The specificity for detection of ductal carcinoma in situ (DCIS) based on micro-calcifications (MC's) is unacceptably low and this leads to many unnecessary biopsies. Diagnostic accuracy might be improved by using dynamic contrast enhanced magnetic resonance imaging (DCEMRI) as an adjunct to X-ray mammography. DCEMRI has excellent soft tissue contrast as well as sensitivity to tumor blood flow. However, recent studies suggest that its specificity is poor, particularly for DCIS. We believe that this is in part because both the uptake and the washout of contrast media are often inadequately sampled, leading to errors in kinetic measurements as well as loss of morphologic information. We will determine whether more complete sampling of contrast media uptake and washout (referred to here as extended temporal sampling - ETS) can identify breast abnormalities in the vicinity of MC's and help to distinguish DCIS from benign conditions. MR images from the area of the MC's will be acquired with temporal resolution of 4 seconds during the first pass of a contrast media bolus following I.V. injection and, subsequently, washout will be followed for 20 minutes. Patients (150 per year) with suspicious MC's on mammography will be scanned as part of routine clinical care. ETS will be incorporated into these MRI examinations without compromising the routine MRI protocol. Several approaches, including a new empirical mathematical model, will be used for analysis of the data. The parameters extracted from the data will be used to detect abnormally high tumor blood flow and vascular heterogeneity. We will determine whether ETS used in combination with conventional MRI increases sensitivity and specificity for DCIS. Clinically important parameters extracted from ETS data will be identified. The proposed research will improve the reliability of MRI as a 'gatekeeper' to avoid unnecessary biopsies, thereby sparing patients discomfort and anxiety while reducing costs. This revised application includes new preliminary results that strongly support the proposed research. In addition, a new Philips clinical MRI scanner will be delivered in early 2005. This will allow us to scan large volumes of tissue with high temporal and spatial resolution. These improvements in protocol address concerns of the previous reviewers.
