Chapter 9 - Functional imaging in brain surgery

Despite the ongoing controversy on the prognostic significance of the extent of resection of brain gliomas and the reliance on nonrandomized retrospective studies, aggressive surgical resection correlates with better long-term survival and improves functional performance (Ciric et al., 1987, Sanai and Berger, 2008). In case of low-grade gliomas, aggressive resection lowers the recurrence rate, prolongs the time to tumor progression, and decreases the chance of malignant tumor transformation. The benefit for patients with high-grade tumors is related to prolonged time to progression and improved neurological performance, as well as longer survival. Further benefits of radical surgery are: more reliable tissue sampling, improved seizure control, immediate reduction of intracranial pressure, and decreased reliance on steroids.

The goal of brain tumor surgery might therefore be defined as preservation of neurological functions while maximizing the extent of tumor resection. Circumscribed brain tumors can be removed completely with a low risk to adjacent cortex. Infiltrative tumors, such as gliomas, may involve the eloquent cortex, even if it functions normally. Hence, postoperative neurological deficits may occur even when the surgical manipulation is limited strictly to the area of abnormality seen on magnetic resonance imaging (MRI). Furthermore, in the case of tumor-associated epilepsy, the epileptogenic zone may overlap with areas essential for crucial functions.

Preoperative and intraoperative visualization of eloquent structures is thus of utmost importance. The motor and sensory cortex can be localized at precentral and postcentral gyri on conventional T2-weighted MRI studies (Berger et al., 1990). Nevertheless, several factors preclude the reliable prediction of eloquent area location based on anatomical imaging. Large differences between individuals with respect to localization of these areas have been observed. Normal cortical anatomical landmarks may be distorted by the tumor (Lee et al., 1999). Tumor growth may lead to changes within the neuronal network as a result of displacement of functional important brain tissue. Due to the phenomenon of functional brain plasticity, new brain areas that are usually not involved in the performance of a certain task may be recruited (Thiel et al., 2001, Jacobs et al., 2005). A preoperative MRI-based definition of the central sulcus, for example, can be hampered in as many as half of the cases (Inoue et al., 1999).

Intraoperative cortical stimulation has been regarded as the gold standard in mapping of eloquent areas, such as those subserving language, motor, and sensory functions. It is performed either by direct cortical stimulation in the awake patient or by recording sensory evoked potentials from electrodes applied to the surface of the brain. The method is widely used and highly reliable. However, it does not allow preoperative planning and prolongs surgery. The set-up is time-consuming, expensive, and difficult to perform in pediatric patients. For mapping of language areas, the patient should be cooperative, with preserved speech function and reading. Furthermore, eloquent language areas may be missed because of the low specificity of the testing paradigm. It is well known that the risk of intraoperative seizures is reduced when low-intensity current stimulation is used. However, low-intensity stimulation may fail to identify some essential areas.

Functional imaging (FI), which includes several techniques – functional MRI (fMRI), positron emission tomography (PET), magnetic source imaging, diffusion tensor imaging, and tractography – is a rapidly advancing alternative that allows for visualization of spatial relationships between eloquent structures and tumor, and provides information essential in planning and performing the surgery, as well as in predicting the outcome.