The Combination Of Shear Wave Elastography And Contrast-enhanced Ultrasound For Diagnosing Breast Tumors:a Mutiple Logistic Regression Analysis

Objective: To assess the value of shear wave elastography(SWE), contrast-enhanced ultrasound(CEUS) and the combination of the two methods in diagnosing breast lesions by multiple logistic regression analysisMethods: BI-RADS 3 or 4 breast lesions categorized by a skilled doctor were detected by SWE and CEUS examination respectively and confirmed by ultrasound guided-core needle biopsy(UG-CNB) and pathological findings. Inclusion criteria:①Age>18 years②Masses were untreated.Exclusion criteria:①Unwilling or unable to provide informed consent②With breast implants③Pregnant or breastfeding④Breast lesions categorized as BI-RADS 1,2,5 or 6⑤With previous ipsilateral breast surgery⑥No visible masses on ultrasound examination⑦Contrast agent allergy⑧Unsuitable for CEUS(ie. cardiopulmonary insufficiency,severe hypertension)⑨Masses with no agent perfusion(ie. cystic masses).SWE quantitative parameters(Emean_avg, Emin_avg, Emax_avg, Emean_d, Emin_d and Emax_d) were calculated and used to create ROC curves for comparing the diagnostic value.The solution of contrast agent(Sonovue,by Co.Bracco, Itally)were bolus injected into bodies with 4.8ml from the median cubital vein of patients.The CEUS parameters(enhancement time, peak intensity, enhancement homogeneousness, centripetal or non-centripetal enhancement, post-enhancement size,Sun Sign) were recorded and used to compare the differences between benign and malignant masses by Chi-Square tests.The diagnostic value of SWE and CEUS paremeters were analyzed by mutiple logistic regression.The area under the ROC curves(AUC) were used to compare the diagnostic efficacy of SWE,CEUS and the combination.Results: The size of 62 breast lesions(30 benign lesions and 32 malignant lesions) were 4.30 to 37.40mm(17.02±7.48 mm for average).6(9.7%)lesions were in category 3,21(33.9%) in category 4A, 26(41.9%) in category 4B and 9(14.5%) in category 4C. ①Univariate analysis:in SWE,all parameters in benign lesions were significantly different(P<0.05) from those in malignant lesions except Emean_avg and Emean_d and Emax_d had the largest AUC(0.882,95%CI=0.798-0.966).In CEUS,the enhancement time, peak intensity, post-enhancement size and Sun Sign were significantly different between benign and malignant lesions. ② Multivariate analysis:in SWE,Emax_avg and Emin_d were selected into the multiple logistic regression equation of which the sensitivity,specificity and accuracy in predicting malignant breast tumors were 81.3%, 86.7% and 83.9%,respectively.In CEUS,the regulated enhancement time, the regulated peak intensity and post-enhancement size were selected into the multiple logistic regression equation of which the sensitivity,specificity and accuracy in predicting malignant breast tumors were 84.4%, 86.7% and 85.5%,respectively.In the combination of SWE and CEUS, post-enhancement size and Emax_avg were selected into the multiple logistic regression equation of which the sensitivity,specificity and accuracy in predicting malignant breast tumors were 87.5%, 90.0% and 88.7%,respectively.③Analysis of ROC curves:the BI-RADS category,AUC of SWE,CEUS and the combination of SWE and CEUS were 0.821, 0.915, 0.920 and 0.941,respectively.The results above demonstrated:①either the diagnostic efficacy of SWE or CEUS was superior to BI-RADS category②the diagnostic efficacy of CEUS was similar to SWE and the combination of the two methods was superior to either one.Conclusion: The multiple logistic regression analysis on SWE and CEUS is operable and practicable.The combination of SWE and CEUS was superior to either one or BI-RADS category in diagnosing malignant breast tumors. Either SWE or CEUS is safe, quick, operable and reproducible and worth promoting in clinical application.