Introduction

Tumor intracellular water lifetime (τ_i) has potential as a biomarker for tumor aggressiveness and treatment response. Previously, we demonstrated that using two flip angles in series during T₁-weighted dynamic contrast enhanced (DCE)-MRI improves precision of τ_i measurement, compared to the conventional DCE-MRI data with single flip angle for the entire dynamic scan.¹ However, it was necessary to measure pre-contrast longitudinal relaxation time (T₁₀) separately before contrast injection. It was also found that τ_i is very sensitive to small variations in T₁₀ value, including mismatch between T₁₀ map and DCE-MRI data.¹ Hence, in this study, we propose to use multiple flip angles in one DCE-MRI acquisition, i.e., active contrast encoding (ACE) MRI, for simultaneous estimation of T₁ and τ_i. The proposed concept of ACE-MRI was tested using murine tumor models.

Methods

Six to eight -week-old Fischer rats (n=3) with 9L gliosarcoma model were included in this study. MRI experiments were performed on Bruker 7T micro-MRI system, with four-channel phased array rat head receive-only MRI surface coil. DCE-MRI scan was performed using 3D-UTE-GRASP pulse sequence² (TR=5ms and TE=0.028ms) to achieve an isotropic spatial resolution and to minimize the T₂^* effect. It was continuously run to acquire 135,800 spokes (9,700 spokes each flip angle segment, total 14 segments, flip angels of segments in series were: 10^o - 10 ^o - 10 ^o - 10 ^o - 2 ^o - 10 ^o - 20 ^o - 10 ^o - 5 ^o - 10 ^o - 15 ^o - 10 ^o - 25 ^o - 10 ^o) for 11 min 19 s. Image reconstruction was conducted to have temporal frame resolution T= 2.5 s/frame, Image matrix = 128x128x128, field of view = 37x37x37 mm and the spatial resolution was 0.289x0.289x0.289 mm. A bolus of gadobutrol (Gadavist, Bayer) in saline at the dose of 0.1 mmol/kg was injected through a tail vein catheter, starting 60 seconds after the start of data acquisition. Prior to the DCE-MRI experiment, a 3D T₁ map with the same isotropic high resolution was obtained using the same 3D-UTE-GRASP sequence³ with variable flip angles (2^o - 4 ^o - 6 ^o - 8 ^o - 10 ^o - 12 ^o, 6,960 spokes for each flip angle, with total acquisition time was 209 s), for comparison with the proposed method that does not use the pre-contrast T₁ mapping data. Arterial Input Function (AIF) was obtained using the median of the 10 most enhancing vessels (selected by initial area under curve, IAUC) from the whole head. For AIF signal intensity to concentration conversion, the T₁ values of those vessels were obtained by minimizing the jumps in the concentration curve as demonstrated before.¹ First, pharmacokinetic model analysis was carried out with the two-compartment exchange model with water exchange^1,4 using pre-contrast separately measured T₁₀, estimating five parameters: PS (permeability surface area product), F_p (blood flow), v_e (extracellular space volume fraction), v_p (vascular space volume fraction) and τ_i (intracellular water lifetime); this approach was denoted as 5P. Then, by incorporating T₁₀ as an additional free parameter in the model instead of using measured T₁₀, pharmacokinetic model analysis was repeated for six parameter estimations, T₁₀, PS, F_p, v_e, v_p, and τ_i; hence this approach was denoted as 6P, i.e., six parameters fitting for simultaneous T₁₀ and τ_i measurement.

Results and Discussions

Figure 1A and 1B show 3D rendering of the pre-contrast T₁ map with the same isotropic resolution as DCE-MRI scan and T₁ maps of one slice in axial/coronal/sagittal planes. Figure 1C showed the maximum intensity projections of the DCE-MRI frames acquired. Figure 2A and 2B show the signal enhancement curves of the selected 10 most enhancing voxels and the converted concentration curve as the AIF used for pharmacokinetic model analysis. Figure 2C and 2D showed example curve fits for one tumor voxel using 5P and 6P methods, respectively. The fitting performance were reasonable despite the signal intensity jumps along the washout portion with multiple flip angles. This was observed with most voxels. Figure 3 shows comparisons of parameter maps between the 5P and 6P methods for tumor center slices in axial/sagittal/coronal planes. It clearly shows that all the parameters are in good agreement between two approaches in terms of absolute value range and spatial patterns. Figure 4 shows the boxplots of the parameters for comparison between the 5P and 6P methods. The median and inter-quartile-range (IQR) of all the parameters were (5P median/IQR, 6P median/IQR): T₁ (1.899/0.179 s and 1.895/0.204 s), v_e (0.157/0.053 and 0.156/0.055), v_p (0.030/0.020 and 0.031/0.022), PS (0.043/0.044 /min and 0.042/0.044 /min), F_p (1.321/2.057 /min and 1.255/2.096 /min), and τ_i (1.181/1.105 s and 1.178/0.996 s). Figure 5 summarized the comparison from all the animals.

Conclusion

This study demonstrates the feasibility of using ACE-MRI with six flip angles for simultaneous estimation of T₁ and τ_i. While the pharmacokinetic parameters were found consistent through the entire tumor volumes between 5P (without T₁₀ estimation) and 6P (with T₁₀ estimation), our results show that the τ_i had a relatively weaker correlation, which was likely caused by small degree of disagreement in T₁ measurement. Future study will focus on optimizing the scan protocol in terms of the number of flip angles, duration of flip angle segment and total scan time toward applying this approach for assessment of tumor aggressiveness and treatment response.