| BackgroundCervical cancer is a malignant tumor in the world. There are about 500000 new cases of cervical cancer and 290000 deaths worldwide every year, a serious threat to the health of women. America as the world’s developed countries, every year there are about 10,500 new cases of cervical cancer and 3900 deaths, cervical cancer ranks third in the female reproductive system cancer. In recent decades, with the rapid development of healthcare, as well as extensive cervical cancer screening programs, overall morbidity and mortality of cervical cancer decline. Nevertheless, the improvement of incidence, mortality and survival does not spread to different races. There are a lot of literature reported the survival differences between American white and black women suffering from cervical cancer. In contrast, among Asian women in the United States and other racial/ethnic groups of women with cervical cancer, survival information is limited. With the rapid growth of Asian populations, and the change of the incidence and mortality of cervical cancer, it is important to seek and improve the prognostic of cervical cancer patients. In addition to the Asian population are less of a concern, the patients with cervical cancer may occur:died of cervical cancer and died of other causes, such as heart disease, circulatory system disease, respiratory system disease and unknown, alive or lost to follow-up, while most research usually combined two causes of death or took death of other reasons as censored data, and less attention has been given to the presence of competing risks.Competing risk model is a process of risk analysis which can be used to analyze potential outcomes of survival data, including the failure time and multiple failure status. Therefore, this article also uses the competing risk model for data exploration. Meanwhile, in exploration process of disease, in addition to focus on possible risk factors, the evaluation for the risk of interesting event (RIE) which is a tumor-specific death risk (the number of patient dead of interesting event divides the number of deaths in given time) worth attention of patients and medical worker. However, most published literature do not take into account the impact of different risk factors in patients with tumor-specific death risk, so this article assess the risk of interesting event under different risk factors.ObjectiveCompeting risk model is a more reasonable analysis method and theory, when competing risk exist, while domestic competing risk model research started late. This study data selected from the Surveillance, Epidemiology, and End Results (SEER) database. The research started froml January 1988 and ended at 31 December 2011 including whites, black and Asian cervical patients’survival data. Tumor-specific death is the interesting event, and death of other causes (such as heart disease, circulatory system diseases, respiratory diseases) are competing risks. This research use competing risk model to analyze the risk factors and the prognosis of patients with cervical cancer, and assess the risk of interesting event in different risk factor to support data for guiding the treatment and prognosis.MethodsUnivariate analysis use the Kaplan-Meier method to estimate the cumulative incidence function of tumor-specific death, whether there are significant differences between races using the Log-rank test; and when considering the competing risks, the cumulative incidence function is calculated by competing risk model and differences between groups were compared using Gray test. Multivariate analysis using cause-specific hazard models and sub-distribution hazard model to explore prognostic factors in white, black and Asian cervical patients. In multivariate analysis, multiple categorical variables change into dichotomous dummy variables, and apply Bayesian Information Criterions (BICcr) in stepwise regression method for screening variables. Meanwhile, the therapeutic effect of different treatment methods is evaluated by number needed to treat (NNT). Finally, use Chen-Nakamura’ model to estimate tumor-specific death risk.Results1.54827 white,7523 black, and 3953 Asian cervical patients included in the study. Censored rate is about 83.6%. The median survival time of patient with tumor-specific death is 1.42 years, while patients dead of other reasons’median survival time is 9.08 years. The mean survival time of patients suffering the interesting event is 2.59 years and the survival time is approximate exponential distribution.2. From the basic data:white cancer patients with a minimum diagnosis age (mean diagnosis age:36.77 years), followed by black (average diagnosis age:40.23 years), Asian’diagnosed age is the highest (average age at diagnosis:45.15 years), the analysis of age difference between the three races was statistically significant. The distribution of patients registry area in different races is differences (P<0.001), with 87.3% of Asian patients and 21.0% of black patients registered in the west, and white’ relatively uniform registration. White, black and Asian patients’ married rate were 59.5%,43.9%,74.3%. The presence of whites, blacks and Asians’ FIGO stage is statistically significant (P<0.001), the highest proportion of white patients with carcinoma in situ, followed by blacks and Asians (71.1% vs.59.1% vs.47.4%). Patients suffering from regional lymph nodes metastasis in the Asian patients accounted for 6.8%, followed by black patients (4.5%). Compared with other ethnic groups, Asians have higher proportion with lymph node metastasis, which may be related to the course staging. Studies show that the earlier diagnosis, the slower of progression and less prone to metastasis. The primary site of cancer are: endometrium, cervix outside the primary site, overlapping parts of the cervix, and white, black and Asian patients with cervical cancer there are also significant differences (P<0.001) in primary site. The cervix is predilection site, in white, black and Asian, and the proportion are:83.9%,88.8%,84.7% respectively. White, black and Asian patients’ tumor differentiation was also statistically significant (P<0.001), but a larger differentiation in this study accounted for the proportion of unknown individuals. In white, black and Asian patients with cervical cancer, the squamous cell carcinoma take the largest proportion, respectively:66.5%,67.1%,64.4%. Tumor size in patients with severe lack of information. There are differences between white black, and Asian patients in treatment method. Asian patients received radiation therapy with most proportion (28.3%), while the proportion of white patients undergoing maximum proportion of surgery (82.8%).3. Univariate analysis:using competing risks model to calculate the cumulative incidence function, the results are similar, but the estimated value of the Kaplan-Meier method were larger, compared with the results obtained from the Kaplan-Meier method. The cumulative incidence function of tumor-specific death in black cervical cancer patients is the highest, followed by Asian patients, and the white cervical cancer patients have the lowest cumulative incidence function of tumor-specific death.4. Multivariate analysis:after screening, diagnosis age, FIGO stage, lymph node metastasis, tumor histological type, treatment and racial are included in multivariate analysis model. Cause-specific hazard regression model shows that white patients have a poor survival condition compared with Asian patients(HR=1.25,95% CI= 1.10-1.44), black patients compared to Asian patients also have a poor prognosis (HR=1.73,95% CI=1.52-1.95),and the research on the treatment effects find that patients being treated have lower risk suffering tumor-specific death compared with untreated patients (HR =0.15,0.69,0.42,surgery vs. no treatment, radiotherapy vs. no treatment, surgery+ radiotherapy vs. no treatment, respectively). When the presence of competing risks, through Sub-distribution hazard regression model, it is not difficult to find that white patients have a poor survival condition compared with Asian patients(HR=1.20,95% CI=1.07-1.35), black patients compared to Asian patients also have a poor prognosis (HR=1.61,95%CI=1.41-1.85), and the white patient live a better survival condition compared with black patients(1.61/1.20>1). Also, when considering the competing risks, the research on the treatment effects find that patients being treated have lower risk suffering tumor-specific death compared with untreated patients (HR =0.18,0.74,0.49, surgery vs. no treatment, radiotherapy vs. no treatment, surgery+ radiotherapy vs. no treatment, respectively).5. Evalute different treatment:the research is a retrospective investigation and does not meet the requirement of RCT, so the propensity score matching is applied to balance the confounding factors in different groups. No treatment is consider as the control group and the study time is 5 years. When not considering the competing risks, the NNT of surgery, radiotherapy and surgery+ radiotherapy are 4,11,19 respectively. And when considering the competing risk the NNT of surgery, radiotherapy and surgery+radiotherapy are 4,9,18 respectively.6. Assess the risk of interesting event:the tumor-specific death risk form 1988-2011 has a decline trend, while it is still a high level with the risk of interesting event 52.72%. when considering the covariates’effect, take the 45～60 Asian patients who had only surgery for example, the risk of interesting event of patients with carcinoma in situ is 7.1%, while the interesting event of IV stage patients is 85.1%.ConclusionCervical cancer is a gynecological malignancies, but also one of the few clear cause of cancer. The prognosis influenced by many factors, when competing risks exist, race, age, the extent of lymph node metastasis, FIGO stage and treatment together affect prognosis. While assess the risk of interesting event, it is found that when the presence of competing risks, the tumor-specific death risk is still a high level but decreasing. When considering the other factors affecting the risk of interesting event, it is not difficult to find the different FIGO stage have great difference in the risk of interesting event, and it means early diagnosis and treatment has a significant effect on women.As a developing country, China’s Cancer Center is not yet mature. China’s large population, vast territory, and a large gap between urban and rural areas means a wide range and long follow-up cancer data are needed by the government and medical sector to develop cancer prevention measures. Cervical cancer is one of the few clear cause of cancer, through Pap smears, HPV testing and other methods can be discoveried and early diagnosis means good prognosis. However, the lack of screening and propaganda for cervical cancer, and lack of cervical cancer vaccine, led to the current screening program bad compliance, also led to the patient has been diagnosed with advanced higher proportion. Therefore it is recommended that the government and medical sectors develops more cost-effective and feasible screening programs, and improve the quality and effectiveness of cervical cancer screening to protect the health of women in general. |
PDF Full Text Request