CT-based Radiomics For Precise Prediction Of Lymph Node Metastasis And Disease-Free Survival In Esectable Esophageal Squamous Cell Carcinoma

ObjectivesTo determine the value of radiomics in predicting lymph node(LN)metastasis and postoperative prognosis in resectable esophageal squamous cell carcinoma(ESCC)patients.Materials and MethodsIn part 1,data of 230 consecutive patients were retrospectively analyzed(154 in the training set and 76 in the test set)to determine the value of radiomics in predicting lymph node(LN)metastasis in ESCC.A total of 1576 radiomics features were extracted from arterial-phase CT images of the whole primary tumor.LASSO logistic regression was performed to choose the key features and construct a radiomics signature.A radiomics nomogram incorporating this signature was developed on the basis of multivariable analysis in the training set.Nomogram performance was determined and validated with respect to its discrimination,calibration and reclassification.Clinical usef-ulness was estimated by decision curve analysis.In part 2:We investigated the prognostic value of pretreatment CT radiomics to predict DFS in ESCC.Data of 124 patients were retrospectively analyzed,with 83 in the training set and 41 in the test set.LASSO Cox regression was performed to choose the optimal features and built a radiomics classifier.A radiomics nomogram incorporating this classifier and pTNM was developed on the basis of multivariable analysis in the training set.Nomogram performance was determined and validated with respect to its discrimination,calibration and reclassification.ResultsIn part 1,the radiomics signature including five features was significantly associated with LN metastasis.The radiomics nomogram,which incorporated the signature and CT-reported LN status(i.e.size criteria),distinguished LN metastasis with an area under curve(AUC)of 0.758 in the training set,and performance was similar in the test set(AUC 0.773).Discrimination of the radiomics nomogram exceeded that of size criteria alone in both the training set(p<0.001)and the test set(p=0.005).Integrated discrimination improvement(IDI)and categorical net reclassification improvement(NRI)showed significant improvement in predictive value when the radiomics signature was added to size criteria in the test set(IDI 17.3%;p<0.001;categorical NRI 52.3%;p<0.001).Decision curve analysis supported that the radiomics nomogram is superior to size criteria.In part 2,the radiomics classifier with six top features was significantly associated with DFS.This classifier predicted DFS with a C-index of 0.822 in the training set,and performance was similar in the test set(C-index 0.795).Discrimination of the radiomics classifier was similar to that of pTNM alone in both the training set and the test set(p both>0.05),The radiomics nomogram,which incorporated the classifier and pTNM,yielded C-indexes of 0.791(0.690-0.891)and 0.699(0.575-0.823)in the training and validation sets,respectively.Integrated discrimination improvement(IDI)and continuous net reclassification improvement(NRI)showed improvement in predicting 3-year DFS when the radiomics classifier was added to pTNM in the test set(IDI 11.2%;p=0.064;Continuous NRI 35.2%;p=0.044).ConclusionsThe radiomics nomogram provides individualized risk estimation of LN metastasis in ESCC patients and outperforms size criteria.The radiomics classifier show the same discrimination as pTNM.Combination of the radiomics signature and pTNM improves reclassification for individualized DFS estimation in patients with ESCC.