The Value Of DCE-MRI And Texture Analysis In Rectal Cancer To Predict The Expression Of Ki67 And P53 State Value

Objective To explore the value of quantitative parameters and texture feature parameters of Dynamic Contrast-Enhanced Magnetic Resonance Imaging（DCE-MRI）to predict the expression status of Ki67 and P53 in patients with rectal cancer.Materials and methods We retrospectively collected 91 patients who underwent DCE-MRI scans in our hospital,were confirmed to be primary rectal adenocarcinoma by surgery and pathology,and were also tested for molecular biology markers Ki67 and P53 by immunohistochemistry,and the data was completed.According to the expression status of markers was divided into:Ki67 high expression group 56 cases,low expression group 35 cases,P53 high expression group 50 cases,and low expression group 41 cases.Before surgery,DCE-MRI was performed.The MR scan sequence included conventional T2WI,DWI and DCE-MRI（injection dose 0.1mmol/kg,rate2.5ml/s,perfusion phase 39）,and record general clinical data（gender,age）,family history of bowel cancer,baseline CEA level,baseline CA19-9 level)and postoperative pathological data（tissue type,degree of differentiation,T stage,lymph node metastasis,vascular invasion,full-thickness invasion,maximum tumor diameter）.Two diagnostic imaging physicians independently performed lesion delineation with unknown groupings,and measured and recorded the DCE-MRI permeability analysis quantitative parameters of the two groups of cases:transport constant K^trans,rate constant K_ep,extravascular cell space percentage V_e;DCE-MRI texture analysis feature parameters extracted from the K^trans,K_ep,and V_emaps of the post-MRI processing image,including the Max Intensity,Median Intensity,Mean Value,Min Intensity,Skewness,Kurtosis,Entropy,Energy,Cluster Prominence,Cluster Shade,Correlation,Inertia.The intra class correlation coefficients（ICC）was used to test the consistency of the data.When the consistency was good（ICC value>0.75）,then the measurement data of senior physicians will be used for subsequent statistics analysis.Chi square test was used to compare the general data of patients between the two groups;Shapiro Wilk test was used to analyze the normality of measurement data.When the normality was satisfied,independent sample t test was used to compare the differences between the two groups;otherwise,Mann Whitney U test was used to compare.Use receiver operating characteristic curve（Receiver Operating Characteristic Curve,ROC）to analyze the predictive performance of differential DCE-MRI quantitative parameters and texture feature parameters in identifying the expression status of Ki67 and P53 in patients with rectal cancer.Logistic regression will be used to obtain statistics.After the parameters of the scientific difference were combined,the effectiveness of joint diagnosis is obtained and ROC analysis is carried out.Results1.The K^transvalue of the Ki67 high expression group（0.297（0.228,0.368）/min）in rectal cancer patients was significantly higher than that of the low expression group（0.167（0.128,0.334）/min）,the difference was statistically significant（P=0.026）,the value of AUC is 0.639,the sensitivity is 85.7%,and the specificity is 51.4%.2.The DCE-MRI texture parameters of the Ki67 high expression group Max Intensity_Ktrans（1.48（0.99,2.71））,Energy_Ktrans（4.76±1.45）,Max Intensity_Kep（10.01±6.31）,Energy_Kep（3.28±2.44）,Entropy_Kep（5.28（4.80,5.52））values are greater than the low expression group Max Intensity_Ktrans（0.92（0.53,1.34））,Energy_Ktrans（1.90±2.49）,Max Intensity_Kep（7.30±5.35）,Energy_Kep（1.66±2.17）,Entropy_Kep（4.65（4.14,5.25））,the differences were statistical significance（P<0.05）,and the AUC values were 0.758,0.826,0.650,0.748,0.716.The combination of texture parameters with statistical differences between the two groups of Ki67 has a higher diagnostic efficiency,with an AUC value of 0.947,a sensitivity of 91.1%,and a specificity of 88.6%.After Delong test,the AUC of the texture parameter joint model with differences between the two groups of Ki67 was greater than that of the single parameter model,and the results were statistically different（P<0.05）.3.The texture parameters of the P53 high expression group,Cluster Prominence_Ktrans（23.00±50.84）×10⁶,Correlation_Ktrans（1.49±1.42）×10^-2,and Inertia_Kep（13.52±22.31）×10⁴,are alllargerthanthelowexpressiongroupCluster Prominence_Ktrans（4.89±5.92）×10⁶,Correlation_Ktrans（0.16±0.17）×10^-2,Inertia_Kep（50.52±61.27）×10³,the differences were statistically significant（P<0.05）,and the AUC values were0.712,0.838,0.638.The combination of P53 texture parameters with differences between the two groups has a higher diagnostic efficiency,with an AUC of 0.914,a sensitivity of 80%,and a specificity of 95%.After Delong test,there is a difference between the two groups of P53.The AUC of the texture parameter combined model is greater than the single parameter model,and the difference is statistically significant.Conclusion The quantitative parameter K^transof DCE-MRI has limited value in predicting the expression status of Ki67 in rectal adenocarcinoma before operation.The quantitative parameters K^trans,K_ep,V_ecannot predict the expression status of P53 in patients with rectal cancer;texture analysis based on DCE-MRI,texture feature parameter Max Intensity,Energy,Entropy can effectively predict the expression status of Ki67 in rectal cancer,the texture feature parameters Cluster Prominence,Correlation,and Inertia can effectively predict the expression status of rectal cancer P53,and the combined analysis of texture analysis feature parameters can obtain higher differential diagnosis performance.