Survival Analysis In Colorectal Carcinoma Based On Individual Characteristics And Bio-molecular Markers

Colorectal carcinoma is one of the common malignant carcinoma that severely harms the public health. In western countries, the mortality of colorectal carcinoma has ranked second among all malignant carcinoma, next only to lung cancer. In China, the mortality is also increasing, currently ranking between the 4th to the 6th. With the advance of the comprehensive treatment plans, mainly involving surgical removals, there has been great progress in the prognosis in the colorectal carcinoma recently. However, the five-year survival rate is still only 60-70%. Although a perfect prognostic evaluation system can be used for guiding clinical treatments and reducing healthcare cost, traditional TNM grading system is still used in the majority of prognostic analysis. Recently, there has been great development in the scientific understanding of the colorectal carcinoma, especially the emergence of many bio-molecular markers. We believe building a prognostic evaluation system based on individual characteristics and bio-molecular markers is an urgent task.Surgical and pathological verified colorectal carcinoma cases were taken from the Xiaoshan tumor registry system during 1990-2000. All clinical and pathological data were individually reviewed. Those died within one month from surgery or died due to causes other than colorectal carcinoma were excluded from the study. Among the total of 320 cases, there were 101 colon carcinoma cases and 219 rectal carcinoma cases. 183 of them were males and 137 were females. There were 226 cases undergone surgical treatments, including 155 colon carcinoma cases and 71 rectal carcinoma cases. 124 of them were males and 102 were females.The survival data was provided by the Xiaoshan Center for Disease Control. All follow-up ended on Dec. 31, 2002. The duration of follow-up varied from 1 month to a maximum of 152 months, with the medium of 45 months and the mean of 52 months. The archival block of representative tumor tissues and corresponding normal tissues were sliced for HE staining and Envision immunohistochemical staining.Statistical analysis was performed using SPSS 10.0 for Windows. Cumulative survival rate was calculated using life-table methods with a one-year step size. Survival curve was drawn using the univariate survival analysis based on Kaplan-Meier methods (the significance level tested by Log-rank method). OR values were calculated by using the univariate survival analysis based on univariate COX proportional hazard model. The statistically significant prognostic factors identified by univariate analysis were then analyzed using multivariate Coxproportional hazard model. The x2 analysis was used for the correlation analysisbetween two factors, and for calculating Spearman's correlation coefficients. a=0.05. First we performed survival analysis on all 320 colon-rectum cancer cases. The five-year cumulative survival rate (FYCSR) for colorectal carcinoma was found to be 62.10%, where the FYCSR for colon carcinoma (71.96%) was higher than the FYCSR for rectal carcinoma (58.18%). The results also showed that the cumulative survival rates decrease over one-year, three-year and five-year periods, with the steepest decrease for rectal carcinoma. There was relatively no change for colon carcinoma. Therefore the percentages of colon cases and rectal cases are important when comparing the FYCSR in research data. When there is more rectal carcinoma in the data, the FYCSR must be low. In addition, when calculating the FYCSR using survival table, there was large discrepancy in the results using one-year interval v.s. five-year interval. Therefore the cumulative survival rates between different research data should not be simply compared.Because the FYCSR for colon carcinoma and rectal carcinoma were different, in order to understand the difference in prognostic factors for colon and rectal cancer, we performed survival analysis on the 101 colon carcinoma cases and 219 rectal carcinoma cases.There was obvious difference in the prognostic factors for colon and rectalcarcinoma. Smoking, lymph...