Evaluation of four injection profiles for uniform contrast‐enhanced signal intensity profiles in MR angiography

Abstract

Background Gadolinium concentration variation during acquisition of contrast‐enhanced MR angiography (CE‐MRA) may lead to artifacts. Purpose To compare signal intensity (SI) profiles of four different contrast agent injection strategies during CE‐MRA with the goal of minimizing SI variation during acquisition. Study Type Prospective. Subjects Forty subjects randomized to receive one of four injection profiles of gadobenate dimeglumine (0.1 mmol/kg), either undiluted (0.5 M) or diluted to 40 ml total volume. Tested profiles: 1) nondiluted single‐phase ("standard" NS; 1.6 ml/s), 2) diluted single‐phase (DS; 1.6 ml/s), 3) diluted biphasic (DB; 9 ml @ 3.3 ml/s, 29 ml @ 1.4 ml/s), 4) patient‐tailored protocol using linear prediction (DT). Field Strength/Sequence Time‐resolved SI measured at 3T with spoiled gradient echo sequences having analogous parameters to those of CE‐MRA. Assessment Plateau arrival time, rise time, duration, peak and tail SI, plateau quality (sum of squared residuals; SSR), average SI for each injection type derived were used. Statistical Test Two‐tailed t‐test. Results Peak SI, arrival, and rise times were not significantly different between groups, excepting peak SI DB slightly > DS (P = 0.042). Duration of NS vs. the diluted groups was significantly shorter (all P < 0.0001), and DS duration was significantly shorter than that of DT and DB (NS 11.4 ± 3.5 vs. DS 22.9 ± 4.3, DB 25.4 ± 2.3, DT 28.3 ± 4.1 sec). Quality (SSR) of the 20‐second plateau was significantly better for DS, DB, DT as compared with NS (all P < 0.001). Data Conclusion Three different strategies to power‐inject diluted gadobenate dimeglumine targeting a 20‐second plateau produced SI profiles with longer duration, more consistent plateau, and no significant loss in peak SI. Such injection profiles may provide more uniform SI during CE‐MRA, potentially reducing blurring artifacts. Level of Evidence: 2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:1808–1816.

Relationship Between Simulated Gadolinium‐Based Contrast Agent Injection Profile and Achievable Resolution Metrics in Contrast‐Enhanced Magnetic Resonance Angiography

“…SI of the simulated gadolinium contrast bolus can be realistically estimated given our understanding of R1 and R2* relaxivity of different GBCAs in blood, physiologic models of arterial contrast concentration evolution following intravenous contrast injection, and experimental data collected from human subjects receiving different contrast injection profiles. 5,[7][8][9]12 For purposes of modeling the effects of the shape of the SI curve over time on CE-MRA acquisition we used an approach with parameters chosen to mimic features of a standard single-bolus GBCA injection of non-diluted gadobenate dimeglumine ("NS" acquisition as described in Maki et al). 5 In particular, in order to approximate the slope of the initial GBCA injection seen in the in vivo data from healthy volunteers in that study, we utilized the Verhoeven model at a simulated contrast injection rate of 1.6 mL/second, with resultant time-variant vascular SI derived from GBCA concentration via the methods of Wilson et al 7,8,12 SI of the simulated contrast injection was chosen to provide a peak signal to noise ratio relative to the background of approximately 15:1.…”

“…5,[7][8][9]12 For purposes of modeling the effects of the shape of the SI curve over time on CE-MRA acquisition we used an approach with parameters chosen to mimic features of a standard single-bolus GBCA injection of non-diluted gadobenate dimeglumine ("NS" acquisition as described in Maki et al). 5 In particular, in order to approximate the slope of the initial GBCA injection seen in the in vivo data from healthy volunteers in that study, we utilized the Verhoeven model at a simulated contrast injection rate of 1.6 mL/second, with resultant time-variant vascular SI derived from GBCA concentration via the methods of Wilson et al 7,8,12 SI of the simulated contrast injection was chosen to provide a peak signal to noise ratio relative to the background of approximately 15:1. After this initial upstroke, we then modeled various lengths of "contrast plateaus," with SI held at a constant level to approximate the effects that may be seen in vivo with alternate contrast injection regimens (eg, more elongated or tailored injections per Maki et al).…”

“…After this initial upstroke, we then modeled various lengths of "contrast plateaus," with SI held at a constant level to approximate the effects that may be seen in vivo with alternate contrast injection regimens (eg, more elongated or tailored injections per Maki et al). 5 Modeled lengths of this contrast plateau were 0%-100% at 1% increments, scaled as a percentage of the total acquisition duration. Following this contrast plateau, we modeled an exponential decay curve to an arbitrary level of SI during the period of contrast recirculation, with the parameters of exponential decay again chosen to mimic that seen in the "NS" injection of Maki et al's Vascular structures modeled such that pre-contrast less intense than background (vessel 50, background 100 arbitrary units).…”

“…shown both theoretically and empirically to influence image quality, mainly through blurring of the vessel edges. [4][5][6] Recent observations, such as those on the unique T1 and T2* behavior of GBCA in blood, particularly at the high concentrations achieved with CE-MRA, have helped refine our understanding of the relationship between how the GBCA is injected and how the resultant intra-arterial signal intensity (SI) evolves. 5,[7][8][9] In particular, rapid GBCA boluses of up to 2-3 mL/second as have often been used for clinical CE-MRA may not increase SI as compared to slower boluses, and instead (for a fixed dose of contrast) only decrease the length of the bolus, thereby leading to increased blurring in the setting of elliptical centric k-space acquisition.…”

“…[4][5][6] Recent observations, such as those on the unique T1 and T2* behavior of GBCA in blood, particularly at the high concentrations achieved with CE-MRA, have helped refine our understanding of the relationship between how the GBCA is injected and how the resultant intra-arterial signal intensity (SI) evolves. 5,[7][8][9] In particular, rapid GBCA boluses of up to 2-3 mL/second as have often been used for clinical CE-MRA may not increase SI as compared to slower boluses, and instead (for a fixed dose of contrast) only decrease the length of the bolus, thereby leading to increased blurring in the setting of elliptical centric k-space acquisition. 4 While these effects can and have been described using mathematical constructs such as the point spread function or modulation transfer function, it can be difficult to translate these concepts into the reality of how clinical CE-MRA images are affected, particularly as viewed by a clinician.…”