Gadobenate dimeglumine (Gd-BOPTA), a high-relaxivity contrast agent, has been recently proposed for dynamic MR imaging of the breast. The objective of this study was to optimize the diagnostic performance of Gd-BOPTA-enhanced dynamic breast MR imaging by using adjusted initial enhancement thresholds.
Thirty-four patients with 36 breast lesions (malignant/benign = 28/8) underwent dynamic breast MRI with 0.1 mmol/kg Gd-BOPTA and 120-second time resolution. A score system based on shape (round/oval/lobular = 0; linear/dendritic/stellate = 1), margins (defined = 0; undefined = 1), pattern (homogeneous = 0; inhomogeneous = 1; rim = 2), kinetics (continuous = 0; plateau = 1; washout = 2), and initial enhancement was used. Initial enhancement was determined with standard (<50% = 0; 50%–100% = 1; >100% = 2) and adjusted (<100% = 0; 100%–240% = 1; >240% = 2) thresholds. Scores of 0 to 3 indicated benign lesions and scores of 4 to 8 malignant lesions. Diagnostic performance was assessed in terms of sensitivity, specificity, positive and negative predictive values, and overall accuracy.
The initial enhancement was >100% for 26 malignant and 7 benign lesions and >240% for 16 and 1 lesions, respectively. The overall score was 5.89 ± 1.34 with standard thresholds and 5.50 ± 1.53 with adjusted thresholds (P = 0.003) for cancers, 4.00 ± 1.93 and 3.25 ± 1.75 (P = 0.028) for benign lesions, respectively. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy was 96%, 13%, 79%, 50%, and 78%, respectively, with standard thresholds and 96%, 75%, 93%, 86%, and 92%, respectively, with adjusted thresholds. A ductal carcinoma in situ was false negative whereas a fat necrosis and a papilloma were false positive with both thresholds. Three fibroadenomas, 1 adenosis, and 1 fibrosis were false positive with standard thresholds but true negatives with adjusted thresholds.
Lesion characterization with Gd-BOPTA requires higher thresholds for initial enhancement than those used with conventional Gd-chelates, leading to improved specificity, predictive values, and accuracy.
From the *University of Milan School of Medicine, Department of Medical and Surgical Sciences, Unit of Radiology, IRCCS Policlinico San Donato, Milan, Italy; and the †Worldwide Medical & Regulatory Affairs, Bracco Imaging Spa, Milan, Italy.
Received July 16, 2007, and accepted for publication, after revision, October 29, 2007.
Reprints: Francesco Sardanelli, MD, University of Milan School of Medicine, Department of Medical and Surgical Sciences, Radiology Unit, IRCCS Policlinico San Donato, Via Morandi 30, 20097 San Donato Milanese, Milan, Italy. E-mail: email@example.com.