Analysis on effect of radiotherapy on the survival of cervical cancer based on SEER database

Objective To investigate the impact of radiotherapy on the survival of patients with cervical cancer based on SEER database. Methods The data of cervical cancer patients from 1973 through 2016 was searched in SEER database and 10 920 valid data was screen out. The medical records of these patients were retrospectively analyzed, including 10 748 patients with radiotherapy and 172 without radiotherapy. Multivariate Cox regression model was constructed to evaluate the influence of different clinicopathological factors on the prognosis of patients with cervical cancer. Nomogram was plotted to evaluate the value of predicting disease progression by multiple indicators. The concordance index(C-index), receiver operating characteristic(ROC)curve and calibration chart are used to evaluate the prediction accuracy and discriminative ability of the model. Survival condition was analyzed by a Kaplan-Meier curve. Results Multivariate regression analysis showed that radiotherapy was an independent prognostic factor for cervical cancer patients(P=0.001). Multivariate regression analysis also showed that age, radiotherapy and TNM stage were all important factors affecting the overall survival of cervical cancer patients, and the factors including ≥60 years(P=0.001), T₄ staging(P=0.019), N₁ staging(P=0.001), N_X staging(P=0.001), M₁ staging(P=0.001)and receiving - no radiotherapy(P=0.001)caused worse prognosis of patients. The results of nomogram showed that age≥80 years, no radiotherapy, tumor progression to T₃ staging and distant metastasis to M₁ staging were independent risk factors for prognosis of cervical cancer patients. The C-index calculated by R software was equaled to 0.745, and standard error(SE)was equaled to 0.004, indicating that the prediction model was moderately accurate and the prediction results is reliable. The ROC curves showed that area under the curve(AUC)at 3 and 5 year was 0.759 and 0.758, respectively, indicating better sensitivity and specificity of the predictive model. The calibration curve fit line in this study has an extremely high overlap with the reference line, indicating that the prediction model was accurate and reliable. Kaplan-Meier survival analysis showed that 80-month overall survival rate was significantly higher in patients who received radiotherapy compared to those who did not, and the survival rate was significantly higher in Ⅰor Ⅱ staging of TNM versus Ⅲ or Ⅳ staging, in younger patients versus older ones. Conclusion Radiotherapy can significantly improve overall survival in patients with cervical cancer. Based on the SEER large-sample population study, this study establishes a plausible Nomogram, which is of value in designing treatment plans and assessing patient prognosis.