Predictive value of enhanced CT radiomics feature model for 5-year overall survival in patients with esophageal squamous cell carcinoma after radiotherapy / 中华放射肿瘤学杂志

ABSTRACT

Objective:To evaluate the value of enhanced CT radiomics feature model for predicting 5-year overall survival (OS) of esophageal squamous cell carcinoma patients after radiotherapy.Methods:Clinical data of 218 patients with esophageal squamous cell carcinoma treated with radical chemoradiotherapy in the Fourth Hospital of Hebei Medical University from July 2016 to December 2017 were retrospectively analyzed. Patients were randomly divided into the training group ( n=153) or a validation group ( n=65) at a 7 vs. 3 ratio. Enhanced CT radiomics features were extracted. The data in the training group was used to construct the prediction model, and the data in the validation group were utilized to validate the efficiency of this model for predicting the 5-year OS of patients. The predictive performance of this model was assessed by the receiver operating characteristic (ROC) curve, consistency index (C-index), and decision curve analysis (DCA). Results:The 1-, 3-, 5-year OS rates were 67.0%, 33.4%, 24.9%. Five radiomic features were selected from extracted features in the training group to construct the radiomic signature (RS) for predicting 5-year OS. The area under the ROC curve (AUC) was 0.760 in the training group and 0.707 in the validation group, and the C-index was 0.680 and 0.684, respectively. The radiomics nomogram, which incorporated the RS with clinical risk factors, were established to predict the 5-year OS of esophageal squamous cell carcinoma patients after radiotherapy. The AUC was 0.782 in the training group and 0.751 in the validation group, and the C-index was 0.708 and 0.688, respectively. According to the optimal cutoff of the model, all patients were divided into the high risk and low risk groups. The 1-, 3-, 5-year OS rates were 86.5%, 65.4%, 28.9% in the low risk group, and 58.4%, 17.8%, 5.9% in the high risk group, and the differences were statistically significant (all P<0.001). Similar conclusions were obtained in the validation group (all P<0.001). Conclusion:Enhanced CT radiomics features can be utilized to construct the prediction model for 5-year OS of esophageal squamous cell carcinoma patients after radiotherapy, which can be applied in clinical practice.