Presurgical evaluation of intracranial tumors with MR elastography has gained significant interest in recent years^1-5. In addition to determining if the lesion is stiff or soft for surgical planning, accurate characterization of mechanical properties could provide a novel avenue for tumor type classification and tumor edge identification. Given the inherently localized nature of the properties of interest, high-resolution solutions are potentially desirable for tumor MRE applications. Here, we present tumor MRE results obtained with a high-resolution imaging protocol⁶, which is known to be critical in recovering local properties⁷, coupled with nonlinear inversion⁸, which is a high-resolution approach that is not burdened by the local homogeneity assumption common to MRE methods. We describe our initial experience in investigating four patients with intracranial tumors.

Four patients volunteered to participate in the imaging study approved by our local Institutional Review Board prior to surgery:

1) 61 yo male with a grade II glioma in the left hemisphere

2) 64 yo female with grade IV glioblastoma in the left hemisphere

3) 73 yo female with tentorial meningioma located in the cerebellum

4) 43 yo female with falcine meningioma located in the frontal lobe

Each patient completed an imaging session on a Siemens 3T Trio scanner with 12-channel head coil. This session included an MRE scan with a 3D multislab, multishot spiral sequence⁶ to generate complex, full vector displacement data at 1.6x1.6x1.6 mm³ resolution. Specific imaging parameters included: FOV = 240 mm; matrix = 150x150; TR/TE = 1800/73 ms; 60, 1.6 mm-thick slices. Vibrations at 50 Hz were generated using a Resoundant pneumatic actuator system. Displacement fields were used to estimate shear modulus maps with nonlinear inversion⁸. From the complex modulus $$$G=G’+iG’’$$$, we calculate the shear stiffness, $$$\mu=2|G|^2/(|G|+G’)$$$, and damping ratio, $$$\xi=G’’/2G’$$$.

Results for each of the four patients are presented in Figures 1-4. Each figure includes the magnitude image (with tumor outline traced by two trained observers), the shear stiffness map, and the damping ratio map. Normal tissue properties for each subject were estimated using the hemispherical mirror of the tumor outlines.

Table 1 collects the mean $$$\mu$$$ and $$$\xi$$$ properties for each tumor and compares them with normal tissue properties. Each of the tumors investigated exhibited $$$\mu$$$ distinct from the surrounding tissues. The glial tumors appeared as soft: the glioma had a stiffness of 1.96 kPa and the glioblastoma had a stiffness of 1.70 kPa, which were 22% and 39% softer than normal tissue, respectively. Conversely, the meningiomas were both very stiff: the tentorial and falcine meningiomas had stiffnesses of 5.29 and 3.76 kPa, respectively, and were 163% and 39% stiffer than normal tissue. These findings are consistent with previous reports that glial tumors appear as soft^2-4, and that meningiomas can be either stiff or soft^1,2,4,5.

We also found that the tumors differed from normal tissue in $$$\xi$$$. The two glial tumors had damping ratios of 0.27 and 0.16, corresponding to 11% and 41% less viscous than normal tissue. This is similarly in agreement with previous work suggesting glial tumors have a low relative viscosity^2-4. The meniningiomas, however, exhibited even lower viscosity than the glial tumors. The tentorial meningioma had a damping ratio of 0.05 – 76% lower than normal – and the falcine meningioma had a damping ratio of 0.09 – 55% lower than normal. This finding contradicts previous reports of meningeal tumors having an above normal viscosity or phase angle^2,3.

In this work we presented our initial experiences in characterizing intracranial tumors using a high-resolution MRE protocol. Our findings suggest agreement with previous MRE reports of brain tumor viscoelasticity, but also reveal a discrepancy in the measured viscosity of meningeal tumors. Further work is needed to assess the interactions between mechanical measures and the imaging and inversion protocols.