Estimation of renal extraction fraction based on postcontrast venous and arterial differential T 1 values: An error analysis

An error analysis for quantifying single kidney extraction fraction (EF) via differential T 1 measurements in the renal vein (RV) and renal artery (RA) is presented. Sources of error include blood flow effects, the effect of a short repetition time (TR), and the impact of uncertainties in the T 1 estimates on the final EF calculations. Blood flow effects were investigated via simulation. For a range of blood velocities in the renal vein that may be found in kidney disease, incomplete refreshment of blood between readouts results in significant errors in T 1 estimation. For a .5‐cm slice, 110‐ms sampling interval, and T 1 of 600 ms, T 1 estimation to within 5% of true T 1 requires an average through‐plane velocity of 6.75 cm/s for parabolic flow, and 3.5 cm/s for plug flow. Improvement can be achieved by accurately estimating the fraction of blood that has not refreshed between readouts ( f old ), while the quality of the T 1 estimate varies with the accuracy of f old estimation. Shortening of the TR was investigated using phantom and in vivo studies. T 1 was estimated to within 3% of the true value on phantoms, and within 5% of the true value for flowing blood for TR = 2 T 1 . The estimated EF is shown to be very sensitive to the difference between T 1RA and T 1RV . To achieve 10% or 20% uncertainty in the EF estimate, T 1 in the renal vein and renal artery must be estimated to within approximately 1% or 2%. Because of limitations on measurement accuracy and precision, this method appears to be impractical at this time. Magn Reson Med 54:309–316, 2005. © 2005 Wiley‐Liss, Inc.