超声人工智能联合剪切波弹性成像用于乳腺结节良恶性鉴别诊断临床研究

Clinical Study of Ultrasonic Artificial Intelligence Combined with Shear Wave Elastic Imaging in the Differential Diagnosis of Benign and Malignant Breast Nodules

目的:探讨超声人工智能(AI)联合剪切波弹性成像(SWE)在乳腺结节良恶性鉴别诊断中的临床效能。方法:回顾性分析200例乳腺良恶性肿瘤患者(共206个结节)的数据,所有病灶均接受SWE及超声AI检查。结果:经病理组织学检查确诊恶性肿瘤98个(47.57%),良性肿瘤108个(52.43%)。乳腺恶性病灶的弹性模量相关指标及AI辅助评分均显著高于良性病灶(P<0.05)。ROC曲线分析显示,SWE与超声AI在乳腺结节良恶性鉴别诊断中均有效,且二者联合使用的效能优于单一指标(P<0.05)。结论:超声AI与SWE联合使用可提高乳腺结节良恶性的鉴别诊断效能。

Objective:To investigate the clinical efficacy of ultrasonic artificial intelligence(AI)combined with shear wave elasto-graphy(SWE)in the differential diagnosis of benign and malignant breast nodules,aiming to provide more references for enhancing the efficiency of ultrasonic differential diagnosis.Methods:The data of 200 patients with benign and malignant breast tumors(totaling 206 nodules)were retrospectively analyzed.All lesions were examined by SWE and ultrasonic AI examinations.Results:Pathological examination confirmed 98 malignant tumors(47.57%)and 108 benign tumors(52.43%).The elastic modulus-related indices and AI-assisted scores of malignant breast lesions were significantly higher than benign breast lesions(P<0.05).ROC curve analysis demonstrated that both SWE metrics(including maximum elastic modulus,average,and standard deviation)and AI-assisted scores were effective in differentiating benign from malignant breast nodules,with the combination of both showing superior performance compared to individual metrics(P<0.05).Conclusion:The combined application of ultrasonic AI and SWE can improve the efficiency of differential diagnosis of benign and malignant breast nodules.