RESERVATION of vital cerebral tissue while maximiz- ing tumor resection is a principal goal in surgical neurooncology. Considering that most patients with tumors in eloquent areas such as the motor cortex are neu- rologically intact or only slightly impaired, it is essential to distinguish functional brain tissue from tumor tissue to avoid causing neurological deterioration during excision of these lesions. Although routine structural MR images can accurate- ly demonstrate brain tumors, they do not give precise in- formation about the involvement and integrity of the white matter tracts in the immediate region surrounding tumors. The high-intensity signal often seen in the white matter ad- jacent to a tumor on T 2 -weighted or FLAIR images may represent either tumor extension or edema in the surround- ing normal white matter tracts. More detailed characteriza- tion of white matter tract integrity surrounding tumors may be helpful in the surgical planning and treatment of patients with intrinsic brain tumors. Several functional approaches such as fMR imaging and intraoperative mapping are used in the presurgical localiza- tion of eloquent cortex near brain tumors. 11,20,24 Functional MR imaging allows for the identification of important func- tional areas of the cerebral cortex that may be invaded by a neoplasm. This imaging modality focuses on cortical struc- tures but does not provide information about subcortical gray matter and white matter, which in many instances may be involved in invasive, intrinsic brain tumors. Recently, DT imaging has been used to map white mat- ter tracts in the brain. 13,22,23 Diffusion-tensor imaging mea- sures the diffusion displacement properties of water in a three-dimensional space as a function of image location. 7,8,23 Water diffusion parallel to the white matter tracts is less re- stricted than water diffusion perpendicular to them. Conse- quently, the measured image signals are higher for diffu- sion-gradient encoding perpendicular to the white matter tracts rather than parallel. 2,16 This directional variation in the signal is termed “diffusion anisotropy.” By acquiring DW images with at least six noncollinear gradient-encoding directions, it is possible to estimate the DT. The principal eigenvector represents the direction of greatest diffusion, which also corresponds to the fiber tract axis. With this di- J Neurosurg 97:568–575, 2002 568 Diffusion-tensor imaging of white matter tracts in patients with cerebral neoplasm BRIAN P. WITWER, M.D., ROHAM MOFTAKHAR, M.D., KHADER M. HASAN, PH.D., PRAVEEN DESHMUKH, M.D., VICTOR HAUGHTON, M.D., AARON FIELD, M.D., PH.D., KONSTANTINOS ARFANAKIS, PH.D., JANE NOYES, N.P., CHAD H. MORITZ, B.S., M. ELIZABETH MEYERAND, PH.D., HOWARD A. ROWLEY , M.D., ANDREW L. ALEXANDER, PH.D., AND BEHNAM BADIE, M.D. Departments of Neurological Surgery, Radiology, Medical Physics, and Psychiatry, University of Wisconsin School of Medicine, Madison, Wisconsin Object. Preserving vital cerebral function while maximizing tumor resection is a principal goal in surgical neurooncol- ogy. Although functional magnetic resonance imaging has been useful in the localization of eloquent cerebral cortex, this method does not provide information about the white matter tracts that may be involved in invasive, intrinsic brain tumors. Recently, diffusion-tensor (DT) imaging techniques have been used to map white matter tracts in the normal brain. The aim of this study was to demonstrate the role of DT imaging in preoperative mapping of white matter tracts in relation to cerebral neoplasms. Methods. Nine patients with brain malignancies (one pilocytic astrocytoma, five oligodendrogliomas, one low-grade oli- goastrocytoma, one Grade 4 astrocytoma, and one metastatic adenocarcinoma) underwent DT imaging examinations prior to tumor excision. Anatomical information about white matter tract location, orientation, and projections was obtained in every patient. Depending on the tumor type and location, evidence of white matter tract edema (two patients), infiltration (two patients), displacement (five patients), and disruption (two patients) could be assessed with the aid of DT imaging in each case. Conclusions. Diffusion-tensor imaging allowed for visualization of white matter tracts and was found to be beneficial in the surgical planning for patients with intrinsic brain tumors. The authors’ experience with DT imaging indicates that anatomically intact fibers may be present in abnormal-appearing areas of the brain. Whether resection of these involved fibers results in subtle postoperative neurological deficits requires further systematic study. KEY WORDS • brain neoplasm • diffusion-tensor imaging • functional magnetic resonance imaging P J. Neurosurg. / Volume 97 / September, 2002 Abbreviations used in this paper: DT = diffusion-tensor; DW = diffusion-weighted; FLAIR = fluid-attenuated inversion-recovery; fMR = functional magnetic resonance.