Neurosurgical Applications of Magnetic Resonance Diffusion Tensor Imaging

Magnetic Resonance (MR) Diffusion Tensor Imaging (DTI) is a rapidly evolving technology that enables the visualization of neural fiber bundles, or white matter (WM) tracts. There are numerous neurosurgical applications for MR DTI including: (1) Tumor grading and staging; (2) Pre-surgical planning (determination of resectability, determination of surgical approach, identification of WM tracts at risk); (3) Intraoperative navigation (tumor resection that spares WM damage, epilepsy resection that spares WM damage, accurate location of deep brain stimulation structures); (4) Post-operative assessment and monitoring (identification of WM damage, identification of tumor recurrence). Limitations of MR DTI include difficulty tracking small and crossing WM tracts, lack of standardized data acquisition and post-processing techniques, and practical equipment, software, and timing considerations. Overall, MR DTI is a useful tool for planning, performing, and following neurosurgical procedures, and has the potential to significantly improve patient care. Technological improvements and increased familiarity with DTI among clinicians are next steps. Introduction Magnetic Resonance (MR) imaging uses magnetic fields to temporarily alter proton (hydrogen atom) orientation and then measures the energy emitted upon proton relaxation, enabling discrimination of tissues with different proton (water) compositions. Water molecules naturally diffuse in accordance with Brownian motion (imagine a drop of dye spreading out in a glass of water). A series of magnetic pulses can be applied to measure the inter-pulse magnitude and direction of proton diffusion. On a pixel-by-pixel basis, this diffusion is described by the Apparent Diffusion Coefficient (ADC), which can be determined in multiple axes. Mori et al1 found that application of the diffusion pulse in a minimum of six directional axes is sufficient to resolve a diffusion vector in three dimensional space describing the overall diffusion for a given pixel, called a tensor (thus the name diffusion tensor imaging (DTI)). This approach has been particularly useful in identifying myelinated axons.The term anisotropy refers to the degree by which protons diffuse predominantly in a single direction. Myelinated fibers are relatively anisotropic with diffusion preferentially along the axis of the fiber. DTI data are depicted in parametric maps that assign colors to different directions (e.g., anterior, posterior, ventral, dorsal, right, left). Thus, MR DTI visually depicts the water molecules within myelinated neurons, crudely outlining WM tracts. DTI has been validated by comparison with experimental histological specimens. Further proof of concept includes experiments where DTI-identified WM tracts were electrically stimulated and produced predicted physiologic responses. Traditionally, subcortical stimulation mapping has served as the gold standard for intraoperative neuronavigation, yet this technique does not visually delineate the intraparenchymal path of WM tracts. In contrast, DTI depicts WM tracts as they course through the central nervous system. Numerous innovative clinical applications of DTI have been described in the literature. Herein we thematically describe them and discuss limitations and future directions. Tumor grading & staging Tumor evaluation with DTI enables discrimination between different types of CNS lesions and visualization of WM tracts depicts WM-tumor interactions. Lazar et al2 evaluated preoperative DTI images of 6 patients with brain lesions and observed various patterns of tumor-induced damage, which were categorized into deviation, deformation, infiltration, or apparent tract interruption. Preoperative knowledge of the WM-tumor interaction contributed to good clinical outcomes, as 4 patients with preoperative impaired motor functioning experienced complete symptom resolution postoperatively. Chen et al3 applied this knowledge in a study of 10 patients with brainstem lesions. Prior to resection, some form of deviation, deformation, infiltration, or apparent tract interruption was diagnosed in each patient. Visualization of the tracts again after surgery ensured the tracts returned to their proper location.The authors concluded that WM tract imaging provided abundant risk stratification and prognosis information. DTI can be used to evaluate specific tumor characteristics including extent of infiltration. One parameter called fractional anisotropy (FA) is a scalar value (ranging from 0-1) and is used to describe the degree of anisotropy of a diffusion process. Deng et al4 found a negative correlation between the FA value and degree of tumor infiltration in twenty patients with gliomas, as lower FA values were observed in the areas of higher glioma infiltration. FA is a promising quantifiable marker of tumor infiltration (that cannot be otherwise determined from conventional MR images). FA values aid differentiation between tumor types. Byrnes et al5 studied 28 patients with either glioblastoma or brain metastases using FA values. Mean FA was significantly lower in the edema surrounding metastatic tumors than surrounding glioblastomas. Imaging was able to accurately discriminate between tumor type for 87.5% (14 of 16) of glioblastomas and 83.3% (10 of 12) of metastases, as validated by histology. Similarly, Tropine et al6 used various DTI metrics to distinguish between fibroblastic and benign meningiomas, concluding that FA values are the valuable predictors. After evaluating 30 patients with WHO grade 1 meningiomas, the authors reported that in comparison to benign subtypes, fibroblastic meningiomas present with higher FA values. Interestingly, the two categories demonstrate different tensor shapes; while tensors formed by benign meningiomas are predominantly spherically shaped (80%), a large amount of fibroblastic meningioma tensors are nonspherically shaped (43%). Jolapara et al7 studied 21 tumor patients using DTI and found that atypical and fibroblastic meningiomas had higher mean FA value than Daniel D. Hirsch, BS1; Benjamin M. Zussman, BS1; Adam E. Flanders, MD2; Ashwini D. Sharan, MD3 1Jefferson Medical College, Philadelphia, Pennsylvania 2Radiology Department, Thomas Jefferson University, Philadelphia, Pennsylvania 3Neurosurgery Department, Thomas Jefferson University, Philadelphia, Pennsylvania Neurosurgical Applications of Magnetic Resonance Diffusion Tensor Imaging benign meningiomas. The authors also evaluated Spherical Anisotropy, another measure of FA looking at the degree to which molecules are traveling in equal directions, and found higher Spherical Anisotropy values in benign meningiomas when compared to atypical and fibroblastic meningiomas. No reliable method of differentiating between atypical and fibroblastic meningiomas was found. Finally, Xu et al8 determined that FA values are useful in differentiating between recurrent tumors and radiation-induced injury. Here, thirty-five glioma patients who had previously undergone radiation therapy underwent DTI. The average FA values were significantly higher in the group of recurrent tumors than that of the radiation-induced injury group. These studies demonstrate the diagnostic power of DTI. Presurgical planning Before a patient’s operation begins, DTI information can assist surgical planning in several ways. It may be used to evaluate tumor respectability and determine surgical feasibility. Setzer et al9 studied 14 patients with intramedullary spinal cord tumors and categorized them according to the interaction between the lesion and the surrounding WMtracts. Lesions were considered resectable (Type 1) when no fibers entered the lesion. Type 2 consisted of lesions that contained only the minority of fibers from a given tract, and was considered resectable only if less than 50% of the tumor, by volume, contained fibers. Lesions were deemed nonresectable (Type 3) when the majority of the lesion contained fibers or the tumor had already demonstrated destruction of fibers. These classifications were clinically translatable: all 5 Type 1 lesions were fully resected, the Type 2 case deemed resectable was fully resected, while 1 of 2 unresectable Type 2 tumors was unresectable, and 5 of 6 Type 3 lesions were unresectable, as evidenced at time of biopsy. Surgical planning is enhanced by preoperative visualization of WM tract location and orientation. Yu et al10 studied 16 brain tumor patients using DTI to reconstruct lesion location and relationship to the surrounding WM, which informed surgical planning that preserved vital tracts and maximized tumor resection. The study group demonstrated a significantly higher extent of tumor removal and postoperative improvement in locomotor function when compared to a control group whose preoperative planning included only conventional MRI methods. Qiu et al11 enrolled 45 patients with suspected gliomas and used DTI to acquire a better understanding of the anatomical relationship between the tumor and pyramidal tract, including the direction of the pyramidal tract to the tumor, how the lesion invaded the pyramidal tract, and the distance between them. The authors noted that because this information was available to them in the planning stage, a surgical approach that was unambiguous and Figure 1 (A) T1 gadolinium-enhanced axial view of right-sided cranial tumor; (B) Axial color-coded DTI image of with tumor circumscribed in red; (C) 3D rendering of tumor/fiber relationship with tumor and fibers as opaque objects; (D) Translucent tumor with cutaway view. A B