New Investigations in Ophthalmology Tanuj Dada, Subrata Mandal
INDEX
×
Chapter Notes

Save Clear

Section 1
1CORNEA2

Confocal Microscopy of the Cornea1

Manotosh Ray
 
INTRODUCTION
Confocal microscopy, one of the most advanced imaging technologies, offers several advantages over conventional wide-field optical microscopy. It has the ability to control the depth of field, eliminate or reduce the background information away from the focal plane and the capability to collect serial optical sections from thick specimens. The basic key to the confocal approach is the use of spatial filtering techniques to eliminate out-of-focus light or glare. There has been a tremendous interest in confocal microscopy in recent years, due in part to the relative ease with which extremely high quality images can be obtained. Confocal microscopy has enhanced the ability to image the cornea in vivo. The application of this technology permits the acquisition of images of high spatial resolution and contrast as compared to conventional microscopy.
Confocal microscope employs an oscillating slit aperture in an ophthalmic microscope configuration, especially suitable for the analysis of cell layers of cornea. It can focus through the entire range of a normal cornea from epithelium to endothelium. A series of scan shows (a) epithelium (b) corneal nerves (c) keratocytes (d) endothelium and a (e) computer generated slice of cornea. There are distinct advantages of confocal microscope over the regular microscope. When focused on a transparent tissue like the cornea with a regular microscope, the unfocused layers affect the visibility of the focused layer. Confocal microscope, on the other hand, can focus on different layers distinctly without affecting the quality of the image.
 
Optics
A halogen light source passes through movable slits (Nipkow disk). A condenser lens (front lens) projects the light to the cornea. Only a small area inside the cornea is illuminated to minimize the light scattering. The reflected light passes through the front lens again and is directed to another slit of same size via a beam-splitter. Finally the image is projected onto a highly sensitive camera and displayed on a computer monitor (Fig. 1.1).
The confocal microscope utilizes a transparent viscous sterile gel that is interposed between the front lens and the cornea to eliminate the optical interface with two different refractive indices. The front lens works on ‘Distance Immersion Principle'. The working distance (distance between front lens and the cornea) is 1.92 mm. The back and forth movement of the front lens enables scanning of the entire cornea starting from anterior chamber and corneal endothelium to most superficial corneal epithelium. Use of standard 40x immersion lens gives magnified cellular detail and an image field of 440 × 330 μm. Other lenses (e.g. 20x) delivers wide field but less distinct cell morphology. Newer models (Confoscan 2.0) capture 350 images per examination at a rate of 25 frames per second.4 Thickness of the captured layers varies from 3-5 microns depending on scanning slit characteristics.
zoom view
Fig. 1.1: Optics of the confocal microscope
In addition, every recorded image is characterized by its position on the ‘Z' axis of the cornea. Every time a confocal scan is performed, a graphic shows the depth coordinate on the ‘Z' axis and the level of reflectivity on the ‘Y' axis. The graphic also displays the distance between two images along the antero-posterior line. This simultaneous graphic recording is called ‘Z' scan graphic. The reflectivity on ‘Z' scan is entirely dependent on the tissue being scanned. A transparent tissue displays low reflectivity whereas a higher reflectivity is obtained from an opaque layer. Therefore, different corneal layers would display different reflectivity on ‘Z' scan. The corneal endothelium displays the maximum reflectivity while stroma is the lowest. An intermediate reflectivity is obtained from epithelial layers. A typical ‘Z' scan of entire normal cornea shows high endothelial reflection curves followed by low stromal reflection and then late intermediate reflectivity from superficial corneal epithelium. Thus confocal miscroscopy enables to perform corneal pachymetry or even measure the distance between two corneal layers.
 
CONFOCAL MICROSCOPY OF NORMAL CORNEA
This is a noninvasive technique of imaging of corneal layers that provides excellent resolution with sufficient contrast. A well-executed scan can visualize the corneal endothelium, stroma, subepithelial nerve plexus and epithelial layers distinctly. The limitations are non-visualization of normal Bowman's layer and Descemet's membrane since these structures are transparent to this microscope. However, it is possible to view these structures when they are pathologically involved. Eyes with corneal opacity or edema can also be successfully scanned.¹ The quality of image depends on (a) centration of the light beam (b) stability of the eye and (c) optimum brightness of the illumination.
 
Epithelium
Corneal epithelium has five to six layers. Three different types of cellular component are recognized in the epithelium.
  • Superficial (2-3 layers): Flat cells
  • Intermediate (2-3 layers): Polygonal cells
  • Basal cells (single layer): Cylindrical cells
The superficial epithelial cells appear as flat polygonal cells with well-defined border, prominent nuclei and uniform density of cytoplasm. The main identifying features of superficial epithelial cells are nuclei, which are brighter than surrounding cytoplasm and usually associated with peri-nuclear hypodense ring (Fig. 1.2). The intermediate epithelial cells are similar polygonal cells as superficial layers but the nuclei are not evident. Basal cell layers are smaller in size and appear denser than other two layers (Fig. 1.3). The nucleus is not evident in basal layers also.5
zoom view
Fig. 1.2: Superficial epithelial cells with prominent nuclei
zoom view
Fig. 1.3: Basal epithelial cells. High cell density with well demarcated cell borders
 
SUBEPITHELIAL NERVE PLEXUS
Corneal nerves originate from the long ciliary nerve, a branch of the ophthalmic division of the trigeminal nerve. Nerve fibers from long ciliary nerve form a circular plexus at the limbus. Radial nerve fibers originate from this circular plexus and run deep into the stroma to form a deep corneal plexus. Now deep vertical fibers derive from deep corneal plexus to run anteriorly to form sub-basal and subepithelial nerve plexus. Small nerve fibers from sub-basal plexus terminate at the superficial epithelium.
This complex anatomy was not possible to visualize in vivo until the advent of corneal confocal microscope. Generally the nerve fibers appear bright and well contrasted against a dark background (Fig. 1.4). Confocal microscopy can visualize the orientation, tortuosity, width, branching pattern and any abnormality of the corneal nerves.2
 
STROMA
Corneal stroma represents 90 percent of total corneal thickness. It has three components:
  1. Cellular stroma: Composed of keratocytes and constitutes 5 percent of entire stroma.
  2. Acellular stroma: Represents the major component (90-95%) of stroma. The main component is regular collagen tissue (Type-I, III, IV) and interstitial substances.
  3. Neurosensory stroma: Represented by stromal nerve plexus and nerve fibers originating from it.
The keratocyte concentration is much higher in the anterior stroma and progressively decreases towards the deep stroma. Generally the keratocyte count is approximately 1000 cells/mm2 in anterior stroma while the average value drops to 700 cells/mm² in the posterior stroma. Confocal image of stroma shows multiple irregularly oval, round or bean shaped bright structures that represent keratocyte nuclei. These nuclei are well contrasted against the dark acellular matrix (Fig. 1.5). Anterior stromal keratocyte nuclei assume rounded bean shaped morphology while the same in the rear stroma are more often irregularly oval. A bright highly reflective keratocyte means a metabolically activated keratocyte of a healthy cornea. In a normal healthy cornea, collagen fibers and interstitial substances appear transparent to the confocal microscope and impossible to visualize.
zoom view
Fig. 1.4: Subepithelial nerve fibers
6
zoom view
Fig. 1.5: Stromal keratocytes with bright oval-shaped nuclei
It is possible to identify stromal nerve fibers in anterior and mid stroma. These nerve fibers belong to the deep corneal plexus and appear as linear bright thick lines. The stromal nerve fiber thickness is greater than the epithelial nerves. Occasionally nerve bifurcations are also clearly visible.
 
ENDOTHELIUM
Endothelium is a non-innervated single layer of cells at the most posterior part of the cornea. Endothelial cell density is maximum at birth and progressively declines with age. Normal endothelial cell count varies from 1600 to 3000 cells/mm2 (average 2700 cells/mm2) in a normal healthy adult.2-4 However, the cornea can still maintain the integrity till the cell count declines below 300-500 cells/mm.2
Homogeneous hexagonal cells with uniform size and shape represent healthy endothelial cells. Increasing age and endothelial assault cause pleomorphism and polymegathism. Confocal microscopy easily identifies endothelial cells. These cells appear as bright hexagonal and polygonal cells with unrecognizable nucleus. The cell borders are represented by a thin, non-reflective dark line (Fig. 1.6). A 20x objective lens provides wide field with less magnification. It is possible to perform cell count and study the minute details of cellular morphology.
 
CONFOCAL MICROSCOPY IN CORNEAL PATHOLOGIES
 
Keratoconus
Keratoconus is a non-inflammatory ectatic disorder of the cornea characterized by a localized conical protrusion associated with an area of stromal thinning. The thinning is most apparent at the apex of the cornea. The steep conical protrusion of the corneal apex causes high myopia with severe irregular astigmatism. Other features include an iron ring, known as Fleischer's ring, that partially or completely encircles the cone.5 The cone appears as ‘oil drop’ reflex on distant direct ophthalmoscopy due to internal reflection of light. Deep vertical folds oriented parallel to the steeper axis of the cornea at the level of deep stroma and Descemet's membrane are known as Vogt's striae. An acute corneal hydrops appears when there is a break in the Descemet's membrane. The corneal edema usually subsides after few months leaving behind a scar and flattening the cornea. The corneal nerves become more readily visible due to thinning of the cornea. High irregular astigmatism precludes adequate spectacle correction. In the early stages, use of contact lenses may improve the visual acuity. However, contact lens fitting can be extremely difficult and in advanced cases, it ceases to improve visual acuity.
The most effective way to identify early cases of keratoconus is computerized corneal topography that has become a gold standard for diagnosis and follow-up of the disease in recent years.6,7
zoom view
Fig. 1.6: Hexagonal endothelial cells in a healthy individual
7
Confocal microscopy is a relatively newer investigative modality to assess the keratoconic cornea. Morphological changes in keratoconus are mostly confined to the corneal apex and depend on the severity of the disease. Rest of the cornea may appear normal. The typical polygonal shape of superficial epithelial cells is lost. They appear distorted and elongated in an oblique direction with highly reflective nuclei (Fig. 1.7). Cell borders are not distinguishable. There may be areas of basal epithelial loss as evident by a linear dark non-reflective patch in confocal microscopy. The subepithelial nerve plexus generally appear normal. However, the sub-basal nerve fibers are curved and take the course of stretched overlying epithelium. Corneal stroma is also affected by keratoconus. The confocal images of stroma are highly specific. The characteristic stromal changes are multiple ‘striae' represented by thin hyporeflective lines oriented vertically, horizontally or obliquely (Fig. 1.8). These are confocal representation of Vogt's striae.8 In advanced stages of keratoconus, the keratocyte concentration is reduced in anterior stroma. The shape of the keratocytes is also altered. Occasionally highly reflective bodies with tapering ends are visible in the anterior stroma near the apex. The nature of these abnormal bodies is not yet known but it may be due to altered keratocytes. The corneal endothelial changes vary from none to occasional pleomorphism and polymegathism.
zoom view
Fig. 1.7: Obliquely elongated superficial epithelium in keratoconus
zoom view
Fig. 1.8: Advanced keratoconus: Vertical striae in the stroma
 
CORNEAL DYSTROPHIES
Corneal dystrophies are inherited abnormalities that affect one or more layers of cornea. Usually both eyes are affected but not necessarily symmetrically. They may present at birth but more frequently develop during adolescence and progress gradually throughout life. Some forms are mild, others severe.
 
Granular Dystrophy
This is an autosomal dominant bilateral non-inflammatory condition that results from deposition of eosinophilic hyaline deposits in the corneal stroma.9 It specifically affects the central cornea and eventually can cause decreased vision and eye discomfort. Initially the lesions are confined to superficial stroma but with progression of the disease, they can involve the posterior stroma as well.
Confocal microscopy reveals highly reflective, bright, dense structures in the anterior and mid-stroma. Keratocytes are not involved. Depth of stromal involvement may be ascertained by using ‘Z’ scan function. This is an added advantage over other contemporary investigations that enables the ophthalmologist to plan for surgical modalities. Confocal microscopy is also useful in differential diagnosis and follow-up of the disease.8
 
POSTERIOR POLYMORPHOUS DYSTROPHY
PPD is a rare inherited disorder of the posterior layer of the cornea. It is a bilateral disorder with early onset, although early stage diagnosis is rare since most of the affected individuals remain asymptomatic. The characteristic endothelial changes are small vesicles or areas of geographic lesions. In fact, endothelial cells lining the posterior surface of the cornea have epithelial like features.10,11 These cells can also cover the trabecular meshwork, leading to glaucoma in some patients. Most severe cases may develop corneal edema due to compromised pump function of the endothelial cells.
Confocal microscopy shows multiple round vesicles at the level of Descemet's membrane and endothelium.12 PPD usually distorts the normal flat profile of the endothelial cells and presents large a dark cystic impressions on the confocal scan. The endothelial cells surrounding the lesion appear large and distorted.
 
FUCH'S ENDOTHELIAL DYSTROPHY
This is chronic bilateral hereditary (variable autosomal dominant or sporadic) disorder of corneal endothelium. It typically presents after the age of 50 and is more common in females. There is loss of endothelial cells that results in deposition of collagen materials in Descemet's membrane (guttata). Corneal guttata is the hallmark of this disease. The integrity of corneal endothelium is essential to maintain the metabolic and osmotic function of the entire cornea. Corneal edema in Fuch's dystrophy initially involves the posterior and mid-stroma. As the disease advances, the edema progresses to involve the anterior cornea. This results in formation of bullous keratopathy.
Confocal microscopy is useful to visualize the corneal guttata. This technique has a distinct advantage over conventional specular microscopy that fails to visualize the endothelium when there is significant corneal edema.13 The corneal guttata appear dark with a bright central reflex14 (Fig. 1.9). In advanced stage the endothelial morphology is altered completely but it is still possible to identify the distorted cell borders.14 In the early stages of bullous keratopathy, intra-epithelial edema is seen as distorted cellular morphology with increased reflectivity. It can also identify the bullae in the basal epithelial layer.
zoom view
Fig. 1.9: Distorted endothelium in Fuch's endothelial dystrophy
 
LASIK
LASIK is one of the latest techniques of excimer laser refractive surgery that is currently being successfully used by refractive surgeons for the correction of various types of refractive errors. LASIK has become the technique of choice to correct myopia and hypermetropia with or without astigmatism.15 LASIK is a modification of PRK where excimer laser is used to ablate superficial corneal stroma after the epithelium has been removed. LASIK involves the use of microkeratome to prepare a hinged corneal flap of uniform thickness. The excimer laser is subsequently used to ablate the mid-corneal stromal bed and thereafter the flap is reposited to its original position without applying any suture. After LASIK, the healing of corneal tissue occurs quickly since there is minimal damage to the corneal epithelium and the Bowman's membrane.
Traditionally the cornea is evaluated with slit-lamp biomicroscopy and computerized corneal topography both pre-and postoperatively. Confocal microscopy adds newer dimensions to the commonly employed investigations. Functional outcome of LASIK depends on many factors including the biomechanics, healing process and the inflammatory response of the flap interface that is created between the epithelial flap and stromal bed.9 Confocal scan is useful in evaluation of following parameters.
  • Corneal flap thickness
  • Interface study
    1. Healing process
    2. Inflammatory response
    3. Abnormal deposits
  • Corneal nerve fiber regeneration
  • Residual stromal thickness
A well-designed flap is the key to a positive outcome of LASIK. Thinner flaps are more at risk from flap complications. Few studies with confocal microscopy have suggested that actual flap thickness after LASIK is consistently lower than predicted thickness.16 The reasons are not yet known. However, corneal edema that may be caused by microkeratome cut and suction may play an important role. Postoperative scarring and tissue retraction could be other possible factors. Using a ‘Z' scan, it is possible to identify the interface that corresponds to a very low level of reflectivity. The flap thickness is obtained by measuring the distance between the high reflective spike from the front surface of the cornea and the low reflective interface (Fig. 1.10).
 
MEASUREMENT OF FLAP THICKNESS IN LASIK
The interface usually appears as a hyporeflective space in between relatively hyper-reflective cellular stroma. The interface can easily be imaged by confocal microscope. Typically the keratocyte concentration is lower than normal in the interface. Bright particles and microstriae are consistently visible in the interface. These bright particles most probably originate from microkeratome blade and represented by highly reflective white bodies (Fig. 1.11). Microstriae are present at the Bowman's layer. Excessive interface microstriae and bright particles may lead to astigmatism and eventually poor outcome after LASIK. These microstriae can be imaged with confocal microscope even when the slit-lamp examination is unremarkable.
zoom view
Fig. 1.10: Measurement of flap thickness in LASIK
zoom view
Fig. 1.11: Bright high reflective particles at the flap-stroma interface
Diffuse lamellar keratitis (DLK) also known as “Sands of Sahara Syndrome”, is a non-infectious inflammation of the interface. The exact etiology is not known but it is assumed to be toxic or allergic in nature. In confocal scan DLK appears as diffuse and multiple infiltrates in the interface with no anterior or posterior extension.
Subepithelial nerve fibers are also affected by LASIK. No nerve is visible in the immediate postoperative period. However, the regenerating nerve fibers appear as a thin irregularly branching line when confocal scan is performed 5-7 days after surgery. The residual stromal thickness can also be measured using ‘Z' scan technique as described while evaluating the epithelial flap.10
 
CORNEAL GRAFTS
Confocal microscope is a useful tool to follow-up the corneal grafts and to diagnose the abnormal changes that may occur various times after surgery. It provides images at the cellular level and thus identifies any pathological changes even before it becomes clinically evident. It can also be used to assess the donor cornea.
Corneal graft survival is entirely dependent on optimum number of healthy endothelial cells. Endothelial cell loss occurs rapidly after corneal transplantation.17 Majority of cell loss takes place during the first two postoperative years.18 Several studies had suggested that endothelial cell loss is much higher after corneal grafting when the primary indications are bullous keratopathy or hereditary stromal dystrophy as compared to keratoconus and corneal leukomas.19,20 Another interesting fact is that endothelial cell loss is greater when corneal transplantation is performed on phakic eyes than on aphakics.21
Confocal microscopy scores over conventional specular microscopy while evaluating endothelial cell characteristics especially in eyes with stromal edema. Endothelial morphology in confocal scan has been described earlier. Immediate postoperatively, the endothelium looks normal and healthy. However, as time progresses, endothelial cell density decreases as evidenced by pleomorphism and polymegathism. Occasionally bright pre-endothelial deposits appear, the significance of which is not yet known (Fig. 1.12).
Re-innervation after grafting is another issue well addressed by confocal microscopy. First sign of innervation that starts few months after keratoplasty is visible at the periphery of the graft stroma. However, complete innervation may take many years to develop. Regenerated nerve fibers look similar to that of a normal cornea. Occasionally, they may take a tortuous and convoluted course depending on age (e.g. older patients) and primary indications of keratoplasty (e.g. bullous keratopathy, corneal dystrophies).
It is well known that allograft rejection is one of the most common causes of graft failure. Graft rejection can be classified as epithelial, subepithelial and endothelial rejection, of which the endothelial rejection is the worst. Confocal features of epithelial rejection are distorted basal epithelial cells with altered subepithelial reflectivity. Subepithelial rejection is identified by discrete opacities underneath the epithelial layer.22 Endothelial rejection, on the other hand, is characterized by coexistence of normal looking and degenerated endothelial cells, focal endothelial cell lesions and bright highly reflective microprecipitates23 (Fig. 1.13).
 
INTRACORNEAL DEPOSITS
Sources of intracorneal deposits can be exogenous or endogenous. They can involve various layers of cornea individually or in combination.
zoom view
Fig. 1.12: Pleomorphism, polymegathism and pre-endothelial deposits of a corneal graft
zoom view
Fig. 1.13: Co-existence of degenerated and normal endothelial cells in early endothelial allograft rejection
11
 
 
Exogenous Sources
  • Long-term use of contact lenses
  • Refractive surgery
  • Vitreoretinal surgery using silicone oil
  • Drugs: Amiodarone, Chloroquine
 
Endogenous Sources
  • Wilson's disease
  • Hyperlipidemia
  • Fabry's disease
  • Hemosiderosis
The clinical diagnosis is based on slit-lamp biomicroscopy and systemic features in selected cases.
 
Vortex Keratopathy
Also known as cornea verticillata is characterized by whorl like corneal epithelial deposits. It can be induced by various drugs, e.g. amiodarone (used for cardiac arrythmias) and anti-malarials (chloroquine, hydroxy-chloroquine). Clinically vortex keratopathy is manifested as golden brown opacities at the inferior corneal epithelium. On electron microscopy, they appear as intracytoplasmic lysosome like lamellar inclusion bodies located at the basal epithelial layer.24 Confocal microscopy adds newer dimensions to the existing knowledge. It demonstrates involvement of entire cornea, although vortex keratopathy is primarily a corneal epithelial pathology. The characteristic features are presence of highly reflective, bright intracellular deposits at the basal epithelial layer (Fig. 1.14). Overlying epithelium is usually normal. In advanced cases these microdeposits may extend to the stroma and eventually to the endothelium.25 Stromal keratocyte density is often reduced.
 
CONCLUSION
Ophthalmic investigations and instrumentations have come long way over the past decades. Confocal microscope is one of those wonderful innovations in recent times. It is becoming more popular everyday and indications are also expanding. Although further studies are necessary to explain the nature of the previously described microscopic corneal lesions, confocal microscopy is truly an exciting tool that can be useful for the clinical diagnosis, follow-up and analysis of the internal corneal lesions.
zoom view
Fig. 1.14: Intracellular deposits at basal epithelial layer in amiodarone toxicity
 
ACKNOWLEDGEMENT
Erlangga A Mangunkusumo
Vanathi Ganesh
REFERENCES
  1. Weigand W, Thaer AA, Kroll P, et al. Optical sectioning of the cornea with a new confocal in vivo slit-scanning videomicroscope. Ophthalmology 1995;102(4):485–92.
  1. Oliveira-Soto L, Efron N. Morphology of corneal nerves using confocal microscopy. Cornea 2001; 20 (4): 374–84.
  1. Tuft SJ, Coster DJ. The corneal endothelium. Eye 1990; 4: 389.
  1. Nucci P, Brancato R, Mets MB, et al. Normal endothelial cell density range in childhood. Arch Ophthalmol 1990; 108:247.
  1. Gass JD. The iron lines of the superficial cornea: Hudson-Stahle line, Stcker's line and Fleischer's ring. Arch Ophthalmol 1964;71:348.
  1. Maguire LJ, Bourne WM. Corneal topography in early keratoconus. Am J Ophthalmol 1989; 108: 107.
  1. Maguire LJ, Lowry J. Identifying progression of subclinical keratoconus by serial topography analysis. Am J Ophthalmol 1991; 112: 41.
  1. Somodi S, Hahnel C, Slowik C, et al. Confocal in vivo microscopy and confoal laser-scanning fluorescence microscopy in keratoconus. Ger J Ophthalmol 1996; 5(6):518–25.
  1. 12Werner LP, Werner L, Dighiero P, et al. Confocal microscopy in Bowman's and stromal corneal dystrophies. Ophthalmology 1999;106(9):1697–1704.
  1. Hirst LW, Waring GO. Clinical specular microscopy of posterior polymorphous endothelial dystrophy. Am J Ophthalmol 1983;95 (2): 143–55.
  1. Mashima Y, Hida T, Akiya S, et al. Specular microscopy of posterior polymorphous endothelial dystrophy. Ophthalmic Paediatr Genet 1986; 7(2): 101–07.
  1. Chiou AG, Kaufman SC, Beuerman RW, et al. Confocal microscopy of posterior polymorphous endothelial dystrophy. Ophthalmologica 1999;213(4):211–13.
  1. Chiou AG, Kaufman SC, Beuerman RW, et al. Confocal microscopy in cornea guttata and Fuch's endothelial dystrophy. Br J Ophthalmol 1999;83(2):185–89.
  1. Rosenblum P, Stark WJ, Maumenee IH, et al. Hereditary Fuch's dystrophy. Am J Ophthalmol 1980;90:455.
  1. Reviglio VE, Bossana EL, Luna JD, et al. Laser in situ keratomileusis for the correction of hyperopia from +0.50 to 11.50 diopters with Keracor 117C laser. J Refract Surg 2000;16(6):716–23.
  1. Durairaj VD, Balentine J, Kouyoumdjian G, et al. The predictability of corneal flap thickness and tissue laser ablation in laser in situ keratomileusis. Ophthalmology 2000; 107(12): 2140–43.
  1. Harper CL, Boulton ML, Marcyniuk B, et al. Endothelial viability of organ cultured corneas following penetrating Keratoplasty. Eye 1998;12(5):834–38.
  1. Vasara K, Setala K, Ruusuvaara P. Follow-up study of corneal endothelial cells, photographed in vivo before eneucleation and 20 years later in graft. Acta Ophthalmol Scand 1999;77(3):273–76.
  1. Obata H, Ishida K, Murao M, et al. Corneal endothelial cell damage in penetrating keratoplasty. Jpn J Ophthalmol 1991;35(4):411–16.
  1. Abott RL, Fine M, Guillet E. Long-term changes in corneal endothelium following penetrating Keratoplasty. A specular microscopic study. Ophthalmology 1983;90(6):676–85.
  1. Ing JJ, Ing HH, Nelson LR, et al. Ten-year postoperative results of penetrating Keratoplasty. Ophthalmology 1998;105(10):1855–65.
  1. Cohen RA, Chew SJ, Gebhardt BM, et al. Confocal microscopy of corneal graft rejection. Cornea 1995; 14(5): 467–72.
  1. Cho BJ, Gross SJ, Pfister DR, et al. In vivo confocal microscopic analysis of corneal allograft rejection in rabbits. Cornea 1998;17(4):417–22.
  1. Ghose M, McCulloch C. Amiodarone induced ultrastructural changes in human eye. Can J Ophthalmol. 1984;19:178–86.
  1. Ciancaglini M, Carpineto P, Zuppardi E, et al. In vivo confocal microscopy of patients with amiodarone induced keratopathy. Cornea 2001;20(4):368–73.