Develop A Simple-to-use Nomogram For Predicting Colorectal Neoplasia And Explore The Application Of Polarized Light Imaging Technology In Colorectal Neoplasia

[Objective]This study aims to identify the risk factors of advanced neoplasia in the domestic population,and construct a nomogram prediction model of advanced neoplasia based on the domestic population after fully evaluating its predictive ability.Explore the potential of Mueller matrix polarization optical imaging technology in assisting the diagnosis of colorectal neoplasia.[Methods]Clinical data of 219 patients diagnosed as advanced neoplasia and 209 patients with non-advanced neoplasia who underwent endoscopy examination at the Digestive Endoscopy Center from July 2019 to October 2020 has been collected.Potential relevant factors of advanced neoplasia with statistically significance were identified by single-factor correlation analysis and multi-factor logistic regression analysis.Then a nomogram was constructed and the ROC curve was drawn to compare the distinguishing ability of the new prediction model and the Asia-Pacific Colorectal Screening(APCS)score.The calibration curve was used to evaluate the calibration of the new prediction model for advanced neoplasia.Finally,the clinical decision curve(DCA)was used to compare the benefits of the two models in clinical use.Fifteen cases of normal,adenoma,colorectal cancer at early stage and advanced colorectal cancer tissues were used to make pathological sections respectively.Each tissue was cut into adjacent 4μm and 12μm sections.The 4μm sections were taken to subsequent HE staining and diagnosed by pathologists,12μm white slices were directly used for polar light imaging.Extract polarized light structure features of colorectal neoplasia.Tamura texture analysis technology were used to analyze texture characteristics of different pathological tissue.[Results]Part 1.Develop a nomogram for predicting advanced neoplasia1.Single factor correlation and multivariate logistic regression to screen advanced neoplasia-related factorsUnivariate correlation analysis showed that gender,BMI,smoking history,drinking history,family history,diabetes,CEA,blood cholesterol,OB test and fatty liver are related factors of advanced neoplasia.Multivariate logistic regression analysis show that male(OR=1.61,95%CI 1.04-2.49,P=0.031),50 years old(OR=2.61,95%CI 1.64-4.15,P<0.001),BMI more than 24(OR=2.02,95%CI 1.25-3.26,P=0.004),smoking(OR=1.97,95%CI 1.26-3.08,P=0.003),family history(OR=1.77,95%CI 1.12-2.80,P=0.014),diabetes(OR=2.35,95%CI 1.22-4.54,P=0.011),CEA(OR=2.51,95%CI 1.16-5.42,P=0.019),positive OB test(OR=3.48,95%CI 2.15-5.64,P<0.001)are predictors of advanced neoplasia for physical examination patients.2.Nomogram for predicting advanced neoplasiaEstablish a nomogram of the prediction model based on the results of multivariate logistic regression analysis The area under the curve(AUC)of the new model is 0.789,95%CI 0.747-0.826,while the APCS is 0.727,95%CI 0.682-0.768.The difference is statistically significant(P<0.001).The calibration curves show before and after calibration are both close to a straight line with a slope of 1,indicating that the new model has good consistency.Part 2.Polarized light imaging technology apply in colorectal neoplasia1.Extract Mueller matrix parameters for polarization imagingMuller Matrix Decomposition(MMPD)method can extract the parameters δ andθ reflecting the structural differences of colorectal tumors and can be used for polarized light imaging.Preliminary simulated experiment results showed that the parameter δ is related to the fiber density,and the parameter θ is related to the fiber angle.Polarized light imaging based on δ parameter can better reflect the structure of colorectal neoplasia.We apply parameter δ in different types of pathological tissues imaging,and use Tamura texture analysis technology to quantitatively analyze the texture characteristics of polarization imaging of different pathological tissues..2.Tamura image texture analysis quantitatively describes tissues of different pathological typesTamura texture analysis in δ parameter imaging shows that the grayscale images of the polarization of the four groups of pathological tissues are distinguishing.The coarseness,contrast,directionality and roughness of four groups are statistically different.Texture analysis based on the quantitative data of the four dimensions of coarseness,contrast,directionality and roughness can be used to assist the identification of benign and malignant colorectal neoplasia.[Conclusion]1.Based on the nomogram,a new advanced neoplasia prediction model was constructed.The new model can better predict the risk of advanced neoplasia than the Asia-Pacific Colorectal Screening(APCS)score.2.Polarized light imaging technology has good diagnostic efficiency for distinguishing benign and malignant colorectal lesions,and has the advantages of non-labeling and quantification.