Optical coherence tomography

Introduction to optical coherence tomography. This imaging technique is similar to ultrasound in that it uses light waves instead of sound waves. Using time-delayed information contained in the light waves reflected from different depths, an OCT system can reconstruct a depth-profile of the structure being analyzed (eg, macula, peripapillary retinal nerve fiber layer, optic nerve) according to a qualitative and quantitative approach.

Qualitative scans. Most OCT devices allow for custom scan capture of single B-scans that can be positioned in the regions of interest with specified lengths and tilt. For example, a single line scan through the foveal pit, with averaging, is often used to determine inner versus outer retinal conditions.

Quantitative scans, retina. The macula thickness can be quantified in micrometers in various areas.

Quantitative scans, optic nerve. The optic nerve images have limited automated quantification that includes cup-to-disc ratio and parameters of neural rim tissue. Many clinicians use instrument-based calipers to determine additional parameters, including the height and width of the neural rim tissue and elevation of the Bruch’s membrane opening that are monitored in conditions resulting in disc edema. Similarly, the rim tissue at the optic nerve head can be monitored for small changes through measures of minimum rim width. However, circumpapillary retinal nerve fiber analysis is an indirect way to quantitatively measure the optic disc swelling and optic atrophy.

Quantitative scans, circumpapillary retinal nerve fiber layer analysis. The optic nerve can appear normal, swollen, or atrophic on OCT. The circumpapillary retinal nerve fiber layer represents the axons of the optic nerve that are adjacent to the optic disc before they enter the nerve (approximately 1.2 million of them in a normal optic nerve). These can indirectly provide useful information about the state of the optic nerve. If the optic nerve is swollen, the retinal nerve fiber layer analysis will show thickening, and if the optic nerve is atrophic, the retinal nerve fiber layer analysis will show thinning. It may take time for the optic nerve fibers to reflect the damage that occurs acutely; therefore, normal-appearing optic nerve on OCT in acute phase does not entirely exclude pathologies. The OCT protocol that measures the thickness of the optic nerve axons can be used as an indirect measurement and determination of optic nerve swelling or atrophy. The mainstay for the retinal ganglion cell associated nerve fiber layer analysis is a 12-degree diameter circular scan centered on the optic nerve head that is shown to have good repeatability. This scan shows a nominal circumference of 10.9 mm in the emmetropic eye and samples most of the retinal ganglion cell axons entering the optic nerve. To quantify the retinal nerve fiber layer, an instrument-based segmentation algorithm identifies the inner limiting membrane and the junction between the nerve fiber layer and ganglion cell layer. The retinal ganglion cells are the neurons, which receive the input from the deeper retina (photoreceptors), and they deliver the visual information to the brain. The retinal ganglion cell axons form the retinal nerve fiber layer surrounding the optic disc. These axons migrate toward the optic nerve and bend nearly 90 degrees to enter the nerve. The axons and the supporting cells and tissues (glia, oligodendrocytes, myelin, and blood vessels) form the optic nerve. Most of the optic nerve fibers synapse in the lateral geniculate body.

Pathologies that affect the optic nerve before the lateral geniculate body will result in atrophy of the retinal nerve fiber layer. Usually, pathologies that affect the visual pathway beyond the lateral geniculate body do not result in retinal nerve fiber changes, unless present at birth or very longstanding. Most commonly, a normal average retinal nerve fiber layer is associated with a normal complement of healthy optic nerve fibers. A normal average retinal nerve fiber layer number is approximately 100, and as one ages this number reduces. However, a normal examination is not equal to a normal average nerve fiber, as there are several factors that determine if the retinal nerve fiber is normal for a particular person (history, prior OCTs, prior optic nerve edema, funduscopic examination findings, congenital variations, right and left comparisons, and reliability of the study with signal strength. Although retinal nerve fiber layer thickness measures are fairly robust, several factors, including scan quality, scan centration, and refractive status, need to be considered when assessing. For the most reliable measures, signal strength should be within the range suggested by the instrument manufacturer, which is typically in the upper 70th percentile.

Quantitative scans, macula volume analysis. The central 20 degrees of the macula region contains approximately 30% to 40% of the retinal ganglion cells, with a peak density of 35,000 retinal ganglion cells/mm2 at a 1 mm eccentricity. Hence, this region is of great interest for disorders that result in loss of retinal ganglion cells. Currently, most spectral domain optical coherence tomography systems provide measures for total retinal thickness, either ganglion cell complex (inner plexiform layer to inner limiting membrane) or ganglion cell and inner plexiform layer thickness measures. The retinal ganglion cell measurement has been helpful in optic neuropathies that affect the thickness of the perifoveal retina by reducing the retinal ganglion cell layer. So there is a way to tell if the perifoveal (retinal) thinning is related to a retinal or an optic nerve cause.

There are two main categories of OCT instrumentation: time-domain OCT and spectral-domain OCT; the latter is much faster and has markedly better resolution and, therefore, now dominates the market (55).

A general outlook of OCT in neuroophthalmology. The structural effects of optic neuropathies have traditionally been identified through a detailed fundus examination and include optic disc edema, optic disc pallor, and retinal nerve fiber layer defects. The retinal nerve fiber layer lacks myelin because it contains the axons of the retinal ganglion cells, so changes in its structure represent visible effects of axonal loss caused by retrograde degeneration. Typically, these changes arise from an afferent visual pathway lesion involving the optic nerve, chiasm, or tracts. In the setting of transsynaptic “neuroaxonal” degeneration, postgeniculate lesions in the afferent visual pathway can also cause optic nerve pallor and corresponding retinal nerve fiber layer defects, which can all be readily captured by OCT, allowing evaluation for axonal loss that may not be obvious by funduscopy.

The indirect quantification of axonal loss (retinal nerve fiber layer thinning) and neuronal damage (ganglion cell inner plexiform layer thinning) is helpful in seeing the effects of optic nerve, chiasm, and optic tract injury. Furthermore, OCT-measured thinning of the macular ganglion cell layer has been found to have a strong relationship with visual loss across a spectrum of optic neuropathies, including glaucoma, optic neuritis, ischemic optic neuropathy, hereditary optic neuropathy, toxic optic neuropathy, optic nerve glioma, and idiopathic intracranial hypertension.

Role in multiple sclerosis. It is important to note that any optic neuropathy can lead to loss of retinal nerve fiber layer, so it is important to conduct an eye examination to exclude all causes of retinal nerve fiber layer loss, including glaucoma.

Reductions in peripapillary retinal nerve fiber layer thickness have been reported in different multiple sclerosis-related subtypes from clinically isolated syndromes to secondary progressive multiple sclerosis. Several studies have also confirmed atrophy of the ganglion cell-inner plexiform layer in multiple sclerosis. Associations between OCT-derived retinal thickness and cerebral atrophy quantified by MRI have also been reported, leading to the suggestion of using OCT as a biomarker of multiple sclerosis disease progression in the clinic (52).

Parisi and colleagues were the first to use OCT in multiple sclerosis when they assessed patients with multiple sclerosis with a history of optic neuritis associated with good visual recovery (48). They noticed a 46% reduction in the average retinal nerve fiber layer of affected eyes and 28% reduction in the unaffected eyes compared to healthy eyes (evidence of subclinical neurodegeneration in patients with multiple sclerosis) (33). According to a study performed by Costello and Burton, the decrease in the peripapillary retinal nerve fiber layer thickness by approximately 10 to 40 um is maximal at 3 to 6 months after the acute episode, and stabilization is observed at 7 to 12 months (29; 07). To place this data into perspective, retinal nerve fiber layer is around 110 to 120 µm thick by age 15 years, and normal healthy individuals lose only 0.27% per year in retinal thickness; thus, by age 40 years, you may have around 99 µm and by age 80 around 89 µm (30).

A recent retrospective study provided evidence for differential rates of ganglion cell layer atrophy unrelated to patients having optic neuritis according to disease-modifying therapy usage. In addition, a recent cohort study identified a cutoff of 88 µm for the retinal nerve fiber layer to predict twice the risk of disability worsening at 2 years and 4 times higher risk for the patients below this cutoff (31). This literature implies a potential role for OCT in the clinical monitoring of multiple sclerosis in the future. Pisa and colleagues extended this work by determining OCT associations with no evidence of disease activity in multiple sclerosis (50). Conventionally applied to active multiple sclerosis amenable to disease-modifying therapies, it relies upon 3 criteria being concurrently met to define “no evidence of disease activity”: no clinical relapse, no disability worsening, and no new MRI activity (no new or enlarging T2 lesions and no gadolinium enhancing lesions). Patients with multiple sclerosis with no evidence of disease activity at follow-up had mean binocular retinal nerve fiber layer thinning of -0.93 um +- 1.35 SD, whereas patients with evidence of disease activity demonstrated greater retinal nerve fiber layer thinning at -2.83 um +- 2 SD (t test p< 0.001). Greater retinal nerve fiber layer loss over 2 years was also associated with worsening EDSS scores (51).

It has been suggested that there is an injury threshold within the retinal nerve fiber layer of 75 µm; therefore, thinning of retinal nerve fiber layer below this level has been associated with impaired visual function. Importantly, OCT studies have shown subnormal values for retinal nerve fiber layer thickness in multiple sclerosis eyes even in the absence of prior optic neuritis. This structural marker can potentially be used to detect clinically silent aspects of disease activity and to predict disability worsening in specific patients. In a study conducted by Martinez-Lapiscina and colleagues, the patients in the lowest peripapillary retinal nerve fiber layer thickness category at baseline had increased risk of disability worsening compared with those in the higher category during the 2nd and 3rd years of the disease (HR of 1.65); this increased by 4 times after 4 to 5 years of follow-up (36).

Role in nonarteritic ischemic optic neuropathy. OCT can be used to monitor the disease course long-term with the assessment of edema and optic atrophy, but nonarteritic anterior ischemic optic neuropathy remains a clinical diagnosis. Some recent studies have demonstrated the use of OCT in the assessment, diagnosis, and management of nonarteritic anterior ischemic optic neuropathy. Acute findings include increased peripapillary retinal thickness associated with optic disc edema, and some cases have subretinal fluid on spectral domain; more chronically, the macular ganglion cell inner plexiform layer thinning is associated with visual field defect. Ganglion cell layer and internal plexiform layer thinning has shown to be better than retinal nerve fiber thinning in indicating early structural loss in nonarteritic anterior ischemic optic neuropathy.

The C/D area ratio and the vertical C/D ratio measured by OCT were significantly higher in control patients than in nonarteritic anterior ischemic optic neuropathy fellow eyes. This finding agrees with previous studies and supports the hypothesis that a crowded nerve head is involved in the pathogenesis of nonarteritic anterior ischemic optic neuropathy because transient hypoperfusion or nonperfusion leads to nerve fiber edema: in a crowded nerve, this edema may compromise the microvasculature of the optic nerve head, leading to more ischemia and finally to nonarteritic anterior ischemic optic neuropathy.

Role in neuromyelitis optica spectrum disorders. The main targets of interest are the ganglion cell bodies associated with axons in the retinal fiber layers. Due to poor differentiability in OCT imaging, the ganglion cell layer is usually measured in combination with the adjacent inner plexiform layer. The retinal nerve fiber layer is regularly affected by swelling for up to 3 months after acute optic neuritis, and the retinal nerve fiber layer does not show loss until swelling resolves; however, the ganglion cell layer serves especially well as a stable parameter to quantify retinal neuro-axonal damage.

Retrograde and anterograde transsynaptic degeneration following optic neuritis potentially causes subsequent alterations in the retina, optic nerve, and anatomically connected tracts. Consequently, a combination of lesion length in the optic nerve measured by MRI and retinal findings by OCT offers the unique possibility of predicting visual outcome after optic neuritis. Recently, a study with a mixed AQP4-ab seropositive and seronegative anti-MOG neuromyelitis optica spectrum disease cohort that had cortical atrophy showed a correlation between retinal nerve fiber layer and pericalcarine cortex thickness, further supporting the concept of transsynaptic degeneration being responsible for some detectable brain atrophy (44).

OCT changes have shown effective in distinguishing multiple sclerosis-associated optic neuritis from optic nerve involvement in the neuromyelitis optica spectrum disease. Multiple studies have shown that eyes in patients with neuromyelitis optica spectrum disease have more pronounced thinning of the peripapillary retinal nerve fiber layer and macular ganglion cell layer than eye in patients with multiple sclerosis optic neuritis eyes. It has also been effective in differentiating neuromyelitis optica spectrum disease subtypes; for example, Martinez-Lapiscina and colleagues demonstrated after comparing patients with anti-MOG and anti-AQP4 that patients with the latter had significant thinning of the retinal nerve fiber layer compared to patients with multiple sclerosis-related optic neuritis, whereas patients with anti-MOG were no different than their multiple sclerosis counterparts (37). However, Pache and colleagues found no difference in the OCT measures when comparing patients with anti-MOG- and anti-AQP4-mediated optic neuritis (47). Although the latter study demonstrated substantial thinning of the retinal nerve fiber layer and ganglion cell layer, the damage in the anti-MOG group was driven predominantly by more frequent optic neuritis attacks, whereas the anti-AQP4 group had more severe attacks leading to the structural OCT changes observed.

Role in papilledema. Many clinicians are confident in their ability to accurately diagnose and grade the severity of papilledema, and most use the accepted standard of the Frisen grading scale. However, even in the original report by Frisen, there was interobserver variability in grading of photographs on repeat testing, whether the grading was done by a medical student, resident, or expert. We do not yet have Frisen grades based on OCT, but this should be possible (60).

A distinctive feature of high-grade papilledema is the presence of peripapillary intraretinal folds described by Paton in 1911 (49). These are occasionally visible on the raster images of the spectral domain optical coherence tomography which were defined by Sibony and colleagues and has allowed us to sub-classify the peripapillary intraretinal folds in 3 types (56). Peripapillary wrinkles are defined on OCT as tightly spaced folds in the nerve fiber bundle layer of the optic nerve head or the juxtapapillary retina or both. They are almost always located temporally at the retinopapillary inflection but can occasionally be located at the vertical poles of the disc and rarely on the nasal side of the optic disc. Retinal folds were most commonly located in the papillomacular bundle, generally horizontal or radial in pattern and occasionally associated with spiral peripapillary wrinkles. The patterns of choroidal folds were quite variable in some cases consisting of coarse, widely spaced, full-thickness folds and in others irregular undulations. These are less common than retinal folds but were associated with higher levels of intracranial pressure (56).

In differentiating papilledema and some other optic neuropathies and neuroretinitis with other optic nerve and retinal-related pathologies, the most obvious sign that can be discerned with OCT is a neurosensory retinal detachment from peripapillary fluid between the retinal pigment epithelium and photoreceptors that tracks into the fovea. On OCT b-scans, the fluid appears dark with low reflectivity. A decline of visual acuity in the absence of macular fluid is usually the most obvious sign of progressive optic neuropathy in this setting. Because retinal nerve fiber layer is thickened in papilledema, a reduction in its thickness, assessed by OCT, may be difficult to interpret and could represent either a reduction in disc edema due to improvement or irreversible axon loss. As an alternative, assessment of ganglion cell loss by OCT in the setting of papilledema may be suitable for early detection of neuron loss in order to identify patients in need of more aggressive treatment because optic disc edema and axon swelling does not appear to directly affect the retinal ganglion cell layer thickness, allowing it to be an effective tool for the early diagnosis of progressive optic nerve failure with retinal nerve fiber layer loss in severe papilledema. The working assumption is that thinning of the ganglion cell layer is a sign of progressive optic neuropathy in the presence of papilledema and may suggest intervention if this is not an artifact (11).

Differentiation between papilledema and pseudopapilledema. The ability to differentiate papilledema due to raised intracranial pressure from other forms of optic disc edema or from pseudopapilledema can be challenging, particularly when the degree of edema is not severe. When calcified optic disc drusen are located superficially, the diagnosis is relatively easy and can be made with careful ophthalmoscopic observation. When calcified drusen are deep and buried under the surface, clinical observation may be equivocal, and the use of autofluorescence have been useful. OCT has been used to differentiate papilledema from pseudopapilledema. Sometimes calcified drusen and their shadows, visualized on OCT, are not easy to distinguish from large, superficial blood vessels. Noncalcified drusen are not usually visualized as they are presumed not to exhibit a significant difference in reflectivity from surrounding disc tissue. Often a patient with pseudopapilledema (with or without calcified drusen) may show visual field loss. In these eyes, the retinal nerve fiber layer may appear thickened in some areas, presumably due to axoplasmic flow stasis, and thin in other areas, corresponding to locations of retinal nerve fiber layer loss associated with visual field loss. However, using a technique called enhanced depth imaging one can see optic disc drusen, which is an OCT approach to differentiate papilledema from pseudopapilledema.

Short wavelength fundus autofluorescence involves stimulating the fundus with short wavelength light and assessing the light emitted by fundus fluorophores at longer wavelengths. This technique is regarded as a valuable diagnostic tool to provide insight into the metabolic status of the retinal pigmentary epithelium by revealing the distribution and accumulation of lipofuscin (the chief fundus fluorophore in the healthy eye). There are similar instruments that allow photography of short wavelength fundus autofluorescence; more commonly used are the confocal scanning laser ophthalmoscope (eg, Spectralis® Heidelberg Retina Angiograph) and fundus cameras (eg, Topcon TRC-NW8F, Canon CR-2 PLUS, Optos 200Tx). Different ocular structures display varying degrees of short wavelength fundus autofluorescence and can be termed as being autofluorescent, which represents higher intensities than the background level (the homogenous level of autofluorescence generated in the posterior pole fundus of a healthy eye) and appears brighter; and vice versa hypofluorescent areas show lower intensities (than the background level) and appear darker. In a normal eye, the optic nerve head is hypoautofluorescent, as there is inherently no retinal pigmentary epithelium layer.

Role in compressive optic neuropathies. Compression of the optic nerve or chiasm usually leads to optic atrophy over time, which can be visualized on OCT. In this setting, OCT serves 2 roles: (1) compression-induced retinal nerve fiber layer loss and (2) prediction of visual outcomes after decompression. Subtle optic nerve damage from compressive optic neuropathy can be appreciated on OCT before it can be seen on ophthalmoscopy. This is especially true for the macular ganglion cell layer analysis, which is very sensitive for detecting compressive optic neuropathy. With compression of the optic nerve, macular ganglion cell thinning precedes peripapillary retinal nerve fiber layer loss, and, occasionally, ganglion cell thinning can be observed even before appreciable changes are detectable with a standard automated perimetry. Although some anterior compressive lesions, such as optic nerve sheath meningioma, will cause visible disc edema, other compressive lesions will sometimes cause mild axoplasmic flow stasis, leading to subclinical retinal nerve fiber layer thickening that can be seen on OCT. In early compression, this can lead to a thickened retinal nerve fiber layer compared to age-matched controls. After further optic nerve damage and consequent atrophy, this will often lead to a “normal” retinal nerve fiber layer thickness, giving a false sense of normality; however visual field would show worsening with significant macular ganglion cell thinning. Further compression then leads to both retinal nerve fiber layer and ganglion cell thinning.

Meningioma. OCT examination is important in predicting the visual prognosis: when a visual acuity decrease occurs in a patient with minimal retinal nerve fiber layer loss, the prognosis is favorable. Loo and colleagues evaluated 14 eyes of 12 patients with pretreatment and posttreatment OCT, along with clinical examination (35). After treatment, the group with normal retinal nerve fiber layer experienced improvement in visual acuity, color vision, and visual field compared with the group with altered retinal nerve fiber layer (28).

Chiasmal lesions. The pattern of macular ganglion cell loss is helpful in detecting some compressive optic neuropathies, especially if they involve the chiasm or optic tract. The macular ganglion cell layer analysis is exquisitely sensitive in detecting chiasmal injury, which causes binasal thinning because the crossing nasal fibers are damaged with compression of the chiasm. Diffuse thinning of the macular ganglion cell layer in one eye with nasal thinning in the contralateral eye will be seen in a junctional scotoma from anterior compression of the optic chiasm. Although binasal ganglion cell loss is seen in chiasmal injury, homonymous macular ganglion cell loss typically indicates injury to the optic tract or lateral geniculate nucleus from various etiologies (eg, tumor, stroke, or demyelination). The OCT measurements of retinal thickness can detect neuronal loss from onset (12). Danesh-Meyer evaluated 40 patients with chiasmatic lesions, with OCT and visual field, pre- and post-decompression treatment (09; 26). Patients with thin retinal nerve fiber layer did not demonstrate significant improvement in visual acuity and visual field as compared to those with normal retinal nerve fiber layer.

Comparison with other imaging tests. OCT may be directly compared with alternative techniques in terms of several different criteria: resolution, imaging depth, acquisition time, complexity, and sample intrusiveness. With regard to the first two, OCT’s imaging depth is typically limited to a few millimeters (less than ultrasound, magnetic resonance imaging, or x-ray computed tomography), but its resolution is greater. This comparison is reversed with respect to confocal microscopy. Like ultrasound, the acquisition time of OCT is short enough to support tomographic imaging at video rates, making it much more tolerant to subject motion than either CT or MRI. It does not require physical contact with the sample and may be used in air-filled hollow organs (unlike ultrasound). OCT uses nonionizing radiation at biologically safe levels, allowing for long exposure times, and its level of complexity is closer to ultrasound than to CT or MRI, allowing for the realization of low-cost portable scanners.

B- and A-scan ophthalmic ultrasonography

Introduction. Ophthalmic ultrasound is an invaluable tool that provides quick, noninvasive and cost-effective evaluations of the eye and the orbit. It not only allows clinicians to view structures not visible with routine ophthalmic equipment or neuroimaging techniques but also provides diagnostic information in various ophthalmic conditions such as papilledema or pseudopapilledema, diplopia, thyroid eye disease, and scleritis as well as retinal detachment or presence and type of an ocular malignancy or foreign body with over 150 entities amenable to analysis.

Two ultrasound modalities.

B-scan mode. B-scan mode, or time-brightness (the more reflective the tissue, the brighter the area) is a 2-dimensional sector image used to characterize ocular structures and determine the status of the globe and orbit. These gray-scale image variations are generated by changes in the tissue medium that reflects sound. Excellent images can be obtained by examining though the eyelids and markedly increases patient cooperation.

Standardized A-scan. Standardized A-scan or time-amplitude (the more reflective the tissue, the higher the A-scan spike) A-scans of the eye and orbit produce 1-dimensional waveforms as they course through tissue. Data from standardized A-scan probes are directly comparable; as the specific gain setting may vary from machine-to-machine, a tissue model is used to establish the center of the S-shaped amplification curve. By reading the fall-off of the sound waves, it is possible to identify its unique internal structure; for example, a melanoma may have linear fall-off in contrast to a hemangioma with alternating high-and-low spikes as the sound waves course through blood vessels. The nonfocused 10 MHz probe is placed on the anesthetized sclera. Although clinically important, note the standardized A-scan is a skill that can take months to master and is difficult to perform reliably, as it is an exam with abstract data requiring several replications. The relationship between A- and B-scan echoes is that a tall echo on A-scan will correspond to a bright echo on B-scan, both indicating high structural density.

Traditional ophthalmic B-scan ultrasound frequencies for globe and orbit imaging are 10 to12 MHz. B-scans are systematically performed to examine all clock hours with both transverse (across several clock hours) and longitudinal (along a single meridian) scans. It is difficult for the clinician to directly visualize tears or other pathologies in the distal retina as the retina thins in the periphery where most tears occur, but longitudinal scans can readily discern these areas and capture images to be included in the chart. Having the eye move back and forth (kinetic scan) can reveal the difference between a retinal tear (moves slowly) and a faster moving posterior vitreous detachment. A combination of longitudinal and transverse scans will allow an estimation of putative or elevated lesion’s dimensions.

V-shaped acoustic shadows are present in both B- and standardized A-scans when examining the optic nerve and extraocular muscles and appear as empty space. Although these structures certainly are not hollow, the perception arises from the fact that the optic nerves fibers and muscles striations are parallel to each other, so the sound waves are not distorted.

It is routine to first examine the retinal surface at the macula, where detailed vision occurs, with B-scan as the foveola affects quality of life. Macula scans do not require extreme gaze and are best obtained with the patient in either primary gaze or a slightly temporal gaze. This permits better centering of the macula within the image for improved resolution.

When considering the optic nerve and orbit, ultrasound has certain advantages over CT and MRI when assessing the optic nerve for calcification (as seen with optic nerve drusen). Ultrasound is cost-effective and more sensitive than other imaging modalities (CT or autofluorescence) in detecting buried optic nerve drusen. Although detectable on CT scans, drusen can be missed between cuts of the scan. Irrespective, patients who have optic nerve drusen and elevated optic nerves may also have superimposed papilledema, which may be difficult to ascertain, so in cases where an underlying neurologic problem may be a cause of raised intracranial pressure, the presence of drusen may not negate the need for further imaging.

It is important to reduce the B-scan gain to observe the characteristically bright reflection and to avoid missing a drusen surrounded by highly reflective orbital tissue. Having an idea of the pathology under investigation, gain is adjusted accordingly. High gain, for instance, should be used for recent vitreous hemorrhage and lower gain for drusen or melanomas.

The 30-degree test developed by Karl Ossoinig and colleagues is performed with the standardized A-scan (61). This technique is used to determine if a larger than normal optic nerve (greater than 4.2 mm measured by ultrasound) is due to increased subarachnoid fluid, as occurs most commonly from raised intracranial pressure, or from a solid or infiltrated optic nerve, as might occur with a meningioma or glioma amongst other etiologies. The optic nerve is first measured when the patient looks straight ahead (primary gaze) then abducting temporally 30 degrees. In a normal optic nerve, the optic nerve measures about the same at ortho/primary and temporal gazes. If the optic nerve is widened from increased CSF, there will be a reduction in its width when the optic nerve and its sheath are stretched at lateral gaze, presumably due to a redistribution of the increased amount of CSF over a greater area. In the presence of a large amount of fluid, a 20% or greater reduction in optic nerve may be observed, suggesting the presence of fluid.

In terms of imaging the orbital component of the optic nerve, the 30-degree test using standardized A-scan is a technically challenging procedure requiring a very well-trained individual. This test allows one to determine if the optic nerve is thickened due to fluid (which often is due to raised intracranial pressure) or if it is thickened due to infiltrative process.

A prospective masked observational study in 65 patients requiring invasive intracranial pressure monitoring reported that optic nerves larger than 4.8 mm showed 96% sensitivity and 94% specificity for detection of intracranial pressure greater than 20 mmHg (53). A patient reporting eye pain and presenting with a red and inflamed conjunctiva may be experiencing scleritis, which is readily detectible via longitudinal B-scans. If Tenon’s space is visible (dark shadow behind the sclera), this might be suggestive of scleritis. Use the calipers to measure from the posterior aspect of Tenon space to the surface of the retina. Measurements exceeding 2.0 mm is indicative of scleritis.

Extraocular muscle enlargement may be associated with diplopia or thyroid eye disease. Initially, a B-scan of the orbit is performed to rule out mechanical movement of the eye by a tumor such that each eye may send a differing signal to the brain, resulting in double vision. If normal, a standardized A-scan is performed to measure the width of the inferior rectus muscle (first to be affected by thyroid eye disease and also the most difficult to measure). Generally, start with the medial rectus then lateral rectus muscle to establish if the muscles exceed the threshold of 5.4 mm as used in our clinic. Each muscle measurement must be replicated several times. When measuring muscles with standardized-A scans look for an initial “dip,” which indicates the insertion of the muscle, then follow it posteriorly to its widest aperture. Measurements between 5.4 mm and 5.8 mm suggest thyroid eye disease whereas 5.9 mm and beyond are more indicative of true muscle enlargement.

Summary. B- and standardized A-scans have a multitude of clinical applications in an ophthalmic or optometric practice as well as in an emergency room, with well over 150 ocular and orbital entities that can be visualized quickly and accurately. Of the multitude of pathologies that can be identified, fewer than 10 occur most commonly. If there are no mentors to guide you, consider a course at national meeting. Regardless, you can learn and glean valuable information from the book, Ultrasound of the Eye and Orbit 2/E (14).

With patience and determination, these valuable skills are readily acquirable, and plug-in B-scans with adequate resolution are available for several thousand dollars.

Fluorescein angiography

Fluorescein dye ranges from yellow to orange-red in color. Its peak excitation is 465 to 490 nm and its peak emission is 520 to 530 nm in physiologic environments. Fluorescein is approximately 80% protein-bound in circulation; the blood-retina barrier prevents it from diffusing into retinal tissue. However, leakage can show in areas with new vessel growth that lack a blood-ocular barrier or in regions with blood-ocular barrier defects induced by inflammation or ischemia. Fluorescein angiography findings are not pathognomonic for any optic nerve condition. However, there are some instances in which fluorescein angiography can be helpful in making a diagnosis. The abnormal areas of fluorescein angiogram of the optic disc are mostly of the hypofluorescent type. There are two possible causes of hypofluorescence: blocked fluorescence or a vascular filling defect. Fluorescein is present but cannot be seen in blocked fluorescence, with vascular filling defects; fluorescein cannot be seen because it is not present. The key in differentiating blocked fluorescence from vascular filling defects is to correlate the hypofluorescence on the angiogram with the ophthalmoscopic findings. If there is a lesion that is ophthalmoscopically visible that corresponds in size, shape, and location to the hypofluorescence on the angiogram, blocked fluorescence is present. If there is no corresponding ophthalmoscopic lesion, it must be assumed that fluorescein has not perfused the vessels and that the hypofluorescence is caused by a vascular filling defect. Vascular filling defects of the disc occur because the capillaries of the optic nerve head fail to fill. This failure can be caused by optic atrophy, vascular occlusion (eg, in cases of ischemic optic neuropathy), and congenital absence of disc tissue as in an optic pit or optic nerve head coloboma. Each one of these conditions is characterized by early hypofluorescence caused by nonfilling and late hyperfluorescence resulting from staining of the involved tissue (65).

Role in nonarteritic anterior ischemic optic neuropathy. There are 3 angiographic features that are well studied in nonarteritic anterior ischemic optic neuropathy:

Autofluorescence

Fundus autofluorescence is a noninvasive imaging technique that detects fluorophobes, naturally occurring molecules that absorb and emit light of specified wavelengths. To produce autofluorescence, a fluorophobe absorbs a photon of the excitation wavelength, which elevates an electron to an excited, high energy state. The electron dissipates some energy through molecular collisions, then emits a quantum of light at a lower energy and longer wavelength as it transitions back to ground state. Classically, fundus autofluorescence utilizes blue-light excitation and then collects emissions within a preset spectra to form a brightness map reflecting the distribution of lipofuscin, a dominant fluorophore located in the retinal pigmentary epithelium. Fundus autofluorescence may use other excitation wavelengths to detect additional fluorophores, such as melanin with near-infrared autofluorescence.

Autofluorescence can be useful in distinguishing drusen of the optic disc from other causes of disc swelling. The true autofluorescence of drusen seen microscopically is more definite. Its origin, however, is unclear. There are many naturally occurring substances that fluoresce. Drusen are cellular calcific deposits that are commonly asymptomatic, and the drusen are generally discovered incidentally during ophthalmologic examination. Diagnosis is often challenging based on clinical examination alone because its appearance can mimic other more common and concerning conditions, such as papilledema from elevated intracranial pressure caused by a space-occupying lesion, ischemic optic neuropathy, and optic nerve head tumors. The primary clinical importance of drusen is that they can mimic true optic disc edema, which may result in extensive, invasive, and unnecessary workup for elevated intracranial pressure, including neuroimaging and lumbar puncture.

Table 1. Imaging Modalities in Optic Nerve Abnormalities

Visual evoked potentials

Overview. Visual evoked potential records electrical responses from the occipital visual cortex to visual stimuli. Pattern-reversal visual evoked potential stimulated by contrast-reversing checkerboard patterns is commonly used for evaluation of optic nerve function. Pattern-reversal visual evoked potential has played a unique role in the diagnosis of optic neuritis and multiple sclerosis because a prolonged visual evoked potential latency is believed to indicate demyelination. Multifocal visual evoked potential simultaneously records visual evoked potentials from multiple locations across the visual field. In this section, we discuss the current application, limitations, and future potentials of pattern-reversal and multifocal visual evoked potential techniques in multiple sclerosis-related ischemic optic neuritis.

Key points.

Historical note and terminology. In 1972, Halliday and colleagues first reported a significantly delayed visual evoked potential response in optic neuritis; such delay persisted after visual acuity had returned to normal (18). In 1973, they further demonstrated the value of delayed visual evoked potential in revealing subclinical lesions in patients with multiple sclerosis without a history of optic neuritis (19). Delayed visual evoked potential has been historically used to fulfill the Poser criteria for definite multiple sclerosis (51). The 2017 revisions to the McDonald criteria state the value of visual evoked potential in confirming a current or historical symptomatic optic neuritis (62). Supporting previous inflammatory demyelinating attack with objective examinations may help reduce multiple sclerosis over-diagnosis (57).

Clinical aspects.

Pattern-reversal visual evoked potential. We recommend following the International Society for Clinical Electrophysiology of Vision standards for recording pattern-reversal visual evoked potential (43). The stimulus consists of a high-contrast (greater than 80%, mean luminance around 50 cd/m2) checkerboard pattern with uniform check sizes. The black and white checks reverse polarity at 1 to 2 Hz (2 to 4 reversals per second). At least 3 different check sizes nominally equivalent to 20/100, 20/50, and 20/25 visual acuity (large 60’, 30’, and small 15’) should be used. A single midline channel recording with active electrode at occiput (Oz), reference electrode at Fz, and ground electrode at forehead or earlobe is typically used for assessing optic nerve function, and low impedance is verified prior to collecting data. Testing is performed monocularly, typically without dilation, and with the best refractive lenses in place for the viewing distance. Patients are instructed to maintain a good fixation at the stimulus center during recording, or the patient can direct a laser pointer to the central fixation spot.

The typical pattern-reversal visual evoked potential waveform consists of an initial negative peak (N75) followed by a large positive peak (P100) then another negative peak (N135). The most commonly used parameters are P100 implicit time (often called latency) measured from the stimulus onset to the positive peak, and P100 amplitude (also called N75-P100 amplitude) measured from the N75 trough to the P100 peak.

Pattern-reversal visual evoked potential responses are affected by many factors, such as visual cortical anatomy, scalp thickness, electrode placement, stimulus condition, gender, age, extreme pupil size, refractive error, and even attention. P100 latency shows less intersubject variability and better reproducibility than amplitude. Interocular comparison in amplitude or latency is particularly useful in detecting monocular disease. Although normative values have been reported, it is important to establish clinic-specific age-matched norms because latency in normal subjects is affected by recording conditions (04). Longer latency is associated with smaller checks, lower mean luminance, or contrast of the stimulus. In general, factors that cause reduced retinal illumination (eg, very small pupil sizes, tinted lenses, or ptosis) increase the response latency. Latency also increases with uncorrected refractive error; therefore, optimal refractive correction is important.

Multifocal visual evoked potential. Pattern-reversal visual evoked potential represents a summed response predominantly from the foveal and inferior field regions; therefore, it does not provide topographic information and may miss localized defects. The multifocal technique, in contrast, permits simultaneous recording of visual evoked potentials from multiple locations (02; 25). Multifocal visual evoked potential uses a dartboard pattern stimulus consisting of many sectors that are scaled according to the cortical magnification for achieving equal effectiveness across eccentricity. The black and white checks in each sector reverse contrast following a fast pseudo-random m-sequence. A mathematical algorithm is used to extract responses corresponding to each sector. Despite relatively large variations of multifocal visual evoked potential amplitude and waveform across individuals and among local regions from the same person, signals from each eye of a normal individual are essentially identical, making interocular comparison a very sensitive test (25; 24; 13).

Since its introduction in 1994, the multifocal visual evoked potential technique has been greatly enhanced by multi-channel recordings, monocular and interocular analyses, signal-to-noise amplitude measurements, and computer-automated measurements of amplitude and latency (02; 25). The customized software developed by Hood and his colleagues also provides “probability plots” that are topographical maps comprised of color-coded points indicating “normal” (black) or “abnormal” amplitude or latency at individual locations compared to a normative database (25; 24; 13). Cluster criteria (the presence of adjacent abnormal points on probability plots) are developed to improve test specificity (25; 33).

Indications. Pattern-reversal visual evoked potential screening is generally recommended for patients with suspected optic neuritis or multiple sclerosis (04). A delayed latency provides evidence for demyelination supporting the diagnosis of optic neuritis or multiple sclerosis. The 2017 revisions of the McDonald Criteria for multiple sclerosis diagnosis discussed objective clinical or paraclinical evidence in considering a current or historical attack, which included “optic disc pallor or a relative afferent pupillary defect, optic nerve T2 hyperintensity on MRI, retinal nerve fiber layer thinning on optical coherence tomography, or P100 latency prolongation on visual evoked potentials in a patient reporting a previous episode of self-limited, painful, monocular visual impairment” (62). When pattern-reversal visual evoked potential findings are normal, the multifocal visual evoked potential test can be performed and may reveal mild localized defects not shown by pattern-reversal visual evoked potential (32).

Other indications for pattern-reversal visual evoked potential and multifocal visual evoked potential include visual functional assessment in those who can’t perform subjective tests or when the subjective tests are inconclusive or inconsistent with other clinical findings. Pattern-reversal visual evoked potential or multifocal visual evoked potential, when used in conjunction with the full-field electroretinogram or multifocal electroretinogram, may help differentiate retinal pathology from disorders at the level of optic nerve or beyond.

Contraindications. Caution should be taken for patients with seizure disorders triggered by flashing lights or patterns (avoid 10 to 13 hertz, which may induce seizures). Patients with poor fixation or those inattentive to the test may provide unreliable results, so a laser pointer may be beneficial.

Results. In optic neuritis, pattern-reversal visual evoked potential typically shows greater than 10 milliseconds delay in latency with less marked amplitude change. However, a delayed latency is neither 100% sensitive nor specific to multiple sclerosis-related optic neuritis; integration of clinical history and other examination findings is crucial. Delayed latency in the same range as that of optic neuritis has been reported in other conditions, eg, neuromyelitis optica and neuromyelitis optica spectrum disorder, compressive optic neuropathy, neurosarcoidosis, Behçet disease, neurosyphilis, hereditary optic neuropathies, and retinal or macular diseases (04; 01). It is important to rule out macular or retinal disease in any patient with a delayed visual evoked potential (22).

Optic neuritis in multiple sclerosis and neuromyelitis optica has indistinguishable clinical presentations although optic neuritis in the latter is often more severe, sequential in rapid succession, or bilateral. One visual evoked potential study reported prolonged P100 latency in 41.9%, nonrecordable findings in 14%, and reduced amplitude in 12.3% of a predominantly Caucasian cohort of definite neuromyelitis optica and anti-aquaporin 4 antibody-seropositive neuromyelitis optica spectrum disorders and found no difference in P100 latency or amplitude between seropositive and seronegative patients (54). Two studies in patients with neuromyelitis optica (predominantly African-Brazilian in one study and Japanese in another) reported that visual evoked potential responses were often nonrecordable, with large reduction in amplitude, less frequency of delayed P100 latency, and subclinical abnormality (42; 45).

Differential diagnosis of optic neuritis from other optic nerve disorders can be challenging in those with less classical presentations eg, older age, male gender, lack of orbital pain, swollen optic disc, severe loss of vision, or bilateral involvement. When optic swelling is present without pain on eye movement, a delayed visual evoked potential latency may aid in the differentiation between optic neuritis and nonarteritic anterior ischemic optic neuropathy (22; 27). In nonarteritic anterior ischemic optic neuropathy, pattern-reversal visual evoked potential usually shows a reduced amplitude without large delay in latency. For instance, Cox and colleagues showed a mean of 21 ms delay in optic neuritis versus 3 ms for nonarteritic anterior ischemic optic neuropathy (08). Visual evoked potential amplitude and latency are normal in the unaffected fellow eye of nonarteritic anterior ischemic optic neuropathy; in contrast, delayed latency may be observed in the asymptomatic fellow eye of optic neuritis, indicating subclinical demyelination.

On the other hand, a delayed visual evoked potential latency is not pathognomonic for optic neuritis. Compressive optic neuropathy is another common pathology associated with increased latency (Ng et al 2001; 22; 21). A recent retrospective study reported an initial diagnostic error in 71.4% of the 35 patients with unilateral optic nerve sheath meningioma (28). Nearly half of the erroneous cases were diagnosed as optic neuritis. Accurate diagnosis depends on adequate imaging and interpretation of brain and orbit MRI with contrast.

Scientific basis. Optic neuritis often refers to an inflammatory demyelinating optic neuropathy associated with multiple sclerosis. Optic neuritis mostly affects young adults, especially women. About 20% of patients with multiple sclerosis present with optic neuritis as an initial demyelinating event. According to the Optic Neuritis Treatment Trial, the risk of developing multiple sclerosis after 15 years of the initial optic neuritis is 72% in those with one or more lesions on brain MRI and 25% in those without (46).

Optic neuritis classically presents with unilateral loss of vision with pain around orbit or with eye movement in about 90% of patients. Although the degree of vision loss varies widely, vision loss in optic neuritis associated with multiple sclerosis is often mild, unilateral, and shows complete or partial recovery of vision within several weeks. Large delay of visual evoked potential latency and some degree of reduction in amplitude are characteristic. An attenuated visual evoked potential amplitude is generally correlated with a worse visual acuity or visual field. In severe cases, visual evoked potential may be nonrecordable initially due to conduction block from acute inflammation. The visual evoked potential amplitude partially recovers or normalizes when the acute optic neuritis resolves. The visual evoked potential latency also shortens, especially during the first 3 to 6 months, although persistent delays are often observed (05).

Delayed visual evoked potential latency has long been considered to reflect the extent of demyelination. In a rat model of lysolecithin-induced optic nerve demyelination, the magnitude of delay correlated with the length of demyelination (66). Also the time frame during which latency shortening occurs after an acute optic neuritis corresponds to the period of remyelination observed in human postmortem tissues (52). In several recent clinical trials of potential remyelination agents, visual evoked potential latency has been used as an outcome measure for remyelination (34; 01).

Amplitude reduction in acute phase is believed to reflect temporal conduction block in optic nerve axons associated with inflammation and edema. Reversal of conduction block leads to amplitude recovery over the following weeks--the same period of time when resolution of MRI enhancement occurs (20). After resolution of acute optic neuritis, amplitude reduction reflects axonal loss or impaired conduction in survived axons. The visual evoked potential amplitude in eyes with last optic neuritis occurring more than 6 months prior correlated with axonal loss measured by OCT (30; 41) and MRI (63).

Multiple sclerosis is a continuously active disease even during remission. Studies have shown that visual evoked potential is more sensitive than MRI and OCT in detecting resolved or subclinical optic neuritis (38; 39). Subclinical visual evoked potential latency prolongation and amplitude reduction in the fellow eye of optic neuritis was found to be proportional to the risk of multiple sclerosis (31). In a study that involved long-term follow up of patients with multiple sclerosis, increased visual evoked potential latency over an 8-year period was correlated with disability score and brain atrophy (64).

Longitudinal studies have shown the value of using visual evoked potential to track latency changes in optic neuritis eyes with recovered, stable visual acuity and visual sensitivity (15). Visual evoked potential latency has served as an outcome measure for evaluation of new remyelination therapies. In a phase 2 clinical trial (RENEW) that examined the effect of opicinumab in acute optic neuritis, the per-protocol analysis showed significantly shorter pattern-reversal visual evoked potential latency in opicinumab group compared to placebo while no change in OCT or low contrast letter acuity was observed (06). In the ReBUILD study, clemastine fumarate improved the visual evoked potential latency in patients with multiple sclerosis and chronic optic neuropathy, demonstrating its potential role of remyelinating (17).

Compared to pattern-reversal visual evoked potential, multifocal visual evoked potential detected 10% to 15% more abnormalities in optic neuritis or multiple sclerosis (32). Prolonged multifocal visual evoked potential latency has also been reported in macular and retinal diseases, in compressive optic neuropathy, and in nonoptimal recording conditions (23; 27). At the current stage, routine clinical use of multifocal visual evoked potential is limited by the longer recording time and analysis. However, compared to pattern-reversal visual evoked potential, the multifocal visual evoked potential is likely a more sensitive research tool in measuring remyelination due to better reproducibility in latency and the localized spatial information provided (40).

Full-field flash electroretinogram

Introduction. Ragnar Granit identified various retinal facets of a mammal (cat) using varying anesthesia depth as well as vessel occlusion, establishing an understanding of the electroretinogram and its underlying components (16). Consequently, Granit won the Nobel Prize for Physiology and Medicine in 1967. The full-field electroretinogram records a massed potential from the whole retina. Unless 20% or more of the retina is affected with a peripheral diseased state, the electroretinograms are usually normal. In other words, a legally blind person with macular degeneration, enlarged blind spot, or other small central scotomas will have a normal full-field electroretinogram.

Each scotopic and photopic test condition evokes a single waveform summed over the peripheral retina representing 90% of the cones (12 million) versus rods (20 million). Clinical flash electroretinogram peripheral retinal components include the A wave, downward from baseline, representing choroidal and photoreceptor activity separately from the B wave, upward from baseline that assesses the middle nuclear layers including bipolar and Mueller cells. Amphibian studies show that the B wave originates from the Mueller cells and that bipolar cells cause ionic changes that induce a passive Mueller cell response. When retinal function deteriorates, the light-induced electrical activity in the retina decreases. Amplitude is measured from the trough of the A wave to the peak of B wave and is more variable than latency due differing eye length and age as well as genetic penetrance variation. Latency or B wave implicit time is measured from onset of stimulus to peak of the B wave.

Typically, the patient is dark adapted for at least 15 to 30 minutes (98% dark adaptation in most subjects at 30 minutes), and an active electrode (contact lens, thread) is placed on an anesthetized cornea or, least sensitive, skin electrodes. Scotopically, using dim blue light at 1 Hz stimulates rods whereas dim red light is employed to assess cone activity. Due to the scotopic low stimulus light level, the A-wave photoreceptors are not present in these two scotopic conditions. With the addition of brighter light stimuli, again in the scotopic condition, both A and B waves are present. The peak B-wave latency shortens, and the waveform amplitudes increase. In normal patients the B wave in the scotopic bright white test condition is 1.5 times greater than the A wave, with attenuated waveforms being below 250 microvolts with the contact lens electrodes. Photopic cone activity requires 5 to 10 minutes of light adaption, and then patients are typically stimulated with a single white flash in the presence of background illumination. Finally, flicker electroretinograms are recorded at approximately 30 flashes per second (30 Hz) to measure a pure cone response as rod photoreceptors cannot recover rapidly enough to respond to the rapid flashes.

Clinical findings.

Retinitis pigmentosa. Retinitis pigmentosa shows loss of both the A and B waves and is extinguished scotopically and photopically. In early retinitis pigmentosa, if waveforms can be recorded, scotopic or photopic B-wave peak implicit times are usually prolonged. In contrast to retinitis pigmentosa, the electroretinograms in a patient with a cone dystrophy exhibit robust rod A and B waves, whereas a rod dystrophy will exhibit normal cone responses only. Inner retinal vascular occlusions may preserve the A wave but markedly reduce the B wave. This same finding is found with retinoschisis, literally a splitting of the retina.

Paraneoplastic syndromes -- full field flash electroretinogram findings.

Paraneoplastic syndromes are rare and occur with cancer due to tumor secretions or the body's response to a tumor. Paraneoplastic syndromes change the immune system response to a neoplasm and cause nonmetastatic systemic effects that accompany malignant disease. The syndromes are varied and result from substances produced by the tumor, occurring in remote locations.

Cancer-associated retinopathy. An autoimmune disease, cancer-associated retinopathy is the most prevalent paraneoplastic retinopathy, with approximately half of the patients reporting visual symptoms (decreased acuity but painless and progressive visual loss), with the onset of visual symptoms and detection of antibodies often preceding the diagnosis of malignancy by months to years (03). Other symptoms include photopsias, ring scotoma, and color impairment before the diagnosis of the underlying malignancy. Importantly, loss of rod and cone function as determined by flash electroretinography is due to the production of autoantibodies against retinal antigens stimulated by a systemic, non-ocular tumor, with the overwhelming majority of cancer-associated retinopathy cases due to small-cell lung carcinoma. The autoimmune reaction itself leads to retinal photoreceptor cell death. Retinal tissue from patients with cancer-associated retinopathy demonstrate extensive degeneration of the outer nuclear layer as well as inner and outer segments as reflected in the electroretinogram of the photoreceptor layer. The inner retinal layers are preserved.

Melanoma-associated retinopathy. Melanoma-associated retinopathy is a paraneoplastic retinal disorder occurring in patients with cutaneous melanoma where patients typically report visual symptoms of night-blindness and photopsias. Visual field testing generally reveals generalized constriction, although arcuate defects and central and paracentral scotomas have also been reported. Melanoma-associated retinopathy has a sex ratio of 4.5:1 skewed towards men with a latency period averaging 3.6 years from diagnosis of the primary neoplasm (ranging 2 months to 19 years). Age of onset is commonly during the sixth decade of life, with a significant male predilection (male to female ratio of 4.7:1) (10).

The electroretinogram in melanoma-associated retinopathy is typically electronegative, recording a dark-adapted A wave only, indicating normal photoreceptor function, followed by a markedly attenuated B wave, reflecting either bipolar cell dysfunction or disruption of transmission to the bipolar cell.

Multifocal electroretinogram (mfERG)

Introduction. A limitation of the traditional full-field flash electroretinogram is that the recording is a single massed waveform from the peripheral retina under scotopic and photopic conditions. Unless 20% or more of the peripheral retina is affected by disease, the electroretinograms are usually normal. In other words, a legally blind person with macular degeneration, enlarged blind spot ,or other small central scotomas will have a normal full-field electroretinogram. However, the multifocal electroretinogram (mfERG) is helpful in the patient with poor vision but a normal-appearing retina. The mfERG enables distinction between optic nerve and retinal disease and has application in patients with early hydroxychloroquine toxicity as well as in diagnosing and distinguishing between the various white dot syndromes and the ability to follow patients with central retinal cone disease who have extinguished full field flash electroretinograms.

An important electrophysiology tool is the multifocal electroretinogram (mfERG) developed by Erich Sutter in 1991 (59). This test rapidly evaluates cones in the central retina using mathematical sequences of binary m-sequences that can extract hundreds of focal electroretinograms from a single electrical signal. This system allows assessment of cone electrical activity and displays the distribution of the central retinal dysfunction. These data can correlate electrophysiologic findings with visual field testing. Small scotomas in the retina can be mapped, the degree of retinal dysfunction can be quantified, and up to 103 separate locations (67 degrees) can be assessed in the central retina. With this method, one can record mfERGs from hundreds of retinal areas in a several minutes. Electroretinogram electrodes (thread or contact lens) are best used to record electroretinograms from an anesthetized cornea and dilated eye.

Black and white and gray achromatic hexagons on the display screen change every 13 milliseconds following a pseudorandom sequence. The data is analyzed via a cross-correlation technique, and individual retinal responses from each of the 103 segments of the retina are displayed with each waveform representing nanovolts per mm squared (nV/mm2) as well as 7 concentric rings denoting amplitude (response density) and latency. Amplitudes are highest at the foveola but may decrease in the first 5 degrees. Note: one finger held at arm’s length subtends one degree on the retina, the width of the foveola. Individual waveforms are biphasic, and data from 7 rings over 67 degrees on the retina can be assessed and averaged to reveal peak latency, which should not vary. Although latency should be the same for each ring, amplitude, nV/deg² may show variability. Note that patient cooperation is essential, so the mfERG may not be appropriate in young or uncooperative patients or in patients with tremors (eg, Parkinson disease).

Clinical features.

Hydroxychloroquine toxicity. Hydroxychloroquine, which is also used to treat discoid or systemic lupus erythematosus, rheumatoid arthritis, dermatological disorders, and Sjogren syndrome can be toxic to the retina, producing ring scotomas as well as foveola damage. Multifocal electroretinograms better quantify and identify retinal toxicity than full-field electroretinograms.

Hydroxychloroquine toxicity is reported to be only 1% after 5 years, and most cases of toxicity occur in patients taking a daily dose 6.5 mg/kg or a cumulative dose of 1000 grams (58). The risk further increases with continued use of the drug. Hydroxychloroquine first affects small areas between 5 and 15 degrees from the fovea, eventually progressing to produce a ring scotoma or foveola impairment.

Abnormal mfERGs may occur in eyes prior to observable structural damage. However, older patients who may be at risk for macular degeneration should have a baseline examination. A clinical question is whether to add a potentially retino-toxic drug to an already compromised retina. Paracentral retinal loss in the second ring (so called “danger zone”) is one of earliest signs of toxicity. Ring 1/Ring 2 is considered normal if under 2.7. Sensitivity and specificity increase if the mfERG is combined with spectral domain OCT. Detection of bull’s eye maculopathy is a clinical sign suggesting retinal damage has already occurred and may be irreversible. Further, after cessation of hydroxychloroquine, there may be no visual recovery and damage may even progress.

Retina versus optic nerve function. In the patient that reports central vision loss, an mfERG in combination with pattern visual evoked response that measures optic nerve conduction time at 20/100, 20/50, and 20/25 visual acuity can be used to discern whether visual loss is due to macular dystrophy (cone dystrophy) of the retina verses an optic nerve problem.

Branch retinal artery occlusion. A localized decrease in response amplitude is seen in branch retinal artery occlusions. A significant decrease follows the arterial supply pattern and continues even after the retina returns to its normal appearance.

White dot syndrome. Multiple evanescent white dot syndrome can present with photopsia and an enlarged blind spot on visual field perimetry. Multifocal electroretinogram abnormalities in multiple evanescent white dot syndrome frequently are reversible, and retinal responses return to normal after a few months.

Table 2. Electrophysiological tests in optic nerve abnormalities