Abstract

Abstract Background and objective: Ethnicity-specific estimates of survival after the diagnosis of invasive cervical cancer in women residing on the Texas-Mexico border are lacking. The objective of this analysis was to examine ethnic differences (Hispanic vs. non-Hispanic) in survival after diagnosis with invasive cervical cancer in women residing in 32 Texas counties located within 100 km of the Texas-Mexico border. Methods: Data from the Texas Cancer Registry's Limited-Use SAS File covering the years 1995-2009 were analyzed. Cases of invasive cervical cancer were identified using the primary site, histology, and tumor behavior variables, which had been coded using the ICD-O-3. The sample was restricted to 1634 women who were residents of Texas border counties at the time they were diagnosed with invasive cervical cancer. Univariate associations were explored initially and followed by survival analyses. Ethnicity hazard ratios (HR) for mortality adjusted for age at diagnosis (65+ years vs. <65 years) were calculated using Cox regression after stratifying by stage at diagnosis (localized vs. other stage) and surgery status (whether or not surgery was performed as a first course of treatment). The other stage category was composed of women who were at a regional stage or had distant metastasis at the time of diagnosis. The censoring variable took on three values: alive (censored), died of causes other than cervical cancer (censored), and died of cervical cancer. HRs are reported along with 95% confidence intervals (CI) and P values. Correlation coefficients of scaled Schoenfeld residuals vs. time indicated that the assumption of proportional hazards was met for both covariates in each of the four stage-surgery strata. Results: A total of 1459 Hispanics and 175 non-Hispanics were identified. Approximately 93% of the non-Hispanics and 99.9% of the Hispanics were of white race. Hispanics were more likely to be under 65 years of age at the time of diagnosis than non-Hispanics: 77.6% vs. 70.9% (P=0.046). Ethnicity was not associated with tumor grade in a univariate analysis (P=0.96). Slightly less than half of the Hispanic cases (48.5%) and 53.1% of the non-Hispanics were at a localized stage at the time of diagnosis (P=0.25). The remaining results are stratified by stage and surgery status: cases who were at a localized stage and had surgery as a first course of treatment (group 1), localized stage and did not have surgery as a first course of treatment (group 2), other stage and had surgery as a first course of treatment (group 3), and other stage and did not have surgery as a first course of treatment (group 4). The total number of cervical cancer deaths in groups 1, 2, 3, and 4 were 38, 46, 53, and 225, respectively. HRs comparing Hispanics with non-Hispanics for groups 1 through 4, respectively, were HR=1.13 (95% CI: 0.40-3.17, P=0.82); HR=0.55 (95% CI: 0.24-1.24, P=0.15); HR=0.70 (95% CI: 0.33-1.45, P=0.35); and HR=0.59 (95% CI: 0.39-0.87, P=0.01). Conclusions: Among women who were not at a localized stage at the time of diagnosis and had not undergone surgery as a first course of treatment, Hispanics experienced a 41% reduction in their hazard of mortality compared to non-Hispanics. Social support may partially explain this protective association. In addition, there could be strong social ties to Mexico that lead women with more advanced disease to return to Mexico thereby limiting the reporting of vital status back to the Texas Cancer Registry. We are exploring how this lack of definitive follow-up data could have also produced the results we see. Citation Format: Thelma Carrillo, Leonid Fradkin, Loretta L. Hernandez, Christina M. Gutierrez, Maria Haynes, Irma Hernandez, Michael Scheurer, Christina Melendez, Michele Follen, Zuber D. Mulla. Ethnic differences in cervical cancer survival in women living along the Texas-Mexico border. [abstract]. In: Proceedings of the Sixth AACR Conference: The Science of Cancer Health Disparities; Dec 6–9, 2013; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2014;23(11 Suppl):Abstract nr B51. doi:10.1158/1538-7755.DISP13-B51

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call