Modern meningioma imaging techniques

Most meningiomas have features that are similar including an extraaxial mass with signal intensity similar to cortex on T1 and T2 MRI sequences, avid homogeneous enhancement following administration of gadolinium contrast, and an enhancing “dural tail” which reflects neoplastic dural infiltration or reactive vascularity, or both, draining into the adjacent dura. Low signal intensity within the tumor may often be due to calcification or to vascular flow voids, a distinction sometimes difficult to make. Meningiomas can be nearly spherical or elongated (en plaque), multiple, and often take origin from a dural sinus, a feature important for surgical planning. These tumors also tend not to respect the dural boundary, which is a distinctive feature not typical of other neoplasms (Fig. 2).

Although most benign meningiomas are innocuous from the standpoint of metastatic potential, they may result in serious complications secondary to dural sinus invasion (Fig. 3), (with or without thrombosis), narrowing and thrombosis of significant arterial structures, and compression of cranial nerves and other important neural structures.

Edema associated with meningioma is thought to be vasogenic in origin, and probably related to tumor secretion of vascular endothelial growth factor (VGEF), rather than a result of direct mass effect on adjacent brain or venous invasion causing vascular congestion [2]. The presence of intra-axial edema is said to predict an increased potential for recurrence [3, 4].

There have been reports that radiolabeled agents, such as ¹¹¹In-octreotide, that have an affinity for somostatin receptors can be useful in detecting and localizing meningiomas [5]. However, this ability is countered by the relative lack of specificity in differentiating meningiomas from other lesions such as high-grade glioma or pituitary adenomas, among others. PET imaging is attractive for investigating the metabolic activity of tumors [6]. However, the role of PET imaging in the evaluation of meningiomas is complicated by the variable metabolic presentation in different meningioma types. In typical benign meningiomas, the usual metabolic marker F-18 fluorodeoxyglucose, presents with isometabolism on PET imaging. Malignant meningiomas may display hypermetabolism which confounds their differentiation from other intracranial tumors [7]. In general, nuclear medicine techniques provide capabilities that are sensitive for meningiomas but that have relatively low specificity.

It is possible using MRI to sensitize the image appearance to the extent to which water can freely diffuse in any volume element (voxel). When the motion of water molecules within a voxel is restricted there is greater magnetization coherence and that voxel will appear bright. This technique is referred to as diffusion weighted imaging (DWI). Reduced water diffusivity (Fig. 4a) has been correlated with more aggressive tumor behavior and is sometimes seen with atypical/malignant meningiomas, high cellular density, and recurrence [8].

The diffusion weighting in DWI acquisitions is encoded on top of the usual T1 and T2 properties of the underlying sequence. It is possible to create images that are insensitive to those underlying T1 and T2 values by performing multiple DWI acquisitions and extracting from them a map of the diffusion effect alone, referred to as an apparent diffusion coefficient (ADC) map. A decrease in ADC values (Fig. 4b) at follow up of a benign meningioma should raise suspicion for dedifferentiation to higher tumor grade [9]. Although diffusion-weighted imaging provides an added tool in the approach to defining meningioma grade a recent report [10] calls into question the predictive ability of DWI methods in grading meningiomas or identifying histological sub-types.

It is possible to perform MRI with acquisition times that are short enough to capture the changes in signal intensity as a bolus of contrast material passes through the brain. These methods, termed perfusion MRI, can provide useful information on the vascular supply of meningiomas, which is only implied from conventional MRI. Typically, the injection is performed at a rate and concentration that causes a loss of signal secondary to perturbation of the magnetic field by the contrast agent. The method used is an echo-planar contrast-enhanced T2* weighted sequence rapidly performed prior, during, and after the bolus infusion of gadolinium contrast material using an intravenous injection. The curves reflect the permeability between intravascular and extravascular compartments, as well as cerebral blood volume. The relative cerebral blood volume (rCBV) is measured within a tumor comparing to the contralateral normal white matter and can be displayed on color maps [11]. Perfusion curves provide additional prognostic information by helping distinguish between benign and atypical/malignant meningiomas.

Benign meningiomas typically derive their blood supply from the external carotid via dural branches. These vessels do not contain a blood brain barrier and are thus quite permeable to gadolinium, which is reflected by a curve with little or no return to baseline following infusion (Fig. 5a, b).

As the meningioma enlarges, it may parasitize pial branches from the brain parenchyma, which do contain a blood brain barrier. The perfusion scans will show an elevated cerebral blood volume with intensities that return to baseline signal levels, reflecting an intact blood-brain barrier of the internal carotid artery supply, as seen in Fig. 6. A high volume of pial-cortical supply (as opposed to dural-meningeal supply) usually predicts an aggressive meningioma with a higher tendency of recurrence.

The cerebral volume of peri tumoral edema was found by Zhang et al. [11] to be elevated surrounding malignant meningiomas compared to the vasogenic edema associated with benign meningiomas, and likely related to angiogenesis/microvascular proliferation in the peri tumoral brain.

As discussed above, MRI offers advantages over X-ray angiography in the evaluation of tissue physiology, including measurement of diffusion and perfusion characteristics that serve as proxies for tissue infarction and vascularity, respectively. In assessing the vascularity of meningiomas, there is substantial interest in clearly defining which specific arterial branch is providing blood supply to the specific regions of interest. For example, preoperative embolization of intracranial meningiomas is performed to reduce tumor vascularity and thus minimize operative blood loss but is not possible if the internal carotid artery (ICA) is the only source of blood supply as embolic material delivered through the ICA would also obliterate normal brain tissue.

Conventional MR perfusion methods are performed with an intravenous injection into an arm vein. The contrast material is then transported back to the heart and ejected into the arterial system. As such, intravenous MR perfusion studies are non-selective and perfusion of a specific region of interest reflects supply from all arterial sources. Furthermore, the temporal passage of the contrast bolus is substantially modulated by significant patient-specific characteristic dynamics that have little to do with tissue perfusion, such as recirculation times in the veins and the lungs, mixing efficiency in the heart, and arterial tortuosity. Selective intraarterial injections performed in an MR suite provide powerful new capabilities in evaluating meningioma perfusion. The ability to investigate the impact of intraarterial procedures on the end organ is now available at a number of sites that have so-called XMR suites, an installation that contains both an X-ray fluoroscopy suite and an MRI scanner in the same room. Although it would not currently be recommended to perform catheterization solely for the purpose of determining perfusion characteristics, intraarterial MR perfusion studies are possible when conducted in the same session that the patient is already undergoing catheterization for pre-surgical embolization. Complete X-ray catheter cerebral angiography is performed in the context of preoperative embolization and, thus, fulfills a secondary goal of identifying all arterial supply to the tumor, whether embolized or not. X-ray angiography is able to qualitatively evaluate regional capillary-level vascularity within the tumor as each supplying artery is separately injected with iodinated contrast material.

At the University of California San Francisco, the XMR suite consists of a combined X-ray angiography and 1.5T MRI suite in which a patient can be slid on a single bed between X-ray angiography and MRI intraprocedurally. That suite has been used to monitor the completeness of tumor embolization in 15 patients [13]. Via a standard transfemoral arterial approach under X-ray guidance nonbraided 5 French (Fr) diameter diagnostic catheters (which have undergone extensive testing for MR safety [14]) are placed in the external carotid artery (ECA) of subjects. Patients are then slid from X-ray to MR nd a baseline MR perfusion study is performed by injecting dilute gadolinium contrast through the ECA catheter. The catheter is then pulled back to the CCA and a similar intraarterial (IA) perfusion study is performed. By subtracting the ECA supply from the CCA supply, the ICA supply to the tumor can be determined without having to do a separate ICA catheterization. Patients are then moved back to the X-ray fluoroscopy suite for the embolization portion of the procedure. The intraarterial study can then be repeated post-embolization (Fig. 8).

The technique of endovascular meningioma embolization is well established [15‐18]. In brief, a microcatheter (1.9–2.3 Fr) is placed through the 5 Fr catheter in the ECA superselectively into the dural vessel supplying the tumor (usually the middle meningeal artery, a branch of the ECA) (Fig. 9). Microcatheter angiograms are performed to confirm that only tumor (and not normal critical structures such as the retina) is supplied by the catheterized vessel. Embolization is achieved by injection of 350–500 μm diameter plastic particles (polyvinyl alcohol, PVA) until stasis is achieved in the tumor-supplying artery. The microcatheter is then cleared by saline injection and pushable platinum coils are often placed more proximally in the feeding artery to achieve complete arterial occlusion. The microcatheter is removed, but the 5 Fr catheter remains in the ECA. The patient is then moved back into the MR scanner and selective IA perfusion studies are repeated with the 5 Fr catheter in the ECA and then pulled back into the CCA. After this second set of IA perfusion studies, the 5 Fr catheter is removed and the femoral arterial access site is closed.

The advanced MRI techniques that are used intraprocedurally during the course of endovascular embolization of meningiomas show promise for guiding both embolization and surgery [13]. Different IA perfusion techniques, including T2* dynamic susceptibility contrast (DSC) and T1 dynamic contrast enhanced (DCE), have been applied. As expected, T2* based techniques appear to be of limited value in heavily calcified meningiomas, but T1 based techniques may be of more use in such lesions. MR perfusion maps appear to be more sensitive in detecting residual tumor vascularity as compared to qualitative analysis of X-ray angiographic images [13]. ECA perfusion usually correlates well with dural tumor supply on X-ray angiography; ICA perfusion (CCA minus ECA) similarly usually correlates with pial supply. Diffusion-weighted imaging (DWI) is also being applied to evaluate for embolization-induced tumor infarction (or unanticipated ischemia of non-tumor tissue). Correlations between MR findings and regional histology of resected tumors are performed to validate MR findings but study numbers are not yet sufficient to define which combination of advanced MR techniques is optimal.

It is hoped that IA perfusion techniques will provide better preoperative insight into tumor vascularity, allowing surgeons to better anticipate which areas of tumor are still likely to bleed intraoperatively when embolization is incomplete. These techniques may also prove valuable in monitoring the effects of new embolic agents, such as liposomes carrying chemotherapeutics or anti-tumor antibodies. As minimally invasive MRI of tumor physiology improves and is validated against resected tumor histology, its ability to monitor nonsurgical treatments for dural and nondural brain neoplasms also comes closer to realization.