Magnetic Resonance Imaging

MRI is essential for evaluating cerebral gliomas both prior to and following their treatment. The detailed definition of normal anatomy and the sensitivity for determining tumor extension provided by MRI are essential for planning surgical resection and subsequent postoperative radiotherapy. Preoperative MRI is excluded only in patients in whom MRI is contraindicated for safety reasons. All other patients are studied with MRI. Those with severe claustrophobia or who are otherwise unable to hold absolutely still for their examination are given appropriate sedation or, rarely, even general anaesthesia to obtain the necessary MRI studies.

Tumor evaluation with MRI requires the intravenous administration of a contrast agent. These are gadolinium (Gd) chelates such as gadolinium diethylenetriaminepentaacetic acid (GdDTPA). Gadolinium agents for MRI are between 50 and 100 times more sensitive to blood-brain barrier breakdown than iodinated contrast agents used with CT. The usual dose of 0.1 mmol Gd/kg is approximately one-tenth that of an iodinated contrast agent. This lower dose decreases the risk of adverse reaction while at the same time providing a level of enhancement several-fold higher than seen with contrast-enhanced CT.

When combined with Gd-based intravenous contrast, MRI is superior to CT for differentiating between tumor and perifocal edema, for defining gross extent of tumor, and for showing the relationship of the tumor to critical adjacent structures. This information is essential for planning stereotactic biopsy or tumor resection and for planning radiotherapy. Often, total resection is precluded by tumor extension into critical structures. Surgery is then planned with the goal of maximum subtotal resection with minimum neurological morbidity. Precise definition of normal and abnormal cerebral anatomy is necessary to achieve this.

The physiologic mechanism of enhancement with Gd, namely disruption of the blood-brain barrier. is the same as the mechanism of enhancement for CT contrast agents. However, there are important fundamental differences in imaging characteristics between iodinated contrast for CT and Gd-based contrast for MRI. Iodinated agents are directly visualized on CT as bright areas due to their increased x-ray absorption. Gadolinium contrast agents are not directly visualized on MR but are indirectly imaged because of their effect on the two modes of MR signal decay, T1 and T2. When Gd atoms are in extremely close proximity (a few nanometres) to water protons excited by an MR pulse. they cause marked shortening of T1 relaxation time and a lesser degree of shortening of T2 relaxation time of these protons. T1 shortening increases signal on T1-weighted images and, hence. we visualize an enhanced area of signal that appears bright. T2 shortening can cause loss of signal on T2-weighted images, but this effect is minimal at clinically used dosages and is of no clinical significance. The major point is that there is little or no effect of Gd contrast agents on T2-weighted images. With an intact blood-brain barrier, Gd remains within the capillary space; there is no enhancement because Gd cannot gain very close access to interstitial water molecules. Also different from contrast-enhanced CT. vessels that contain rapidly flowing blood are not enhanced with gadolinium-enhanced MRI because the flowing protons do not remain in the MR slice volume long enough to be imaged. However, slowly flowing blood, as occurs in veins and venous sinuses, may enhance. With MRA there are circumstances in which Gd can improve the detectability of slowly flowing blood and improve visualization of small vessels.

In MRI and CT of adult gliomas the degree and pattern of tumor enhancement roughly correlates with tumor grade. This is a rough correlation and not an absolute determinant. Furthermore, this correlation only applies to adult gliomas and is not predictive for other primary intracerebral tumors. for paediatric cerebral tumors, or with extra-axial cerebral masses such as meningiomas.

Heavily T2-weighted sequences are the most sensitive for the detection of tumor and edema extent, but the tumor focus is not well separated from surrounding edema. T1-weighted images following contrast enhancement generally provide better localization of the tumor nidus and improved diagnostic information relating to tumor grade, blood-brain barrier breakdown, hemorrhage, edema, and necrosis. Contrast-enhanced T1-weighted images also better show small focal lesions such as metastases, small areas of tumor recurrence, and ependymal or leptomeningeal tumor spread because of improved signal contrast. T1-weighted images without contrast are less sensitive to tumor and edema but are necessary for comparison with postcontrast images and for characterization of enhancement pattern, hemorrhage, cysts, and necrosis. Proton density images are useful for distinguishing tumor and edema from adjacent cerebrospinal fluid (CSF), which may have a similar appearance as high-signal areas on heavily T2­weighted images.

We obtain both T1- and spin echo proton density and T2­weighted images without contrast followed by postcontrast T1­weighted images after the intravenous injection of gadolinium. Imaging is done in sagittal, axial, and coronal planes, which provides optimal detail for initial treatment planning. Following treatment, these sequences provide the most sensitive and accurate method for determining tumor response to therapy or for detecting early tumor recurrence.

In selected cases. additional MR sequences are used to clarify specific diagnostic questions or to provide additional information. For example, gradient echo imaging can be used to detect occult hemosiderin from prior subclinical bleeding. In a case where surgical access to the tumor may be problematic, three-dimensional gradient echo imaging permits very thin sections that can be reformatted at any desired plane of obliquity. Combined with the appropriate software, this information can be used to construct surface maps of the brain overlying a tumor that will provide sulcal and gyral anatomy preoperatively, or to display "cutaway views" that can provide views of different operative approaches for surgical planning.

There are also important limitations to MRI that one must be aware of. Neither CT nor MRI can distinguish peritumoral edema from nonenhancing infiltrating tumor. Furthermore, even when there is a well-defined enhancing tumor nidus, infiltrating tumor and isolated tumor cells can extend several centimetres beyond the enhancing region into the surrounding "edematous" zone and, in some cases, beyond any abnormality seen on CT or MRI. Finally, for all practical purposes, bulk calcium emits no MR signal, making tumor calcification difficult or impossible to detect unless present in large amounts.