基于人工智能的多模态磁共振成像对鉴别乳腺良恶性病变的价值研究  

Value of Multimodal Magnetic Resonance Imaging Based on Artificial Intelligence in Differentiating Benign and Malignant Breast Lesions

作　　者：徐文华 刘晋亮 吴勇兴 XU Wenhua;LIU Jinliang;WU Yongxing(Imaging Department of Fuzhou Jianqiang Fifth Hospital,Fuzhou 344000,Jiangxi,China)

机构地区：[1]抚州健强第五医院影像科,江西抚州344000

出　　处：《医学信息》2024年第24期144-147,共4页Journal of Medical Information

基　　金：抚州市指导性科技计划项目(抚科社字【2023】7号序号59)。

摘　　要：目的探讨基于人工智能的多模态磁共振成像对乳腺良恶性病变的诊断价值。方法选择2021年5月-2023年7月抚州健强第五医院收治的120例进行手术治疗的乳腺良恶性病变患者为研究对象,将患者分为良性病变组72例和恶性病变组48例。患者均在本院完成基于人工智能的多模态磁共振成像与高频彩色多普勒超声检查,比较两组不同检查方法的准确度、灵敏度和特异度,以及两组sADC、ADC、f指标。结果基于人工智能的多模态磁共振成像对乳腺良恶性病变诊断灵敏度、特异度、准确率、阳性预测值、阴性预测值分别为95.83%、98.61%、97.50%、93.88%、95.95%,高于高频彩色多普勒超声的83.33%、80.56%、81.67%、64.52%、72.50%(P<0.05);恶性病变组sADC低于良性病变组(P<0.05);恶性病变组ADC以及f高于良性病变组(P<0.05)。结论基于人工智能的多模态磁共振成像对乳腺良恶性病变诊断灵敏度较高,有助于提高乳腺肿物诊断的准确率,可提供更加准确的检查结果。Objective To explore the diagnostic value of multimodal magnetic resonance imaging based on artificial intelligence in benign and malignant breast lesions.Methods From May 2021 to July 2023,120 patients with benign and malignant breast lesions who underwent surgical treatment in Fuzhou Jianqiang Fifth Hospital were selected as the research objects.The patients were divided into benign lesion group(n=72)and malignant lesion group(n=48).All patients completed multimodal magnetic resonance imaging and high-frequency color Doppler ultrasound examination based on artificial intelligence in our hospital.The accuracy,sensitivity and specificity of different examination methods,as well as sADC,ADC and f indexes of the two groups were compared.Results The sensitivity,specificity,accuracy,positive predictive value and negative predictive value of multimodal magnetic resonance imaging based on artificial intelligence in the diagnosis of benign and malignant breast lesions were 95.83%,98.61%,97.50%,93.88%and 95.95%,respectively,which were higher than 83.33%,80.56%,81.67%,64.52%and 72.50%of high frequency color Doppler ultrasound(P<0.05).The sADC in malignant group was lower than that in benign group(P<0.05).The ADC and f in the malignant lesion group were higher than those in the benign lesion group(P<0.05).Conclusion Multimodal magnetic resonance imaging based on artificial intelligence has high sensitivity in the diagnosis of benign and malignant breast lesions,which is helpful to improve the accuracy of breast mass diagnosis and provide more accurate examination results.

关 键 词：人工智能 多模态磁共振成像 高频彩色多普勒超声 乳腺良恶性病变 

分 类 号：R445.2[医药卫生—影像医学与核医学] R655.8[医药卫生—诊断学]