Risk Prediction Model For Esophageal Squamous Intraepithelial Neoplasia And Esophageal Squamous Cell Carcinoma Based On Routine Blood Indicators

Background and Objective:Esophageal cancer(EC)is one of the most aggressive tumors of the gastrointestinal tract and accounts for more than 50% of new cases and deaths in China,posing a serious threat to the health of our population.More than 85% of esophageal cancers are squamous cancers and 14% are adenocarcinomas.Studies have pointed out that proper diagnosis and intervention 5-10 years before the onset of squamous epithelial carcinoma in the esophagus can significantly reduce the incidence of esophageal cancer.In addition to invasive tests such as endoscopy,hematological markers have gained interest because of their less invasive and less restrictive advantages.However,there is a lack of markers with high sensitivity and specificity.Therefore,the development of expanded tumor-related markers is a key issue in the prevention and treatment of esophageal cancer at this stage.The aim of this study is to construct a risk prediction model for esophageal squamous intraepithelial neoplasia and esophageal squamous carcinoma using hematological indicators to assist clinicians in improving the early diagnosis and treatment of esophageal lesions and to predict the progression of lesions in a timely manner.Methods:A total of 129 patients diagnosed with esophageal squamous intraepithelial neoplasia and 573 patients with esophageal squamous carcinoma at the affiliated hospital of Yangzhou University from January 2016 to December 2021,and a total of1492 patients who underwent physical examination at the health management center during the same period were collected retrospectively.General clinical data,routine blood tests,tumor markers and lesion data were collected from the patients.SPSS and R software were applied for statistical analysis.The t-test or Mann-Whitney U-test was selected according to the positive skewed distribution for continuous variables and the chi-square test was used for categorical variables.Independent predictors of esophageal squamous intraepithelial neoplasia,esophageal squamous carcinoma and lymph node metastasis were determined by logistic regression analysis.Based on the results of the multi-factor logistic regression analysis,column line graph prediction models were constructed on R software.Subject operating characteristic(ROC)curves were plotted,and the accuracy of risk prediction was judged by the area under the curve(AUC),and the discrimination ability of the model was assessed using the Harrell consistency index(C-index).Bootstrap self-sampling was used to validate the data set with 1000 resamples of bootstrap for internal validation,and the predicted probabilities were compared with the actual probabilities to plot calibration curves.Decision curve analysis(DCA)was used to assess the clinical validity of the model.The test levelα=0.05,with P<0.05 being considered statistically significant.Results:To construct a risk prediction model for esophageal intraepithelial neoplasia,a total of 129 patients with esophageal intraepithelial neoplasia and 1492 healthy controls were included in this study.Compared to healthy controls,patients with esophageal squamous intraepithelial neoplasia were older,with a median age of 67 years,a higher proportion of patients with diabetes(12.4%)and a lower median body mass index(BMI)(23.67).Multi-factor logistic regression analysis showed that age,BMI,lymphocyte(LYM)count and platelet distribution width(PDW)were independent predictors of esophageal squamous intraepithelial neoplasia.The AUC of the model was 0.921,and the C-index of the Bootstrap internal validation was 0.921,suggesting good consistency and net clinical benefit of the risk prediction model for esophageal intraepithelial neoplasia.A total of 573 patients with esophageal cancer and 1492 healthy controls were included in order to construct a predictive model for the risk of esophageal carcinogenesis.Compared to healthy controls,patients with esophageal cancer had a higher median age(69 years),a higher proportion of patients with a history of smoking(48.5%)and diabetes(8.4%),and a lower median BMI(22.92).Multi-factor logistic regression analysis suggested that age,smoking history,BMI,LYM,PDW and systemic immune inflammation index(SII)were independent predictors of esophageal squamous cell carcinoma.The AUC of the column line graph prediction model constructed in this way was 0.974,and its C-index value was 0.974 after internal validation by Bootstrap,suggesting that the model has good consistency and net clinical benefit.Further prediction of risk of lymph node metastasis in patients with esophageal squamous carcinoma suggested that LYM,PDW,and lesion size were independent predictors of lymph node metastasis in esophageal squamous cell carcinoma.The AUC of the prediction model for the risk of metastasis was 0.684 and the C-index value was 0.684 after internal validation by Bootstrap,and the model had good consistency.Conclusions:(1)Age,BMI,LYM,PDW are predictors of esophageal squamous intraepithelial neoplasia.(2)Age,smoking history,BMI,LYM,PDW,and SII were predictors of esophageal squamous cell carcinoma.(3)LYM,PDW and tumor lesion size were independent risk factors for lymph node metastasis in esophageal squamous carcinoma.(4)In this study,a prediction model for esophageal intraepithelial neoplasia and esophageal squamous carcinoma was successfully constructed and the model was in good agreement.The prediction model can help clinicians and endoscopists to identify risk groups early,intervene and treat them as early as possible,thereby reducing the incidence and mortality of esophageal cancer.