Different MR systems yield variable kinetic characteristics of malignant breast lesions.

Abstract

#4025 Background: Dynamic contrast enhanced magnetic resonance imaging (DCEMRI) has demonstrated superior sensitivity for detecting earlier cancers compared with x-ray mammography, and is being used increasingly for high-risk screening, diagnostic imaging and to evaluate extent of malignant disease. When assessing lesion malignancy both the morphology and contrast uptake and washout—or kinetics—of the lesion are important. At our institution DCEMRI breast examinations have been performed on three different MR systems. The purpose of this study was to compare the MR kinetic curve data of malignant lesions acquired by these systems.
 Methods: 445 patients with 485 malignant lesions were selected for an IRB approved review. The lesions were classified as ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC) and 'other'. Dynamic MR protocol: 1 pre and 3-7 post-contrast images, acquired on a system using a non fat-suppressed dynamic sequence (NFS) and 2 newer systems by different manufacturers using fat suppressed dynamic sequences (FS1 and FS2). Kinetic curve data was processed and displayed on a CADstream workstation. Analysis of kinetic curve shape was made by an experienced radiologist according to the BI-RADS lexicon. Several quantitative kinetic parameters were calculated, both directly from the curve data and after fitting to an empirical mathematical model (EMM). The kinetic parameters of malignant lesions were compared between the three systems.
 Results: 299 malignant lesions (185 IDC, 57 DCIS) were imaged on NFS, 104 lesions (69 IDC, 21 DCIS) on FS1, and 82 on FS2 (61 IDC, 17 DCIS). Compared to both systems NFS and FS1, IDC lesions acquired on FS2 demonstrated significantly lower initial enhancement, longer time to peak enhancement and slower washout rate (p < 0.0004). 80% of IDC lesions acquired on FS1 were classified as 'washout', compared with only 46% of IDC lesions on FS2. On both FS1 and FS2, we did not find a difference in the kinetic parameters of IDC vs. DCIS lesions. However, IDC lesions imaged on NFS exhibited significantly higher contrast uptake, shorter time to peak and stronger washout compared to DCIS lesions (p < 0.0001).
 Discussion: The kinetic curve data of malignant lesions acquired by one system exhibited significantly lower initial contrast uptake and different curve shape compared with the other two. In addition, on both newer systems, the kinetic parameters of DCIS were comparable with IDC, which is contrary to what was found on the older system. Differences in k-space sampling, T1 weighting or magnetization transfer effects may be possible explanations. The results of this study underscore the importance of developing standardized acquisition and analysis methods, to ensure that across all available systems (i) malignant lesions are sufficiently conspicuous and thus reliably detected and (ii) interpretation of kinetic data is consistent. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4025.