A new magnetic resonance method designed specifically for breast imaging has demonstrated a significant improvement in lesion conspicuity. This method utilizes a new pulse sequence called RODEO (Rotating Delivery of Excitation Off-resonance) that provides T1 weighting for excellent Gadolinium contrast enhancement with fat suppression to aid in the visualization of the contrast enhancing tumors which could be obscured by hyperintense fat on conventional T1 weighted MRI. A high resolution 3D acquisition produces a 128x256x256 image matrix for voxel resolution of 1.4x0.7x0.7 mm in an efficient 5 minute scan time. Interpretation is aided by computerized image processing on the workstation to yield high resolution, reformatted images in any plane and maximum intensity projection (MIP) images at any angle. This improved technology can demonstrate cancers that cannot be seen with mammography or sonography. Breast MRI may have the greatest benefits in the evaluation of certain clinical problems that cannot be addressed adequately by conventional breast imaging. In this study, specific clinical problem groups will be evaluated by blinded reading of the imaging studies by breast imaging experts to determine if the new MRI methods provide substantial changes in lesion detection, lesion margin identification, or characterization compared to conventional breast imaging methodology. A total of 1200 examinations will be performed in 1100 patients. The estimated occurrence of pathology and likelihood of surgical confirmation is used to define sufficient patient numbers to assure adequate statistical revelance. Clinical problem groups to be studies include: pre-menopausal dense breasts (400 patients), post-menopausal dense breasts (200 patients), silicone implants/injections (100 patients for each group), scar (100 patients), and chemotherapy response (100 patients prior to and following chemotherapy). MRI sensitivity and specificity will be determined by correlation with histology determined by a rigorous pathologic analysis of mastectomy specimens using serial thin sectioning of the frozen specimen. This approach allows direct correspondence between MRI locations and the serial pathologic sectioning to define MRI detected lesions that are not identified on less sensitive conventional imaging methods. This special analysis is also needed to determine the presence of any false negatives. Studies that depend upon lumpectomies or biopsies for histology are biased since the only lesions that can be sampled are either palpable or visible with imaging technology. To predict the true clinical value of MRI, an extensive analysis of potential clinical outcomes will be undertaken. A group of blinded clinical experts will determine patient management with and without MRI information. These data will be categorized by patient diagnostic problem group to predict the potential economic benefits of breast MRI determined by the analysis of clinical outcome.
