Maximizing Accuracy and Precision on Pharmacokinetic Parameter Estimates in DCE-MRI: What is the Optimal Flip Angle?

Purpose: Dynamic Contrast Enhanced (DCE) Magnetic Resonance Imaging (MRI) is a promising tool to investigate physiological tissue properties such as blood flow, blood volume, vessel permeability and leakage space. It has been applied in various clinical studies to assess cancer stage or response to therapy. However the accuracy and reproducibility of the method remains a subject for debate since numerous sources of error may bias the pharmacokinetic (PK) parameter estimates [1-4] regardless of the measurement protocol. The standard protocol consists of measuring baseline tissue T1s (called T10s), and then using a fast spoiled gradient echo (SPGR) sequence, time-dependent MR signal changes are measured after the bolus injection of a low molecular weight Gd-based contrast agent. Although the flip angle (FA) uncertainty can be a major source of error, especially at high field (≥ 3T) [5, 6], its impact on PK parameter estimate error has neither been studied [3] nor has the error been propagated to the PK parameter estimates [2]. Here we propose to address this issue and study the propagation of measurement noise and FA uncertainty all the way through the model to the PK parameter estimates. MR signal and PK models are implemented in this study jointly in a Monte-Carlo simulation that includes a realistic DCE-MRI extraction procedure.