Combination of two-dimensional and three-dimensional super microvascular imaging for correcting S-Detect classification of benign and malignant complex cystic and solid breast lesions
Objective To observe the value of two-dimensional and three-dimensional super microvascular imaging(2D-SMI and 3D-SMI)for correcting S-Detect classification of benign and malignant complex cystic and solid breast lesions.Methods Data of 146 complex cystic and solid breast lesions confirmed by pathology(78 benign and 68 malignant)in 143 female patients were retrospectively analyzed.S-Detect classification results based on 2D-ultrasound were recorded.Then characteristics of blood supply,vascular architecture and blood flow of lesions showed on 2D-SMI and 3D-SMI were used to correct S-Detect classification.The sensitivity,specificity and accuracy of S-Detect classification were calculated before and after 2D-SMI and 3D-SMI correction.The diagnostic efficacy was analyzed with receiver operating characteristic(ROC)curve,and the area under the curve(AUC)was calculated and compared using DeLong test.Results The sensitivity,specificity,accuracy and AUC of S-Detect classification of benign and malignant complex cystic and solid breast lesions was 61.76％,58.97％,60.27％and 0.604 before and 92.65％,58.97％,74.66％and 0.758 after 2D-SMI and 3D-SMI correction,respectively.AUC after 2D-SMI and 3D-SMI correction was higher than that before correction(P＜0.05).Conclusion Evaluating complex cystic and solid breast lesions with combination of 2D-SMI and 3D-SMI was helpful for correcting towards benign S-Detect classification of complex cystic and solid breast lesions.
breast diseasesultrasonographyartificial intelligence
NETL
NSTL
万方数据
