The Relationship Between The Radiomics Features Of Rectal Cancer And Lymph Node Metastasis

Purpose:Recent studies have shown that magnetic resonance(MR)radiomic analysis is feasible and has some value in identifying tumor characteristics,but there are few data regarding the role of MR-based radiomic features in rectal cancer.The aim of this study was to determine whether radiomic features extracted from T2-weighted imaging(T2WI)can identify pathological features in rectal cancer.Study Type:Retrospective study.Methods:A cohort comprising 91 rectal cancer patients who underwent surgery between March 2016 and January 2018.Assessment:According to the patient's postoperative lymph node metastasis,the patients were randomly divided into two groups:training set(n=61)and validation set(n=30).The total volume of the tumor was identified by two radiologists on the high resolution T2WI.1314 kinds of radiomics features were extracted.To achieve reliable results,least absolute shrinkage and selection operator(LASSO)were implemented.The area under the curve(AUC)was calculated to evaluate the predictability of the model in the LASSO analysis.Results:1.There were no statistically significant in sex ratio,age among the two groups(all P>0.05).2.There were no statistically significant in the lymph node metastasis among the two groups(all P>0.05).3.The radiology feature data of the training set were statistically analyzed,and five radiology features were screened and labeled by the LASSO model,and the radiology formula was obtained:Rad-score =-1.045004900-0.136274513×log-sigma-5-0-mm-3D_glcm_Correlation+0.008792035 ×log-sigma-5-0-mm-3D_glcm_Imc1-0.046635220 ×wavelet-HHL_firstorder_Skewness+0.126077129 ×wavelet-HLH_gldm_SmallDependenceLowGrayLevelEmphasis-0.015831674× wavelet-HLH_glcm_JointAverageThe optimal boundary value was obtained by the area under the curve(AUC)of the subject curve:1.004.4.In the training set and test set respectively using the subjects(ROC)curve expression of radiology,using the area under the curve(AUC)in the evaluation set of radiology label to predict the efficiency of rectal cancer lymph node?? In the training set:AUC = 0.87(>0,70)(95%Cl:0.7871-0.9602,sensitivity:87.5%,specificity:82.2%)? In the test set:AUC = 0.88(>0.70)(95%CI:0.7545-1,sensitivity:87.5%,specificity:81.8%)Conclusion:The use of MRI-derived radiomic features to identify the Lymph Node Metastasis is feasible in rectal cancer.