Construction and efficacy test of a survival prediction model for locally advanced cervical cancer based on anti-angiogenesis
This study aimed to develop and evaluate an anti-angiogenesis-based model for predicting the survival and the potential benefits of targeted therapy for patients with localized advanced cervical cancer. We collected clinical data from 163 patients with cervical cancer who received paclitaxel and cisplatin (TP) or TP plus bevacizumab during or after radiotherapy from June 2017 to February 2023. We analyzed the clinical measures of recent efficacy and overall survival (OS) using univariate and logistic multivariate and Cox regression methods, respectively. We constructed a nomogram model and evaluated its efficacy using the c-index, the area under the curve (AUC), a calibration curve, and the clinical decision curve (DCA). We found that targeted agents and hemoglobin were independent determinants of near-term efficacy (P < 0.05), while targeted agents and stage were independent factors of OS (P < 0.05). We developed a predictive model for an OS prognostic nomogram and performed internal validation 1000 times using the Bootstrap re-sampling method. The c-index was 0.81, and the AUC was 0.84 (P < 0.01).The calibration curves showed a good agreement between the projected and actual values. The DCA curve indicated that the model had a high positive predictive accuracy. We developed a novel anti-angiogenesis-based survival prediction model for patients with locally advanced cervical cancer. This model could estimate the benefit of targeted therapy before treatment, and it had good validation.