Identification and validation of serum tumor-markers based nomogram to predict the prognostic value of patients with cervical adenocarcinoma

Abstract

Unlike cervical squamous cell carcinoma (CSCC), no uniform standard has been implemented to identify serum biomarkers for adenocarcinoma of the cervix (ADC). In the present study, we aimed to determine whether pretreatment serum tumor markers were of prognostic value in patients with ADC and constructed and validated the novel accurate nomogram for stratifying the risk groups. Patients with ADC who underwent curative hysterectomy or definitive radiotherapy from January 2011 to December 2016 were included. Significant factors independently predicting prognosis were selected by univariate multivariate Cox proportional hazard regression models and adopted for constructing the overall survival (OS) and progression-free survival (PFS) prediction nomograms. The receiver operating characteristic (ROC) curve and concordance index (C-index) with calibration curve was used to determine the accuracy of the nomogram in the prediction and determination of performance. We enrolled a total of 295 samples and randomized them as the training set (n = 207) or validation set (n = 88). Federation of Gynecology and Obstetrics Staging Guidelines (FIGO) stage, para-aortic lymph node (PALN), carcinoembryonic antigen (CEA), neuron-specific enolase (NSE), and HCG-β were assessed as the common factors independently predicting OS and PFS. For our constructed nomograms, its C-index values in OS and PFS prediction were 0.896 (95% CI, 0.879-0.913) and 0.895 (95% CI, 0.878-0.912) in training set, whereas 0.845 (95% CI:0.796-0.894) and 0.846 (95% CI:0.797-0.895) in validation set. ROC and calibration curves for our constructed nomograms predicted the excellent consistency of nomogram-predicted values with real measurements of 1-, 3-, and 5-year OS. We explored novel prognostic serum tumor markers of ADC and constructed effective nomograms comprising NSE, HCG-β, FIGO stage, PALN, and CEA, which could estimate OS and PFS for patients with ADC. These nomograms performed well in predicting patient prognosis, which was a potentially useful approach for stratifying ADC risk, thus contributing to clinical decision-making and individualized follow-up planning.