Spectrum of Magnetic Resonance Imaging Features in Unilateral Optic Tract Dysfunction

Background: Optic tract dysfunction may be the predominant or only clinical manifestation of an intracranial disorder including mass legion, ischemic infarct, inflammatory disease, and trauma. Documentation of the neuroimaging features of these lesions is limited to reports mostly published before the availability of MRI. This study was undertaken to document the spectrum of MRI features in patients presenting with optic tract dysfunction. Methods: A retrospective study from 2004 to 2015 at a single tertiary care neuro-ophthalmology service of 24 patients who had unilateral optic tract dysfunction defined by a homonymous hemianopia and a relative afferent pupil defect that could not be attributed to optic neuropathy or retinopathy. Two institutional neuroradiologists, who were privy to the presence of optic tract dysfunction but not to its side or cause, independently documented the MRI abnormalities on a standard data collection form and then convened for a consensus review of the imaging abnormalities with the 2 clinician authors. Results: The clinical diagnoses were 6 ischemic strokes, 5 malignant brain tumors, 5 postoperative neurosurgical cases, 4 intracranial hemorrhages, 2 traumatic brain injuries, 1 midbrain/optic tract primary demyelination, and 1 temporal lobe herpes simplex encephalitis. In their independent review, both neuroradiologists identified MRI abnormalities in 20 (83%) cases that were extrinsic to the optic tract in the neighboring temporal lobe, midbrain, thalamus, basal ganglia, or suprasellar space. In 5 of those cases, the extrinsic abnormality included features suggesting compression of the optic tract, but these compressive features were not appreciated by either neuroradiologist until the consensus conference. In 15 cases, MRI abnormalities intrinsic to the optic tract itself were eventually identified, including T2 or fluid-attenuated inversion recovery image (FLAIR) hyperintensity (9 cases) or hypointensity (1 case), thinning (6 cases), thickening (2 cases), and contrast enhancement (1 case). However, none of these intrinsic MRI abnormalities was identified during the independent review, being detected only in the consensus conference. Conclusions: Neuroradiologists aware of unilateral optic tract dysfunction but not of its side detected extrinsic (neighborhood) MRI abnormalities in most cases but did not appreciate that these extrinsic features sometimes included compression of the optic tract. MRI abnormalities intrinsic to the optic tract were entirely overlooked during independent review, being recognized only in a consensus conference with clinician authors. Neuroradiologists are more likely to detect MRI abnormalities pertinent to optic tract dysfunction once they have more complete clinical information and with higher resolution imaging, especially T1 postcontrast axial and coronal sequences and T2 or FLAIR coronal scans.

Original Contribution Spectrum of Magnetic Resonance Imaging Features in Unilateral Optic Tract Dysfunction Kristopher M. Kowal, MD, Francisco F. Rivas Rodriguez, MD, Ashok Srinivasan, MD, Jonathan D. Trobe, MD Background: Optic tract dysfunction may be the predominant or only clinical manifestation of an intracranial disorder including mass legion, ischemic infarct, inflammatory disease, and trauma. Documentation of the neuroimaging features of these lesions is limited to reports mostly published before the availability of MRI. This study was undertaken to document the spectrum of MRI features in patients presenting with optic tract dysfunction. Methods: A retrospective study from 2004 to 2015 at a single tertiary care neuro-ophthalmology service of 24 patients who had unilateral optic tract dysfunction defined by a homonymous hemianopia and a relative afferent pupil defect that could not be attributed to optic neuropathy or retinopathy. Two institutional neuroradiologists, who were privy to the presence of optic tract dysfunction but not to its side or cause, independently documented the MRI abnormalities on a standard data collection form and then convened for a consensus review of the imaging abnormalities with the 2 clinician authors. Results: The clinical diagnoses were 6 ischemic strokes, 5 malignant brain tumors, 5 postoperative neurosurgical cases, 4 intracranial hemorrhages, 2 traumatic brain injuries, 1 midbrain/optic tract primary demyelination, and 1 temporal lobe herpes simplex encephalitis. In their independent review, both neuroradiologists identified MRI abnormalities in 20 (83%) cases that were extrinsic to the optic tract in the neighboring temporal lobe, midbrain, thalamus, basal ganglia, or suprasellar space. In 5 of those cases, the extrinsic abnormality included features suggesting compression of the optic tract, but these compressive features were not appreciated by either neuroradiologist until the consensus conference. In 15 cases, MRI abnormalities intrinsic to the optic tract itself were eventually identified, including T2 or fluid-attenuated inversion recovery image (FLAIR) hyperintensity (9 cases) or hypointensity (1 case), thinning (6 cases), thickening (2 cases), and contrast enhancement (1 case). However, none of these intrinsic MRI abnormalities Departments of Ophthalmology and Visual Sciences (KMK, JDT), Radiology (Neuroradiology) (FFRR, AS) and Neurology (JDT), Kellogg Eye Center, University of Michigan, Ann Arbor, Michigan. The authors report no conflicts of interest. Address correspondence to Jonathan D. Trobe, MD, Kellogg Eye Center, 1000 Wall Street, Ann Arbor, MI 49105; E-mail: [e-mail] Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 was identified during the independent review, being detected only in the consensus conference. Conclusions: Neuroradiologists aware of unilateral optic tract dysfunction but not of its side detected extrinsic (neighborhood) MRI abnormalities in most cases but did not appreciate that these extrinsic features sometimes included compression of the optic tract. MRI abnormalities intrinsic to the optic tract were entirely overlooked during independent review, being recognized only in a consensus conference with clinician authors. Neuroradiologists are more likely to detect MRI abnormalities pertinent to optic tract dysfunction once they have more complete clinical information and with higher resolution imaging, especially T1 postcontrast axial and coronal sequences and T2 or FLAIR coronal scans. Journal of Neuro-Ophthalmology 2017;37:17-23 doi: 10.1097/WNO.0000000000000411 © 2016 by North American Neuro-Ophthalmology Society T he optic tract carries axons from the optic chiasm principally to the lateral geniculate nucleus. It may be damaged by lesions that arise from the sellar, suprasellar, or anterior medial temporal regions, particularly neoplasms, vascular malformations, and carotid aneurysms (1-4), and also by head trauma (5,6) and inflammatory disorders (7). With a blood supply from the anterior choroidal artery, the optic tract is surrounded by cerebrospinal fluid proximally, where it is visible only on a single axial MRI slice of conventional thickness. Because it is a relatively thin structure wedged between the cerebral peduncle and anterior temporal lobe, its distal segment is not visible on axial sections and hardly visible on coronal sections (Fig. 1); therefore, it may escape the attention of neuroradiologists. Reports documenting the associated optic tract MRI abnormalities (6-14) are limited and clinical correlation has been even more limited. The goal of this study was to extend the documentation of MRI features of optic tract dysfunction by formally analyzing a large series of patients. 17 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 1. Anatomy of the optic tract in axial (top) and at designated coronal sections A-D (bottom). Note that the proximal optic tract is likely to be more visible on magnetic resonance imaging because it is surrounded by the cerebrospinal fluid. The more distal portion is barely visible, being wedged between the midbrain medially and the temporal lobe laterally. (Modified by David Murrel from Nolte J, Angevine JB. The Human Brain in Photographs and Diagrams. Third edition. Philadelphia, PA, Mosby Elsevier, 2007). 18 Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution METHODS We identified patients with clinically defined optic tract dysfunction from the case records of one of the authors (J.D.T.) dating back no farther than 2004 to encompass the modern imaging era in which MRI protocols did not substantially change. Patients had to display a homonymous hemianopia on reliable automated (Humphrey) visual field testing, together with an afferent pupil defect on the side of the hemianopia that could not be attributed to an ipsilateral optic neuropathy or retinopathy. Bow-tie optic disc pallor supported localization to the optic tract in 50% of cases. We included patients who had undergone a motion-free MRI brain or orbit study before or after the clinical diagnosis. Twenty-four cases met these criteria. Two board-certified neuroradiologists independently reviewed the MRI scans closest to the date of clinical diagnosis. One reviewer had 10 years' experience (A.S.), the TABLE 1. Demographics, diagnoses and neuroimaging findings in patients with optic tract syndrome Case Age (years)/ Gender Diagnosis Intrinsic MRI Signs 1 2 57/F 48/F Posterior Cerebral Artery Infarct Posterior Cerebral Artery Infarct None None 3 4 48/F 28/F Anterior Choroidal Artery Infarct Posterior Cerebral Artery Infarct T2 hyperintensity None 5 65/F Middle Cerebral Artery Infarct 6 56/M Middle Cerebral Artery Infarct 7 20/M 8 51/M 9 49/M 10 21/F Deep Cerebral Venous Sinus Thrombosis Temporal Lobe Grade IV Astrocytoma Temporal Lobe Grade IV Astrocytoma Temporal Lobe Oligoastrocytoma 11 12 63/F 52/F 13 53/M Craniopharyngioma Primary Central Nervous System (CNS) Lymphoma Biopsy of Primary CNS Lymphoma 14 57/F Biopsy of Temporal Lobe Metastasis 15 17/M 16 20/M 17 18 35/M 66/M 19 42/M 20 34/F 21 22 23 63/M 47/F 57/M 24 51/F Resection of Middle Cranial Fossa Tumor (Pathology Unknown) Resection of Temporal Lobe for Intractable Seizures Hemorrhagic Midbrain Cavernoma Hemorrhagic Basal Ganglia Cavernoma Arteriovenous Malformation of Mesial Temporal Lobe Postpartum Basal Ganglionic Hemorrhage Traumatic Brain Injury Traumatic Brain Injury Idiopathic Midbrain and Optic Tract Demyelination Herpes Simplex Encephalitis None Thinning, T2 hyperintensity None None Obliteration Thickening, T2 Hyperintensity None T2 Hyperintensity Enhancement, T2 hyperintensity Thickening, T2 hyperintensity None Thinning Location of Extrinsic MRI Findings Temporal lobe, noncompressive Temporal lobe and lateral geniculate nucleus, noncompressive None (Fig. 9) Temporal lobe and lateral geniculate nucleus, noncompressive Temporal lobe and midbrain, noncompressive Temporal lobe, noncompressive Lateral geniculate nucleus, noncompressive Temporal lobe and midbrain, compressive (Fig. 4) Temporal lobe, compressive Temporal lobe, compressive Suprasellar, compressive (Fig. 5) Temporal lobe and midbrain, noncompressive (Fig. 3) Midbrain, noncompressive Temporal lobe and midbrain, noncompressive (Fig. 2) Temporal lobe, noncompressive None (Fig. 8) None Midbrain, noncompressive T2 hypointensity and Basal ganglia, noncompressive (Fig. 6) hyperintensity Enhancement Temporal lobe and midbrain, noncompressive Thinning Temporal lobe and midbrain and basal ganglia, noncompressive T2 hyperintensity None (Fig. 7) Thinning None Thinning Midbrain, noncompressive T2 hyperintensity Temporal lobe, compressive F, female; M, male; MRI, magnetic resonance imaging. Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 19 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 2. Case 14. Left optic tract dysfunction associated with anterior temporal lobe breast cancer metastasis. Axial T2 magnetic resonance imaging shows subtle left optic tract thickening and signal hyperintensity (arrow). Note hyperintensity in the adjacent left temporal lobe. other had 7 years' experience (F.F.R.R.). In 17 cases (70%), the MRI had been performed within 3 months of the diagnosis of optic tract dysfunction. In 7 cases, the MRI had been performed 6, 11, 12, 14, 14, 15, and 16 months after the clinical diagnosis. During that interval, there had been no change in the patients' clinical findings. The neuroradiologists had been informed that all cases had optic tract dysfunction on examination but were not told of the side of the lesion or the clinical diagnosis. They were charged with reviewing particularly the imaging that covered the area between the retrobulbar optic nerves and visual cortex. They documented their findings on a spreadsheet corresponding to the following pulse sequences: precontrast and postcontrast axial T1, axial T2, axial fluid-attenuated inversion recovery image (FLAIR), axial T2*, pre-contrast coronal T1, coronal T2, postcontrast axial, and coronal T1. The following sequences were available on all patients: axial FLAIR, axial T2, precontrast axial T1, and postcontrast axial and coronal T1. The slice thickness of the sequences ranged from 3 to 5 mm (brain studies 4-5 mm; orbit studies 3 mm). Studies were performed on a 1.5-T magnet (16 cases) or 3-T magnet (4 cases). Four studies performed at another institution were of unknown field strength. Because we were interested in comparing radiologic diagnosis reached after independent review by each neuroradiologist and after consensus review of neuroradiologists and clinician authors, we conducted a second phase of radiologic interpretation. After the neuroradiologists had completed their independent assessments of the imaging features and selected a favored diagnosis, all 4 authors gathered for a consensus conference in which the side of the lesion and its diagnosis were revealed to the neuroradiologists, who were invited to reassess the imaging features. At the consensus conference, the imaging abnormalities were divided into 2 groups: those in which the imaging abnormalities did not involve the optic tract but only its neighboring structures ("extrinsic abnormalities") and those that involved the optic tract itself ("intrinsic abnormalities"). The extrinsic abnormalities were further divided into those that indented, displaced, or obliterated the optic tract ("extrinsic compressive") or had no apparent compressive effect on the optic tract ("extrinsic noncompressive"). RESULTS Diagnoses Diagnoses were based on clinical and pathologic data. There were 7 ischemic strokes (4 posterior cerebral artery, 2 middle cerebral artery, and 1 deep venous sinus thrombosis), 5 tumors (3 astrocytomas, 1 craniopharyngioma, and 1 FIG. 3. Case 12. Right optic tract dysfunction after biopsy of right anterior temporal lobe lymphoma. Axial T2 scan (A) shows signal hyperintensity in right optic tract (arrow) compared with normal signal intensity in unaffected left optic tract. Postcontrast coronal T1 magnetic resonance imaging (B) shows enhancement of lymphoma (arrow) and involvement of the right optic tract; compare with lack of enhancement in unaffected left optic tract (arrowhead). 20 Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution suprasellar space. In 5 of these 20 cases (Cases 8, 9, 10, 11, and 24), extrinsic imaging abnormalities suggested compression of the optic tract, including indentation, displacement, or obliteration. These compressive features were caused by 3 temporal lobe lesions, 1 craniopharyngioma, and 1 herpes encephalitis. In the independent review, neither neuroradiologist had noted any of these compressive features. FIG. 4. Case 8. Left optic tract dysfunction caused by compression from temporal lobe malignant astrocytoma. Axial T2 scan shows that the left optic tract disappears (arrow) as it curves around the midbrain in the region of the adjacent temporal lobe mass; compare with normal size of the right optic tract (arrowhead). lymphoma), 4 neurosurgical procedures (2 biopsies and 2 resections), 4 hemorrhages (2 cavernomas, 1 arteriovenous malformation, and 1 postpartum), 2 traumatic brain injuries, 1 midbrain/optic tract primary demyelination, and 1 temporal lobe herpes simplex encephalitis (Table 1). Imaging Features Extrinsic Imaging Abnormalities During the independent review, imaging abnormalities involving the neighborhood of the optic tract ("extrinsic imaging abnormalities") were identified by both neuroradiologists in 20 (83%) of 24 cases. These extrinsic abnormalities included signal alteration in the nearby temporal lobe, midbrain, basal ganglia, lateral geniculate nucleus, or Intrinsic Imaging Abnormalities Intrinsic imaging abnormalities involving the optic tract itself were identified in 15 (62%) cases, including hyperintensity on T2 or FLAIR scans in 9 cases, thinning in 5 cases, thickening in 2 cases, T2 hypointensity in 1 case, and contrast enhancement in 1 case. These intrinsic abnormalities escaped detection during the independent review, coming to attention only in the consensus conference. Postcontrast coronal T1 sequences, available in most cases, increased detection of these intrinsic abnormalities. In some cases, intrinsic abnormalities were visible only on 3-mm slice thickness coronal T2 sequences. Among the 15 cases with intrinsic imaging abnormalities, 4 had no accompanying extrinsic MRI abnormalities (2 cases of traumatic brain injury, 1 case after temporal lobectomy, and 1 case from presumed spontaneous anterior choroidal artery infarction). Thus, in 4 (16%) of the cases, the neuroradiologists reviewing independently did not identify any imaging abnormality pertinent to the optic tract. Given that neither the compressive features of the extrinsic imaging abnormalities nor the intrinsic imaging abnormalities were identified during independent review, the neuroradiologists would not have necessarily surmised that optic tract dysfunction had occurred in any of the cases in this series. Even after they convened with the clinicians, they would not have surmised optic tract dysfunction in the 7 (29%) cases (Cases 1, 2, 4, 5, 7, 15, and 17) that lacked extrinsic compressive or intrinsic imaging abnormalities. FIG. 5. Case 11. Left optic tract dysfunction from suprasellar craniopharyngioma. Coronal T2 (A) shows hyperintensity of a cystic mass (arrows) involving the left side of the chiasm. Postcontrast coronal T1 image at a slightly posterior location (B) demonstrates upward displacement of the left optic tract (arrow); compare with the normal position of unaffected right optic tract (arrowhead). Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 21 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 6. Case 18. Right optic tract dysfunction from hemorrhagic arteriovenous malformation of basal ganglia (not visualized). Axial T2 scan (A) shows signal hypointensity of right optic tract (arrow) from hemosiderin; compare with normal signal intensity of unaffected left optic tract (arrowhead). Axial T2 image at a slightly different level (B) shows signal hypointensity in the periphery of the right optic tract (arrow) and signal hyperintensity in the center of the right optic tract (arrowhead) from methemoglobin, reflecting different ages of blood products. Postcontrast coronal T1 scan (C) shows hook-shaped enhancement within the right optic tract (arrow); compare with lack of enhancement of unaffected left optic tract. Examples of intrinsic and extrinsic noncompressive MRI abnormalities are provided in Figures 2 and 3. Examples of extrinsic compressive MRI abnormalities are provided in Figures 4 and 5. An example of extrinsic and intrinsic MRI abnormalities is provided in Figure 6. Examples of purely intrinsic MRI abnormalities are provided in Figures 7-9. DISCUSSION Our study extends the documentation of MRI abnormalities in patients with optic tract dysfunction and provides FIG. 7. Case 21. Right optic tract dysfunction from closed head trauma. Precontrast axial T1 MRI shows thinning of the right optic tract (arrow); compare with normal size of unaffected left optic tract (arrowhead). 22 more definitive clinical correlation. Previous reports have documented increased T2 or FLAIR signal in the optic tract of patients with sellar region tumors, mostly craniopharyngiomas (11-14). Single case reports of metastases (8) have also described high T2 or FLAIR signal. A single case of syphilis (10) reported optic tract enhancement. A single case of pituitary apoplexy (9) documented decreased T2 or FLAIR signal from optic tract hemorrhage. A single case of head trauma (6) revealed optic tract thinning. In our study of 24 cases, the largest reported series of clinically delineated unilateral optic tract dysfunction, consensus review by 2 neuroradiologists and 2 clinicians revealed 2 types of imaging abnormalities: 1) alterations in brain tissue adjacent to the affected optic tract (extrinsic abnormalities) and 2) alterations within the optic tract itself (intrinsic abnormalities). FIG. 8. Case 16. Left optic tract dysfunction after left temporal lobectomy for intractable seizures. Coronal T2 image shows thinning of the left optic tract (arrow); compare with normal bulk of unaffected right optic tract (arrowhead). Presumptive mechanism of optic tract injury is inadvertent occlusion of left anterior choroidal artery. Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution traumatic brain injury cases, identification of the optic tract T2 or FLAIR hyperintensity (Case 21) and optic tract thinning (Case 22) allowed a definitive explanation for what had been attributed initially to malingering. We surmise that without being alerted to the side of optic tract dysfunction, neuroradiologists independently reading MRI scans might overlook many of the pertinent imaging abnormalities. With this extra clinical information and appropriate imaging sequences, they might detect the intrinsic optic tract signal abnormalities, including altered signal intensity, position, and thickness, which constitute stronger evidence of optic tract dysfunction. Identifying these abnormalities is critical in providing an organic explanation for vague visual complaints or homonymous hemianopias that might otherwise go unrecognized or be falsely attributed to nonorganic causes. FIG. 9. Case 3. Right optic tract syndrome caused by ischemia from proximal posterior cerebral artery occlusion. Axial T2 MRI shows increased signal in right optic tract (arrow). REFERENCES The extrinsic abnormalities were identified in a majority of cases on independent review by both neuroradiologists, but neither linked them to optic tract dysfunction. Compression of the optic tract, present in one quarter of the cases with extrinsic signal abnormalities, was appreciated only after joint review with the clinicians. Intrinsic MRI abnormalities were not appreciated at all in the independent review by the neuroradiologists, being identified only in the consensus conference. Such abnormalities were eventually identified in 15 (63%) cases, including mostly T2 or FLAIR hyperintensity, and also thinning or thickening of the optic tract, hypointensity, and enhancement. The intrinsic signs were more easily seen in the proximal portion of the optic tract, where it is outlined by the cerebrospinal fluid (Fig. 1). Postcontrast coronal T1 and thin slice coronal T2 were helpful sequences in detecting these intrinsic abnormalities. The neuroradiologists believed that their detection of the extrinsic compressive and intrinsic imaging abnormalities improved in the consensus conference once they were informed of the side of the lesion and were able to confer with each other and the clinicians. In the presumed anterior choroidal artery infarct cases (Cases 3 and 16), identifying the intrinsic imaging abnormalities was especially helpful. In those patients, clinicians had been seeking the more common explanation for homonymous hemianopia in a lesion of the posterior optic radiation or visual cortex, but there were no MRI abnormalities in that region. The optic tract signal abnormalities were clues that occlusion of an arterial branch to the optic tract was likely responsible for the visual field loss. In the 2 1. Savino PJ, Paris M, Schatz NJ, Orr LS, Corbett JJ. Optic tract syndrome: a review of 21 patients. Arch Ophthalmol. 1978;96:656-663. 2. Newman SA, Miller NR. Optic tract syndrome. Arch Ophthalmol. 1983;101:1241-1250. 3. Bender MB, Bodis-Wollner I. Visual dysfunction in optic tract lesions. Ann Neurol. 1978;3:187-193. 4. Cogan DG, Neurology of the Visual System. Springfield, IL: Charles C. Thomas, 1967. 5. Lindenberg R, Walsh FB, Sacks JG. Neuropathology of Vision: At Atlas. Philadelphia, PA: Lea & Febiger, 1973. 6. Al-Zubidi N, Ansari W, Fung SH, Lee AG. Diffusion tensor imaging in traumatic optic tract syndrome. J Neuroophthalmol. 2014;34:95-98. 7. Ono S, Koide R, Warabi Y, Yagishita A, Yoshia H. Homonymous visual field defect due to optic tract involvement in a patient with multiple sclerosis. Brain Nerve. 2007;59:1390-1391. 8. Saeki N, Murai H, Kubota M, Fujimoto N. Oedema along the optic tracts due to pituitary metastasis. Br J Neurosurg. 2001;15:523-526. 9. Kim HJ, Cho WH. Optic tract hemorrhage after pituitary apoplexy. AJNR Am J Neuroradiol. 2007;28:141-142. 10. Iwamoto K, Aoyagi J, Kiyozuka T, Iwasaki Y, Fujioka T. Neurosyphilis with unilateral optic tract lesion causing homonymous hemianopia. Neurologist. 2009;15:345-346. 11. Nagahata M, Hosoya T, Kayama T, Yamaguchi K. Edema along the optic tract: a useful MR finding for the diagnosis of craniopharyngiomas. AJNR Am J Neuroradiol. 1998;19:1753-1757. 12. Saeki N, Uchino Y, Murai H, Kubota M, Isobe K, Uno T, Sunami K, Yamaura A. MR imaging study of edema-like change along the optic tract in patients with pituitary region tumors. AJNR Am J Neuroradiol. 2003;24:336-342. 13. Hirunpat S, Tanomkiat W, Sriprung H, Chetpaophan J. Optic tract edema: a highly specific magnetic resonance imaging finding for the diagnosis of craniopharyngioma. Acta Radiol. 2005;46:419-423. 14. Shizukuishi T, Abe O, Haradome H, Fukushima T, Katayama Y, Sugitani M. Granular cell tumor of the neurohypophysis with optic tract edema. Jpn J Radiol. 2014;32:179-182. Kowal et al: J Neuro-Ophthalmol 2017; 37: 17-23 23 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.