Assessment of the effect of haematocrit‐dependent arterial input functions on the accuracy of pharmacokinetic parameters in dynamic contrast‐enhanced MRI

Abstract

The detection and prognosis of prostate cancer in its early stages are critically important. It is therefore essential to improve the existing dynamic contrast-enhanced MRI (DCE MRI) techniques commonly used for the assessment of the tumour vascular environment. The goal of this study was to describe a method for the estimation of the arterial input function (AIF) in DCE MRI by measuring R(2) * values in the femoral artery of patients with early-stage prostate cancer. The calculation of contrast agent concentrations was based on calibration curves determined in whole blood samples for a range of normal haematocrit (HCT) values (HCT = 0.35-0.525). Individual AIFs corrected for HCT were compared with individual AIFs calibrated with a mean whole blood [R(2)*-Gd-DTPA-BMA] [Gd-DTPA-BMA, gadolinium diethylenetriaminepentaacetate-bis(methylamide) (gadodiamide)] curve at an assumed HCT = 0.44, as well as a population AIF at an assumed HCT = 0.45. The area under the curve of the first-pass bolus ranged between 0.6 min mM at HCT = 0.53 and 1.3 min mM at HCT = 0.39. Significant differences in magnitude at peak contrast agent concentrations (HCT = 0.36, [Gd-DTPA-BMA](max) = 9 ± 0.4 mM; HCT = 0.46, [Gd-DTPA-BMA](max) = 4.0 ± 0.2 mM) were found. Using model-based simulations, the accuracy of the kinetic parameters estimated using individual AIFs corrected for HCT demonstrated that, for the use of individual calibration curves with HCT values differing by more than 10%, K(trans) and k(ep) values were largely underestimated (up to 60% difference for K(trans)). Moreover, blood volume estimates were severely underestimated. Estimates of kinetic parameters in early-stage prostate cancer patients demonstrated that the efflux rate constant (k(ep)) was influenced significantly by the definition of AIF. Regardless of whether an individually calibrated AIF or a population AIF (average of all individually calibrated AIFs) was used, pixel-by-pixel mapping of k(ep) and v(b) in the prostate gland appeared to be more sensitive than with the usual biexponential approach.

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

“…(7) and (8) and then fitted to (9), and, through this relation, show a dependence on errors in T 1blood and α [7,34]. However, as flow effects in blood affect the measured T 1 and α in a complicated way, for a correct determination of the C p (t) Eq.…”

“…(7)(8)(9)(10) highlight the stages where errors can be introduced on the path to the calculation of the PMPs.…”

“…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].…”

Arterial input functions in dynamic contrast-enhanced magnetic resonance imaging: which model performs best when assessing breast cancer response?

“…Biexponential approximations to the following input AIFs were used: Fritz-Hansen, 6 modified Fritz-Hansen, 7 Weinmann, 10 Parker, 4 femoral artery 8 and a computationally efficient raised cosine model AIF approximating Parker 9 ( Figure 1), in order to calculate the following quantitative kinetic parameters: K trans (measured in min 21 ), the volume of extracellular extravascular space per unit volume of tissue (v e ), or the leakage space, in % and rate constant (k ep ) (measured in min 21 ). Changes (in %) in median parameter values of all fitting voxels containing tumour were calculated after two cycles of NAC for each tumour.…”

“…There are a number of assumed AIFs in current use. These include the data reported by Fritz-Hansen, 6 a combination of the data of Weinmann and Fritz-Hansen known as the modified FritzHansen, 7 the biexponential approximated form of Parker 4 ("biexponential Parker"), femoral artery, 8 the mathematical equivalent AIF to the Parker ("Cosine Bolus Parker") 9 and the now rarely used Weinmann function 10 ( Figure 1). Using differing AIFs will have effects on the DCE-MRI modelling process and the outputs generated.…”