| BackgroudThe most common carcinoma in the endocrine system is thyroid cancer,whose incidence has increased dramatically in recent years.Papillary thyroid cancer(PTC)is the most common type,follicular thyroid cancer(FTC)is the second most common form of thyroid cancer and accounts for about 10-15%of differentiated thyroid cancers.The incidence of FTC has remained stable worldwide.Comparing to papillary thyroid cancer,it is considered a more aggressive disease with a poorer prognosis due to its tendency to vascular invasion and distant metastases at first presentation.The TNM cancer staging system of American Joint Committee on Cancer(AJCC)is the most common guidelines for the prognosis of FTC.This classification system consists of three factors,T(depth of invasion),N(number of metastatic nodes),and M(the status of distant metastasis).This system is suitable for patient populations but is not very useful in predicting individual patient outcomes.Moreover,some other factors including sex,race,marital status,multifocality,surgery,presence of vascular invasion,margin status,and radioiodine therapy,may be important for determining outcomes in individual patients,Therefore,a prognostic predictive evaluation system for individual patients with follicular thyroid cancer is needed.ObjectiveTo identify the clinicopathological factors affecting the prognosis of patients with follicular thyroid cancer.To construct nomograms to predict the overall survival(OS)and cancer-specific survival(CSS)of patients with follicular thyroid cancer,and compare them with the 6th TNM cancer staging system of American Joint Committee on Cancer(AJCC).The aim is to provide a more individualized assessment method for patients with follicular thyroid cancer,thus providing a guidance for clinical decision-making to some degree.MethodsIn total,the clinicopathological data of 4323 eligible patients diagnosed with follicular thyroid cancer from 2004 to 2016 were extracted from the Surveillance,Epidemiology,and End Results(SEER)database established by the National Cancer Institute.The enrolled patients were randomly assigned into two groups,70%were into the training set and 30%were into validation set The chi-square test was applied to compare the clinicopathological characteristics of the patients.A prediction model was developed based on the training set,and risk factors affecting survival prognosis were first analyzed using univariate Cox proportional risk regression.Only variables found to be associated with survival in the univariate analysis were included in the multivariate analysis(p<0.05).Variables were selected through the backward stepwise selection method using the Cox proportional hazard regression model.Nomograms were developed based on the significant factors,and R software was applied to plot the nomograms of overall survival(OS)and cancer-specific survival(CSS)of 5 and 10 years,respectively.The calibration curve was obtained by the Bootstrap free sampling method with 1000 repetitions,and the calibration curve was used to assess the calibration of the prediction nomogram.Calculate the Area Under Curve(AUC)of the receiver operating characteristic curve(ROC)and the concordance index(C-index)in the training and validation sets to evaluate the nomogram discrimination and compared them with the 6th TNM cancer staging system of American Joint Committee on Cancer(AJCC).ResultsA total of 4323 eligible patients with follicular thyroid cancer were screened from the SEER database,3027 patients were randomized to the training group and 1296 patients were randomized to the validation group.After univariate and multivariate Cox proportional risk regression analyses,five factors including age,sex,marital status,extrathyroidal extension,and M stage were used to construct nomogram for overall survival(OS),and five factors including age,sex,marital status,N stage and M stage were used to construct nomogram for specific survival(CSS).The calibration curves showed good predictive calibration of the predictive nomogram in both the training and validation groups for overall survival and specific survival.Further regression model discrimination analysis was performed to assess the predictive accuracy of the predictive nomogram by C-index.The internal validation was performed in the training set,the C-index of the prediction nomogram was 0.764(95%CI,0.727-0.801)and 0.853(95%CI,0.794-0.912)for OS and CSS,respectively.In addition,the C-index of the 6th edition TNM cancer staging system was 0.678,(95%CI:0.631-0.725)and 0.818(95%CI:0.757-0.879)for OS and CSS respectively.The external validation was performed in the validation set,the C-index of the prediction nomogram was 0.786(95%CI,0.725-0.847)and 0.889(95%CI,0.766-1.000)for OS and CSS,respectively.In addition,the C-index of the 6th edition TNM cancer staging system was 0.725(95%CI:0.658-0.792)and 0.887(95%CI:0.762-1.000)for OS and CSS,respectively.The two ROC models of the 5-and 10-year survival were compared in training set.The AUC values for predicting the 5and 10-year OS were 0.767 and 0.784,whereas the AUC values of the 6th edition TNM cancer staging system were 0.693 and 0.674,respectively.Regarding the prediction of 5-and 10-year CSS rates,the AUC values of the nomogram were 0.886 and 0.819,while the AUC values of the 6th edition TNM cancer staging system were 0.868 and 0.779.ConclusionsFor patients with follicular thyroid cancer,the results of this study showed that age,gender,marital status,extrathyroidal extension and M stage are independent prognostic factors on overall survival,and age,marital status,extrathyroidal extension,N stage,and M stage are independent prognostic factors on cancer-specific survival.Compared with the 6th edition TNM cancer staging system,the prediction nomogram we constructed has better prediction accuracy.The nomogram developed in this study may be an alternative tool for predicting prognosis,which may improve the risk of survival of individualized assessment and creating more clinical treatments. |
PDF Full Text Request