GO
GO5. Clinical examination of the cornea with the confocal microscope
5.1 The normal cornea
5.2 Investigation of the cornea in some selected conditions
5.3 The cornea with known pathologies
5.4 The post surgical cornea
5.5 Confocal microscopy in eye banking
5. Clinical Examination of the Cornea with the Confocal Microscope
Allocation of confocal images to the structures of the normal cornea (fig.3) :
The cornea consists of the following layers: superficial epithelial
cells, wing cells, basal epithelial cells, Bowman’s layer, stroma, Descemet’s
membrane, and endothelial cells, from the tear film on the anterior side
of the cornea to the posterior side of the cornea adjacent to the aqueous
humor (Maurice, 1984). The unique optical properties of the cornea
are consistent with its morphology (McCally and Farrell, 1990). The
normal cornea is avascular. At the anterior side is the corneal epithelium
(about 50 microns thick at the central corneal region), under the epithelium
is Bowman’s layer. It is 10-16 microns thick and is acellular except for
the nerves which perforate it. It separates the epithelium from the stroma.
The basal lamina of the epithelium is located on Bowman’s layer. The stromal
region is about 450 microns thick and contains large nerves, stromal
keratocytes, and orthogonal layers of collagen fibers. The morphology
of the stromal keratocytes has been described with electron microscopy
(Müller et al., 1995). Posterior to the stroma is Descemet’s
membrane, which is an acellular layer 15-20 microns in thickness.
The limiting layer on the posterior side of the cornea is a single layer
of corneal endothelial cells. Corneal thickness increases in the normal
cornea from the center to the periphery (Hitzenberger et al. 1982) and
with increasing age.
For detailed review of the normal and pathological corneal morphology
with standard light and electron microscopy, the reader is referred to
some excellent papers on this subject (Waring 1978, Waring and Rodriguez
1987, Versura et al.1985).
The standard histological sections give a sagittal view of the
corneal and, in this respect, are quite similar to the clinical biomicroscopy
with the slit lamp. As the corneal layers and pathological findings extend
laterally, the clinician is used to move the slitlamp laterally as well,
change the width and illumination angle of the slit and thus collect information
about changes in tissue reflectivity and their location. In contrast, the
confocal microscope collects optical sections parallel to the corneal
surface; all layers of the cornea can be imaged by changing the z-position
of the objective. Consequently, the confocal flying slit microscope is
an instrument which (i) allows an immediate tangential view of the corneal
layers and (ii) extends the diagnostic principle of biomicroscopy
into the microscopic range.
As the ophthalmologist is usually not trained to see and interpret tangential sections of the cornea, in the following we will give in parallel in vivo optical sections with the confocal microscope and paraffin embedded histological sections with the light microscope. The morphology as viewed on tangential sections will be briefly discussed by the various layers for the normal human cornea (fig.3).
Epithelium: From the tear film surface to Bowman’s layer the corneal
epithelium consists of the following layers: superficial epithelial cells,
wing cells, and basal epithelial cells. The superficial cells are
40 to 50 microns in length and about 4 microns in thickness. They can be
visualized very nicely in the confocal microscope (fig. 4),
showing a multicornered shape with high and low reflectivity, resembling
of the dark and bright cells of the epithelial surface seen in scanning
electron microscopy (SEM). In contrast to the SEM images, the confocal
image also displays a dark (condensed?) nucleus, sometimes with a small
bright reflex.
Three layers of intermediate or wing cells lie between the superficial cells and the basal epithelial cells (fig. 5). In the confocal microscope, the more forward layers are poorly visible with faintly reflecting nuclei. The bottom layer of the intermediate cells may display no reflectivity of the nucleus, however, the cell borders are somewhat more reflective, especially in hypoxic or edematous corneas.
The basal epithelial cells (fig.6) are about 20 microns in height and have a diameter of about 10 microns. Basal epithelial cells can slide in the plane of the basal epithelium and proliferate. They also differentiate into wing cells. During the process of epithelial renewal, basal epithelial cells detach from the basal lamina and differentiate into wing cells. Wing cells are progressively displaced anteriorly. They eventually become superficial epithelial cells, which are eventually desquamated from the corneal surface (Maurice, 1984). The kinetics of cell proliferation, cell differentiation, and cell migration from the basal epithelium to the superficial epithelium in the normal human cornea are not known and are under investigation (Beebe and Masters, 1996).
In tangential light - microscopic sections, the basal cells are visible
as small, near to round cells which are smaller than intermediate cells
and localize directly on Bowman's membrane. In the confocal slit microscope,
basal cells are nicely visible with dark non - reflective cell bodies with
reflective cell borders (fig. 6). In slightly oblique sections, these cells
are found to sit on the non - reflective Bowman's layer (fig.7 and 8).
Occasionally, highly reflective round structures in this layer (which have
the size of one cell) may indicate mitosis or cell division of an individual
cell. In corneas that have had photorefractive keratectomy (PRK),
the basal cells are localized directly on a layer of scar tissue and the
most anterior keratocyte layer (see fig. 55).
Bowman's layer stains uniformly in light microscopy and, except for
some perforating nerves, contains no further structures. In the confocal
microscope, this layer can be detected on the basis of its non-reflectivity
which contrasts the overlying subepithelial corneal nerves (fig. 7), anteriorly
the basal cells and posteriorly the first keratocyte layer (fig.8). The
keratocyte nuclei of this layer have a different morphology as compared
to their counterparts from the deeper stromal layers (Böhnke et al.
1997).
Stroma: The shape and spatial arrangement of keratocytes of the corneal
stroma is specific for their location (Müller et al 1995). The first
keratocyte layer is located directly in the stromal lamella posterior to
Bowman’s membrane. In the confocal microscopic image, these cells have
a characteristic morphology with multi-angulated nuclei (fig. 9), which
do not occur in any other corneal layer.
Anterior and mid - stroma: Behind the first keratocyte layers, the stroma of the healthy human cornea shows a rather homogeneous type of keratocyte nuclei regarding their arrangement and their morphology. The mid - stroma has somewhat elongated cell nuclei which sometimes show a clear indentation. In tangential sections for light microscopy, the collagen bundles stain distinctly, showing some artifacts like shrinking and curling up of dissected fibers (fig.10a). The confocal image shows oval to round nuclei with occasional indentations. The cell bodies, the keratocyte processes, and the stromal collagen and ground substance are not visible in healthy corneas (fig.10b). In between the layers of the stromal collagen, branching nerve fibers of variable size can be seen with a rather high reflectivitiy.
Posterior stroma: Corneal tangential sections for light microscopy do
not show very specific findings. The confocal microscope allows for a differentiated
morphology, showing a specific type of keratocyte nuclear shape with a
elongated appearance in the last or second to last keratocyte layer (fig.11).
Descemet's membrane, similar to Bowman's layer, may give a light reflection
in the confocal microscope and serves to contrast the pre-Descemet keratocytes
or the corneal endothelium. In tangential light microscopic sections, Descemet's
membrane stains uniformly positive (fig.12).
The corneal endothelium is not easily sectioned with routine light
microscopic techniques. However, it can be identified as a cellular layer
lining the posterior surface of descemet's membrane. Due to a considerable
portion of cutting artifact, subcellular details are not clearly distinguished
(fig.13a). In confocal microscopy, the endothelium is clearly visible in
normal corneas with all objectives from 25x to 50x magnification when the
focusing in the Z-axis is performed properly (fig.13b). The endothelial
intercellular borders sometimes appear as double contours delineating a
slender intercellular space (fig.14, fig.64f), especially in corneas with
a low cell count. Individual endothelial cells occasionally contain phagocytosed
pigment granules, which are stationary intracellular deposits of high reflectivity
(fig.15).
Corneal nerves: Large nerves are located in the anterior corneal stroma.
They form a nerve plexus (the subepithelial nerve plexus) immediately under
Bowman’s layer (Müller, et al., 1996, 1997). Nerves traverse Bowman’s
layer at numerous perforation points. At each perforation point, they branch
to form a basal epithelial nerve plexus in the region just anterior to
the basal lamina. From the basal epithelial nerve plexus additional
branches penetrate between the basal epithelial cells and wing cells.
The branching nerves of the corneal epithelium terminate in free nerve
endings which have been observed just below the superficial epithelial
cells. These fine nerve endings are only poorly visible in light microscopy.
Each subepithelial nerve plexus is a unique shape when observed with the real-time scanning slit confocal microscope. The large stromal nerve trunks (fig.16), subepithelial nerve plexuses (fig.7 and 8), and perforation points, are stationary in time over a period of weeks (Auran et al., 1995). Therefore the subepithelial nerve plexuses, as well as the perforation points of Bowman’s layer serve as fiduciary points for relocating the same regions of the cornea and the same fields of basal epithelial cells and wing cells (Masters and Thaer, 1994a, 1996). In contrast to the stationary subepithelial nerve plexuses, the nerve fiber bundles of the basal epithelial nerve plexus, slide centripetally at 10 to 20 microns per day (Auran et al., 1995). The scanning slit confocal microscope has been used to investigate corneal innervation in normal volunteers, donor corneas, and in eyes after enucleation for ocular tumors (Richter et al., 1997).
In summary, the real-time, scanning-slit, in vivo, confocal microscope
is very useful to image the normal in vivo human cornea. The use of high
numerical aperture water immersion microscope objectives (NA 1.0) and a
slit scanning system with high light throughput gives this microscope the
capability to uniquely image the layer of wing cells immediately adjacent
to the layer of basal epithelial cells, which confers this unique capability
to this microscope. The layer of epithelial cells has not been imaged by
any other in vivo, real-time, confocal microscope. This weakly reflecting
cell layer is a benchmark for comparing different types of in vivo, real-time
confocal microscopes (Masters and Thaer, 1994a, 1994b, 1995, 1996).
All of the confocal images shown in this sections are images
of single video frames which have been digitized and exposed to photographic
film.
5.2
Investigation of the clinically normal cornea in some selected conditions
Aging
The cornea is known to undergo changes with aging, which can be measured
as an increase in corneal thickness, a decrease in endothelial cell count,
and other less well defined changes in stromal composition. With the confocal
microscope, epithelial changes found in some elderly persons can be interpreted
as a reduction of epithelial cell regenerative capacity, resulting in elongated
epithelial cells, and possibly enlarged basal epithelial cells. The subepithelial
corneal nerve plexus looks normal, although morphometric analyses have
not been carried out.
The stromal keratocytes look normal in number and shape of the nucleus. In the stromal matrix, small highly reflective micro-dots are occasionally found probably representing remnants of past metabolic activities (fig. 17). The association of the increased thickness with changes in cellular or matrix composition will be answered in future 3-D reconstructions. The endothelial cell numbers slowly declines throughout life. Accordingly, the endothelial cells show an increase in cell size, as well as an increased incidence of secondary changes like guttae and intracellular deposits.
Contact lens wear
Contact lens wear induces short term and long term changes, which can
be attributed to mechanical and hypoxic damage. As with modern rigid gas
permeable lenses acute changes are hardly detectable we fitted a young
individual with a perfectly clear cornea a large rigid gas impermeable
lens to induce hypoxic changes. After 7 hours, the epithelium showed a
marked increase in intercellular space size and reflectivity (fig. 18).
The stroma in parallel developed an edema, which resulted in visibility
of the keratocyte cell bodies (which are normally not visible) in addition
to the nicely visible cell nuclei (fig. 19). This dissociation of reflectivity
between stroma and cytoplasm may indicate either increased cell reflectivity
due to subcellular changes like vacuolization, or simply be result of a
decreased stromal reflectivity due to the increased hydration. After the
contact lens trauma, the epithelium undergoes an increased desquamation
of epithelial surface cells with stabilization of the epithelium 1 hour
after discontinuing the contact lens wear (fig.20). The endothelium in
this setup did not develop "dark spots" as a sign of hypoxia (Holden et
al 1985a), as the cornea was perfectly healthy and the hypoxia time had
not been long enough. These acute changes are fully reversible when the
contact lens is discontinued.
In long term contact lens wear, a number of chronic changes of
all corneal layers have been previously described (Holden et al 1985b).
Investigating long term contact lens wearers with normal clinical and slit
lamp findings, we recently described a new type of corneal stromal deposit
(fig. 21), which may reflect a degenerative corneal change, possibly a
panstromal lipofuscinosis (Böhnke and Masters, 1997). Since then,
we have continued to study contact lens wearers, and we succeeded to replace
the semiquantitative dot gradings by a calculation of intrastromal microdots
frequencies, which may occur in the range of 0 to 150,000 per mm3 (Cadez
and Böhnke 1998). Although the visual function at this stage may not
be impaired, the clinical importance of these findings may be the estimation
of the individual dot grade and progression dynamics, which may render
a considerable number of individuals to stop wearing contact lenses or
reduce the total life contact lens wearing time before massive and confluent
deposits, which may then correspond to the clinical picture of deep stromal
degeneration in contact lens wearers (Kilp 1982) is irreversibly encountered.
Using the confocal microscope, a comparison of different contact lens types
and of individual contact lens tolerance can be performed with significantly
refined sensitivity.
5.3 The Cornea with Known Pathologies
Aside from histological specimens, the corneal pathology can be visualized with the limited magnification of the slit lamp in the living eye. For selected cellular layers like the epithelial surface and the corneal endothelium, the confocal microscope has furnished us with a means to study cellular and subcellular details of all corneal layers, which otherwise cannot be visualized with any other method in the living eye. From the large number of corneal dystrophies, degenerations and inflammatory changes, some selected examples will be given below sorted by the layer of their predominant occurrence.
Epithelium:
The epithelial physiology is characterized by a state of constant renewal,
which requires continuous centripetal cell movement as described with the
X-Y-Z theory (Thoft and Friend 1983).
A disturbance of this sliding process may lead to a decreased adhesion
of the epithelial sheet to the basal lamina and Bowman’s layer, resulting
in spontaneous or easily provocated corneal erosions. As an example of
a disorganized basal epithelial layer, a clinical case is briefly
presented:
Map dot fingerprint dystrophy:
A 69 year old male patient was admitted for cataract surgery.
Preoperative examination revealed no pathological finding except for a
dense nuclear cataract. The phacoemulsification and intraocular lens insertion
went uneventful. At the end of surgery, upon wiping some debris from the
epithelium with a surgical sponge, the central corneal epithelium was observed
to be loose and developed a central erosion during this manipulation. The
epithelial defect closed within 3 days, showing intraepithelial fine map
and dot changes in the corneal center (fig. 22). Confocal microscopy showed
intraepithelial disorganization of the epithelial layers, as well as very
few cyst - like formations of basal cells bulging up into more superficial
layers (fig. 23). The central intraepithelial changes slowly subsided over
a course of 6 weeks, and the uncorrected visual acuity reached 20/20. The
patient has been ever well thereafter.
The group of epithelial and anterior stromal dystrophies, which are usually diagnosed on the basis of their (low magnification) slit lamp picture, have been well studied with light and electron microscopy (Waring et al. 1978). The confocal microscope allows for an in vivo observation of the associated changes far beyond the level of the slit lamp findings, which can lead to a much earlier diagnosis of the disease and recurrences after treatment (fig. 24). With confocal microscopy, pathological changes and possibly their response to treatment could be quantified with a much higher resolution.
Filiform keratitis: In dry eyes and associated with rheumatic disease, epithelial extrusions from the superficial layer of the cornea develop, giving the corneal surface the appearance of being stippled with fine thread-like ("filiform") structures. With the confocal microscope it can be demonstrated in vivo that these are indeed composed of epithelial surface cells, which in response to the pathologic state of the corneal surface do form a tunnel - shaped "filum" of cells extruding from the corneal surface (fig. 25).
In the confocal image, a oblique section of one of these structures shows the borders of many cells organized in a pipe-like fashion.
In various conditions, the epithelium may accumulate substance of either exogenous or endogenous origin. The cardiac drug Cordarone induces an epithelial deposit, which can be observed with confocal microscopy. A degeneration affecting visual acuity is the subepithelial deposit of calcium-phosphate named calcific band keratopathy (fig. 26), which with confocal microscopy can be visualized to be located in Bowman's layer just below the basal epithelial cells (fig. 27). This disease is associated with chronic intraocular inflammation and other conditions and can temporarily be relieved by a chemical removal of the calcium salts from Bowman's layer.
Epithelial infection can occur with parasites, bacteria, Chlamydia, and herpes or adenovirus. Human herpes simplex virus may infect all corneal layers, leading to pronounced morphological changes from the cytopathic virus as well as from the resulting immune response (fig. 28 a - d). Except for larger infectious parasitic organisms like Acanthamoeba, a direct visualization of the agents involved is not possible. In acute viral infection like herpes simplex virus, epithelial necrosis is less specific at the cellular level than the morphological picture of the entire dendritic lesion. Secondary to ocular surface infections with Adenovirus, the subepithelial immune reaction called nummular keratitis may be visualized as a patchy infiltrate of (mononuclear) cells, sometimes associated with a thin layer of increased stromal reflectivity indicating discrete scar formation.
Stroma:
Stromal dystrophies: The slit lamp and microscopical finding of the
epithelial and anterior stromal dystrophies have been previously described
(Waring et al., 1978). Confocal microscopy does not further resolve the
lesions which are dense and nicely visible with the slit lamp like in granular
corneal dystrophy. Beyond the stromal deposits that are visible with the
slit lamp (fig.29) , the confocal microscope detects smaller deposits of
high reflectivity in stromal layers which appear not involved in slit lamp
examination (fig. 30). A further description of the vast amount of in vivo
microscopical findings in the various forms of stromal degenerations is
beyond the scope of this paper. However, it may be stated that in the named
stromal dystrophies confocal microscopy does not add too much information
to the diagnosis. As these patients (except for family studies) are usually
diagnosed at a stage which is rather late from the microscopical point
of view, the role of confocal microscopy for these diseases still has to
be established.
Keratoconus, which involves progressive weakening, thinning and
ectasia of the corneal stroma, involves a typical slit lamp finding consisting
of fine striae in the deep corneal layers. With the resolution of confocal
microscopy, the substructure of these striae, which itself consists of
fine parallel lines, may be observed in the uninvolved fellow cornea of
an involved eye in cases that are considered to have "unilateral"
disease previously (fig. 31).
Stromal degeneration: In long - standing stromal inflammation or other
toxic challenges, the corneal stroma may react with degenerative changes.
These may be discrete, giving the cornea a normal appearance in slit lamp
microscopy, an at the same time be easily detectable with confocal microscopy
similar to the condition described above in long term contact lens wear.
Presumably in a vast number of challenges a nonspecific corneal stromal
degeneration may occur, which only can be observed with confocal microscopy.
In chronic (herpetic) keratitis, these changes are centered around the
diseased area. In other cases, they may nonspecifically involve the entire
stroma, showing enhanced keratocyte processes, increased reflectivity of
scar tissue, and highly reflective stromal (lipofuscein) deposits. A chronic
inflammatory stimulus to the cornea (plus long term treatment with eye
drops containing whatsoever drug and preservatives like Benzalkonium) may
slightly increase the corneal light scatter and reflectivity in the slit
lamp examination. With confocal microscopy, these near to normal corneas
may contain numerous highly reflective structures which resemble those
found after photorefractive keratectomy (fig. 32). As these findings have
not been studied in detail yet, and as histological specimen currently
are not available, the nature of these changes may only be clarified in
the future. Currently the highly reflective portion of these changes is
believed to be corneal lipofuscin deposits, which thus may be much more
common than previously thought (Hidayat et al. 1992) and present an universal
pathway of slow corneal stromal degeneration.
Stromal deposits can occur from a variety of internal and external causes,
including metabolic and immunologic diseases as well as various types of
ocular trauma. The plant Dieffenbachia ejects oxalate crystals when their
leaves are broken off, which may penetrate and deposit in all corneal layers
(fig. 33) (Seet et al. 1995). There they slowly resolve over two months,
a process which can monitored with the confocal microscope (Chiou et al.
1997).
Stromal infection may be caused by bacterial, fungal, viral and parasitic
agents.
Bacterial infection of the stroma may be overwhelming like in bacterial
corneal ulceration or of slowly progressive type like in crystalline keratopathy
(Reiss et al., 1986, Townshend et al. 1989, Sutphin et al. 1997). In the
former we have found confocal microscopy to show predominately the inflammatory
cellular infiltrate, while in the latter the "crystalline" changes are
further resolved into their substructures (fig. 34). While in established
crystalline keratopathy confocal microscopy does not really add further
information, the observation of sub-crystalline structures in corneas not
suspected to have bacterial infection may accelerate the decision for antibiotic
treatment. In fully established bacterial corneal ulcers, a cellular infiltrate
differing from the keratocyte morphology can be observed. However, confocal
microscopy does not really add to the etiologic diagnosis. The situation
may be different in Borrelia keratitis, which has been investigated with
confocal microscopy (Linna et al. 1996).
Fungal infection may be visualized with the confocal microscope,
given the inflammatory infiltrate is not masking the causative agent like
in the full clinical picture of infiltrates and satellite lesions. Again,
at very early stages or in a later phase under treatment, the visualization
of fungal structures may be of clinical value (fig. 35).
In viral infection like in herpetic or meta herpetic stromal disease, the immune response to the virus is far more destructive to the tissue than the virus itself. In long term clinical follow up of patients with (treated) herpetic keratitis, hypercellularity of the stroma, increased formation of scar tissue with secondary lipofuscinosis, and an altered morphology of the stromal nerves may be observed. As the stromal scar usually has a high reflectivity, other tissue details with low to moderate reflectivity are frequently obscured (fig.28).
The infection of the cornea with acanthamoeba has a typical clinical picture which does not require confocal microscopy for diagnosis. In non - typical or pretreated cases however, the observation of the organism in the corneal stroma can be of clinical usefulness. We have found it very hard to make a decisive observation of the trophozoite in the corneal stroma, as this may be variable in shape and look very similar to inflammatory or damaged cells. In contrast, the acanthamoeba cysts display a very typical morphology with a round to oval shape and a high reflectivity. We have used confocal microscopy to study the clinical course in Acanthamoeba infection under topical treatment, determine the infectious status of the stroma before grafting, and follow the patient for a recurrence after penetrating keratoplasty. With this approach, we feel that the indication for an anti-Acanthamoeba treatment can be decided on the basis of microscopic findings, limiting the total time and number of topical drug applications.
Endothelium
The corneal endothelium has been studied with specular microscopy,
which can be regarded as a specific application of confocal microscopy
(Maurice, 1974; Laing et al 1975, Laing et al. 1976, Laing et al 1979),
since more than 20 years. In contrast to the early instruments requiring
a contact of the lens with the corneal surface, non - contact video based
systems (Hartmann, 1984) with automatic analysis of morphological cell
parameters are commercially available.
Most of the clinical conditions affecting the corneal endothelium have
been studied using this well established method. The specular microscope
today is a routine equipment for the assessment of the endothelial cell
status, especially in patients with low endothelial cell counts and borderline
compensation of the cornea. The adjacent cell layers like keratocytes anterior
or corneal precipitates posterior to the endothelium can be studied only
to a limited extent using indirect illumination techniques (Hartmann 1985).
With confocal microscopy, the endothelial morphology can be studied
as with specular microscopy (fig. 36). In contrast to the latter, confocal
microscopy reveals associated changes of the adjacent tissue layers, as
the imaging of these layers is performed by direct and not by an indirect
reflection of light. Due to this, details like endothelial precipitates
and the replacement of the endothelium with fibrous tissue or with cells
of a different origin can be studied and followed over time (fig. 37).
The keratocyte layers just before Descemet's membrane, which have a morphology
different from those of other stromal layers, can be studied in detail
with great resolution, allowing for the recognition of very discrete tissue
changes considered to be only very subtle with standard biomicroscopy (fig.
38).
Corneal surgery has evolved during the last 90 years from simple suturing
of a corneal graft into a recipient bed of a diseased cornea (Zirm ,1906)
to a large number of highly sophisticated surgical procedures including
manipulations of the refractive power of the otherwise healthy cornea (Waring,
1985). We have used flying slit confocal microscopy to study the corneal
morphology in many of these conditions, and we will give some examples
of these findings in the following.
Corneal grafting for visual rehabilitation is performed to replace
diseased corneal tissue that has poor optical properties because of changes
in tissue shape or transparency. Corneal grafting can performed as a penetrating
full thickness or a lamellar corneal graft.
In penetrating keratoplasty, the corneal healing takes longer than a year, until the grafted button has been integrated into the corneal architecture. Morphological aspects of the corneal tissue after penetrating keratoplasty include endothelial morphology and cell count (fig. 39), viability of stromal keratocytes (fig. 40), stability of the epithelium, and the regrowth of corneal nerves into the grafted tissue (fig. 41). Corneal graft rejection is considered a major problem of penetrating keratoplasty with rejection rates up to 20% in the first year (Yamagami et al., 1994). In our five - year follow up routine after penetrating keratoplasty, we have found a rejection rate below 3 percent in the first year. Whilst others have reported on findings with confocal morphology of this condition (Cohen et al., 1995), we have only rarely been able to investigate such a patient in the acute phase. In cases of previous reaction, which may be controlled by topical steroid administration, the study of endothelial precipitates may help to differentiate between inactive remnants of previous disease in lesions containing pigment and debris and endothelial precipitates composed of mononuclear cells and adjacent endothelial cell necrosis. In previous rejection or otherwise troubled grafts, the endothelium can be visualized even in slightly hazy corneas, allowing for a study of the endothelial cell status. (fig. 42). Other changes associated with corneal wound healing after surgery like fibrocellular tissue reactions to the corneal suture can be studied as well (fig. 43).
In lamellar corneal grafting procedures, which replace only a predefined portion of the anterior stroma, the specific advantages (faster healing and unimpaired ocular integrity) and disadvantages (debris and scar formation in the graft interface) can be studied. The recipient corneal nerves grow into the grafted tissue much faster than in penetrating keratoplasty and can be visualized with the confocal microscope. In cases with chronic herpetic disease and significant damage to the corneal innervation, the level of corneal sensitivity can be correlated with the amount of subepithelial nerve fibers (fig. 44). The most significant optical phenomenon of lamellar corneal grafting is the fate of the lamella and the formation of scar tissue in the interface, which can harbor debris, scar tissue, and degenerative stromal cells (fig. 45). Additionally, the optical significance of the scar tissue can be estimated by a confocal Z-scan, which records the intensity of reflected light along the sagittal motorized scan in the optical center of the cornea (fig. 46).
A specific variant of lamellar corneal grafting named epikeratoplasty
has been advocated for the correction of an abnormal curvature of the cornea
like in keratoconus or even for the correction of refractive errors of
the eye (Werblin, 1983). In this procedure, corneal donor tissue is molded
to the desired geometrical shape and grafted onto the deepithelialized
recipient cornea. The graft is then enclavated and sutured in a circular
keratotomy of the recipient cornea, where through scar formation the graft
is fixed to the recipient cornea. The epithelium covers the graft within
a few days after surgery, an appreciable innervation occurs only at a much
later stage. Although these grafts look well in slit lamp biomicroscopy,
the optical properties of these corneas and the visual performance of the
patients are not as good as might have been expected.
With the confocal microscope, we have been able to study some of these
patients. The subepithelial nerve plexus is found to be present in an altered
morphology (fig. 47). The transplanted lenticule contains numerous keratocytes
(fig. 48), which appear quite normal in shape and number. The optical problem
in these grafts can be demonstrated by inspecting the interface between
the posterior graft surface and Bowman's layer of the recipient. Here we
find a condensation and obvious irregularities in the optical properties
of the grafted tissue (fig. 49), which were found in all patients seen
by us. With the Z-scanning function of the confocal microscope, an increased
reflectivity in this layer can clearly be demonstrated (fig. 50).
Corneal refractive surgery involves a number of different methods,
which alter the refractive power of the cornea by either
- changing the central anterior corneal curvature
- by manipulating the peripheral cornea stromal curvature (like in
radial keratotomy, "RK")
- by removing a defined spherical volume from the anterior surface
of the corneal stroma (like in photorefractive keratectomy, "PRK")
- by removing a defined volume from the intermediate corneal stroma
(laser in situ keratomileusis, "LASIK")
- by implanting circumferential structures in the peripheral stroma
of the cornea (intracorneal ring, "ICR")
- changing the curvature and refractive index of the central corneal
stroma by implanting biocompatible optic lenses into the central cornea
(intracorneal lens, "ICL"). As these procedures are performed to "only"
alleviate the patient from wearing spectacles or contact lenses in a normal
and healthy cornea, the immediate and long term side effects are under
thorough observation by some members of the ophthalmologic community.
Confocal microscopy has been used to study the process of wound healing in RK in rabbits (Jester et al., 1992), but not in humans. As this type of surgery due to its late effect of hyperopic shift (Waring et al.,1994) is not performed frequently any more, confocal microscopic images of recent human cases are not available to us.
Quite different is the situation with PRK, which is frequently performed
and can be followed with confocal microscopy. In this procedure, the preoperative
status of the cornea (including changes from long term contact lens wear),
the early postoperative phase of wound healing and the long term postoperative
course are being studied by us.
As this procedure removes Bowman’s layer in the center of the cornea
and exposes the cornea to ultraviolet radiation, late effects on the cornea
may be expected.
In the first days after PRK, we have been able to document the process of epithelial wound healing (fig. 51) and of the keratocyte reaction to the ablation process (fig. 52). In this early phase, the stromal hydration is increased due to the uptake of water via the anterior stromal wound. This effects a decreased stromal reflectivity, making the keratocyte bodies and even the long keratocyte processes of the anterior stroma visible in confocal microscopy (fig. 53). Weeks to months after the surgery, the epithelium continues with wound healing activity in an effort to establish firm contact to the corneal stroma. In some patients, an increased number of highly reflective structures, probably representing dividing epithelial cells, can be found even months to years later (fig. 54).
The superficial corneal nerves, which have been ablated along with the anterior stroma, return weeks to months later and slowly resume a near to normal morphology (fig. 55). In patients with clinically visible haze a reticulate layer of scar tissue is visible (fig. 56) months after surgery. Even in patients without clinically visible haze, a very thin superficial scar located at the epithelial - stromal junction is visible (fig. 57).
A new finding after PRK has been the observation of elongated structures in the corneal stroma, which have a rod and needle like morphology. These changes are, in lateral extension, limited to the site of the PRK, being more abundant in the anterior stromal layers than in the deeper ones (fig. 58). Although the intensity of these changes varies in different patients, they can be observed in all patients. Furthermore, they seem to be stable and non reversible up to 4 years after the photoablation. As the patients with these findings usually have excellent visual acuity, the clinical significance of these findings currently is unknown. We speculate, however, that these findings represent a degenerative postinflammatory response after PRK, which may ultimately affect corneal physiological function.
In LASIK procedures, late stromal changes of the PRK type plus a midstromal scar can be regularly observed with the confocal microscope. In postoperative complications like epithelial cell implantation into the anterior / posterior stromal ablation interface, the further course of these cells can be followed (fig. 59). We assume that, using this technique, an early decision can be made in case of infiltrates of the interface, which may have therapeutic consequences.
A rather new technique, the intrastromal corneal ring (ICR), is used to correct moderate myopia. In this method, pieces of PMMA are inserted into the corneal stroma, where they are tolerated quite well (Grabner, 1997). As this procedure is claimed not to alter the optical center of the cornea and also to be fully reversible, it may offer significant benefits over the current methods. We have studied patients with the ICR located in the corneal stroma and compared central and peripheral corneal confocal microscopic findings. The corneal center appears not altered throughout all corneal layers (fig. 60), whereas in the peripheral cornea the stroma undergoes some changes. The stroma adjacent to the anterior and posterior surface of the intracorneal ring appears quite normal with only minute amounts of scar tissue formed during the first postoperative year (fig. 61), whereas lateral (peripheral and central) to the ring intrastromal deposits are observed. These are already visible with the slit lamp, and in confocal microscopy they appear as extremely reflective structures probably consisting of scar tissue, calcified degenerative cells and other debris (fig. 62). Surrounding these lesions, additional scar tissue can be seen.
We recently had the opportunity to study a patient, in whom the ICR was removed due to an undercorrection. In this patient, a near to normal central cornea and the described degenerative stromal changes were observed in the peripheral cornea. The corneal endothelium under the former ring implantation site was normal (fig. 63).
In conclusion, these illustrations show that confocal microscopy is
a powerful tool to investigate corneal refractive procedures in patients.
Doing so, an improved level of awareness for discrete and otherwise non
detectable changes of the cornea can be achieved.
Corneal tissue can be preserved with various methods, enabling for serological testing of the donor, ensuring endothelial cell viability, and organizing the steps to perform corneal grafting. A number of methods is available corneal preservation, each with specific advantages and disadvantages (Böhnke et al. 1984, Wilson and Bourne 1989, Böhnke 1991). All methods have in common that they are supposed to support the viability of the corneal endothelium. Recently, the attention of the eye banks has been focused on maintaining viability, integrity and differentiation of the other corneal layers as well (Böhnke 1991). The endothelial integrity is in most American eye banks evaluated with the specular eye bank microscope. Some time ago, we introduced the routine use of phase contrast microscopy in eye banking, which we considered to be an ideal tool for detecting degenerative changes in all corneal layers (Böhnke 1983, Böhnke et al 1984, Böhnke 1991). A perfectly viable corneal tissue, however, will only give a poor contrast in organ cultured corneoscleral discs. For this latter reason, long term storage methods may require a better means to study the morphology of the freshly excised as well as the long term organ cultured donor cornea.
We routinely investigate the donor tissue after enucleation of the donor eye. In this phase, the cornea shows some degenerative changes like epithelial desquamation, an altered morphology of the corneal nerves, stromal edema (fig.64 a - f), and some (reversible) endothelial changes like single cell necrosis and general status of the morphology. These investigations can be performed on the intact donor eye with the clinical confocal microscope.
To study the excised corneoscleral disc, the use of water - immersion objectives requires an immersion of the objective into the culture medium. Consequently, the microscope has to assume a vertical position and appropriate measures to maintain sterility of the culture must be established. We have also tried to use non - contact objectives with a longer working distance, however, we found the resolution not sufficient for stromal and epithelial cell layers.
With the immersion optics, the organ cultured shows a multilayered
epithelium with a normal layer of basal cells (fig. 65). The stroma appears
similar to the post mortem status before the cornea is dehydrated in dextrane
- containing culture medium (Böhnke et al. 1984). A variety of conditions
like prolonged incubation in dextrane medium, as it has been advocated
in the early days of organ culture (Sperling 1978), induces a vacuolic
degeneration of the keratocytes (fig. 66) as it has been previously been
shown in electron microscopy (Pels and Schuchard, 1984). The endothelium
can also be visualized with the confocal microscope, however, for that
purpose the established routine method of specular microscopy works as
well.
We have also used confocal microscopy to evaluate other methods
like corneal cryopreservation. A detailed discussion of these morphological
findings would be beyond the scope of this chapter.
GO