to 59 years 182 patients, ? 60 years 201 cases, with a median age of 57.12 years.
1.2 Statistical Methods According to the standard diagnosis and treatment of common malignant tumors in China  and clinical observations, selected 11 pairs of colorectal cancer surgery may be an impact on the characteristics of factors for each of the factors relevant information encoded in Table 1. To all the data into the computer, the survival time in months, calculated at the end of second follow-up
surgery to receive based on the time cut-off tail and died in
it go lost to disease. Statistical analysis included univariate analysis, COX proportional hazard model multivariate analysis. Calculate cumulative survival rate, all data are used SAS8.1 software system for statistical analysis to P <0.05 as statistically significant.
2.1 survival 523 cases of colorectal cancer after resection of one-year survival rate was 85.9%, 3-year survival
rate was 54.9%, 5-year survival rate was 42.7%.
2.2 Single-factor analysis of prognostic tumor depth of invasion, Ducks' stage, tumor length tracks, stump the situation, organization type, number of lymph node metastases affect the prognosis, statistically significant (P <0.01); and
gender , age, tumor location, cutting edge length, and preoperative radiotherapy on the prognosis was not statistically significant (P> 0.05, Table 2). Reposted elsewhere in the paper for free download http://
Table 1 affect the prognosis of colorectal cancer
resection factors (omitted)
Table 2 523 cases of colorectal cancer after surgical resection of a single prognostic factor analysis (omitted)
2.3 affect the prognosis COX multivariate analysis model of each variable by COX multivariate analysis, drawn and
prognosis-related factors are: the number of lymph node metastasis, Duke's stage, tumor depth of invasion, tumor diameter and histological type length (P < 0.0001, Table 3).
Table 3 COX regression model of the variable table
Total χ2 = 102.37, ν = 5, P <0.0001
In China, the incidence of colorectal cancer showed an upward trend year by year. Despite the application of the extended radical surgery, but the colorectal cancer after 5-
year survival rate has remained at around 50% , this study 5-year survival rate of 42.7%, slightly lower than average. And foreign scholars in patients with colorectal cancer prognosis after resection of single-factor analysis, the
impact of significant prognostic factors include lymph node
metastasis, depth of invasion, tumor length and whether there are stump cancer, and so [2,3]. With regard to multi-factor
study, reported less at home and abroad. We are 523 cases of colorectal cancer patients after resection of study results
show that the main factors influencing the prognosis lymph node metastasis, Duke's stage, tumor depth of invasion, tumor diameter and histological type length (P <0.0001), in which lymph node metastasis is the most important factor, consistent
with the reported in the literature [4,5]. Therefore, to improve colorectal cancer resection 5-year survival rate, the
need to strengthen regional lymph node treatment.
This group of 523 cases of colorectal cancer cases, 5-
year follow-up rate was 91.9%, data reliability, using COX proportional hazard model multivariate analysis, the effective control of the confounding factors and dealing with the end of
the data, and quantitative analysis of the observed indicators the role of the strength and direction, which is a single-
factor analysis and other multi-factor analysis method is
difficult to solve, and its conclusions more realistic objective circumstances, in order to predict prognosis and guide treatment provides a reliable basis.
 Tang ZY. Modern oncology [M]. Shanghai: Shanghai Medical University Press, 1993. 539 540.
 Qin Haixia, Nancke Jun, Yang. Elderly colorectal cancer survival analysis [J]. Modern Oncology, 2003,11 (5): 357 359.
 Galindo Gallego M, Fernandez Acenero MJ, Sanz Ortega J, et al. Vascular enueration as a prognsticator for colorectal carcinoma [J]. Eur J Cancer, 2000,36 (1): 55 60.
 Jiang Bo, Qie Yan Feng, Hong-Xia Guo, et al.
Colorectal cancer and related factors in the survival time of COX regression analysis [J]. Chinese Journal of Epidemiology, 2002,23 (2): 157.
 Massacesi C, Norman A, Price T, et al. A clinical nomogram for predicting long term survival in advanced
colorectal cancer [J]. Eur J Cancer, 2000, 36 (16): 2044
2052. Reposted elsewhere in the Free Paper Download Center http://