Zaher Sbeity MD Department of Ophthalmology Einhorn Clinical Research Center New York Eye and Ear Infirmary 310 East 14th Street New York New York 10003 USA Tel: + 1 212 477 7540 Fax: + 1 212 420 0067 Email: email@example.com
Purpose: This study aimed to evaluate the efficacy of non-contact confocal laser microscopy in detecting structural alterations of the cornea, iris and lens in fellow eyes of patients with clinically unilateral exfoliation syndrome (XFS) and XFS suspects.
Methods: In a prospective observational case series, eyes of 12 patients with clinically unilateral XFS, six XFS suspects and six age-matched controls were divided into four groups. Group A included eyes with clinically visible exfoliation material (XFM) on the pupillary border or anterior lens capsule (n = 12); group B included fellow eyes of clinically unilateral XFS patients without visible XFM (n = 12); group C comprised eyes in XFS suspects with signs of pigment dispersion without visible XFM (n = 10), and group D consisted of control eyes with no evidence of XFM or pigment dispersion (n = 12). The cornea, iris and lens were imaged using a non-contact lens prototype for the Rostock Cornea Module (HRT II). Images were analysed by two observers masked to the clinical findings.
Results: Visible XFM (group A) on the lens capsule was characterized by hyperreflectivity in the granular and central disc areas and hyporeflective spaces in the intermediate zones. Similar hyperreflectivity was noted in four of 12 and one of 10 eyes in groups B and C, respectively, but in none in group D. Corneal endothelial hyperreflective deposits were found in eight of 12, four of 12 and one of 10 eyes in groups A, B and C, respectively, and no eyes in group D.
Conclusions: This technology permits visualization of XFM and/or XFM-related changes in the cornea and lens in the unaffected eyes of patients with clinically unilateral XFS. It may allow early detection of XFS and impact glaucoma screening and clinical surveillance decisions.
Exfoliation material (XFM) has been found by electron microscopy and immunohistochemistry in the stromal connective tissue of the conjunctiva, ciliary body and iris, as well as in the epithelial basement membrane region of the cornea, ciliary body and lens (Dark et al. 1977; Morrison & Green 1988; Asano et al. 1995; Schlötzer-Schrehardt et al. 1997). Most patients with exfoliation syndrome (XFS) in many series have been described as having clinically unilateral involvement (Ritch & Schlötzer-Schrehardt 2001). However, there is increasing evidence that XFS is not a truly unilateral disease, but, rather, an asymmetric bilateral disease. Deposits of XFM have been found in the conjunctiva, iris and lens capsule of XFS suspects and unaffected eyes of patients with clinically unilateral XFS (Speakman & Ghosh 1976; Takei & Mizuno 1978; Prince et al. 1987; Oliveira et al. 2006; Parekh et al. 2008). Microscopic alterations in the iris pigment epithelium and ciliary epithelium have also been reported in unaffected fellow eyes (Hammer et al. 2001). Zonular abnormalities detected by ultrasound biomicroscopy (Sbeity et al. 2008a) and the presence of pigment-related signs may indicate preclinical XFS. These signs include pigment dispersion in the anterior segment after pupillary dilation, iris pigment dotting, increased trabecular pigmentation, loss of pigment ruff, and iris sphincter region transillumination defects (Prince et al. 1987; Ritch & Schlötzer-Schrehardt 2001). Although these signs are suggestive, they are not sufficient to confirm a diagnosis of XFS. Early diagnosis of XFS may allow for the early detection of elevated intraocular pressure and glaucomatous optic neuropathy and has important implications for glaucoma screening and clinical surveillance decisions. For instance, patients with evidence of preclinical XFM could be followed more closely to watch for the development of clinically evident XFS because this is a major risk factor for the development of ocular hypertension and exfoliative glaucoma, which carries a more severe prognosis than does primary open-angle glaucoma. Currently, histology is the only tool available for diagnosing preclinical XFS (Prince et al. 1987). However, this requires a surgical biopsy. To date, there is no imaging tool that can detect preclinical XFS.
Recent advances in imaging technology have improved diagnostic modalities in ophthalmology. The Rostock Cornea Module (RCM) originally consisted of a contact lens system attachment to the Heidelberg Retina Tomograph (II). It utilizes laser scanning technology and has been used to visualize cellular changes in various corneal and conjunctival pathologies (Garibaldi et al. 2005; Guthoff et al. 2006). In this study, we examined the ability of a non-contact lens prototype of the RCM to detect XFM and/or XFM-related changes in the cornea, iris and lens capsule of the clinically unaffected eye of patients with unilateral XFS and in eyes of XFS suspects.
Materials and Methods
Patients with clinically unilateral XFS, XFS suspects and controls were enrolled. All patients underwent a complete ophthalmic examination and eyes were then divided into four groups. Group A (n = 12) comprised eyes of patients with unilateral XFS with visible XFM on the pupillary border or anterior lens capsule on slit-lamp biomicroscopy following pupillary dilation. These eyes displayed the three distinct zones characteristic of XFS (homogeneous central disc, intermediate clear zone and peripheral granular zone). Group B (n = 12) included the contralateral clinically uninvolved eyes of subjects in group A. Group C (n = 10) consisted of eyes of XFS suspects which showed at least two of the five clinical signs of pigment dispersion (loss of iris pigment ruff, iris transillumination defects in the sphincter region, increased trabecular pigmentation, pigment dispersion after dilation, iris pigment dotting) without visible XFM. Group D (n = 12) comprised eyes of controls (age-matched to XFS suspects and unilateral XFS patients), in which anterior segment examination was normal and showed no evidence of XFM and/or pigment-related signs. All eyes enrolled had no history of intraocular surgery, trauma or ocular disease other than cataract.
The Rostock Cornea Module (RCM; Heidelberg Engineering GmbH, Heidelberg, Germany) of the Heidelberg Retina Tomograph II with a novel non-contact lens prototype attachment was used to image the anterior segment (Nikon 50x CF Plan/NA 0.45 EPI SLWD, working distance 13.8 mm, estimated digital lateral/axial resolution 1–2/10 μm, field of view [FOV] 500 × 500 μm). The lateral (δr = 0.8 μm) and axial (δz = 8 μm) optical resolution were calculated using the following equations:
respectively, where λ = 670 nm, NA (numerical aperture) = 0.45 and n (cornea) = 1.376. The digital resolution is inferior to the theoretical optical resolution and is defined as FOV/number of pixels (in this case: 500 μm/384 = 1.3 μm).
Patients were asked to focus on a target 1 m behind the examiner. Serial transverse section images of the cornea, iris and lens of all patients were taken. Images of the corneal layers were taken first with the focal plane parallel to the central cornea. A z-axis drive knob was used to advance the focal plane from the corneal epithelium to endothelium. The focal plane was then advanced to visualize the central and mid-peripheral iris, and moved further posteriorly to screen the anterior lens capsule and lens. While positioned on the anterior lens capsule (mid-pupillary level), the focal plane was moved horizontally and vertically in all four directions to screen for XFM. A minimum of three scans consisting of 100 transverse section images were taken. All captured images were saved digitally in video mode and were analysed by an examiner masked to the clinical data. For optimal visualization of the lens, patients were imaged with a dilated pupil (minimum size 5 mm). Alternatively, a darkened room was used for those patients who had reactive undilated pupils. The study was approved by the New York Eye and Ear Infirmary Institutional Review Board and followed the tenets of the Declaration of Helsinki.
We imaged 46 eyes (groups A, B and D: 12 eyes each; group C: 10 eyes). The mean age and corresponding demographic data of all patients are shown in Table 1. All eyes enrolled had unremarkable corneas on slit-lamp biomicroscopy. There was 92% agreement between the two masked observers when evaluating RCM images (κ = 0.83, 95% confidence interval [CI] 0.66–0.99).
Table 1. Subject demographics.
Mean age, years
Ethnic groups (n)
F = female; M = male; XFS = exfoliation syndrome; W = White; H = Hispanic; B = Black.
*p-value (mean age of controls and XFS suspects versus the unilateral XFS group), Mann–Whitney test.
Unilateral exfoliation syndrome (groups A and B)
The corneal endothelium of eight eyes in group A and four eyes in group B (12 of 24) revealed different degrees of scattered round, pleomorphic, hyperreflective deposits (Fig. 1, Table 2) with sporadic polymegethism and guttata. Two affected eyes demonstrated additional hyperreflective fibrillar-like particles in the subepithelial anterior stroma. The endothelial images of the remaining corneas were unremarkable.
Table 2. Eyes with hyperreflective material detected by Rostock Cornea Module. Values are n unless otherwise stated.
Exfoliation material on the pupillary margin was hyperreflective (Fig. 2). The anterior iris surface of 10 eyes in group A and three eyes in group B revealed sparse, scattered, hyperreflective deposits of different sizes corresponding to dispersed pigment granules seen clinically as iris pigment dotting (Fig. 2). Iris thinning in the sphincter region was displayed as large, hyporeflective dark areas signifying reduced connective tissue when imaged by the RCM.
Imaging of the anterior lens capsule in all affected eyes revealed a relatively homogenous, hyperreflective material in the central and peripupillary area corresponding to the central disc and granular peripheral zone of XFM, respectively. Uniform epithelial cells were clearly seen in the intermediate zone (Fig. 3). While imaging the anterior capsule in 10 eyes (six affected and four unaffected eyes), we noted multiple hyperreflective dots of different sizes floating in the aqueous humour of the posterior chamber (pupillary area). These may correspond to liberated XFM or pigment granules. None of these eyes had signs of pigment or cells in the posterior chamber when examined by slit-lamp biomicroscopy.
In 33% (95% CI 9.9–65.1%) of unaffected eyes, hyperreflective fibrillar-like material was seen in the pupillary area of the anterior lens capsule (Fig. 4). This was statistically significant when compared with age-matched controls (p < 0.005, Fisher’s exact test). The epithelial cells of the lens were only partially seen near the pupillary area. These eyes also had hyperreflective deposits in the cornea and iris, as well as hyperreflective material floating in the posterior chamber (Fig. 4). Interestingly, three of these patients had ocular hypertension and two of them were using ocular hypotensive medications.
Exfoliation syndrome suspects and controls (groups C and D)
Cornea, iris and lens
In group C (XFS suspect eyes), one of 10 eyes showed hyperreflective material in the cornea, iris and lens, which appeared to be very similar to that found in eyes with XFS. In addition, the pupil margin showed hyporeflective areas corresponding to pupillary ruff loss. Clinically, this patient had asymmetric hyperpigmented angles and iris dotting, and a positive family history of XFS. The remaining eyes in this group showed no evidence of similar hyperreflective deposits throughout the anterior segment. Imaging of the cornea and lens capsule in group D (control eyes) revealed no hyperreflective deposits. In all groups, cataractous lenses demonstrated hyporeflective cavitations corresponding to vacuoles seen clinically.
Clinically visible XFM and pigment granules on the iris or lens were hyperreflective when imaged by the RCM. One third of the clinically unilateral XFS subjects demonstrated bilateral capsular deposition of hyperreflective material in both the affected and unaffected eyes. This hyperreflective material was similar in form and reflectivity and was located predominantly in the peripupillary area, coinciding with the typical location of pregranular XFM (Bartholomew 1971; Dark et al. 1977; Zhang et al. 2000). These changes are unlikely to be solely age-related because our age-matched controls did not show similar findings. Therefore, we believe this hyperreflective material most likely represents preclinical XFM, an entity that can only be detected through histology. There was no significant difference in mean age between study and control groups (70.8 years versus 66.7 years; p = 0.22, Mann–Whitney test). Unlike the XFS group, the control group comprised mainly non-White subjects who are less likely to develop XFS.
Exfoliative keratopathy is characterized by endothelial cell loss and subsequent oedema (Stefaniotou et al. 1992; Naumann & Schlötzer-Schrehardt 2000). We found the clinically unremarkable corneas of patients with XFS to have round and small hyperreflective endothelial deposits, probably corresponding to subclinical XFM and/or pigment granules. It is crucial to differentiate XFM from pigment granules when evaluating confocal images. Although pigment granules appear more hyperreflective than XFM, we were unable to detect any measurable difference in the intensity of their reflectivity. Alternatively, we relied on indirect signs such as size, shape and homogeneity to differentiate between these two entities. The pleomorphic and irregular deposits found on the corneal endothelium (smaller than an endothelial cell) are likely to represent XFM rather than pigment granules because the latter usually appear round and regular in size. Martone et al. (2007) found similar small hyperreflective deposits in the clinically unremarkable cornea of a patient with XFS.
This finding agrees with our results of a previous prospective study where we imaged 30 XFS eyes with clinically unremarkable corneas using the non-contact RCM (Sbeity et al. 2008a, 2008b). We found 19 eyes (63%) to have variable degrees of similar hyperreflective deposits in the endothelium and less frequently in the anterior stroma. In our current study, these deposits were not found in all controls or in nine of 10 eyes in XFS suspects and were more frequently seen in the affected, rather than the unaffected, XFS eyes (eight of 12 versus four of 12, respectively). Therefore, we believe the presence of endothelial deposits may be directly related to the clinical manifestations of XFM. The presence of fibrillar-like material in the anterior corneal stroma of our patients with XFS serves as evidence of the intrinsic production of anomalous proteins and/or XFM by the subepithelial basement membrane (Hiscott et al. 1996; Schlötzer-Schrehardt et al. 1997).
The fact that not all corneas in group A showed these hyperreflective deposits can be explained either by the reduced sensitivity of this technology or by a non-central location of these deposits that was not imaged. A large population study is needed to determine the sensitivity and specificity of this new technology and should ideally be confirmed by immunohistochemistry. Complementary genotyping for single-nucleotide polymorphisms (SNPs) in the lysyl oxidase-like 1 (LOXL1) gene may assist in the future validation of our in vivo confocal findings in XFS suspects.
A limitation of this study is our small sample size. However, the sample sizes of other reports in the literature regarding the diagnosis of subclinical XFS in unaffected eyes of XFS subjects or XFS suspects are comparable. The prevalence of preclinical microscopic changes found in the fellow eye in unilateral XFS has been variably reported. These XFM-related changes were absent in a case report involving a single patient, but were present in the majority of fellow eyes in multiple case series (Layden & Shaffer 1974; Prince et al. 1987; Schlötzer-Schrehardt et al. 1992; Hammer et al. 2001; Parekh et al. 2008). Parekh et al. (2008) used transmission electron microscopy to examine the conjunctiva and lens capsule of both eyes in patients with clinically unilateral XFS and found microscopic evidence of bilateral involvement in the lens capsule and/or the conjunctiva in 26 of 32 patients. In their study population, the probability that a seemingly unaffected eye of XFS was involved ultrastructurally was 81%. Using in vivo confocal microscopy, we found a lower prevalence (25%) of XFM or XFM-related changes in the clinically unaffected eye. In another immunohistochemical study, five affected eyes with clinically unilateral XFS were evaluated following autopsy, revealing no deposits of XFM in the corneal endothelium and stroma (Kivelä et al. 1997). However, we found hyperreflective deposits in the anterior stroma of the cornea in two affected eyes and in none of the fellow eyes. Of note, endothelial deposits found by RCM may be a mixture of XFM and pigment granules; however, this needs to be confirmed histologically.
The ability of the RCM to detect XFM or XFM-related material without having to remove tissue for biopsy represents progress toward the diagnosis of preclinical XFS. The suboptimal resolution and absence of confirmatory tissue staining are limitations of this in vivo technology. This may help to explain the disparity in sensitivity between the two technologies. The inability to determine the exact depth of the focal plane adds another limitation to the non-contact prototype. We used common ocular structures such as the central cornea and pupillary borders to assist in image localization. Therefore, electron microscopy remains the reference standard for the diagnosis of ultrastructural XFM.
In conclusion, the non-contact RCM permits in vivo imaging of microstructural changes in the cornea, iris and lens of patients with XFS. To our knowledge, this is the first study demonstrating the ability of an in-vivo imaging device to detect subclinical material deposited on the anterior lens capsule in unaffected eyes of patients with clinically unilateral XFS. This promising, rapid, non-contact imaging technique needs further improvement and may aid in detecting subclinical XFS. This approach needs to be further studied and optimized for imaging anterior segment pathology in patients with XFS and can be complementary to genetic approaches such as that of identifying SNPs in the LOXL1 gene.
This study was supported in part by the Linda and Stuart Nelson Research Fund, New York Glaucoma Research Institute, New York, NY.