Patient-specific calibration for breast MRI: breast-coil insertable reference phantom

Introduction: The routine use of a calibration phantom that fits easily into a breast coil could provide the basis for standardization and patient-specific calibration of breast MRI. Here, we use a phantom to correct the variable flip angles in a T1 measurement. This increases the accuracy of T1 and contrast media concentration measurements obtained with a clinically employed T1-weighted dynamic contrast-enhanced protocol. Methods: A unique phantom was designed, with a comfortable breast-coil insert mounted with color-coded circumferential and vertical tubes filled with distilled water solutions containing Gd-DTPA (0.0, 0.1, 0.2, 0.3, 0.4, 0.5 mM, Omniscan). The geometry of the phantom was adjusted so that multiple phantom compartments were present in many slices, irrespective of slice orientation. A healthy volunteer (no contrast media injected, University of Chicago Institutional Review Board approval) was scanned in a standard bilateral breast coil at 1.5 T with one breast placed in the phantom cup. Clinical T2-weighted (2D turbo spin echo, TR/TE = 5647/120 ms) and T1-weighted (fat-suppressed ultrafast gradient echo, flip angle (FA) = 12°, TR/TE = 12.6/3.7 ms) images were acquired. A variable flip angle protocol was used for T1 measurement (3D FSPGRE, nominal FA’s = 3, 5, 10, 15, 20, 30°, TR/TE = 5.0/1.5 ms, NA = 3). The clinical images were evaluated by an experienced breast radiologist to assess image quality in the presence of the phantom. In addition, the clinical T1-weighted image was used to measure phantom signal intensities as a function of T1. Based on theoretical T1 values in the phantom, the true flip angles were estimated and these values were used to calculate accurate T1’s for breast tissue. Results: The circumferential phantom compartments (vertical compartments are not visible in this projection; they are on the backside) and their appearance on the T2and T1-weighted MR images are displayed in the figure below. In the graph, the measured phantom signal intensities are plotted as a function of the theoretical T1.