Introduction

Glioma is an important and heterogeneous class of primary brain tumors originated from glial cells that surround neurons in the brain tissue, representing about 50 % of all cerebral neoplasm (1)⁠. The WHO classification system subdivides gliomas by histopathological grade and molecular profile. It is well known that grade IV and IDH wildtype tumors are more aggressive and have worse prognosis. Among low grade gliomas (II and III), such classification encompasses a more varied group of tumors with different clinical impact. A lot of research has been done trying to subdivide them into more homogeneous subgroups.
Emerging noninvasive imaging techniques have been increasingly used for better characterization of brain tumors, in order to predict grade and biological behaviour. Perfusion imaging, for example, has been added to contrast enhanced sequences for better characterization of tumor cells’ hemodynamic, which directly reflects tumor's aggressiveness. Also, it better evaluates tumors microvasculature, as contrast agents are composed of relatively big molecules and may not arrive at the distal thin vessels, unless the blood-brain barrier (BBB) is disrupted. Also, another sequence recently used to analyze gliomas microvasculature is susceptibility weighted images (SWI).
Intravoxel incoherent motion (IVIM) is a new MRI method which explore an scaling in diffusion weighting to infer diffusion and perfusion information simultaneously (2)⁠. As it explores the diffusion movement of water molecules, it is possible to assess both the macro- and microvascular perfusion within the tumor. Therefore, the aim of this study was to evaluate how the IVIM perfusion outputs correlates with the classical measurements of dynamic susceptibility contrast (DSC) and the microvascular signal of SWI and how it can be complementary to both. Furthermore, we also assessed how the microvasculature of tumors differ in correlation with tumor grade.

Methods

Four patients with diffuse glioma (two low grade and two high grade) were scanned in a 3T Philips system and a 32-channel head coil. IVIM images were acquired using 15 different b-values and a SE-EPI sequence (TR/TE = 1500/106 ms, FOV = 240 x 240 mm2, matrix = 288 x 288). DSC was acquired using a GRE-EPI sequence, TR/TE = 1700/25 ms, FOV = 240 x 240 mm2, matrix 160 x 160, 80 temporal acquisition. SWI was acquired using a GRE sequence, TR/TE = 1400/20 ms, FOV = 240 x 240 mm2, matrix 512 x 512.
IVIM data processing was performed in MATLAB using the NNLS fitting model (3)⁠. DSC maps and SW images were corregistered to the IVIM space using the ANTs toolbox (4)⁠. Statistical analysis was performed in MATLAB.

Results

From the high grade gliomas (HGG) example shown in figure 1 it is possible to observe a complementary pattern among the DSC CBF, CBV (FIg. 1a, 1b) and the IVIM fD* (Fig. 1c) maps, which is also well correlated to the microvascular signal in SWI. The low correlation coefficient between CBF and fD* highlight this pattern.
We also identified different IVIM maps and values among low grade gliomas (LGG) examples (Figures 2 and 3), reflecting the heterogeneity of biological behaviour in this subgroup. The LGG tumor with a more benign molecular profile in figure 2 (astrocytoma grade II IDH-mutated), that didn’t show any aggressive imaging feature, also presented less foci of high IVIM. In contrast, the tumor which presented heterogeneous T2 and FLAIR signal, foci of dark SWI and contrast enhancement and an area of high rCBV, presented high IVM values within the area of high FLAIR signal (Figure 3).

Discussion

The HGG has a higher cellularity level, which increases the demand for blood flow and perfusion. This behavior can be captured by traditional contrast enhanced perfusion images, in which the tumor periphery has a high rCBV and CBF. The perfusion signal in the tumor's interior, on the other hand, does not follow such a high perfusion pattern, since the contrast agent cannot reach the microvasculature (Fig. 1a), unless the BBB is disrupted. However, by employing the IVIM method it was possible to detect the microvascular perfusion in the interior of the tumor, in agreement to a visual inspection of SWI maps.
In LGG, however, it is not expected to find disruption in BBB, which hampers the measurement of perfusion using contrast enhanced methods. In the LGG case shown in figure 3, the foci of dark SWI suggests that the glioma’s aggressiveness is increasing, which was also captured by fD* IVIM maps, reporting an increase of microvascular perfusion. The null correlation between fD* and DSC perfusion in the tumor’s interior (Fig. 3h) enhance the better usage of IVIM in this case.
More interestingly, however, is the analysis of the LGG shown in figure 2. In this case of a less aggressive tumor, neither the DSC maps nor the SWI were able to capture changes in the microvascular perfusion. The fD* map had the sensitivity to detect small focus of increased perfusion within the areas of high FLAIR signal.

Conclusion

The present results for HGG suggested the IVIM maps as a complementary imaging modality to analyse the tumor’s hemodynamic. More interesting, however, was the IVIM application in LGG as a possible early biomarker for microvascular perfusion in less aggressive tumors.

Acknowledgements

CNPq - process number 151245/2019-3. FAPESP - process number 2019/06148-6.