The Utility of Magnetic Resonance Imaging in Assessing Patients With Pituitary Tumors Compressing the Anterior Visual Pathway

Pituitary tumors are one of the most common types of intracranial neoplasms, and can cause progressive visual loss. An ongoing challenge in the management of patients with pituitary tumors is the cost, availability, and reliability of current magnetic resonance imaging (MRI) techniques to capture clinically significant incremental tumor growth. The purpose of this study was to evaluate the various MRI-based structural analyses and to explore the relationship between measures of structure and function in the afferent visual pathway of patients with pituitary tumors. We performed a critical review of literature on MRI-based structural analyses of pituitary adenomas using PubMed, Embase, Cochrane Library, and Google Scholar. In addition, preoperative structural characteristics of the optic apparatus, optic nerve compression, and optic chiasm elevation identified as important in the literature review, were examined in 18 of our patients from October 2010 to January 2014. In our review of literature, a total of 443 citations were obtained from our search strategy and review of bibliographies. Eight of these studies met inclusion/exclusion criteria and were retrieved for critical review. Of the 8 included studies, only 2 studies examined the relationship between MRI-based structural measurements and postoperative visual recovery. In our small case-series, MRI analysis of chiasm elevation, severity of optic nerve compression, chiasm position, height of chiasm, tumor height, and tumor volume failed to differentiate patients with postoperative visual dysfunction vs those with visual recovery (P > 0.05). Although MRI-based structural analysis is an important and useful tool for managing patients with pituitary tumors, there are limited objective measures shown to be predictive of postoperative visual recovery.

Original Contribution The Utility of Magnetic Resonance Imaging in Assessing Patients With Pituitary Tumors Compressing the Anterior Visual Pathway Won Hyung A. Ryu, MD, MSc, Yves Starreveld, MD, PhD, Jodie M. Burton, MD, MSc, Junjie Liu, COMT, Fiona Costello, MD, the PITNET Study Group Background: Pituitary tumors are one of the most common types of intracranial neoplasms, and can cause progressive visual loss. An ongoing challenge in the management of patients with pituitary tumors is the cost, availability, and reliability of current magnetic resonance imaging (MRI) techniques to capture clinically significant incremental tumor growth. The purpose of this study was to evaluate the various MRI-based structural analyses and to explore the relationship between measures of structure and function in the afferent visual pathway of patients with pituitary tumors. Methods: We performed a critical review of literature on MRI-based structural analyses of pituitary adenomas using PubMed, Embase, Cochrane Library, and Google Scholar. In addition, preoperative structural characteristics of the optic apparatus, optic nerve compression, and optic chiasm elevation identified as important in the literature review, were examined in 18 of our patients from October 2010 to January 2014. Results: In our review of literature, a total of 443 citations were obtained from our search strategy and review of bibliographies. Eight of these studies met inclusion/exclusion criteria and were retrieved for critical review. Of the 8 included studies, only 2 studies examined the relationship between MRI-based structural measurements and postoperative visual recovery. In our small case-series, MRI analysis of chiasm elevation, severity of optic nerve compression, chiasm position, height of chiasm, tumor height, and tumor volume failed to differentiate patients with postoperative visual dysfunction vs those with visual recovery (P . 0.05). Conclusions: Although MRI-based structural analysis is an important and useful tool for managing patients with pituitary Department of Clinical Neurological Sciences (WHAR, YS), Division of Neurosurgery, University of Calgary, Alberta, Canada; Department of Clinical Neurological Sciences (JMB, FC), Division of Neurology, University of Calgary, Alberta, Canada; and Department of Surgery (JJ, FC), Division of Ophthalmology, University of Calgary, Alberta, Canada. The authors report no conflicts of interest. Address correspondence to Fiona Costello, MD, Foothills Medical Centre, Area 3, #1120H Health Sciences Centre, 3350 Hospital Drive, Calgary, AB T2N 4N1, Canada; E-mail: fi[e-mail] 230 tumors, there are limited objective measures shown to be predictive of postoperative visual recovery. Journal of Neuro-Ophthalmology 2017;37:230-238 doi: 10.1097/WNO.0000000000000408 © 2016 by North American Neuro-Ophthalmology Society P ituitary tumors are the third most common primary intracranial neoplasm with an annual incidence of 0.8- 8 per 100,000 (1). Often pituitary tumors remain clinically occult, unless patients develop neuroendocrine abnormalities and/or visual dysfunction. For lesions with suprasellar extension, vision loss may manifest if there is mass effect on the anterior visual pathway. Yet, the pathophysiology of compressive optic neuropathy in this patient population is not fully understood. The classical teaching has been that compression of the optic chiasm is associated with bitemporal hemianopia that will likely necessitate either medical or surgical decompression. Recently published studies, however, suggest marked variability in the type and range of visual dysfunction experienced by patients with pituitary adenoma (2). Specially, some patients remain asymptomatic even with the tumor abutting the optic apparatus and can be potentially managed conservatively with serial assessments (2). Currently, the treatment algorithm for patients with pituitary tumors remains institution dependent or based on the surgeon's preference. One of the many challenges in detecting lesions of the pituitary gland that grow to greater than 10 mm (macroadenomas) and compress the anterior visual pathway, is that vision loss is insidious and many patients are unaware of their deficits until central visual function is affected (Fig. 1). Thus, affected individuals with the Snellen equivalent of 20/20 vision may present with severe visual field defects and dyschromatopsia. Normal visual function may be restored after decompression of the optic chiasm by surgical or medical means. However, the optimal treatment-timing Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 1. A 28-year-old man reports a 6-month history of "visual blurring." Visual acuity is 20/20 in both eyes. A. Automated (Humphrey) perimetry shows a bitemporal hemianopia. B. There is temporal optic disc pallor bilaterally. Postcontrast sagittal (C) and coronal (D) T1 magnetic resonance imaging reveals a pituitary macroadenoma. window is not known, and there are no validated predictors of visual recovery (3). Various factors proposed to impact postoperative visual outcomes include rate of tumor growth, severity of chiasm compression, tumor size, retinal nerve fiber layer thickness, and duration of visual dysfunction (3-8). Currently, magnetic resonance imaging (MRI) and visual field testing with static or kinetic perimetry techniques are the primary diagnostic tools that guide the treatment planning in patients with pituitary tumors. Yet, the cost, availability, and reliability of MRI in capturing clinically significant incremental tumor growth represent significant challenges. Establishing the ideal timing of surgical intervention to prevent persistent visual dysfunction in patients with pituitary tumors is of vital importance. The management of patients presenting with severe vision loss is relatively clear; yet determining the optimal timing of surgical intervention for patients with mild to no vision loss, who have the potential for progressive visual decline over time, is less certain. The first goal of this study was to critically review the available literature regarding the proposed MRI-based structural analyses related to visual function in patients with pituitary tumors. Second, we aimed to determine the relationship between visual function and MRI-based structural analysis in patients with pituitary tumors treated in our Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 institution. Specifically, we explored whether MRI-measured compression of the optic apparatus and tumor characteristics correlated with visual recovery. METHODS Literature Review A critical review of the literature was performed that discussed surgical treatment of pituitary adenoma using PubMed, Embase, Cochrane Library, and Google Scholar. The review process was performed with the assistance of a University of Calgary health science librarian. All studies published from January 1989 to August 2013 were eligible for inclusion. Search terms used were: "pituitary gland," "pituitary neoplasm," "pituitary surgery," "magnetic resonance imaging," "visual field," "visual acuity," "vision," "visual test," "vision disorders," "optic nerve," "optic chiasm," and "visual pathway." Titles and abstracts of the retrieved citations were searched and included/excluded according to the following criteria (Fig. 2). Inclusion criteria were: 1) structural analyses using MRI, 2) visual assessment by an ophthalmologist, including perimetry for visual field analyses, and 3) publication from January 1990 until August 2013. Exclusion criteria were: 1) studies published on nonhuman data, 231 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 2. Flow chart of process of literature review dealing with surgical management of pituitary tumor. 2) non-English studies, 3) noninvestigative studies (e.g., technical reports, case reports, letters, and comments). Included studies were then retrieved and references were reviewed to identify any further articles that were not identified in the initial PubMed search but which may have been relevant to the current review. Current Study Method This study was approved by the University of Calgary Conjoint Research Ethics Board. Records of 18 consecutive patients with a parasellar/intrasellar mass between October 2010 and December 2014 from the Department of Clinical Neurological Sciences (Division of Neurosurgery) in Foothills Medical Center (a tertiary care academic institution in Calgary, Canada) were retrospectively reviewed. All 18 patients had histologic diagnosis of pituitary adenoma and underwent complete surgical resection. Inclusion criteria were: 1) new radiological diagnosis of pituitary tumor, 2) access to magnetic resonance imaging (MRI) confirming contact between the tumor and the optic apparatus, 3) preoperative and 6-month postoperative assessment by a neuro-ophthalmologist, 4) preoperative visual field or/ and visual acuity dysfunction. Exclusion criteria included amblyopia or the presence of any anterior segment, posterior segment, or optic nerve related pathology other than compressive pituitary tumor that could impair visual function. Neuro-ophthalmological evaluation included: Snellen high contrast letter acuity, slit lamp biomicroscopy, pupil assessment, ocular motility evaluation, intraocular pressure quantification, color vision (Hardy Rand Rittler pseudo-isochromatic plates) testing, dilated ophthalmoscopy, and optical coherence tomography (OCT) (Cirrus OCT; Carl Zeiss Meditec, Dublin, CA). Visual field mean deviation and foveal threshold (measured in decibels [dB]) were determined with automated perimetry (Humphrey 232 central 30-2 full threshold testing; Carl Zeiss Meditec). Perimety results were included if the fixation losses and false negative errors were less than one-third; and, the false positive responses measured less than 20%. The presence of ocular motility deficits, afferent or efferent pupil abnormalities, external ocular findings (ptosis), optic disc cupping, edema, or pallor was described, if present. During the course of the routine pituitary tumor work-up, all patients were clinically reviewed by an interdisciplinary team consisting of an endocrinologist, neuro-ophthalmologist, neuroradiologist, neurosurgeon, and otolaryngologist. At our institution, surgical management is reserved for patients with visual dysfunction and patients with hormonally active tumors refractory to medical treatment, or individuals who cannot tolerate medical management. All MRI studies were done using a 1.5T scanner and were reviewed by a neurosurgeon and radiologist. There were minor differences in the imaging sequences, but all MRI studies included precontrast and postcontrast T1 scans done in the coronal and sagittal planes. The MRI scans were assessed for various imaging characteristics that were identified as potential predictors of visual recovery based on our review of literature. These variables included chiasm elevation, chiasm position, chiasm height, tumor height, and tumor volume. The chiasm elevation (E chiasm) was measured by connecting the most inferolateral aspects of the chiasm (standard line) and measuring the perpendicular distance between the standard line and the inferior aspect of the chiasm at maximum compression (2). This measurement method was illustrated in a previous publication (2). Chiasm position was measured using the coronal plane with a standard line between the upper surfaces of the bilateral internal carotid arteries. The chiasm position was defined as the distance between this standard line and the optic chiasm (9-11). Chiasm height was defined as the maximal central Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution height of the chiasm on the coronal plane (10). Tumor height was defined as the maximal central height of the tumor (8). Tumor volume was measured using the maximal diameter in the x, y, z dimensions with the formula for volume = 4/3 p (x/2$y/2$z/2) (12). Analysis of optic nerve compression was performed by computing a ratio of the cross-sectional area of optic nerve at the point of maximum compression to the intra-orbital optic nerve (Fig. 3). This is a novel MRI-based structural model used in this study to quantify the severity of optic nerve compression. Statistical analyses using Student's t test were performed comparing the MRI analyses in patients with postoperative improvement in vision with Snellen high contrast visual acuity equivalent better than 20/40 in either eye and/or Humphrey visual field mean deviation better than 25.00 dB (good outcome) and patients persistent visual deficits defined as visual acuity equivalent worse than 20/40 in either eye and/or Humphrey visual field mean deviation worse than 25.00 dB, with a pattern of visual field loss consistent with optic nerve/chiasm compression (poor outcome). For the comparative analyses, the P value less than 0.05 was set as statistically significant. (Table 1). All included studies were retrospective in design; 6 were cohort studies, and 2 were case-controlled studies. The sample sizes ranged from 19 to 201 subjects. The MRIbased structural analyses used in the studies included both qualitative and quantitative methods. Most commonly used measurements focused on quantifying optic chiasm deformation, such as height of chiasm and grade of chiasm compression. Other measurements included tumor volume, optic nerve signal intensity, and cross-sectional area of the optic chiasm. The most commonly used assessment of visual function was visual field, as measured by either Goldmann perimetry or Humphrey perimetry. In contrast, Tokumaru et al (14) focused on visual acuity as the primary measure of visual function. Of the 8 included studies, only 2 studies examined the relationship between MRI-based structural measurements and postoperative visual recovery. Tokumaru et al reported the degree of improvement in visual acuity correlated only with disease duration, and not with any structural measurements. Monteiro et al (11) found that the 3 predictors of postoperative visual recovery were the presence of optic atrophy, grade of visual loss, and coronal measurement of chiasm elevation. RESULTS Current Study Results Literature Review A total of 443 citations were generated from our search strategy and review of bibliographies. Eight of these studies met inclusion/exclusion criteria and were retrieved for critical review. All of these studies were published between 1995 and 2011; most studies were published after 2006 The mean age of the 18 patients (13 females) included in the study was 51 (range 31-92). Of the 18 patients, 6 patients were initially misdiagnosed with other presumed causes of vision loss before the discovery of the pituitary tumor including: keratoconus (n = 1), cataracts (n = 3), glaucoma (n = 1), and macular degeneration (n = 1). On preoperative neuroophthalmic assessment, all 18 patients had clinically FIG. 3. MRI-based analysis of optic nerve compression using a ratio of the cross-sectional area of optic nerve at the point of maximum compression (B) to the cross-sectional area of the intra-orbital optic nerve (A). Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 233 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Authors Type of Study Ikeda and Yoshimoto (9) Retrospective case-control Eda et al (8) Retrospective cohort Tokumaru et al (14) Retrospective case-control Carrim et al (10) Retrospective cohort Sample Size Visual Assessment MRI Analyses Conclusion Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Position of chiasm significantly elevated 50 (26 with visual 1. Goldmann perimetry Chiasm position: Elevation above compared to control (P , 0.005). symptoms) (presence vs absence of VF standard lines between 1) frontal/ defect) posterior clinoid process (sagittal); 2) between upper surface of bilateral ICA (coronal) 17 controls 2. Cushing's grading of visual All patients with chiasm elevation disturbance greater than 8 mm (sagittal) and 13 mm (coronal) had VF deficit 28 1. Goldmann perimetry 1. Relative position of tumor to chiasm Cushing's grade of visual field (presence vs absence of VF (superior, anterior, posterior) disturbance (13) correlated defect) significantly with tumor size in superior type (r = 0.685) 2. Cushing's grading of visual 2. Tumor size (height, width, disturbance thickness) 27 1. VF (test not specified) 1. Optic nerve signal intensity Optic nerve hypterintensity correlated (categorical grading) with degree of optic chiasm compression and VA (P , 0.01) 10 controls 2. VA (test not specified) 2. Grade of chiasm compression No correlation between optic nerve (1-3 categorical grading) signal intensity and tumor size/ location Degrees of improvement in VA correlated with disease duration 19 (15 with visual 1. Goldmann perimetry 1. Cross-sectional area of chiasm Significant correlation between chiasm symptoms) (Goldmann units)* height and visual field defect. (r = 20.69; r = 20.63; r = 20.52) 2. VA 2. Central height of chiasm Significant correlation between tumor height and bitemporal and binocular visual field defect. (r = 0.55; r = 0.46 respectively) 3. Height of tumor above standard line No correlation between area of chiasm between right and left ICAs and visual field No relationship between VA and VF Original Contribution 234 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. TABLE 1. Published studies of MRI-based structural analyses related to visual function in patients with pituitary tumors Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Authors Frisen and Jensen (4) Type of Study Retrospective cohort Sample Size 31 (15 w/visual symptoms) Wang et al (15) Retrospective cohort 201 (83 w/visual symptoms) Monteiro et al (11) Retrospective cohort 30 (23 w/visual symptoms) Lee et al (12) Retrospective cohort 39 (29 w/visual symptoms) Visual Assessment MRI Analyses 1. Automated perimetry (High- 1. Chiasm appearance (1-4 categorical grading) pass Resolution perimetry, RareBit perimetry: raw score of correct stimuli detection) 2. Elevation of chiasm above standard line between olfactory bulbs and ponto-mesencepahlic junction (BPheight index) 3. Elevation of chiasm above standard line between intracranial apertures of optic nerve canals 1. Goldmann perimetry (0-3 Hardy classification of suprasellar categorical grading) extension (0, A-D)† 2. VA (0-4 categorical grading of standard log VA) 1. Automated (Humphrey) 1. Chiasm position: Elevation above perimetry (1-4 categorical standard lines between 1) frontal/ grading of mean deviation) posterior clinoid process (sagittal); 2) between upper surface of bilateral ICA (coronal) 2. VA 2. Optic atrophy (1-4 categorical grading) Tumor volume: 4/3 p (a/2$b/2$c/2) (a, b, c: diameter in x, y, z dimensions) 50% patients with chiasm elevation of 6 mm had visual field defect 90% of patients with chiasm elevation of 11 mm BP-height index best correlated with VF (r = 0.75). Significant correlation between VA and Hardy grading (P = 0.000). No relationship between duration and severity of visual symptom Significant association with chiasm position and visual field defect (r = 20.624). Predictors of visual recovery were degree of optic atrophy, grade of visual loss, and coronal measure of chiasm elevation (P , 0.001). Optic atrophy was the best predictor. Significant association with tumor volume and degrees of visual field defect (r = 20.693). Original Contribution 1. Automated (Humphrey) perimetry (mean deviation and pattern standard deviation) 2. VA Conclusion *Goldmann unit = area within each subdivision of the Goldmann chart defined by 15° meridians and 10° isopters. BP-height index = Elevation of chiasm above standard line between olfactory bulbs and ponto-mesencepahlic junction. † Hardy grading = Classification of suprasellar tumor extension (0 = no suprasellar extension, A = suprasellar cistern only, B = recesses of third ventricle, C = whole anterior third ventricle, D = intracranial extradural). ICA, internal carotid artery; MRI, magnetic resonance imaging; VA, visual acuity; VF, visual field. 235 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. (Continued ) Original Contribution TABLE 2. Preoperative MRI-based structural analyses in relation to postoperative visual outcome in patients with pituitary adenoma Mean Mean Mean Mean Mean Mean optic nerve compression (SD) chiasm elevation (SD) chiasm position (SD) chiasm height (SD) tumor height (SD) tumor volume (SD) Good Visual Outcome (N = 9) Poor Visual Outcome (N = 9) T value P value 45% (23%) 0.5 cm (0.31) 2.4 cm (0.50) 0.14 cm (0.03) 2.79 cm (0.70) 9.98 cm3 (6.40) 40% (15%) 0.43 cm (0.19) 2.26 cm (0.50) 0.13 cm (0.06) 2.85 cm (0.68) 9.32 cm3 (4.34) 0.57 0.62 0.64 0.32 0.17 0.26 0.58 0.53 0.53 0.75 0.87 0.80 Poor outcome = Snellen high contrast visual acuity equivalent worse than 20/40 in either eye and/or Humphrey visual field mean deviation worse than 25.00 dB. MRI, magnetic resonance imaging. significant visual dysfunction (visual field deficit based on Goldmann or automated perimetry ± decreased bestcorrected visual acuity). The preoperative visual acuity ranged from 20/40 to 20/400, whereas the visual field mean deviation ranged from 25.31 to 222.87 dB. Once the diagnosis of pituitary macroadenoma was confirmed, all 18 patients underwent endoscopic transnasal transsphenoidal surgery. This was done once it was established with neuroendocrine testing that none of the patients had a hormone-secreting lesion, which could be treated with medical therapy. Complete resection (or near complete resection) with elimination of compression of anterior visual pathway structures was achieved in all cases by postoperative MRI, and no postoperative complications were noted during the 6-month follow-up period. Preoperative MRI analysis of optic nerve compression did not differentiate patients with postoperative visual recovery vs those with persistent visual dysfunction at 6-month followup (P = 0.21; Table 2). Similarly, preoperative structural analyses of the optic chiasm, including chiasm elevation, chiasm position, and chiasm height did not show significant correlation with visual outcome (Table 2). Preoperative tumor characteristics, including tumor height and tumor volume also failed to differentiate between patients with positive visual outcome and patients with postoperative dysfunction in either visual field or visual acuity (Table 2). DISCUSSION The classical teaching in neurosurgery and ophthalmology has been that anterior visual pathway compression from a pituitary macroadenoma is associated with bitemporal hemianopia, which necessitates surgical or medical decompression (16). As more data on the pathophysiology of pituitary tumors are accrued, questions have been raised about the relationship between compression of the anterior visual pathway and visual function (4,17-20). A growing number of published reports have highlighted the highly variable clinical consequences of compressive pituitary lesions (2,4,21). Interestingly, this was documented almost 100 years ago by Harvey Cushing with his accounts 236 of minimally symptomatic patients with severe deformation of the optic chiasm (22). Although surgery remains a safe and effective standard of care for patients with pituitary tumor, the decision to undergo an invasive procedure must always balance the risks and benefits of intervention. It is a continuing goal to optimize the management of pituitary lesions in the most safe, efficient, and cost-effective manner. Ideally, surgery should be considered for patients who are at a high risk of developing persistent visual dysfunction. The ability to predict visual recovery would be a powerful tool in managing patients with pituitary tumors, including the optimal timing and indications for surgical treatment, triaging of patients, and perioperative patient counseling. The results of our study show that the most commonly used neuroradiologic assessments, optic chiasm, and optic nerve deformation, did not consistently identify patients with preoperative visual defects or correlate with successful postsurgical outcome. This finding further emphasizes the individual variability within the structural- functional paradigm of compressive neuropathy caused by growing pituitary tumor. Moreover, it highlights potential risk in managing patients with pituitary tumor if we rely solely on neuroimaging for clinical decision making. Improvements in the quality and availability of highresolution MRI, have resulted in its integral role for managing patients with pituitary tumor. Various radiographic models have been proposed for quantifying deformation of the optic apparatus in predicting visual dysfunction and surgical outcome (4,10,15). These include cross-sectional area of chiasm, height of chiasm, tumor height, spatial relationship between the chiasm and tumor, the presence of optic atrophy and signal change in optic nerve (4,8,10-12,14,15). Unfortunately, these models have either shown mixed results or were limited by study methodology. For example, Eda et al (8) reported spatial relationship between the chiasm and tumor to be correlated with Cushing's grade of visual disturbance, whereas Ikeda and Yoshimoto (9) proposed that relative position of chiasm and tumor (i.e., anterior vs posterior vs superior) did not play a role in visual dysfunction. Another commonly used radiographic measurement is tumor size (8,12). A potentially Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution significant limitation in this structural model is that the height and volume of pituitary tumors are both affected by the individual anatomical variability of the sellar and the variable direction of tumor growth. In our review of literature, only 2 reports specifically examined the role of MRI-based structural analyses as potential predictors of visual outcome (11,14). Monteiro et al (11) examined chiasm position and optic atrophy in determining visual function and surgical outcome and found, using a multivariate analysis, that the best predictor of visual recovery was severity of optic atrophy compared with severity of visual field loss and measured chiasm elevation. Unfortunately, the clinical utility of optic atrophy is limited by the subjective nature of its determination and the quality of the MR images. Tokumaru et al (14) also examined the role of optic nerve signal intensity as a potential predictor of visual recovery, but found that only disease duration correlated with postoperative improvement in visual acuity. This structural analysis of qualifying signal change shares similar challenges as mentioned above regarding optic atrophy. The lack of objective radiographic models with predictive ability present an on-going dilemma in evaluating and triaging patients with an enlarging pituitary tumor based on structural characteristics of the tumor. Ideally, the treatment decision algorithm should only include information on the structural changes of the visual pathway that is objective, easily reproducible, cost-effective, and predictive of posttreatment visual recovery. Furthermore, a thorough assessment of visual function should remain the primary driver in clinical decision making with neuroimaging as a complementary source of information. Our study had several limitations. The literature review was restricted by the quality of the papers included. A second issue to consider was the fact that only English language articles were included, which may have introduced a language, and potentially an ethnic, bias. However, the studies reviewed were published from a variety of centers internationally. With regards to the current study, we performed a retrospective review of 18 patients and thus the generalizability of our findings to a broader population is uncertain. In addition, visual field assessment by Humphrey automated perimetry may be influenced by impairment in visual acuity, co-operation of the patients, and learning effects (23). Previous reports have raised some concern regarding the Humphrey automated perimetry not being able to differentiate reliable vs unreliable patients; this point is particularly salient if explicit criteria are not used to differentiate reliable from unreliable visual field results (24). Lastly, the proposed optic nerve compression analyses used in our study is novel, and needs to be validated in a larger patient population. Although MRI-based structural analysis is an important and useful tool for managing patients with pituitary tumors, there are limited objective measures shown to be predictive Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 of postoperative visual recovery. Our findings indicate the need for continued research into new putative markers of anterior visual pathway structural and functional integrity, including optical coherence tomography, diffusion tensor imaging, and functional MRI. These measures may complement standard MRI techniques, and provide insights into the complex structure-function relationships that underpin vision loss caused by compressive lesions. STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: W. H. A. Ryu, Y. Starreveld, J. M. Burton, and F. Costello; b. Acquisition of data: W. H. A. Ryu, J. Liu, and F. Costello; c. Analysis and interpretation of data: W. H. A. Ryu, Y. Starreveld, J. M. Burton, J. Liu, and F. Costello. Category 2: a. Drafting the manuscript: W. H. A. Ryu, Y. Starreveld, J. M. Burton, J. Liu, and F. Costello; b. Revising it for intellectual content: W. H. A. Ryu, Y. Starreveld, J. M. Burton, J. Liu, and F. Costello. Category 3: a. Final approval of the completed manuscript: W. H. A. Ryu, Y. Starreveld, J. M. Burton, J. Liu, and F. Costello. ACKNOWLEDGMENTS PITNET study group is a research collaboration group from University of Calgary. (The Pituitary program: An Interdepartmental, multi-disciplinary Team-based approach to optimizing Neurosurgical, Visual, and Endocrinological Treatment). In addition to the co-authors, the other members of PITNET are Dr. Alim Mitha, Dr. Shelly Bhayana, Dr. Brad Goodyear, and Dr. Stefan Lang. REFERENCES 1. Surawicz TS, McCarthy BJ, Kupelian V, Jukich PJ, Bruner JM, Davis FG. Descriptive epidemiology of primary brain and CNS tumors: results from the central brain tumor registry of the United States, 1990-1994. Neuro-oncol. 1999;1:14-25. 2. Ryu WH, Tam S, Rotenberg B, Labib MA, Lee D, Nicolle DA, Van Uum S, Duggal N. Conservative management of pituitary macroadenoma contacting the optic apparatus. Can J Neurol Sci. 2010;37:837-842. 3. Danesh-Meyer HV, Papchenko T, Savino PJ, Law A, Evans J, Gamble GD. In vivo retinal nerve fiber layer thickness measured by optical coherence tomography predicts visual recovery after surgery for parachiasmal tumors. Invest Ophthalmol Vis Sci. 2008;49:1879-1885. 4. Frisen L, Jensen C. How robust is the optic chiasm? Perimetric and neuro-imaging correlations. Acta Neurol Scand. 2008;117:198-204. 5. Gnanalingham KK, Bhattacharjee S, Pennington R, Ng J, Mendoza N. The time course of visual field recovery following transphenoidal surgery for pituitary adenomas: predictive factors for a good outcome. J Neurol Neurosurg Psychiatry. 2005;76:415-419. 6. Honegger J, Zimmermann S, Psaras T, Petrick M, Mittelbronn M, Ernemann U, Reincke M, Dietz K. Growth modelling of non-functioning pituitary adenomas in patients referred for surgery. Eur J Endocrinol. 2008;158:287-294. 7. Dekkers OM, Hammer S, de Keizer RJ, Roelfsema F, Schutte PJ, Smit JW, Romijin JA, Pereira AM. The natural course of non-functioning pituitary macroadenomas. Eur J Endocrinol. 2007;156:217-224. 8. Eda M, Saeki N, Fujimoto N, Sunami K. Demonstration of the optic pathway in large pituitary adenoma on 237 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 9. 10. 11. 12. 13. 14. 15. heavily T2 weighted MR images. Br J Neurosurg. 2002;16:21-29. Ikeda H, Yoshimoto T. Visual disturbances in patients with pituitary adenoma. Acta Neurol Scand. 1995;92:157-160. Carrim ZI, Reeks GA, Chohan AW, Dunn LT, Hadley DM. Predicting impairment of central vision from dimensions of the optic chiasm in patients with pituitary adenoma. Acta Neurochir (Wien). 2007;149:255-260. Monteiro ML, Zambon BK, Cunha LP. Predictive factors for the development of visual loss in patients with pituitary macroadenomas and for visual recovery after optic pathway decompression. Can J Ophthalmol. 2010;45: 404-408. Lee JP, Park IW, Chung YS. The volume of tumor mass and visual field defect in patients with pituitary macroadenoma. Korean J Ophthalmol. 2011;25:37-41. Cushing H, Heuer GJ. Distortions of the visual fields in cases of brain tumors. Dyschromatopsia in relation to stages of choked disk. JAMA. 1911;57:200-208. Tokumaru AM, Sakata I, Terada H, Kosuda S, Nawashiro H, Yoshii M. Optic nerve hyperintensity on T2-weighted images among patients with pituitary macroadenoma: correlation with visual impairment. AJNR Am J Neuroradiol. 2006;27:250-254. Wang H, Sun W, Fu Z, Si Z, Zhu Y, Zhai G, Xu S, Pang Q. The pattern of visual impairment in patients with pituitary adenoma. J Int Med Res. 2008;36:1064-1069. 238 16. Kosmorsky GS, Dupps WJ Jr, Drake RL. Nonuniform pressure generation in the optic chiasm may explain bitemporal hemianopsia. Ophthalmology. 2008;115:560-565. 17. Collette JM, Francois J, Neetens A. Vascularization of the optic pathway. V. Chiasma. Br J Ophthalmol. 1956;40:730-741. 18. McIlwaine GG, Carrim ZI, Lueck CJ, Chrisp TM. A mechanical theory to account for bitemporal hemianopia from chiasmal compression. J Neuroophthalmol. 2005;25:40-43. 19. Kanamori A, Nakamura M, Matsui N, Nagai A, Nakanishi Y, Kusuhara S, Yamada Y, Negi A. Optical coherence tomography detects characteristic retinal nerve fiber layer thickness corresponding to band atrophy of the optic discs. Ophthalmology. 2004;111:2278-2283. 20. Sade B, Mohr G, Vezina JL. Distortion of normal pituitary structures in sellar pathologies on MRI. Can J Neurol Sci. 2004;31:467-473. 21. Karavitaki N, Collison K, Halliday J, Byrne JV, Price P, Cudlip S, Wass JA. What is the natural history of nonoperated nonfunctioning pituitary adenomas? Clin Endocrinol (Oxf). 2007;67:938-943. 22. Cushing HW. Distortions of the visual fields in case of brain tumor. Brain. 1915;37:341-400. 23. Werner E, Adelson A, Krupin T. Effect of patient experience on the results of automated perimetry in clinically stable glaucoma patients. Ophthalmology. 1988;95:764-767. 24. Heijl A, Lindgren A, Lindgre G. Test-retest variability in glaucomatous visual fields. Am J Ophthalmol. 2000;108:130-135. Ryu et al: J Neuro-Ophthalmol 2017; 37: 230-238 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.