Background: artifacts of respiratory motion are one of the well-known limitations of dynamic contrast-enhanced MRI (DCE-MRI) of the kidney.
Purpose: to propose and evaluate a retrospective triggering approach to minimize the effect of respiratory motion in DCE-MRI of the kidney.
Material and methods: nine consecutive patients underwent renal perfusion measurements. Data were acquired with a 2D saturation-recovery TurboFLASH sequence. In order to test the dependence of the results on size and location of the manually drawn triggering regions of interest (ROIs), three widely differing triggering regions were defined by one observer. Mean value, standard deviation, and variability of the renal function parameters plasma flow (F(P)), plasma volume (V(P)), plasma transit time (T(P)), tubular flow (F(T)), tubular volume (V(T)), and tubular transit time (T(T)) were calculated on a per-patient basis.
Results: the results show that triggered data have adequate temporal resolution to measure blood flow. The overall average values of the function parameters were: 152.77 (F(P)), 15.18 (V(P)), 6,73 (T(P)), 18.50 (F(T)), 35.36 (V(T)), and 117.67 (T(T)). The variability (calculated in % SD from the mean value) for three different respiratory triggering regions defined on a per-patient basis was between 0.81% and 9.87% for F(P), 1.45% and 8.19% for V(P), 0% and 9.63% for T(P), 2.15% and 12.23% for T(F), 0.8% and 17.28% for V(T), and 1.97% and 12.87% for T(T).
Conclusion: triggering reduces the oscillations in the signal curves and produces sharper parametric maps. In contrast to numerically challenging approaches like registration and segmentation it can be applied in clinical routine, but a (semi)-automatic approach to select the triggering ROI is desirable to reduce user dependence.