First order correction for T₂*‐relaxation in determining contrast agent concentration from spoiled gradient echo pulse sequence signal intensity

Abstract

Purpose: To investigate the accuracy of a method neglecting T 2 *-relaxation, for the conversion of spoiled gradient echo pulse sequence signal intensity to contrast agent (CA) concentration, in dynamic contrast enhanced MRI studies. In addition a new closed form conversion expression is proposed that accounts for a first order approximation of T 2 *-relaxation. Materials and Methods:The accuracy of both conversion methods is compared theoretically by means of simulations for four pulse sequences from literature. Both methods are tested in vivo against the numerical conversion method for measuring the arterial input function in mice.Results: Simulations show that the T 2 *-neglecting method underestimates typical tissue CA concentrations (0 mM to 2 mM) up to 6%, while the errors for arterial concentrations (0 mM to 10 mM) range up to 43%. The results from our first order method are numerically indistinguishable from the simulation input values in tumor tissue, while for arterial concentrations the error is reduced up to a factor 10. In vivo, peak Gd-DOTA concentration is underestimated up to 14% with the T 2 *-neglecting method and up to 0.9% with our first order method. Conclusion:Our conversion method reduces the underestimation of CA concentration severely in a broad physiological concentration range and is easy to perform in any clinical setting.

Simulating the effect of input errors on the accuracy of Tofts’ pharmacokinetic model parameters

“…The causes of inaccuracy in the AIF measurement have been thoroughly studied in the literature [6,8,16,17,38] and may be identified in many sources. Examples are the T 2 * saturation due to high contrast agent concentration in vessels, flow, or partial volume effects.…”

“…Moreover, because of the high concentrations of CA in large vessels, where the plasma signal is sampled, the approximation of negligible T 2 * in Eq. (4) might not be valid [6]. It is therefore not possible to model errors on C p (t) in the same way errors are modeled in C t (t).…”

“…[21]. While in tissue the T 10 can be derived using a T 1 mapping sequence, in blood the T 1 is often entered as an assumed value, because of the challenge of measuring T 1 in flowing spins [6,7,17,34].…”

“…Its calculation is hampered not only by flow effects and saturation due to the high contrast agent concentration, but also by partial volume effects, and by insufficient temporal sampling, affecting especially the description of the CA first pass [6,7]. Furthermore, other factors such as the unknown hematocrit (Hct) also affect the amplitude of the AIF and can be a significant source of error in the calculated PMPs [8].…”

Effect of $$ T_{2}^{*} $$ correction on contrast kinetic model analysis using a reference tissue arterial input function at 7 T

“…The underestimations were reduced by correction with T 2 * measurements. De Naeyer et al [21] also showed that neglecting T 2 * could lead to errors in arterial concentration estimation up to 43%. In contrast, the corresponding underestimation of the AIF in our study ranged from 6.2% to 37.1% with a mean of 21.3%, indicating that the T 2 * effect in reference muscle tissue may be smaller than in direct arterial measurements.…”