Presurgical fMRI and DTI for the Prediction of Perioperative Motor and Language Deficits in Primary or Metastatic Brain Lesions Paul D. Bailey, Domenico Zac` a, Mahmud Mossa Basha, Shruti Agarwal, Sachin K. Gujar, Haris I. Sair, John Eng, Jay J. Pillai From the Division of Neuroradiology, Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins Hospital & The Johns Hopkins University School of Medicine, Baltimore, MD (PDB, SA, SKG, HIS, JJP); Center for Mind/Brain Sciences, University of Trento, Trento, Italy (DZ); Department of Radiology, University of Washington, Seattle, WA (MMB); and Division of General and Diagnostic Radiology, Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins Hospital & The Johns Hopkins University School of Medicine, Baltimore, MD (JE). ABSTRACT BACKGROUND AND PURPOSE: To determine whether lesion to activation distance (LAD) on presurgical blood-oxygen- level-dependent functional magnetic resonance imaging (fMRI) and degree of white matter involvement by primary or metastatic brain lesions predict perioperative motor and language deficits. METHODS: We retrospectively evaluated 76 patients with intra-axial brain lesions referred for presurgical fMRI and diffusion tensor imaging (DTI). We measured expressive, receptive, global language and motor LAD and assessed degree of involvement of the corticospinal tract (CST) and the superior longitudinal fasciculus (SLF). We performed a Wilcoxon rank-sum test to determine whether the LAD and the degree of CST/SLF involvement were statistically significantly different between patients with and without preoperative or postoperative neurological deficits. RESULTS: In preoperatively symptomatic patients, motor and expressive language LAD were significantly lower (z = –3.78, P = .0002, and z = –2.51, P = .01, respectively) than in asymptomatic patients. No significant difference was noted in LAD between postoperative symptomatic and asymptomatic patients, except for a trend level effect for motor LAD (P = .07). The degree of CST involvement was significantly different between symptomatic and asymptomatic patients (z = 3.40, P = .0007 and z = 2.97, P = .003, respectively, for pre- and postoperative motor deficits).The degree of SLF involvement was significantly different between preoperatively (but not postoperatively) symptomatic and asymptomatic patients (z = 2.85, P = .004). CONCLUSION: Presurgical motor and expressive language LAD as well as degree of tract involvement on DTI are predictive of preoperative but not postoperative deficits, except for CST DTI and (trend level) motor LAD; inability of language LAD to predict postoperative deficits suggests that preoperative fMRI is valuable to neurosurgeons in avoiding resection of eloquent cortex. Keywords: Functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), brain tumor, neurological deficits, lesion to activation distance (LAD), white matter tract involvement. Acceptance: Received January 1, 2015. Accepted for publication May 21, 2015. Correspondence: Address correspondence to Jay J. Pillai, M.D., Division of Neuroradiology, Russell H. Morgan Department of Radiology and Radio- logical Science, The Johns Hopkins Hospital, Phipps B-100, 1800 Orleans Street, Baltimore, MD 21287. E-mail: jpillai1@jhmi.edu. Portions of this paper were presented in a talk at the 7th annual meeting of the American Society of Functional Neuroradiology (Charleston, SC 2013). J Neuroimaging 2015;00:1-9. DOI: 10.1111/jon.12273 Introduction The goal of neurosurgery in patients with brain tumors is to maximize the extent of resection while minimizing morbidity. In pursuing this aim, it is valuable to preoperatively determine the spatial relationship between cortical and subcortical elo- quent areas and lesion margins. Blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) are able to reliably and nonin- vasively provide this information. 7,11,19 Studies comparing the results of preoperative fMRI with intraoperative electric cortical stimulation have reported very high concordance (between 87 and 100%) for localization of motor and language areas between the two techniques. 4,8 Excellent agreement also has been re- ported between the preoperative BOLD imaging-based assess- ment of language lateralization and the results of the more in- vasive Wada test. 1,25 Preoperative DTI also has been validated (although less extensively than BOLD fMRI) by comparison to the gold standard, intraoperative subcortical stimulation, with good agreement for white matter tract localization. 3,14 Other studies have shown beneficial effects of fMRI/DTI on presur- gical planning and postoperative outcome, including reduction of postsurgical functional deficits, more effective intraoperative mapping and improved survival. 2,15,18,22,31 Several previous studies have investigated the relationship between pre- and postoperative neurological deficits and lesion proximity to eloquent cortex or eloquent white matter tracts. 5,9,23,24,26,32 These studies have demonstrated that greater proximity of lesions to eloquent cortex or greater degree of eloquent white matter involvement result in higher likelihood of functional deficits. However, the currently available studies have limited the investigation to either fMRI or DTI, have been performed on small patient samples 24,26 or have investigated only global language or motor function. 5,7,9,23,24,26,32 In this study, we investigated how the lesion to activation distance (LAD) determined by BOLD fMRI and degree of eloquent white matter tract involvement on DTI predict perioperative Copyright ◦ C 2015 by the American Society of Neuroimaging 1