The Application Of Ultrasound Radiomics In Predicting The Aggressiveness Of Papillary Thyroid Carcinoma

Objectives: This study aimed to investigate the application value of ultrasound-based radiomics in the preoperative prediction of the aggressiveness in papillary thyroid carcinoma(PTC).This included(1)constructing and validating a clinical-radiomics nomogram based on ultrasound(US)to predict cervical lymph node metastasis(LNM)in PTC;(2)establishing and verifying a clinical-radiomics nomogram based on Bmode ultrasound(BMUS)and Contrast-enhanced ultrasound(CEUS)for prediction of extrathyroidal extension(ETE)in PTC.Methods:(1)In the study to predict cervical LNM in patients with PTC,we collected 211 patients with PTC between June 2018 and April2020,then we randomly divided these patients into the training set(n =148)and the validation set(n = 63).Radiomics features were extracted from BMUS images and CEUS images of the primary tumor.The MaxRelevance and Min-Redundancy(m RMR)algorithm,least absolute shrinkage and selection operator(LASSO)algorithm,and backward stepwise logistic regression(LR)were applied to select key features and establish the radiomics score(Radscore),including BMUS Radscore and CEUS Radscore.The clinical model and the clinical-radiomics model were established using the univariate analysis and multivariate backward stepwise LR.The clinical-radiomics model was finally presented as a clinical-radiomics nomogram,the performance of which was evaluated by the receiver operating characteristic(ROC)curves,integrated discrimination improvement(IDI),Hosmer–Lemeshow test,calibration curves,and decision curve analysis(DCA).(2)In the study to predict ETE in PTC,a total of 216 patients with PTC between January 2018 and June2020 were collected and divided into the training set(n = 152)and the validation set(n = 64).The LASSO algorithm was applied for radiomics features selection.Univariate analysis was performed to find clinical risk factors for predicting ETE.The BMUS Radscore,CEUS Radscore,clinical model,and clinical-radiomics model was established using multivariate backward stepwise LR based on BMUS radiomics features,CEUS radiomics features,clinical risk factors and the combination of those features,respectively.The diagnostic efficacies of different models were assessed using ROC curves and the area under the ROC curve(AUC)of different models were compared by the Delong test.The model with the best performance was then selected to develop a nomogram.The calibration and clinical utility of the nomogram was evaluated and validated.Results:(1)In the study to predict cervical LNM in patients with PTC,the clinical-radiomics nomogram was constructed with four predictors,including gender,age,US-reported LNM,and CEUS Radscore.The clinical-radiomics nomogram performed well in both the training set(AUC= 0.820)and the validation set(AUC = 0.814).The significantly improved IDI demonstrated that the addition of CEUS Radscore,a very helpful marker for predicting the LNM in PTC,significantly enhanced the reclassification ability over the clinical model.The Hosmer–Lemeshow test and the calibration curves demonstrated good calibration.The DCA showed that the clinical-radiomics nomogram had satisfactory clinical utility.(2)In the study to predict ETE in PTC,the clinical-radiomics model,which is constructed with age,CEUS-reported ETE,BMUS Radscore and CEUS Radscore,showed the best diagnostic efficiency in both the training set(AUC = 0.843)and the validation set(AUC = 0.792).The Delong test showed that the clinical-radiomics model had significantly higher AUC values than the clinical model in both the training(P = 0.004)and validation sets(P = 0.044).Moreover,a clinical-radiomics nomogram was established for easier clinical practice.The Hosmer-Lemeshow test and the calibration curves demonstrated satisfactory calibration.The DCA showed that the clinical-radiomics nomogram had high clinical benefits.Conclusions:(1)The clinical-radiomics nomogram constructed by CEUS Radscore and key clinical features can be used to preoperatively predict the cervical LNM status of PTC patients with high diagnostic efficacy.(2)The clinical-radiomics nomogram integrating dual-modal ultrasound radiomics features and clinical risk factors can effectively predict ETE in PTC,and it is a robust supplement to the traditional clinical strategy.In conclusion,the presented 2 clinical-radiomics nomograms can be used as effective tools for individualized prediction of cervical LNM and ETE in PTC.