Impact of precontrast T10 relaxation times on dynamic contrast-enhanced MRI pharmacokinetic parameters: T10 mapping versus a fixed T10 reference value

Purpose To investigate variation in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) pharmacokinetic parameter measurements between different methods of precontrast tissue relaxation (T-10) estimation: pixel-based mapping versus a fixed reference value. Materials and Methods In 15 DCE-MRI studies the female pelvis, uterine fibroids, the left psoas muscle, and the fifth lumbar vertebral body were chosen to represent tissues with varying perfusion characteristics. All DCE-MRI studies were processed using a variable flip angle T-10 map and a fixed T-10 reference value of 1000 msec. A subset of five DCE-MRI studies were each processed multiple times using the fixed T-10 method with the reference T-10 ranging from 0-2000 msec in 100-msec increments. Pharmacokinetic measurements of K-trans, k(ep), v(e), and initial area under the gadolinium curve (iAUGC) were performed maintaining the identical position for region of interest placement on each structure. Results The mean difference in pharmacokinetic output between the pixel-based T-10 map and the fixed T-10 reference value ranged from 6.6% for k(ep) in the muscle to 54.9% for iAUGC in the vertebral body. At lower T-10 (<1000 msec) aberrations in T-10 estimation resulted in a larger error. Accurate measurement of T-10 for each structure subsequently incorporated as a fixed T-10 reference value yielded relative differences from -41.8% to 22.3% compared to the pixel-based T-10 map. Conclusion Direct comparison of pharmacokinetic parameters derived from a pixel-based approach versus a reference value uniformly applied to all pixels for T-10 estimation is impeded by the inherent spatial heterogeneity of T-10 within tissues. J. Magn. Reson. Imaging 2014;39:1136-1145. (c) 2013 Wiley Periodicals, Inc.