摘要
目的:评估乳腺MRI对初诊乳腺癌患者对侧乳腺可疑病灶的诊断效能.方法:回顾性分析2017年1月—2023年5月在本院初诊为乳腺癌且对侧存在可疑病变的137例患者的临床、病理及影像学资料.所有患者行乳腺MRI(T2WI、DWI、DCE-MRI序列)、超声及乳腺X线摄影(CC位及MLO位).由2位具有5年以上乳腺 MRI诊断经验的放射科医师独立评估每例患者对侧乳腺可疑病变的MRI征象并进行BI-RADS分类.以病理结果作为金标准,比较乳腺MRI、超声、乳腺X线摄影(MG)检查对对侧乳腺癌的检出效能.同时,比较初诊乳腺癌(第一癌)和对侧乳腺癌(第二癌)的临床、病理及乳腺MRI特征的异同.结果:经病理证实,对侧病变为乳腺癌者33例,为良性病灶者104例.多模态MRI对初诊乳腺癌患者对侧可疑病灶的诊断具有高敏感度(94％)、较高的特异度(72％)和符合率(77％),AUC(0.83)高于乳腺US(AUC=0.60)和MG(AUC=0.57).双侧和单侧乳腺癌患者的年龄分别为(48.5±8.8)和(46.0±8.5)岁.双侧乳腺癌均以浸润性导管癌(IDC)居多(第一癌占48.5％,第二癌占36.4％),第二癌病灶的直径显著小于第一癌(P=0.033),孕激素受体(PR)阳性(P=0.040)、Ki-67低表达(P=0.017)及同侧腋窝淋巴结转移(P=0.011)的占比均高于第一癌且差异均具有统计学意义.第二癌的MRI表现中,非肿块型病灶较多(57.6％);肿块型病灶多表现为不规则形(78.6％)、边界不光整(78.6％)和不均匀强化(64.3％),非肿块型病灶则以区域性分布(47.4％)、强化不均匀(42.1％)和集簇状强化(42.1％)为特征,这些特征的出现率与第一癌之间的差异均无统计学意义(P＞0.05).第二癌与第一癌的时间-信号强度曲线(TIC)类型的差异具有统计学意义(P=0.017).结论:术前MRI对初诊乳腺癌患者对侧乳腺癌的检出具有较高的敏感性和准确性,AUC高于乳腺US及MG.双侧乳腺癌患者的双侧病灶在T分期、N分期、PR表达状态、Ki-67水平及同侧腋窝淋巴结转移等特征方面存在明显差异.
Abstract
Objective:To assess the diagnostic performance of breast magnetic resonance imaging(MRI)in detecting contralateral suspicious lesions in patients with newly diagnosed breast cancer.Methods:A retrospective analysis was conducted on 137 patients with newly diagnosed breast cancer and contralateral suspicious lesions at our hospital from January 2017 to May 2023.Their clinical,pa-thological,and imaging data were reviewed.All patients underwent breast MRI(including T2WI,DWI and DCE-MRI sequences),ultrasound(US)and mammography(CC position and MLO position)examinations.Two radiologists with over 5 years of breast MRI diagnostic experience independently e-valuated MRI images of contralateral suspicious lesions in patients with newly diagnosed breast cancer and classified them using the BI-RADS system.Taken the biopsy or postoperative pathology results as the gold standard,the detecting efficacy of MRI,US,and mammography(MG)for contralateral breast cancer was compared.Additionally,clinical,pathological,and breast MRI features were compared be-tween the initially diagnosed breast cancer(first cancer)and contralateral breast cancer(second can-cer).Results:Pathologically confirmed contralateral breast cancer was identified in 33 patients,while 104 patients had benign lesions.Multi-modal MRI showed high sensitivity(94％),substantial specifici-ty(72％)and accuracy(77％)for detecting contralateral lesions in breast cancer patients,with higher AUC value(0.83)than that of breast US(AUC=0.60)and MG(AUC=0.57).The average age of pa-tients with bilateral breast cancer was(48.5±8.8)years,while unilateral breast cancer patients ave-raged 46.0±8.5 years.In bilateral breast cancer,invasive ductal carcinoma(IDC)predominated(first cancer 48.5％,second cancer 36.4％).The breast cancer lesions of the secondary cancer were signifi-cantly smaller than those of the primary cancer(P=0.033),and features such as PR positivity(P=0.040),low Ki-67 expression(P=0.017),and ipsilateral axillary lymph node metastasis(P=0.011)were statistically significant compared to initially diagnosed breast cancer.In the MRI presentation of the secondary breast cancer,non-mass lesions were more common(57.6％).Mass lesions exhibited ir-regular shapes(78.6％),irregular margins(78.6％),and heterogeneous enhancement(64.3％).Non-mass lesions were characterized by regional distribution(47.4％),heterogeneous enhancement(42.1％),and clustered enhancement(42.1％).However,these differences between the secondary breast cancer and initially diagnosed breast cancer of MRI presentation were not statistically signifi-cant.Notably,there was a significant statistical difference in the types of time-signal curves(TIC)be-tween the secondary breast cancer and initially diagnosed breast cancer(P=0.017).Conclusion:Preo-perative MRI demonstrates high sensitivity and accuracy in detecting contralateral breast cancer in pa-tients with newly diagnosed breast cancer,and the AUC value is higher than that of breast US and MG.Differences in tumor T stage,N stage,PR expression status,Ki-67 levels,and ipsilateral axillary lymph node metastasis are statistically significant between bilateral breast cancer lesions.
维普
万方数据