Three-dimensional High-speed Optical Coherence Tomography Imaging of Lamina Cribrosa in Glaucoma Pf.박찬기/R4 조현경 Ophthalmology. 2009 Feb;116(2):214-22.

Introduction Optic neuropathy owing to glaucoma is believed to begin within the lamina cribrosa, based on findings in histopathologic studies of cadaver or enucleated human eyes and animal studies. Changes seen in human eyes with glaucoma include deformation and compression, larger pores, and less connective tissue associated with this structure.

Introduction Quigley et al found significant thinning of the lamina cribrosa on vertical sections of eyes with various stages of glaucoma compared with normal eyes when examined by electron microscopy. Jonas et al found thinning on morphometric analysis of light microscopic sections. These changes in the lamina cribrosa are thought to promote optic neuropathy by decreasing support for optic nerve fibers.

Introduction However, histopathologic studies are limited for investigations of glaucoma’s effects on soft tissue dimensions because tissues can swell or shrink during preparation, the influence of intraocular pressure is altered in enucleated eyes, and too few cadaver or enucleated eyes are available for large-scale studies.

Introduction On standard-resolution time-domain optical coherence tomography (OCT) cross-sectional images of the optic nerve head, the lamina cribrosa appears with less clarity than is desirable for morphometric analysis of the lamina cribrosa structures. Only with spectral-domain OCT (SD-OCT), which provides considerably improved imaging speed and higher axial resolution views compared with time- domain OCT, has it been possible to depict structures of the optic nerve head in 3 dimensions.

Introduction However, 3-dimensional (3D) imaging of lamina cribrosa using SD-OCT has not been reported. To determine which structures are visible in the lamina cribrosa on 3D SD-OCT images in eyes with glaucoma To test for a correlation between lamina cribrosa thickness measured on SD-OCT images and visual field loss

Materials and Methods Prospective observational study 30 Japanese adults with glaucoma (39 eyes of 24 patients) or ocular hypertension (OHT; 13 eyes of 10 patients) Kyoto University Hospital between November 1, 2005, and April 30, 2006

Materials and Methods Comprehensive ophthalmologic examination, best-corrected visual acuity, slit-lamp biomicroscopy, gonioscopy, Goldman applanation tonometry, dilated stereoscopic examination of the optic nerve head and fundus, and achromatic automated perimetry using the 30-2 SITA (Humphrey Visual Field Analyzer; Carl Zeiss-Meditec, Inc., Dublin, CA) Experienced ophthalmologists performed standard OCT using a Stratus OCT (Carl Zeiss-Meditec) and a prototype SD- OCT system Perimetry and OCT examinations were performed within 6 months of each other.

Materials and Methods <Inclusion criteria> best-corrected visual acuity of ≥20/40, spherical refraction within ±6.0 diopters and cylinder correction within ±3.0 diopters. <Exclusion criteria> Those with intraocular disease (e.g., diabetic retinopathy or retinal vein occlusion) or neurologic disease (e.g., pituitary tumor) that could cause visual field loss

Materials and Methods Visual field loss was expressed as mean deviation (MD) from normal A visual field measurement was defined as reliable when false-positive and false-negative results were <25% and fixation losses were <25% A visual field change owing to glaucoma was defined as (1) abnormal range on glaucoma hemifield test; or (2) 3 abnormal points with P<5% probability of being normal, 1 with P<1% by pattern deviation; or (3) pattern standard deviation of <5% if the visual field was otherwise normal, confirmed on 2 consecutive tests.

Materials and Methods <Spectral-domain System for 3-Dimensional Optical Coherence Tomography Imaging> A prototype SD-OCT system was fabricated for patient examinations at Kyoto University Hospital, Japan, by the Computational Optics Group at the University of Tsukuba (Ibaraki, Japan) and TOPCON Corporation (Tokyo, Japan). This prototype consisted of a SD-OCT system and a nonmydriatic fundus camera (TRCNW200, TOPCON) controlled by a computer software program (LabVIEW, National Instruments Corporation, Austin, TX).

Materials and Methods Sensitivity of 98 decibels (dB) Tissue axial resolution of 4.3 µm Acquisition rate of ~18,700 axial scans per second The total data acquisition time for a single 3D-OCT (volumetric) image was 3.5 seconds. light source has a center wavelength of 830 nm and bandwidth of 50 nm For 3D analyses, a raster scan protocol of 256 x 256 axial scans covering a 2.8 x 2.8 mm disc area was used

Materials and Methods Volumetric rendering of the 3D-OCT data set was performed and longitudinal and en face cross-sections were constructed using image processing software (Amira 3.1, Mercury Computer Systems Inc., Chelmsford, MA). A fundus image was generated as an en face projection image from the 3D data set.

Materials and Methods <Measurement of Lamina Cribrosa Thickness> We defined the thickness of the lamina cribrosa as the distance between the anterior and posterior borders of the highly reflective region visible beneath the optic disc cup in longitudinal SD-OCT cross-sections of the optic nerve head. Because thick retinal vessels cause “fringe washout” (large shadows in OCT cross-sectional images) that prevent visualization of structures underlying the vessels, we measured lamina cribrosa thickness as centrally as possible, where there is the least likelihood of vascular shadows.

Materials and Methods <Interobserver Reproducibility> 12 randomly selected SD-OCT cross-sectional images of optic nerve heads were evaluated by 6 examiners and the intraclass correlation coefficient was calculated. According to Fleiss JL, scores ≥ 0.75, between 0.40 and 0.75, and ≤ 0.4 are termed as excellent, moderate, and poor, respectively.

Materials and Methods <Statistical Analysis> SPSS version 11.1 (SPSS Inc., Chicago, IL) One-way analysis of variance and Scheffe’s multiple comparison were used to compare the mean thickness of the lamina cribrosa among 4 groups: eyes with normal visual field (OHT) eyes with early glaucoma (MD >- 6) eyes with moderate glaucoma (-12 ≤ MD ≤ -6) eyes with advanced glaucoma (MD < -12)

Materials and Methods <Statistical Analysis> Because age differed significantly between the groups, age was included as a covariant in the analysis of variance model. The Spearman t test was calculated to study the relationship between lamina cribrosa thickness and the MD value of the visual field. The level of statistical significance was set at P<0.05.

Results The reconstructed 3D volume of the optic disc was observed in 52 eyes of 30 patients. In 3 of the 52 eyes, vascular shadowing prevented clear imaging of lamina cribrosa structures and in 19 of the 52 eyes, the posterior borders of the highly reflective region beneath the optic disc cup were unclear owing to poor image contrast. -> excluded from analyses of lamina cribrosa thickness. The remaining 30 eyes of 22 patients (1 eye each of 14 patients and 2 eyes of each of 8 patients) were considered for the measurement and analysis of the lamina cribrosa thickness.

Results The patients ranged in age from 47 to 77 years (mean standard deviation, 62.1 ± 8.87 years; 11 men and 11 women) Of the 30 eyes, 24 had glaucoma (15 primary open-angle glaucoma, 7 normal tension glaucoma, and 2 secondary glaucoma), and 6 had OHT Values for visual field MD ranged from -29.29 to +1.04dB (-6.46±7.26)

Results The mean duration between the visual field and OCT examinations was 2.58±1.74 months, ranging from 0 to 6 months No progression of visual field loss as observed from the comparison of the pre- and post-OCT examination MD values. Spherical refraction ranged from -4.00 to +2.75 diopters (-0.28±1.97 diopters).

Results <3D OCT Imaging of Optic Nerve Head and Lamina Cribrosa> 3D volume-rendered image of the optic nerve head Red arrow points to an OCT fundus image produced from the 3D data set

Results On longitudinal cross-sectional SD-OCT images (B-scan images), the lamina cribrosa appears as a highly reflective layer beneath the optic disc cup Clear changes in reflectivity in the anterior and posterior borders of the lamina cribrosa Posteriorly concave platelike structure : low & highly reflective parts vascular shadowing effect nasally

Results En face cross-sectional images at the levels indicated by orange lines in D-1. On en face cross-sectional images through the lamina cribrosa, multiple dot areas of low reflectivity were seen within the highly reflective lamina cribrosa

Results The positions of dots of low reflectivity in the en face cross-sectional images at the level of anterior border of the lamina cribrosa corresponded with laminar pores in the color fundus photographs

A. F/62 OHT, MD 0.23 dB B. M/62 Moderate POAG MD -8.14 dB C. F/69 Advanced MD -19.98 dB

Results Orthogonal (longitudinal and en face) cross-sectional images within a 3D volume-rendered image

Results 3D volume and cross-sectional images of the optic nerve head axially enlarged by 3.5-fold. C-4; dot areas of low reflectivity (green arrowhead) at all levels form a column of low reflectivity (yellow arrowheads) extending axially in the longitudinal cross-sectional image.

Results The mean thickness of the lamina cribrosa in the 30 eyes was 190.5±52.7 µm (range, 80.5–329.0).

Results There were significant differences in the mean thicknesses of the lamina cribrosa between the 4 groups (P<0.001) the mean thickness was significantly less in eyes with advanced glaucoma than in eyes with OHT (P<0.001) or early glaucoma (P< 0.019)

Results Lamina cribrosa thickness showed significant positive correlation with MD by the Spearman test ( σ= 0.744; P<0.001) and linear regression analysis (r2= 0.493; P<0.001)

A. F/53 OHT, MD 1.01dB B. Lamina cribrosa thickness 266µm C. M/63 NTG, MD -14.06dB D. Lamina cribrosa thickness 137µm

Results Lamina cribrosa thickness showed significant negative correlation with age (σ= 0.386; P= 0.035) and no significant correlation with spherical refraction (σ= 0.047; P=0.806) There was excellent interobserver reproducibility (intraclass correlation coefficient =0.784) in measurement of lamina cribrosa thickness.

Discussion Quigley et al used EM to find that greater severity of glaucoma was correlated with greater thinning of the lamina cribrosa Jonas et al found on LM that the lamina cribrosa was significantly thinner in eyes with glaucoma compared to control eyes 3D SD-OCT imaging enabled better visualization of the lamina cribrosa and measurement of the lamina cribrosa thickness in living human eyes

Discussion In tree shrews, a comparison of SD-OCT cross-sectional images and histologic sections of the optic nerve head showed that the highly reflective region underlying optic disc cup corresponds with the lamina cribrosa Polarization-sensitive OCT showed that there is high birefringence in the regions that correspond to the highly reflective layer beneath the optic disc cup in healthy humans, collagen tissue has high birefringence -> highly reflective region underlying optic disc cup corresponds to the lamina cribrosa

Discussion Low-reflectivity dots represent laminar pores The presence of low reflectivity dots throughout all layers of the highly reflective region on longitudinal sections and in orthogonal crosssections allowed us to identify the low-reflectivity dots with low reflective, columnlike structures, possibly representing the passage for optic nerve fiber bundles, traveling axially through all levels in lamina cribrosa

Discussion It is uncertain whether the dots of low reflectivity we noted in our study represent nerve fiber bundles within laminar pores or empty pores The retinal nerve fiber layer appears as a highly reflective layer in OCT cross-sectional images However, the intrapapillary retinal nerve fiber layer, which changes direction at the disc margin to run parallel to the optical axis of the OCT probe beam, has low reflectivity dots of low reflectivity in the lamina cribrosa were seen not only in eyes with glaucoma, but also in eyes with OHT

Discussion The hyporeflectivity of the intrapapillary retinal nerve fiber layer contrasting with the high reflectivity of the laminar plate visualization of laminar pores in 3D SD-OCT images, whether there is a decrease in the nerve fiber bundles or not

Discussion The axial resolution of SD-OCT images is poorer than that achieved with LM or EM on our SD-OCT images the anterior and posterior borders of the highly reflective region representing the lamina cribrosa were not sharp enough to allow our measurements Ultra-high-resolution SD-OCT, speckle noise-free SD OCT, or SD-OCT using a 1040-nm wavelength light source may improve the visualization of the borders of the lamina cribrosa 1040-nm light source -> enhance penetration into the deeper tissues

Discussion However, measurements on histologic specimens may not be accurate reflections of in vivo measurements Despite the limitations of each method, the mean thicknesses of the lamina cribrosa in eyes with normal visual field and with glaucoma in our study were almost the same as the mean thicknesses measured on histologic sections by Quigley et al.

Discussion The mean patient age was significantly different between groups; patients with advanced glaucoma were significantly older than those with early glaucoma A study using LM on cadaver eyes found that the thickness of the human lamina cribrosa increased with increasing age In our study, lamina cribrosa thickness decreased with increasing age -> the thinning of lamina cribrosa with glaucoma progression does not result from the age difference between the groups

Discussion Our SD-OCT system has 2 limitations with regard to the selection of points for the measurement of the lamina cribrosa thickness 1) Lamina cribrosa beneath the rim was invisible 2) Vascular shadowing of the intrapapillary central retinal vessels obscured many parts of the lamina cribrosa so that its thickness could not be measured in its actual central region of all eyes

Discussion 1) lamina cribrosa beneath the rim was invisible in OCT imaging, the area where the lamina cribrosa is visible depends on the cup size -> cannot measure the lamina cribrosa thickness in the periphery in eyes with a small cup we selected the measurement points as centrally as possible in this study

Discussion The previous studies by Quigley et al and Jones et al showed that a greater severity of glaucoma correlated with a greater thinning of the lamina cribrosa in the center -> the selection of the central points for measurement owing to this limitation is relevant as long as one intends to study the relationship between the lamina cribrosa thickness and the severity of glaucoma in patients

Discussion 2) vascular shadowing of the intrapapillary central retinal vessels -> lamina cribrosa thickness could not be measured in its actual central region In our study, measurements were limited to areas where the lamina cribrosa was clearly visible from the center to the midperiphery

Discussion Jonas et al showed that the lamina cribrosa varied little in thickness between regions; it was only 0.81% to 2% thinner in healthy eyes and 6.3% to 13.7% thinner in eyes with glaucoma at the midperipheral point of lamina cribrosa than at the thickest central point In our study, the thickness of lamina cribrosa was 47% thinner in eyes with advanced glaucoma than in eyes with OHT, a much larger difference than variation between regions

Discussion Despite these limitations, we successfully obtained the 3D images of all layers of the lamina cribrosa and measured its thickness in patients with glaucoma, and we found good correlation between lamina cribrosa thickness measured with our SD-OCT system and MD in visual field measured by static automatic perimetry

Discussion These warrant larger scale clinical studies using 3D SD- OCT imaging to confirm the relationship between the thinning of the lamina cribrosa and the extent of visual field loss and to further determine if, when, and with what types of glaucoma deformation and compression of the lamina cribrosa occur

Thank you for your attention~