Analysis Of Risk Factors And Establishment Of Predictive Model: Sound Damage After Esophageal Cancer Operation

Objective:To retrospectively analyze the risk factors associated with sound damage after radical resection of esophageal cancer,and establish a predictive model based on independent risk factors to predict the concomitant sound damage after radical resection in patients with clinical esophageal cancer.Materials and methods:A retrospective analysis was conducted based on the data of patients who suffered from thoracoscopic and laparoscopic radical resection of esophageal cancer in our department from June 2016 to November 2017.According to the inclusion and exclusion criteria,111 patients were enrolled in the study.According to the presence or absence of concurrent sound damage,the sample data were divided into 29 cases with sound damage and 82 cases without sound damage.The potential factors that collected from data were divided into two aspects: preoperative and intraoperative,and a total of 23 clinically relevant indicators were included.Among them,there were18 preoperative indicators(sex,age,BMI,smoking history,drinking history,hypertension,diabetes,preoperative electrocardiogram,pulmonary function VC%,pulmonary function FEV1%,preoperative albumin,preoperative anemia,clinical Tumor T stage,clinical tumor N stage,tumor site,tumor length,tumor thickness,neoadjuvant chemoradiation or chemotherapy).There were 5 intraoperative factors(whether robotic assisted surgery,lymph node dissection,operation time,intraoperative blood loss,intraoperative blood loss/body weight).Univariate correlation analysis and multivariate correlation analysis were used to investigate the correlation between each index and the sound damage of esophageal cancer after operation,to explore the independent risk factors of postoperative sound damage in esophageal cancer,and finally to establish a predictive model to assess the accuracy of the prediction.Results:Univariate analysis showed that postoperative sound damage in esophageal cancer was associated with hypertension,clinical tumor stage T ≥ 2,and tumor site;Multivariate correlation analysis suggested hypertension OR(5.4),clinical tumor staging T ≥ 2 OR(5.0),tumor site OR(upper thoracic segment 11.4,middle thoracic segment 3.3)is an independent risk factor for postoperative sound damage.Based on the identified three risk factors,the prediction model equation S=X1+X2+X3 was established(X1 is the value of hypertension,X2 is the assignment of clinical tumor T stage,X3 is the assignment of tumor site),and the S value was treated as the characteristics of each patient.The ROC curve was constructed to evaluate the accuracy of the prediction model,and the area under the curve was AUC=0.757.According to the total score of the prediction model,S>10 points were classified as high-risk group(n=19);S≥5 points and ≤10 points were classified as middle-risk group(n=47);S<5 points were classified as low-risk group(n=45).Verification of the subgroups was available.In the 111 patients,the true postoperative sound impairment rate was 63.2% in the high-risk group,27.7% in the intermediate-risk group,and 8.9% in the low-risk group.Conclusion:Hypertension,clinical tumor stage T≥2,tumor site is an independent risk factor for postoperative sound damage.The predictive model based on risk factors can be used as a reference for predicting postoperative sound damage in esophageal cancer.