Objective To construct a combined diagnostic model based on two-dimensional shear wave elastography(2D-SWE)and conventional ultrasound parameters,and to explore its clinical application value in predicting axillary lymph node metastasis in breast cancer.Methods A total of 285 patients with pathologically confirmed breast cancer from our hospital were selected and divided into a modeling set(199 cases,with 85 cases of axillary lymph node metastasis and 114 non-metastatic cases)and a validation set(86 cases,with 43 cases of axillary lymph node metastasis and 43 non-metastatic cases)in a 7∶3 ratio.The differences of conventional ultrasound and two-dimensional shear wave elastography parameters between axillary lymph node metastatic and non-metastatic cases in the modeling set were compared,including tumor location,maximum diameter,morphology,margin,echogenicity,calcification,mean Young's modulus value and lymph node hilum structure,cortical echogenicity,longitudinal diameter,transverse diameter,lymph node cortex-to-medulla area ratio(LCMR),as well as blood flow parameters for both tumor and lymph node,such as peak systolic velocity(PSV),end-diastolic velocity(EDV),PSV-to-EDV ratio(S/D),resistance index(RI),pulsatility index(PI).Binary Logistic regression analysis was used to screen the independent influencing factors for predicting axillary lymph node metastasis in breast cancer.Based on these factors,a conventional ultrasound model,a 2D-SWE model,and a combined diagnostic model were constructed,respectively.Receiver operating characteristic(ROC)curve,calibration curve,and decision curve analysis were used to evaluate the predictive performance,calibration degree and clinical applicability of each model in the modeling and validation sets,the results was compared.Results There were significant differences in tumor maximum diameter,PSV,RI,mean Young's modulus value and lymph node portal structure,cortical echo,LCMR,PSV and S/D between axillary lymph node metastatic and non-metastatic cases in the modeling set(all P<0.05).Binary Logistic regression analysis showed that the tumor maximum diameter,the mean Young's modulus value,PSV and lymph node PSV,LCMR were independent influencing factors for predicting axillary lymph node metastasis in breast cancer(all P<0.05).ROC curve analysis showed that the area under the curve of 2D-SWE model,conventional ultrasound model and combined diagnostic model for predicting axillary lymph node metastasis in breast cancer were 0.75,0.77,0.91 and 0.73,0.77,0.90,respectively.The area under the curve of combined diagnostic model was higher than that of the other two models in the modeling set and verification set,the difference were statistically significant(both P<0.05).Calibration curve showed that the calibration curve of the combined diagnostic model was consistent with the ideal curve,while the calibration curves of the 2D-SWE model and the conventional ultrasound model deviates from the ideal curve.Compared with the 2D-SWE model and conventional ultrasound model,the combined diagnostic model had higher H-L value and lower Brier score,with statistically significant differences(all P<0.05).Decision curve showed that the clinical net benefit of combined diagnostic model for predicting the axillary lymph node metastasis in breast cancer in the modeling set and the validation set were higher.The net reclassification index and comprehensive discrimination improvement index of the modeling set and the validation set were 15.18％,12.36％and 17.14％,14.20％,respectively,which were significantly higher than those of 2D-SWE model and conventional ultrasound model in the same data set(all P<0.05).Conclusion The combined diagnostic model based on 2D-SWE and conventional ultrasound parameters significantly improves the accuracy of predicting axillary lymph node metastasis in breast cancer,and has better clinical application value.
UltrasonographyShear wave elastographyBreast cancerAxillary lymph node metastasisPredictive model
NETL
NSTL
万方数据
