Nomogram Based on Conventional Ultrasound Combined with Contrast-Enhanced Ultrasound for Predicting Central Lymph Node Metastasis in Clinical Lymph Node-Negative Papillary Thyroid Carcinoma / 中国医学影像学杂志

ABSTRACT

Purpose To establish a nomogram based on conventional ultrasound combined with contrast-enhanced ultrasound(CEUS)for predicting the probability of cervical central lymph node metastasis(CLNM)in clinical lymph node-negative(CN0)papillary thyroid carcinoma(PTC)patients.Materials and Methods A retrospective study was performed on 359 patients with single CN0 PTC,all of whom underwent thyroid surgery and prophylactic central compartment neck dissection in the First Affiliated Hospital of Shihezi University from September 2015 to March 2022.According to the postoperative pathological results,there were 116 cases with CLNM(+)and other 243 cases with CLNM(-).The indicators of gender,age,conventional ultrasound and CEUS were recorded,and multivariate stepwise Logistic regression was performed to screen out risk predictors to construct prediction models for CLNM in CN0 PTC.The receiver operating characteristic curves of prediction models were drawn,and the area under the curve(AUC)was further compared.The preferable prediction model was selected to establish the risk probability nomogram,and the prediction performance and clinical applicability of the nomogram model were assessed.Results Multivariate analysis showed that gender,age,the maximum diameter of nodule,capsule invasion and enhancement pattern on CEUS were risk factors for CLNM in CN0 PTC(all P<0.05).The AUC of prediction model 1 including the above five indicators was 0.753,and the AUC of prediction model 2 excluding CEUS indicator was 0.704.There were statistically significant difference in AUCs between the two models(Z=2.473,P=0.013).Prediction model 1 was selected to construct a risk probability nomogram for predicting CLNM in CN0 PTC.The nomogram had a C-index of 0.753 and showed well consistency on the calibration curve.Clinical decision curve analysis indicated that the nomogram could achieve ideal net benefit when the threshold probability was between 10.7%to 81.5%.Conclusion Gender,age,the maximum diameter of nodule,capsule invasion and enhancement pattern on CEUS may be the risk predictors for CLNM in CN0 PTC.The nomogram model based on the above indicators can predict the probability of CLNM effectively,and the CEUS indicators can substantially improve the prediction performance of the model.