[1] 8w Does IIH Alter Brain Microstructures? -A DTI-Based Approach

[1] 238w Introduction.-To investigate the correlation of microstructural parameters with CSF pressure and macroscopic changes assessed by diffusion tensor imaging (DTI) in patients with idiopathic intracranial hypertension (IIH).Methods.-Twenty-three patients with IIH as well as age-, sex-, and body mass index (BMI)-matched controls underwent high resolution MR imaging of the optic nerve sheaths (ONS), pituitary gland, and ventricles. For DTI data a voxelwise permutation analysis was performed for the whole brain and ROI analysis was performed for the optic nerve and optic radiation. DTI measurements were correlated to morphometric measurements, CSF opening pressure, and headache intensity. The reliability of diagnostic performance of DTI parameters was assessed using ROC analysis.Results.-Analysis of DTI metrics revealed a significant reduction in the fractional anisotropy (FA) of the optic nerve in patients with IIH. In contrast, systematic regional variations between IIH patients and controls were neither observed in the whole brain analysis nor in the optic radiation. FA values of the optic nerve show significant correlations with the optic nerve sheath diameter (P 5 .003, r 5 2.589). The correlation of the alterations of the FA values of the optic radiation and the whole brain do not show a significant association to morphometric alterations in the ONS diameter and hypophysis height as well as to CSF opening pressure and headache intensity.Conclusions.-The results indicate that IIH is associated with microstructural changes in the optic nerve. These alterations may be the direct consequence of chronically elevated intracranial pressure.

[1] 120w Idiopathic intracranial hypertension (IIH) is an uncommon headache syndrome characterized by elevated intracranial pressure (ICP), most commonly in obese women of childbearing age. Patients typically suffer from chronic disabling headache and visual disturbances. 1 Impairment of visual function is often progressive and permanent in up to 25% of all cases. 2,3 The diagnosis of IIH and the monitoring of its treatment require the assessment of CSF pressure by lumbar puncture. IIH diagnosis is supported by characteristic MRI findings, including an empty sella, distended optic nerve sheath (ONS), and a posterior flattening of the optic globe. 1,[4][5][6][7][8][9] However, due to the sensitivity and specificity of currently known neuroimaging signs, neuroimaging techniques are not able to replace invasive assessment of CSF pressure.

[2] 142w The pathophysiology of IIH remains largely unknown, though disturbed CSF dynamics seem to trigger ICP elevation. 10 Data on effects of IIH on white matter tracts and brain connectivity are sparse. Diffusion tensor imaging (DTI) is an established MRI technique for assessment of microstructural tissue properties by defining mean diffusivity (MD, spatially averaged mobility of water molecules) and/or fractional anisotropy (FA, mobility differences of water molecules between the three spatial directions). 11,12 The aim of the study was to analyze the influence of increased ICP on microstructural brain tissue properties acquired by DTI in IIH patients and correlate the values with morphometric abnormalities observed in the MRI as well as with clinical findings, in particular CSF opening pressure and headache intensity as the potential detection of microscopic alterations would improve the understanding of the disorder and perhaps allow to improve diagnostic accuracy.

[1] 48w Analysis of the MD values of the whole brain as well as of the optic nerve and optic tract revealed no significant alterations between patients and healthy controls (P > .05). This is in line with the results from the T2w-images, which do not show signs of edema.

[2] 60w Correlation of DTI and Morphometric Parameters.-No normal distribution for mean ONS diameter was found (P 5 .584). The mean ONS diameter of IIH patients (Fig. 1A,B; Table 1) was significantly larger (median 6.00 mm, 5th percentile 3.54 mm, 95th percentile 7.26 mm]) than in healthy controls (median 4.35 mm, 5th percentile 3.45 mm, 95th percentile 5.49 mm; P < .0001]).

[3] 49w A normal distribution for values of the mean height of the hypophysis was identified (P 5 .045). The mean height of the hypophysis was found significantly smaller (3.21 6 1.55 mm [1.10-6.50 mm]) in patients (Fig. 1C,D) than in controls (5.50 6 1.28 mm [2.20-8.50 mm]) (P < .0001).

[4] 23w The ONS diameter showed an inverse correlation with the FA values of the optic nerve (r 5 2.589, P 5 .003; Fig. 3A).

[5] 72w No correlations were found between FA values of the optic nerve and the height of the hypophysis (r 5 .094, P 5 .668; Fig. 3C). Also no correlations could be identified for ONS diameter and pituitary height with FA values of the whole brain (r 5 2.216, P 5 .323; r 5 .213, P 5 .329) or the optic radiation (r 5 2.005, P 5 .980; r 5 2.050, P 5 .821).

[6] 105w Correlation of DTI and Clinical Parameters.-FA and MD values of whole brain, optic nerve and optic radiation ONS did not correlate with maximum CSF opening pressures (r 5 .053, P 5 .814; r 5 .230, P 5 .302). No correlations were found between FA values of the optic nerve and the maximum headache intensity (r 5 .193, P 5 .378; Fig. 3B) or the maximum CSF opening pressure (r 5 .078, P 5 .378). Furthermore, no correlations were found for maximum headache intensity with FA values of the whole brain (r 5 2.006, P 5 .980) and optic radiation (r 5 .251, P 5 .248).

[7] 352w Cut-off-Values.-The ROC analysis (Fig. 4; Table 2) revealed an optimal cuf-off-value of 0.266 for optic nerve FA (sensitivity 0.783, 95% CI [0.575, 0.906], specificity 0.739, 95% CI [0.531, 0.876], accuracy 0.761). For the FA of the optic radiation the optimal cut-off-value was 0.481 (sensitivity 0.609, 95% CI [0.407, 0.778], specificity 0.652, 95% CI [0.447, 0.812], accuracy 0.630) without significant differences (AUC 0.603; P 5 .216; Fig. 4; Table 2). For whole brain FA (AUC 5 0.535, P 5 .691; Fig. 4) and MD (AUC 5 0.500, P 5 1) as well as for the MD of the optic radiation (AUC 5 0.389, P 5 .190) no significant group differences were found (Table 2). 1 4.9 2.9 0.25 0.49 0.21 4 F 30 28.28 -1.89 2 6.0 4.8 0.16 0.46 0.22 5 F 25 31.22 40.0 9.80 1 6.1 2.7 0.23 0.51 0.21 6 F 28 31.71 43.0 9.30 1 6.8 1.9 0.22 0.53 0.21 7 F 20 33.56 42.0 9.86 1 3.3 4.8 0.28 0.49 0.21 8 F 41 25.88 43.0 7.84 1 6.4 1.9 0.19 0.47 0.20 9 F 27 30.46 30.0 7.57 1 6.7 2.9 0.15 0.50 0.20 10 F 21 24.09 39.0 7.97 1 7.6 2.4 0.20 0.48 0.20 11 M 32 28.41 31.0 4.00 1 5.3 1.1 0.20 0.49 0.20 12 M 61 33.65 31.0 4.86 1 5.3 1.6 0.21 0.45 0.19 13 F 20 29.40 50.0 9.05 1 5.4 6.5 0.20 0.47 0.20 14 F 40 40.90 35.0 9.59 1 5.1 3.4 0.23 0.49 0.20 15 F 41 32.32 40.0 9.73 1 5.6 3.2 0.18 0.52 0.19 16 M 27 29.04 31.0 2.70 1 5.9 1.8 0.24 0.48 0.21 17 F 20 31.64 31.0 6.00 1 7.0 5.3 0.19 0.52 0.21 18 F 44 32.02 50.0 8.65 1 7.3 1.65 0.19 0.44 0.18 19 F 35 36.76 36.0 6.49 1 5.6 4.0 0.24 0.53 0.20 20 F 63 26.29 50.0 5.68 1 6.4 1.6 0.17 0.50 0.18 21 F 33 26.61 29.0 7.70 1 6.0 3.4 0.20 0.46 0.19 22 F 56 30.85 42.0 9.00 1 6.2 3.4 0.20 0.50 0.20 23 F 48 46.02 40.0 8.00

[8] 40w 5.6 1.6 0.24 0.49 0.19 Ratio/ Mean (6SD)/ Percentage F:M 5 20:3 37.04 6 13.72 33.51 6 7.67 37.61 6 6.93 7.09 6 2.62 91.30% 5.94 6 0.90 3.21 6 1.55 13.34 6 3.49 0.48 6 0.03 0.1990 6 0.0099

[1] 29w Our objective was to investigate whether IIH is associated with microstructural abnormalities that may result from chronically increased ICP and analyze a potential correlation to morphometric and clinical findings.

[2] 42w MR-imaging is of increasing significance for the diagnostic workup and clinical follow-up of IIH but Fig. 4.-ROC curves for DTI measurements. ROC curves of the FA values for whole brain, optic nerve, and optic radiation. [Color figure can be viewed at wileyonlinelibrary.com]

[3] 71w currently it does not replace the measurement of CSF opening pressure by lumbar puncture as imaging signs such as structural abnormalities of the hypophysis and ONS are commonly associated with IIH but do reach a specificity that would reliably allow to draw clinical conclusions. Moreover it has not been entirely clarified if and to which extent some or all of the morphometric abnormalities can improve after a therapeutic normalization of ICP.

[4] 121w Visual disturbances represent one of the main symptoms of IIH. In most cases they represent the direct consequence of papilledema, which results from elevated ICP. The deterioration of visual acuity and of functional conductivity of the optic nerve are usually assessed by ophthalmologic measurements and the assessment of visual evoked potentials. We demonstrate in the present study that these macroscopic alterations of the optic nerve and the ONS are accompanied by microstructural abnormalities. It may be speculated that these microstructural abnormalities may be the underlying cause for functional impairment of the optic nerve. In line with morphological measurements that do not show abnormalities in the optic tract, we did not observe any significant changes in this area using a DTI-based approach.

[5] 101w The study results show that microstructural abnormalities in the optic nerve correlate with the morphological alteration of the ONS diameter. Interestingly, we did not identify a correlation between microstructural abnormalities and other morphometric alterations such as the hypophysis height or clinical symptoms including headache intensity. These results suggest that microstructural and clinical abnormalities may follow increased CSF pressure with a significant delay. Nevertheless we cannot exclude the possibility that the lack of correlation of macro-and microstructural abnormalities with CSF opening pressure and headache intensity may result from the fact that these parameters may vary substantially over relatively short periods of time.

[6] 269w While the study demonstrates microstructural abnormalities that most probably are the result of increased CSF pressure and lead to the functional deterioration of visual acuity, it does clarify the mechanism by which these microstructural alterations are elicited. It is unclear to what extent the brain is compressible and how CSF spaces are affected in response to increasing ICP. MD changes in normal pressure hydrocephalus (NPH) patients suggest a higher risk of microstructural alterations when ICP changes are present. 24 Given that IIH is associated with an increased water diffusion within the brain suggesting that a transependymal flow may induce intracellular water accumulation and interstitial brain edema, 25,26 it may be postulated that these changes may induce the microstructural alterations observed in our study. This concept is supported by a study showing that lower FA values are measured during vasogenic brain edema that do not recover after resolution of the edema. 27 However, other mechanisms such as an abnormal cerebral blood flow (CBF) and blood volume (CBV) may be of significant relevance but findings remain controversial. [28][29][30] A potential limitation of our study is the relatively small sample size. However, it is unlikely that a larger sample would have revealed a different result. In addition to the sample size the relatively low spatial resolution does not allow a better distinction of anatomical structures surrounding the optic nerve in DTI. In particular a distinction between optic nerve and optic nerve sheath is not possible. Nevertheless, a higher resolution may offer a better anatomical allocation of the microstructural abnormalities but it is unlikely that an improved resolution may yield normal microstructural properties.

[7] 88w Another limitation could be that among the healthy volunteers in the control group we may have included a subject that in fact could have an elevated intracranial pressure, in particular as patients and healthy volunteers were matched for BMI and a lumbar puncture was not performed in the control group due to ethical considerations. However, this is highly unlikely as all healthy volunteers did not show any sign for an elevation of intracranial pressure (headache, visual disturbance) and had an unremarkable OCT scan with no signs of papilledema.

[8] 41w Finally, due limitations imposed by the ethics committee, in particular in healthy volunteers, we used TOF-MRV instead of contrast-enhanced methods that require the use of intravenously administered contrast medium. Despite being highly unlikely, we may therefore have missed small venous pathologies.

[1] 66w In the present study we show in our DTI analysis that IIH with visual disturbances is associated with microstructural abnormalities. Based on the findings these are likely to be the result from chronically elevated CSF pressure. Future studies will have to elucidate the role of pressure-induced edema as the potential cause of the microstructural damage and a potential reversibility of the alterations by normalization of ICP.

[1] 46w The study was approved by the local ethics committee of the Charit e -Universit€ atsmedizin Berlin (EA1/266/09). A statement of written informed consent was obtained from all patients and healthy volunteers participating in the study. The MRI measurements were performed between June 2010 and April 2011.

[2] 78w Patients and Healthy Volunteers.-Twenty-three patients (3 male, 20 female, age 37.0 6 13.7 years, BMI 33.5 6 7.7 kg/m 2 ), diagnosed with definite IIH according to the diagnostic criteria, 13,14 were compared with 23 healthy volunteers (3 male, 20 female, age 37.2 6 12.7 years, BMI 34.0 6 7.0 kg/m 2 ) matched with respect to sex (P 5 1.0), age (P 5 .973) and BMI (P 5 .816). All patients and controls were of Caucasian descent.

[3] 192w For the identification of suitable patients, medical history files of 190 patients with IIH referred to our University Medical Center between November 2005 and May 2010 were screened via retrieval algorithms for the ICD-10 classification code G93.2. Only patients with a definite diagnosis of IIH according to the diagnostic criteria and a minimum age of 18 years were included in the study. Exclusion criteria comprised systemic conditions or medications that may have an effect on ICP, inadequate follow-up, pregnancy, or postpartumstatus, body weight above 160 kg, clinical history of procedures affecting CSF circulation, depression and claustrophobia. We identified 71 patients that definitely met the diagnostic criteria. Of these, 15 patients were excluded due to the following exclusion criteria: shunt surgery (n 5 6), body weight > 160 kg (n 5 3), pregnancy (n 5 1), claustrophobia (n 5 4), and major depression (n 5 1), 18 patients could not be contacted and 15 patients refused to participate in the study. As a result, 23 patients took part in the study. As the study was conceived as an exploratory pilot study the sample size was chosen based on our experiences with previous studies.

[4] 127w Clinical Assessment.-Sixteen patients (69.6%) presented with headache. Mean and standard deviation of reported pain intensity were 2.2 6 2.5 [range 0.0-7.5] according to the visual analog scale (VAS) and mean of maximum pain intensity amounted to 7.1 6 2.6 [1.0-9.9]. Visual disturbances were present in 95.7% of patients. Tinnitus was reported by 39.1% of patients, of which 44.4% characterized their tinnitus as pulsatile. Light sensitivity (60.9%) and noise sensitivity (52.2%) as well as vertigo (60.9%) were also common symptoms. All patients underwent lumbar puncture and CSF opening pressures were obtained. The mean CSF opening pressure was 37.6 6 6.9 cm H 2 O [29.0 cm H 2 O-50.0 cm H 2 O]. All patients and healthy volunteers underwent an OCT scan to assess the presence of papilledema.

[5] 326w MR Imaging (MRI).-MRI was performed on a 1.5 T scanner (Siemens Avanto Magnetom, Erlangen, Germany). A commercially available loop surface coil (Siemens) with a diameter of 7 cm was used in addition to the circularly polarized head coil. The surface coil was placed over the eye of maximum visual disturbances within the head coil and fixed with tape. A coronal turbo spin echo (TSE) sequence was used to evaluate the optic nerve and ONS (repetition time [TR] 6960 ms, echo time [TE] 99 ms, field of view [FOV] 85 3 85 mm 2 , matrix size of 256 3 256 [in plane resolution 0.332 3 0.332 mm 2 ], slice thickness 2 mm and acquisition time [TA] 7 min 20 s). Axial and sagittal T2weighted TSE sequences were used to evaluate the hypophysis and ventricle size (TR 5170 ms, TE 99 ms, FOV 170 3 170 mm 2 , matrix size of 256 3 320 [in plane resolution 0.664 3 0.531 mm 2 ], slice thickness 3 mm and TA 5 min 27 s. DTI images were recorded using a single-shot echo planar (EPI) sequence in axial orientation with (TR 5448 ms, TE 88 ms, FOV of 192 3 192 mm 2 , matrix size of 128 3 128 [in plane resolution 1.5 3 1.5 mm 2 ], slice thickness 3 mm, TA 11 min 44 s). The diffusion weighting (b value) was set to 0 and 1000 s/mm 2 . Images were acquired with diffusion gradients in 126 directions with a number of acquisitions of 6, comprising 41 slices covering the whole brain. 15 A 2D time-of-flight (2D TOF) venography (TR 23 ms, TE 6.5 ms, FOV 250 3 250 mm 2 , matrix size 512 3 512 [in plane resolution 0.488 3 0.488 mm 2 ], slice thickness 2 mm, TA 7 min 40 s) was obtained to evaluate sinus vein stenosis (SVS) and to exclude sinus vein thrombosis (SVT) as secondary IIH causes.

[6] 111w Morphometric Analysis.-For morphometric measurements raters were blinded for the participants' condition. Maximum ONS diameters were measured on coronal T2w images perpendicular to the ON in the slice with the maximum ONS diameter (Fig. 1A,B). An independent analysis on volumetric assessment of the optic nerve sheath and the hypophysis has been performed in the same study population and published separately. 5 However, two-dimensional morphometric analysis was included in this manuscript to allow a correlation between macroscopic findings and potential microscopic alterations. Maximum height of the hypophysis was measured on the midsagittal T2w images (Fig. 1C,D). Cerebral SVS was defined by the degree of stenosis on axial images and maximum intensity projection images.

[7] 515w DTI Analysis.-DTI data were converted from DICOM into NIFTI format using the DICOM to NIFTI format converter dcm2nii (Dicom office toolkit DCMTK, http://dicom.office.de, Oldenburg, Germany). Voxelwise statistical analyses of FA and MD data were performed using FSL 4.1.8 (FMRIB Software Library, Oxford, UK). [16][17][18] DTI data were corrected for eddy current distortion and motion. b0 data were used for brain segmentation using BET (Brain Extraction). 19 An automatic segmentation tool (FAST) was used for segmentation of white and grey matter. FA and MD values were then calculated by reconstruction of diffusion tensors using FDT (FMRIB's Diffusion Toolbox) and DTIFIT. Via tract-based spatial statistics (TBSS) 20 all subjects' FA and MD data were adjusted into a 1 3 1 3 1 mm standard space using FNIRT (FMRIB's nonlinear registration tool), which uses a b-spline representation of the registration warp field. 21 Next, mean FA and MD images were generated and thinned to create a mean FA (Fig. 1G,H) and mean MD skeleton representing the centers of all tracts common to the group. Each subject's aligned FA and MD data were then projected onto this white matter tract skeleton and the resulting data were compared using a voxelwise cross-subject threshold-free cluster-enhanced permutation analysis. 20 In addition to the voxelwise analysis a global analysis of whole brain FA and MD values of patient and control group were compared using Amira 5.3.2 (Visage Imaging Inc., San Diego, CA, USA). For ROI analysis of the optic pathway, optic nerve and optic radiation were manually outlined on MPRAGE and transferred to coregistered FA images using Amira 5.3.2 (Fig. 1E,F). Additionally the optic radiation was identified by setting thresholds -voxels with FA values below 0.3 were excluded from the analysis, as the optic radiation is a large white matter fiber structure with a predominantly anterior-posterior diffusion flow. 22 Statistical Analysis.-Data were analyzed using XLSTAT Version 2016.5 (Addinsoft SARL, New York, NY, USA). Normality of the data was calculated using the Shapiro-Wilk test. The Mann-Whitney U-test was used to compare the group means and data was expressed as median with 5th and 95th percentiles, based on the quantiles estimation function of XLSTAT. Statistical significance was assumed at P < .05. Pearson's correlation coefficient and correlation matrices were calculated to identify the strength of the different parameters. Receiver-operating characteristic (ROC) analysis was used to select an optimal cut-point for prediction and to assess the predictive value in terms of sensitivity, specificity, and accuracy. 23 Selection of the optimal cut-point was based on the Youden index, that is, the maximum sum of sensitivity and specificity. The 95% confidence interval for AUC was calculated by bootstrapping. data using FSL revealed no significant differences between patients and controls. Median [5th, 95th] FA values of the whole brain (0.201, [0.181, 0.213] in patients vs (0.199, [0.182, 0.208] in controls with a difference of 0.002 (95% CI: 20.006 to 0.010; P 5 .693; Fig. 2A) and the optic radiation (0.487 [0.444, 0.526] in patients vs 0.475 [0.450, 0.512] in controls with a difference of 0.011 (95% CI: 20.003 to 0.028; P 5 .235; Fig. 2C) showed no significant differences.

[1] 136w Category 1 (a) Conception and Design Edzard Wiener, Jan Hoffmann (b) Acquisition of Data Acquisition of Data Christoph Schmidt, Edzard Wiener, Lutz L€ udemann, Hagen Kunte, Katharina Maria Kreutz, Nils Becker, Lutz Harms, Randolf Klingebiel, Jan Hoffmann (c) Analysis and Interpretation of Data Christoph Schmidt, Edzard Wiener, Lutz L€ udemann, Lutz Harms, Randolf Klingebiel, Jan Hoffmann Category 2 (a) Drafting the Manuscript Drafting the Manuscript Christoph Schmidt, Edzard Wiener, Jan Hoffmann (b) Revising It Intellectual Content Revising It for Intellectual Content Christoph Schmidt, Edzard Wiener, Lutz L€ udemann, Hagen Kunte, Katharina Maria Kreutz, Nils Becker, Lutz Harms, Randolf Klingebiel, Jan Hoffmann Category 3 (a) Final Approval of the Completed Manuscript Final Approval of the Completed Manuscript Christoph Schmidt, Edzard Wiener, Lutz L€ udemann, Hagen Kunte, Katharina Maria Kreutz, Nils Becker, Lutz Harms, Randolf Klingebiel, Jan Hoffmann