Optical Coherence Tomography
MICHAEL IP, OLIVIA C. LIAO and JAY S. DUKER
Table Of Contents
PRINCIPLES OF OPERATION|
CLINICOPATHOLOGIC CORRELATION: NORMAL EYES
|Introduced in 1991, optical coherence tomography (OCT) is a new technique
for high-resolution cross-sectional imaging of various ocular structures. With
an axial resolution of 10 microns, it provides the best resolution
of retinal architecture of any imaging technique currently available.1–4 In contrast to OCT, conventional ultrasound has a spatial resolution of 150 microns. The
recently introduced ultrasound biomicroscopy has a
spatial resolution of 20 microns but is limited to imaging of the anterior
segment by its penetration depth of 4 mm. Scanning laser ophthalmoscopy
and scanning laser tomography are two other recently developed
retinal imaging modalities, but neither one approximates the axial resolution
and sectioning capability of OCT. Traditional imaging systems
such as computed tomography and magnetic resonance imaging have limited
utility in ophthalmic imaging because the spatial resolution of these
systems are on the order of several hundred microns.|
The advantages of OCT go beyond its extremely high resolution. Information may be stored on computer disks (Bernoulli disks) and easily accessed for future retrieval. Additionally, the use of a computer-controlled fixation light in the eye being studied provides a reproducible means of acquiring OCT images of the same retinal cross-section on follow-up studies. Thus, comparisons of the same retinal area can be made over subsequent visits. This feature makes OCT particularly useful in the diagnosis and treatment of macular disease. Quantitative measurements of nerve fiber layer thickness can be obtained with OCT, making this a useful test in the management and diagnosis of glaucoma. Other techniques such as stereoscopic videographic imaging, nerve fiber layer photography, or measuring the polarization rotation of reflected light have been used in the assessment of glaucomatous changes. However, these modalities do not provide direct measurements of nerve fiber layer thickness.
The limitations of OCT include the inability to obtain high-quality images through media opacities such as dense cataract or vitreous hemorrhage. The use of OCT is also limited to cooperative patients who are able to maintain fixation for the full acquisition time of 2.5 seconds per section. Computer image-processing techniques can eliminate eye movements from fluctuations in the intraocular pressure produced by variations in pulse pressure, microsaccades, and mild tremor but cannot eliminate image degradation from patients with excessive movement.
|PRINCIPLES OF OPERATION|
|The operation of OCT is based on the principle of low coherence interferometry.5–7 In this technique, the distances and sizes of different structures in
the eye are determined by measuring the “echo” time it takes
for light to be backscattered from different structures at various
axial distances. This is analogous to A-mode ultrasound, whereby the axial
length of the eye is measured using sound rather than light.|
The OCT hardware can be assembled as an adjunct to a standard slit-lamp or incorporated into a specially designed fundus camera and allows for structures to be imaged without direct contact with the ocular surface. A fiberoptically integrated light source consisting of a superluminescent diode projects a partially coherent, continuous, near-infrared light at a wavelength of 810 nm. The power of the light source is approximately 200 uW, well below acceptable levels set by the American National Standards Institute for retinal radiation. For slit-lamp use, the light is focused with a standard mounted + 78 diopter condensing lens and computer-operated scanning mirrors onto the area of interest. The optical beam from the light source is directed onto a partially reflective mirror (optical beam splitter). This mirror splits the light into two beams; one beam is directed into the patient's eye and is reflected from intraocular structures at various distances. This reflected beam consists of multiple echoes and provides information about the distance and thickness of various intraocular tissues. The second beam is reflected from a reference mirror at a known spatial location. This beam travels back to the beam splitter, where it combines with the optical beam reflected from the patient's eye. When the two light beams coincide, they produce a phenomenon known as interference, which is measured by a photodetector.
The beam of light reflected from the reference mirror coincides with the beam of light reflected from a given structure in a patient's eye only if both pulses arrive at the same time. This occurs when the distance that light travels to and from the reference mirror equals the distance that light travels when it is reflected from a given intraocular structure. The position of the reference mirror can be adjusted so that the time delay of the reference light beam matches the time delay of light echoes from various intraocular structures. Thus, the interferometer can precisely measure the echo delay of intraocular structures.8–11
By measuring the echo delay of various intraocular structures, a translated axial image similar to A-mode ultrasonography is produced. When successive axial measurements at different transverse points are combined, a tomographic or cross-sectional image of tissue is obtained. OCT images are displayed in false-color to enhance differentiation of structures. Bright colors (red to white) correspond to tissues with high relative optical reflectivity, whereas darker colors (blue to black) correspond to areas of minimal or no optical reflectivity. Cursors may be placed on the image to give exact measurements to the nearest micron. The image produced is enlarged by a factor of two in the vertical direction for better image readability; this may in some cases distort the image (e.g., exaggerate the depth of the nerve cup).
The instrumentation for OCT operates analogously to a video fundus camera. The transverse position of the OCT beam is controlled by a mechanical scanning mirror. When the beam is scanned, it produces a pattern on the retina visible to both the patient and the operator. Thus, the exact location of the topographic image on the fundus can be determined. The field of view at lowest magnification is 30°. The live fundus image may be viewed either directly through the ocular or via a video camera.
|CLINICOPATHOLOGIC CORRELATION: NORMAL EYES|
|Initial studies using postmortem eyes showed good correlation between OCT
images and histologic sections. Subsequent in vivo analysis demonstrated the ability of OCT to image the substructure of
the retina. Figure 1 shows an OCT image of a normal fovea and optic disc taken along the papillomacular
axis. Anatomic features such as the fovea, optic disc, and
retinal profile are evident. The vitreoretinal interface is noted by
the contrast between the nonreflective vitreous and the reflective surface
of the retina. The foveal center demonstrates normal retinal thinning
and has a characteristic pit to its contour. The optic disc demonstrates
normal nerve head contour and cupping. The retinal nerve fiber
layer, inner plexiform layer, outer plexiform layer, photoreceptor layer, choroid, and
sclera are all well delineated.|
The posterior aspect of the neurosensory retina is bounded in the OCT images by a highly reflective red layer about 70 microns thick that represents the choriocapillaris and retinal pigment epithelium (RPE) layer. In vivo choriocapillaris thickness by OCT is greater than that of histologic sections because of postmortem blood depletion and artifacts in tissue processing. The high contrast between the choriocapillaris/RPE layer and the neural retina in OCT images provides a useful boundary for use in measurements of retinal thickness. Retinal blood vessels are evident in OCT images by their shadowing of deeper retinal structures. The region just anterior to the choriocapillaris/RPE layer is typically weakly reflective and corresponds to the photoreceptor layer. The highly reflective red layer at the inner margin of the retina corresponds to the nerve fiber layer. In the OCT image taken along the papillomacular axis (see Fig. 1), the thickness of the nerve fiber layer increases from the fovea to the optic disc. Circular tomographs taken around the disc demonstrate modulations in thickness consistent with the superotemporal and inferotemporal bundling of the nerve fibers. Because retinal detail is so exquisitely imaged by OCT, this imaging modality can be applied to a large number of clinical entities.
Central Serous Chorioretinopathy
Central serous chorioretinopathy (CSCR) is characterized by detachment(s) of the neurosensory retina caused by one or more focal leaks at the level of the RPE. When small or shallow, these serous detachments may be difficult to detect clinically. OCT images of such areas demonstrate elevation of neurosensory retina by the presence of subretinal fluid.12 The well-defined contrast in optical reflectivity between the nonreflective serous fluid and the more highly reflective posterior boundary of the neurosensory retina allows OCT images to be highly sensitive to even small neurosensory detachments. Indeed, OCT images may show the presence of neurosensory detachments not detectable by clinical examination. The ability of OCT to image the same retinal area on subsequent visits allows for the longitudinal monitoring of the clinical course of the serous detachment in this disease (Figs. 2 and 3). OCT is particularly useful when thisdisease presents in older patients. The presence of drusen and pigmentary changes in these patients may lead to the erroneous conclusion that a choroidal neovascular complex is the cause of the neurosensory detachment. OCT may be able to provide additional diagnostic information in these patients by excluding the existence of a choroidal neovascular membrane or abnormalities in the choriocapillaris/RPE layer.
Serous Macular Detachment Secondary to Optic Nerve Pit
Optical coherence tomography images of this clinical entity clearly demonstrate the relation between the optic nerve pit and serous macular detachment (Fig. 4). These images support the concept that fluid from the optic pit directly enters the neurosensory retina and not the subretinal space.13
Optical coherence tomography offers an objective test for serial, quantitative evaluations of retinal thickness.14 Increases in retinal thickness are usually caused by accumulation of intraretinal fluid, widening the distance between the well-delineated anterior and posterior boundaries of the neurosensory retina. Because of the high axial resolution of OCT, retinal thickness can be measured to within 10 microns and followed serially.
Macular edema is the leading cause of decreased vision in patients with diabetic retinopathy. Although “clinically significant macular edema” continues to be a clinical diagnosis, OCT can provide the clinician with additional useful information. OCT images can quantitatively measure the amount of retinal thickening present; the amount of thickening has been shown to correlate with visual acuity.15 Additionally, OCT can be used to follow the clinical response to focal laser treatment for clinically significant macular edema, and successful resolution of macular edema after laser treatment may be observed (Figs. 5 and 6).
Cystoid macular edema is characterized by fluid accumulation in intraretinal cystic spaces. This may be from a variety of causes, including diabetes, venous occlusion, cataract surgery, and inflammatory disease. OCT images demonstrate cystic areas of decreased reflectivity within the neurosensory retina consistent with the known histopathology of this entity. Again, the ability to quantitate the extent of thickening is useful not only in the diagnosis of this disease but also in assessing the response to treatment (topical medications, periocular steroids, vitrectomy).
Age-Related Macular Degeneration
In the imaging of macular pathology from age-related macular degeneration, OCT has clinical utility in several areas. As described above, OCT can help distinguish central serous retinopathy from exudative age-related macular degeneration in older patients. Retinal edema from choroidal neovascularization, particularly when it involves the fovea, can have a profound effect on visual function. Because of its ability to measure retinal thickness accurately, OCT is useful in assessing the response to laser treatment for choroidal neovascularization.
OCT can also differentiate among serous, hemorrhagic, and fibrovascular RPE detachments.16,17 Serous RPE detachments present as focal elevations of the reflective RPE band over an optically clear space (Fig. 7). The angle of the edge of the detachment is typically acute, probably because of tight adherence of RPE cells to Bruch's membrane at the edge of the detachment. Hemorrhagic RPE detachments are distinguished by a moderately reflective layer directly beneath the detached RPE, corresponding to the sub-RPE blood (Fig. 8). Fibrovascular RPE detachments, in contrast, demonstrate moderate reflectivity throughout the entire sub-RPE space (Fig. 9). This is probably the result of the lower scattering coefficient of the fibrovascular proliferation as compared to blood.
OCT may serve as an adjunct to fluorescein and indocyanine green angiography in the classification of choroidal neovascular membranes. Choroidal neovascular membranes typically have one of three presentations on OCT: fibrovascular RPE detachment, well-defined, or poorly defined. The OCT presentations typically correspond to the angiographic classifications. Well-defined choroidal neovascular membranes on OCT appear as a fusiform thickening of the reflective band corresponding to the RPE/choriocapillaris. The thickening extends anteriorly on the OCT image, creating an elevation in the normally smooth contour of the RPE band. Poorly defined choroidal neovascular membranes on OCT images present as ill-defined, diffuse areas of choroidal reflectivity that blend into the normal contour of the RPE band; a distinct boundary cannot be ascertained.
Perhaps of greatest interest is the ability of OCT to localize a choroidal neovascular membrane to either the subretinal or sub-RPE space. Thus, OCT may have utility in defining surgically approachable membranes in age-related macular degeneration as well as from other causes of choroidal neovascularization. Figure 10 shows a choroidal neovascular membrane that has penetrated the RPE to lie primarily in the subretinal space.
Optical coherence tomography facilitates the ability to stage macular holes according to the Gass classification: stage I, foveal detachment; stage II, small, full-thickness hole; stage III, fully developed, full-thickness hole; and stage IV, fully developed, full-thickness hole with vitreous detachment.18,19 In a stage I hole, the OCT image shows loss of the foveal depression, a cystic space in the fovea, and vitreous fibrils inserting obliquely onto the fovea. The micron scale resolution of OCT has greatly facilitated our understanding of stage I holes, demonstrating that it is oblique vitreous traction rather than tangential traction on the fovea (as traditionally believed) that is responsible for the pathologic changes in this entity. In a stage II hole, an anvil- or flask-shaped full-thickness retinal defect is present. In an eccentric stage II hole, an anterior flap of attached retina is present. In contrast, OCT images through stage III holes demonstrate an anvil- or flask-shaped defect without a retinal flap at the mouth of the flask. In stage I through III holes, the posterior hyaloid face is typically visualized inserting into the foveal or perifoveal region, supporting the vitreomacular traction theory of macular hole development. In a stage IV macular hole, a full-thickness retinal defect is noted in addition to complete separation of the posterior hyaloid face from the retina. Lamellar macular holes, in which there is partial loss of inner retinal tissue, are also clearly imaged by OCT. OCT enables the clinician to monitor patients longitudinally and document hole progression and can assist in the timing of surgical intervention. Successful hole closure after surgery may also be documented with OCT (Figs. 11 and 12). Finally, periodic OCT examination of fellow eyes may be performed to identify impending macular holes, because patients with idiopathic macular hole may be at risk for bilateral disease.
Epiretinal membranes from trauma, inflammatory disease, proliferative disease, intraocular surgery, or idiopathic causes may be clearly imaged by OCT.20 Epiretinal membranes appear as a band of moderate to high reflectivity anterior to or contiguous with the retinal surface. In some cases epiretinal membranes barely detectable clinically are well detailed on OCT images. In addition to direct visualization of the epiretinal membrane, secondary retinal thickening from traction can be observed.
OCT imaging of epiretinal membranes is useful in several respects. As with macular holes, longitudinal observation can help the clinician with the timing of surgery (cases where serial OCT images demonstrate progressive retinal thickening) and the assessment of surgical outcomes. Macular pseudohole from an epiretinal membrane can sometimes be difficult to differentiate from a true macular hole on ophthalmoscopic examination alone. OCT is useful in this respect: macular pseudohole appears as a thickened band of moderate reflectivity on the retinal surface, with a steepened foveal pit contour (Fig. 13). Full-thickness retinal tissue is present at the base of the apparent hole formed by the epiretinal tissue.
Retinal Detachment and Retinoschisis
Full-thickness retinal detachment can usually be distinguished from degenerative retinoschisis on the basis of clinical features alone. However, in some cases this is difficult; various ancillary tests such as laser photocoagulation, visual field evaluation, and B-scan ultrasonography can be helpful but are not always definitive. OCT is an objective and reliable method to distinguish the two entities.17 In retinoschisis, OCT images show splitting of the neurosensory retina consistent with the known histopathology of a separation at the outer plexiform layer (Fig. 14). Retinal detachment presents as a separation of full-thickness neural retina from the underlying RPE band (Fig. 15). Although lesions anterior to the equator cannot be imaged by OCT, most lesions that are posterior to the equator, or that have a component posterior to the equator, can be effectively imaged.
The diagnosis and management of glaucoma remains a difficult clinical problem. Intraocular pressure measurements do not always adequately predict the extent of glaucomatous change. Optic nerve head and gonioscopic evaluation by slit-lamp biomicroscopy is subjective. Visual field loss and optic nerve head cupping are late clinical findings, detected only after up to 50% of retinal nerve fibers have been lost.
OCT, because of its high resolution, is able to detect nerve fiber layer thinning before the onset of visual changes.21 Nerve fiber layer thickness, as measured by OCT, has been shown to correspond to visual function. As expected from the histology of the peripapillary retina, the nerve fiber layer is thickest in the inferior and superior quadrants. The nerve fiber layer has been demonstrated to be significantly thinned in areas corresponding to visual field loss.
Typically, the scans are performed radially around the optic nerve for 360° using two radii of curvature (2.25 and 3.37 mm), and the nerve fiber layer thickness is plotted schematically (Fig. 16). Normal nerve fiber layer thickness is a mean of 148.6 microns for superior nerve fibers, 143.5 microns for inferior nerve fibers, 66.9 microns for temporal nerve fibers, and 117.2 for nasal nerve fibers. The direct measurement of the nerve fiber layer thickness by OCT is an objective assessment of glaucomatous progression. OCT shows promise in the early diagnosis of glaucoma before visual field defects, optic nerve head changes, and ophthalmoscopically visible nerve fiber layer loss are evident.
Optical coherence tomography of the cornea has not been widely investigated. Currently, information concerning corneal thickness and profile can be obtained with pachymetry and topography. However, there is potential for OCT to aid in the evaluation of corneal thickness by specific corneal layer.
Optical coherence tomography can demonstrate consistent changes in reflectivity in experimentally induced cataract. However, this technique is not used in the clinical evaluation of patients with cataract.
|The high axial resolution and cross-sectional capability of OCT make it
a useful technique in the evaluation of various retinal disorders and
glaucoma. The technology and clinical application of OCT continue to
evolve. Future innovations include the use of imaging wavelengths further
in the infrared, which should provide increased penetration of the
probe light through pigment and hemorrhage, potentially allowing imaging
of choroidal neovascular membranes beneath thick intraretinal or subretinal
blood.15 A broad-bandwidth titanium:sapphire laser has been used in preliminary
studies to create cross-sectional images with a resolution of 3 microns. Additionally, decreasing the acquisition time of 2.5 seconds is under
investigation. This would facilitate imaging of poorly cooperative
Topographic maps using multiple cross-sectional tomograms have been developed. The use of these maps is being investigated as a method to delineate the borders of choroidal neovascular membranes. Finally, three-dimensional images of the retinal surface are being developed. These images may prove to have higher sensitivity than single axial tomograms or topographic reconstructions in the detection of retinal thickening or choroidal neovascularization.