Comparison of models and contrast agents for improved signal and signal linearity in dynamic contrast-enhanced pulmonary magnetic resonance imaging.

OBJECTIVES: The objectives of this study were to compare pulmonary blood flow (PBF) measurements acquired with 3 previously published models (low-dose "single bolus," "dual bolus" and a "nonlinear correction" algorithm) for addressing the nonlinear relationship between contrast agent concentration and magnetic resonance signal in the arterial input function (AIF) and to compare both lung signal and PBF measurements obtained using gadopentetate dimeglumine (Gd-DTPA, Magnevist) with those obtained using the high-relaxivity agent gadobenate dimeglumine (Gd-BOPTA, Multihance).
MATERIALS AND METHODS: Ten of 12 healthy humans were successfully scanned on 2 consecutive days at 1.5 T. Contrast-enhanced pulmonary perfusion scans were acquired with a 3-dimensional spoiled gradient echo pulse sequence and interleaved variable density k-space sampling with a 1-second frame rate and 4 × 4 × 4-mm resolution. Each day, 2 perfusion scans were acquired with either Gd-DTPA or Gd-BOPTA; the order of the administered contrast agent was randomized. Region of interest analysis was used to determine PBF on the basis of the indicator dilution theory. Linear mixed-effects modeling was used to compare the AIF models and contrast agents.
RESULTS: With Gd-DTPA, no significant differences were observed between the mean PBF calculated for the single bolus (323 ± 110 mL/100mL/min), dual bolus (315 ± 177 mL/100mL/min), and nonlinear correction (298 ± 100 mL/100mL/min) approach. With Gd-BOPTA, the mean PBF using the dual bolus approach (245 ± 103 mL/100mL/min) was lower than with the single bolus (345 ± 130 mL/100mL/min P < 0.01) and nonlinear correction (321 ± 115 mL/100mL/min; P = 0.02). Peak lung enhancement was significantly higher in all regions with Gd-BOPTA than with Gd-DTPA (P << 0.01).
CONCLUSIONS: The dual bolus approach with Gd-BOPTA resulted in a significantly lower PBF than did the other combinations of contrast agent and AIF model. No other statistically significant differences were found. Given the much higher signal in the lung parenchyma using Gd-BOPTA, the use of Gd-BOPTA with either single bolus or the nonlinear correction method appears most promising for voxelwise perfusion quantification using 3-dimensional dynamic contrast-enhanced pulmonary perfusion magnetic resonance imaging.