Introduction

The blood-brain barrier (BBB) play an important role in the transfer of solutes and essential nutrients into the brain. More and more research show that many brain diseases are related to the BBB dysfunction. The water exchange across the BBB, which can be measured by MRI, maybe a highly sensitive marker of BBB dysfunction. On the other hand, water exchange across the BBB is also be reported as a biomarker of metabolic activity¹. There are several MRI methods to measure the vascular water efflux rate constant (k_bo), including shutter speed (SS) DCE-MRI², arterial spin labeled (ASL) based methods³, and the recently developed filter-exchange imaging (FEXI) ⁴. However, there still lacks cross-validation of these different methods.

In this study, we aim to explore the relation of k_bo obtained from different methods, including SS DCE-MRI and FEXI. We performed the MRI comparison on high-grade glioma (HGG) patients, whose k_bo shows significant reduction in tumor region in previous study⁵.

Methods

Four HGG patients were recruited for this study with approval from the ethic committee of Shangdong Province Hospital. All MRI data were acquired on a 3.0T MRI instrument (Magnetom Skyra, Siemens Healthcare, Erlangen, Germany). DCE-MRI data were acquired with 3D CAIPIRINHA-Dixon-TWIST sequence: FOV, 340×340×120 mm; resolution, 0.8×0.8×1.5 mm³; flip angle, 10°; TR, 6 msec; TE, 1.3 msec; temporal resolution, 4.5 seconds; 120 frames (~ 9 min). The shutter-speed model (SSM) is used to analyze the DCE-MRI data to obtain k_bo⁵. As shown in Fig. 1, the key point of SS DCE-MRI is to use contrast agent to create enough transendothelial SS (defined as the longitudinal relaxation rate constants’ difference between intravascular and extravascular space). The SSM further considers the transmembrane water exchange rate constants as fitting parameters in the model⁵. Before SSM model fitting, a 3D gauss filter and kinetics-induced bilateral filter (KIBF) were performed to improve the image SNR⁶.

On each subject, FEXI data were acquired before DCE-MRI. The FEXI sequence (Fig. 2a) containing two PGSE blocks separated by a mixing block (mixing time, t_m). We followed the recommended FEXI protocol for k_bo measurements by Bai et al⁴. (Fig. 2b): diffusion filter (b_f) = 250 s/mm² and two b values in the detection block (b = 0 s/mm² and 250 s/mm²), TE in the filter block (TE_f) 26ms, TE in the detection block 37ms, the directions of b_f and b were always kept the same, 3.0×3.0 mm² in plane resolution, slice thickness = 5 mm, 20 slices. Images were acquired for three t_m : 25ms, 200ms, and 400ms. FEXI were also acquired with the filter inactive (b_f = 0 s/mm²) and shortest t_m (25ms), as the equilibrium data in the model fitting. For each t_m , ADC'(t_m) were computed from the two b values,$$ADC'(t_{m}) = -\frac{1}{b_{2}-b_{1}}\ln(\frac{s(t_{m},b_{2})}{s(t_{m},b_{1})}) [1]$$where b₁ and b₂ were the two b values in the detection block, and S(t_m, b₁) and S(t_m, b₂) are the FEXI signal at b₁ and b₂ , respectively. Then, the ADC'(t_m) calculated at the three t_m and b_f = 0 were fitted to $$ADC'(t_{m}) = ADC(1-\sigma\exp(-t_{m}AXR)) [2]$$with the trust-region nonlinear least-squares algorithm in MATLAB to obtain AXR, the filter efficiency (σ) and equilibrium ADC.

Results and Discussions

Examples of structural MRI images along with DCE-MRI parametric maps of one HGG subject are shown in Fig. 3. The tumor is located in the right temporal lobe (red arrow on enhanced image), which can be visualized in the enhanced image, T1 image, T2-weighted image, and ADC image. The tumor also shows typical hyperintensity on the SSM DCE-MRI K^trans map and the water mole fraction for blood (p_b) map. The high K^trans in tumor reflect that the tumor BBB has been damaged causing the increased contrast agent leakage into brain tissue. And the vascular density has a significant increase in tumor as shown in p_b map.

Fig. 4 shows the k_bo map from DCE-MRI and AXR map from FEXI of the same slice shown in Fig. 3. In visual inspection, both k_bo and AXR shows significant reduction from normal-appearing tissue to tumor region. We can also observe that k_bo and AXR shows similar spatial patterns as pointed with red arrows. On the other hand, AXR map shows much lower SNR than k_bo map, which is due to the stimulated echo used in FEXI resulting with low SNR.

We further calculated the averaged k_bo and AXR in tumor ROI and normal-appearing white matter (NAWM) ROI of the four HGG subjects (Fig. 5). k_bo and AXR shows similar values in NAWM (2.26 s^-1 and 2.67 s^-1, respectively), which also agrees with previous findings⁴. In addition, both k_bo and AXR shows significant reduction in tumor ROI with k_bo = 0.149 s^-1 (P < 0.01) and AXR =0.165 s^-1 ( P < 0.01). These results suggest that these two methods report similar vascular water exchange speed in different pathological conditions and both these methods could reveal the k_bo contrast in HGG.

Conclusion

By comparing AXR from FEXI with the k_bo from SS DCE-MRI, we find consistent values in both NAWM and tumor ROIs in HGG patients, suggesting the consistence of these two MRI methods in measuring vascular water exchange.

Acknowledgements

No acknowledgement found.