Dynamic contrast-enhanced MRI data in vertebral bone marrow (vBM) are currently analyzed with descriptive indices. The purpose of this study was to develop and evaluate a quantitative approach, considering the tissue composition of vBM. Therefore, a measurement of the water fraction, f(wat), and the precontrast relaxation times, T(10 wat), T(10 fat), was added to the routine protocol. Signal analysis was generalized by allowing for an arbitrary fraction of fat. Plasma flow, plasma volume, extraction flow, and interstitial volume were determined from dynamic contrast-enhanced-MRI data. Simulations were used to determine the sensitivity to the precontrast values and to retrospectively verify the choice of the sequence parameters. Measurements were performed in healthy vertebral bodies (n = 30) and lesions of 15 patients with vertebral fractures. Extraction flow (milliliters per 100 mL/min) provided the strongest normal/abnormal separation: mean (standard deviation) was 0.3 (0.8) in healthy vBM and 6(4) in the fractures. Neglecting the fat component and the approximated signal analysis using relative signal enhancement produced significant differences. We conclude that correcting for the fat component in the signal and parametrization by tracer-kinetic analysis is necessary to avoid misinterpretation and/or systematic errors. The quantitative analysis is equally well suited as a descriptive parameter for the differentiation between normal and abnormal vertebral bone marrow.
(c) 2010 Wiley-Liss, Inc.