Sardanelli Francesco, Newstead Gillian M, Putz Barbara, Jirakova Trnkova Zuzana, Trimboli Rubina M, Abe Hiroyuki, Haverstock Daniel, Rosenberg Martin
From the *University of Milan, Department of Biomedical Sciences for Health, Radiology Unit, IRCCS Policlinico San Donato, Italy; †Department of Radiology, The University of Chicago, Illinois; ‡Bayer HealthCare, Bayer Pharma AG, Berlin, Germany; §Unit of Radiology, IRCCS Policlinico San Donato, Milan, Italy; and ∥Bayer HealthCare Pharmaceuticals Inc, Whippany, NJ.
Invest Radiol. 2016 Jul;51(7):454-61. doi: 10.1097/RLI.0000000000000254.
The aim of this study was to evaluate the diagnostic efficacy of gadobutrol enhanced preoperative breast magnetic resonance imaging (MRI) in 2 prospective studies.
Approval of ethics committees and informed consent from patients were obtained. Both Gadobutrol-Enhanced MR Mammography (GEMMA) trials followed a standardized protocol using 1.5 T scanners. After unenhanced scans, patients received 0.1 mmol/kg of gadobutrol for the dynamic study. Six independent blinded readers, 3 for GEMMA1 and 3 for GEMMA2, assessed unenhanced images and, 2 or more weeks apart, contrast-enhanced plus unenhanced breast MRI images (CE-BMRI), using a standard 5-region scheme. Another 6 independent readers (3 for each study) evaluated mammograms alone. Sensitivity was calculated taking into account the identification of regions harboring malignancies (within-patient sensitivity), whereas specificity was based on cancer-free breasts. The first patient from each center was used for site qualification and blinded reader training and excluded from the efficacy analyses. Reference standard was pathology for regions harboring malignancy and a combination of negative pathology, mammography, and ultrasound for cancer-free regions.
Of 906 breast cancer patients enrolled in 13 countries in the 2 studies, 865 received gadobutrol and 787 were evaluated for diagnostic performance (390 in GEMMA1 and 397 in GEMMA2). Within-patient sensitivity, that is, the detection rate of malignant disease extent per patient, ranged from 80% to 89% for CE-BMRI and was significantly superior to unenhanced breast MRI alone (37%-73%) and to mammography alone (68%-73%) for all readers in both trials. Specifity of the CE-BMRI ranged from 83% to 95%.
In a very large multicenter preoperative setting, gadobutrol-enhanced breast MRI demonstrated high levels of sensitivity and specificity, consistent with published data on breast MRI.
本研究旨在通过两项前瞻性研究评估钆布醇增强术前乳腺磁共振成像（MRI）的诊断效能。
获得伦理委员会批准并取得患者知情同意。两项钆布醇增强乳腺磁共振成像（GEMMA）试验均遵循使用1.5T扫描仪的标准化方案。在进行平扫后，患者接受0.1 mmol/kg钆布醇进行动态研究。6名独立的盲法阅片者，其中3名负责GEMMA1，3名负责GEMMA2，评估平扫图像，并在间隔2周或更长时间后，使用标准的5区方案评估对比增强加平扫的乳腺MRI图像（CE-BMRI）。另外6名独立阅片者（每项研究3名）仅评估乳腺X线摄影。敏感性计算时考虑了对含有恶性肿瘤区域的识别（患者内敏感性），而特异性基于无癌乳房。每个中心的首例患者用于场地资格认定和盲法阅片者培训，并排除在效能分析之外。对于含有恶性肿瘤的区域，参考标准为病理检查；对于无癌区域，参考标准为病理检查阴性、乳腺X线摄影和超声检查结果的综合判断。
在两项研究中，来自13个国家的906例乳腺癌患者中，865例接受了钆布醇检查，787例接受了诊断效能评估（GEMMA1中390例，GEMMA2中397例）。患者内敏感性，即每位患者恶性疾病范围的检出率，CE-BMRI为80%至89%，在两项试验中对所有阅片者而言均显著优于单纯平扫乳腺MRI（37% - 73%）和单纯乳腺X线摄影（68% - 73%）。CE-BMRI的特异性为83%至95%。
在非常大型的多中心术前检查中，钆布醇增强乳腺MRI显示出高水平的敏感性和特异性，与已发表的乳腺MRI数据一致。
