|Home | About | Journals | Submit | Contact Us | Français|
Visual dysfunction is one of the most common clinical manifestations of multiple sclerosis (MS). Just over a decade ago, MS clincial trials did not include visual outcomes but experts recognized the need for more senstive measures of visual function. Low-contrast letter acuity emerged as the leading candidate to measure visual disability in MS and subsequent studies found low-contrast acuity testing to correlate well with brain MRI lesion burden, visual-evoked potentials (VEPs), quality of life (QOL), and retinal nerve fiber layer (RNFL) loss as measured by optical coherence tomography (OCT). OCT in MS has allowed for assessment of structure-funciton correlations that make the anterior visual pathway and acute optic neuritis (ON) ideal models for testing novel agents for neuroprotection and repair. New therapies that reduce axonal loss by neuroprotective or myelin repair mechanisms can now be assessed non-invasively by OCT and coupled with visual function data. Based on OCT studies in MS, RNFL thickness is reduced significantly among patients (92 µm) versus controls (105 µm), and is particularly reduced in MS eyes with a history of ON (85 µm). Worsening of visual function by a clinically significant ≥7 letters or ~1.5 lines for low-contrast acuity is associated with approximately 4.5 µm reductions in RNFL thickness in MS eyes. Longitudinal studies of OCT have also shown RNFL axonal loss over time that occurs even in the absence of acute ON, and that correlates with clincially meaningful worsening of vision and QOL, even in patients with benign MS. The latest OCT investigations involve high-resolution spectral-domain (SD) OCT with segmentation and measurement of specific retinal layers using computerized algorithms. These methods allow quantitation of ganglion cell (neuronal) layer loss as well as axonal degeneration in MS in vivo. In this review, we examine the data from these studies and ongoing trials that highlight the entity of ON as a model to investigate neuroprotection and neurorepair. In doing so, we also present representative group data from studies that have examined visual function, OCT measures, and QOL scales in patients with MS, ON, and disease-free controls. These data, as well as those from recent meta-analyses, may be used to provide reference values for the development of clinical trial protocols.
Visual loss has been recognized for decades as one of the most common and disabling clinical manifestations of multiple sclerosis (MS). It was not until a decade ago, however, that MS clinical trials began to include sensitive measures of visual function. Inspired by the effective use of contrast sensitivity (Pelli-Robson charts) to demonstrate visual abnormalities in the Optic Neuritis Treatment Trial (ONTT) (1–6), low-contrast letter acuity was introduced into MS trials as an exploratory outcome to better capture the often subtle visual symptoms experienced by patients with acute optic neuritis (ON) (7–10).
Visual symptoms can result from a variety of pathological processes, including inflammation, demyelination, and axonal degeneration in the afferent visual pathway, (retina, optic nerves, chiasm, and tracts) (11–15). During the past decade, the contributions of axonal and neuronal loss to disability in MS have been increasingly recognized (12, 16–19). Developed initially as a non-invasive tool to monitor retinal disease and glaucoma, optical coherence tomography (OCT) was first applied to cohorts of patients with MS and ON over a decade ago to provide validation of clinical visual outcomes and to capture anterior visual pathway axonal loss (15, 20–25). As a result of these early studies and >100 subsequent published investigations of OCT in MS and ON (26), the visual pathway is now recognized as a model for structure-function correlation in MS for studying the pathophysiology of disease and for testing both novel and standard therapies that involve neuroprotection and repair. OCT can be used non-invasively to capture reductions in axonal loss that may be associated with neuroprotective or myelin repair therapies; this information can be correlated with visual function and QOL data to further establish the clinical relevance of these structural outcomes.
OCT enables investigators to rapidly and reproducibly evaluate the structural composition of the retina. Ultimately, OCT could substantially increase our understanding of the mechanisms of tissue injury in MS, ON and other optic neuropathies.
A decade after publication of the first studies of low-contrast acuity in MS, this test as well as OCT measures are commonly used in MS clinical trials. They have formed the foundation for new trials that use acute ON as a model for identifying novel compounds for neurorepair (27–32).
This review describes a number of advancements in the development of visual function outcomes in MS as well as OCT as a novel technology that enables objective analysis of the processes of neurodegeneration. In addition, we present representative group data from studies that have examined visual function, OCT measures, and QOL scales in patients with MS, ON, and disease-free controls (Table 1). These data may be used to provide initial reference values and should be viewed in the context of 1) the continually-evolving field of vision in MS; 2) the potential challenge inherent in applying group data to individual patients; and 3) the perspective that even the small observed differences in mean OCT values have been shown to correlate with clincially meaninful changes in visual funciton and QOL in MS patients.
In the late 1990’s, the National Multiple Sclerosis Society Clinical Outcomes Assessment Task Force developed the MS Functional Composite (MSFC) in response to the need for more sensitive neurologic outcome scale (33, 34). Designed to be a battery of performance measures complementary to the Expanded Disability Status Scale (EDSS), the MSFC includes the 25-Foot Timed Walk, the 9-Hole Peg Test, and Paced Auditory Serial Addition Test (PASAT) (35–38). While multidimensional, the MSFC did not include a measure of visual function. In the evaluation of candidate MSFC visual components from MS clinical trial data used to develop the MSFC, Snellen high-contrast visual acuity (VA) did not change over time or demonstrate concurrent changes with EDSS scores (33). Contrast sensitivity, as tested by line gratings and letter charts in MS and by Pelli-Robson charts in the Optic Neuritis Treatment Trial (ONTT), had been shown to be a sensitive measure of afferent visual function, even among patient with Snellen acuities of 20/20 or better (1–6, 39–43). Importantly, measures of low-contrast vision are predictive of “real-world” visual tasks such as reading rate, facial recognition, and driving (44).
While used successfully in the ONTT to demonstrate persistent visual abnormalities beyond recovery of high-contrast visual acuity, Pelli-Robson contrast sensitivity charts were not available for purchase in 1998 when the first MS clinical trial to incorporate a low-contrast visual measure as an exploratory outcome was begun. This trial, IMPACT (International MS Progressive Avonex Clinical Trial), incorporated low-contrast Sloan letter charts (7, 8, 45, 46). These charts are the low-contrast (gray letters on white) “cousin” of the ETDRS high-contrast VA charts used in ophthalmology clinical trials (Fig. 1). Sloan charts have a standardized format based on the ETDRS VA charts. Three contrast levels have been used in MS trials and research studies, including 100% (high-contrast, used to measure VA as a descriptor of the study cohorts), 2.5%, and 1.25% (lightest contrast level). Charts are scored letter by letter, and numbers of letters identified correctly constitute the score for each chart (Table 1). Most recently, visual improvement and loss by the low-contrast acuity chart has been defined as a 7-letter change in score, while 5-letter changes in high-contrast VA are now considered clinically significant for patients with good visual acuity (47). This threshold represents a change that exceeds that which would be expected from repeated testing when there was no real change (45, 48), and to correlate with RNFL axonal loss in patients with MS (47). As shown in Table 1, differences in mean letter scores for MS vs. disease-free control groups are consistently ≥5 letters for VA and ≥7 letters for low-contrast acuity. Since the 5- and 7-letter criteria are meant to be applied to individual eye or patient differenes, differences in mean scores of such magnitude are therefore likely to be clinically meaningful. Furthermore, these 5- and 7-letter mean changes have been shown to correlate signifcantly with vision-specific and overall QOL as well as structural measures of OCT RNFL thickness. These associations of OCT and QOL measures with clinically meaninful changes in visual function provide perspective for the data in Table 1 despite the seemingly small magnitudes of differences in means for OCT values.
In the IMPACT trial and in heterogeneous MS cohorts (7, 8, 45), low-contrast letter acuity was shown to be a highly reliable (high intraclass correlations of 0.86–0.95) and practical method that was superior in identifying MS-related visual loss compared to other available tests (Fig. 2). Correlations of Sloan chart scores with MS Functional Composite (MSFC) and Expanded Disability Status Scale (EDSS) in these studies were significant and moderate in magnitude (rs=0.56 vs. MSFC; rs= −0.43 vs. EDSS, P<0.0001). These studies provided the first ever data demonstrating that low-contrast acuity scores capture unique aspects of dysfunction not captured by standard neurologic scales of EDSS or the MSFC, and suggested that adding a visual component to the MSFC would increase the applicability of this measure.
In the AFFIRM trial of natalizumab vs. placebo for relapsing-remitting MS, low-contrast letter acuity was able to demonstrate treatment effects and changes over time both with respect to reduction of the cumulative probability of sustained visual loss (by 47%, P<0.001 for 2.5% contrast level) and sustained clinically meaningful visual improvement (by 57%, P=0.01) in the treatment group (9, 10). Similar to previous studies and to MS trial datasets used originally to develop the MSFC, high-contrast VA did not show treatment differences or detect sustained visual loss or improvement over time. The analyses of visual loss in AFFIRM used initially a 2-line (10-letter) criterion for clinically significant change; subsequent ophthalmologic studies for VA, however, suggested use of a one-line (5-letter cutoff). Examination of inter-rater and test-retest reliability data for low-contrast acuity revealed 7 letters to represent 2 standard deviations on the inter-rater differences; this 7-letter criterion was subsequently incorporated into the AFFIRM analyses of visual improvement as well as subsequent longitudinal studies of optical coherence tomography (OCT).
The relevance of low-contrast letter acuity measurements to patient functioning has been underscored by studies linking these scores to vision-specific quality of life (QOL). Scores for the 25-Item National Eye Institute Visual Functioning Questionnaire (NEI-VFQ-25) are reduced among patients with MS (Table 1) (49–51). A 10-Item Neuro-Ophthalmic Supplement to the NEI-VFQ-25 has also been designed using MS cohorts to capture symptoms relevant to neurologic disease. In a recent study of 167 patients with MS, two-line differences in visual function by low-contrast acuity are associated, on average, with >4-point (6.7 – 10.9-points, P<0.001 accounting for age) worsening in NEI-VFQ-25 composite (overall) score, reductions that are considered to be clinically meaningful based on ophthalmic epidemiologic studies (52–55). In Table 1, the differences in mean scores between MS patients and disease-free controls were ≥4 points; such differences may be considered clinically significant for both individual changes and group data. Binocular scores for low-contrast acuity also correlated significantly with score for the 10-Item Neuro-Ophthalmic Supplement (P<0.001), the Impact of Visual Impairment Scale (IVIS, P<0.001) and the SF-36 Physical Components Summary (P<0.001) in this MS cohort. Collectively, these data demonstrated that low-contrast acuity testing provides information on patient-reported aspects of vision. This is a key feature for clinical measures and a prerequisite for their use as primary outcomes in clinical trials.
While the earliest studies of low-contrast acuity and QOL examined binocular vision scores (vision testing with both eyes together), monocular measurements (with each eye separately) have more recently been introduced into MS trials and vision research with the inclusion of optical coherence tomography (OCT) scanning as a structural correlate to vision (Table 1). Given the importance of both of these types of acuity to measurements of function, analyses of MS cohort data including both binocular and monocular scores were undertaken to examine the relation between the two and to determine the potential roles for binocular summation and inhibition (56). Binocular summation of acuity occurs when vision is improved under binocular viewing conditions (binocular score is greater than the scores for either eye alone) (57). In contrast, patients with binocular inhibition have worse binocular vision compared to the better eye alone. Among 1,007 patients with MS and 324 disease-free controls, binocular summation was substantial (5–7-letter increases over better eye acuity, 28–52% with >7 letters, P<0.001) for low-contrast acuity at the 2.5% and 1.25% levels. For high-contrast VA, only 3.0–3.4% of patients showed similar degrees of summation. Increasing age (P<0.0001), greater interocular differences in acuity (P<0.0001), and prior history of ON (P=0.015) were associated with lower magnitudes of binocular summation; in fact, some of these patients had binocular inhibition. Importantly, greater degrees of binocular summation were predictive of better QOL by the NEI-VFQ-25 (P=0.02) and 10-Item Supplement (P=0.03), indicating that the capacity to use both eyes together is an important factor in determining how well patients with MS can perform daily activities (56).
Correlation with biological markers as well as electrophysiological measures of disease is one of the most important factors in the evaluation of clinical outcome measures such as low-contrast letter acuity. The first such study determined the relation of binocular low-contrast scores to brain MRI measures of T2 lesion burden (58). This study of 45 patients with MS found that those with worse low-contrast acuity scores had greater T2 lesion volumes in whole brain (3 mm3 increase for every 1-line worsening, P=0.002–0.004, accounting for age and disease duration). Area 17 and optic radiation white matter lesion burden specifically correlated with visual function, but non-vision-related white matter did not show associations. These findings were important in providing a structural correlate to low-contrast acuity scores, and also in emphasizing a potential post-geniculate pathway component to binocular low-contrast vision.
Standard brain MRI techniques have provided information regarding disease burden in MS. However, the capacity for these techniques to precisely quantify axonal and neuronal loss, particularly in the anterior visual pathway, has been limited. Electrophysiologic markers such as the visual-evoked potential (VEP) can provide additional information on anterior visual pathway integrity. Recent studies have combined assessments of vision loss with VEP. Conventional VEPs measure the cortical response to monocular stimulation in the central 30 degrees of the visual field. In MS, the latency characteristically is delayed with normal amplitude. Axonal loss, however, can reduce this amplitude (59). Abnormal VEPs in unaffected eyes provided evidence for clinically silent lesions in the optic nerve that might help identify dissemination in space and help establish the diagnosis of MS (60). Various studies have shown that contrast sensitivity can be measured by VEP (43), and VEP in MS using low-contrast stimuli show increased latencies or absent waveforms when compared with high-contrast stimuli VEPs. Low-contrast VEP may prove to be helpful in identifying demyelination (61). Other studies have concluded that multifocal VEP provides higher sensitivity and specificity in detecting abnormalities in visual function in MS and in ON (62).
Naismith et al. (63) systematically evaluated the utility of OCT and VEPs to detect the presence of clinical and subclinical ON, and examined the relation of these measures to visual function. This retrospective, cross-sectional study evaluated 65 subjects (n=96 eyes) with MS (n=40), clinically isolated syndrome (CIS, n=1), neuromyelitis optica (NMO, n=20) and idiopathic demyelination (n=4). Patients had at least one episode of ON ≥6 months prior to enrollment. VEPs detected ON in 81% of patients (32% of subclinical ON in unaffected eyes and 75% of all subclinical ON). In contrast, using their criteria, they found that OCT identified 60% of eyes with ON, and less than 20% of subclinically affected eyes. The authors concluded that OCT is less sensitive than VEPs in optic neuritis. On the other hand, to the extent that this study focused on patients with relatively poor vision, there were likely floor effects with regard to OCT RNFL thickness.
Although acute ON and demyelination are important contributors to visual dysfunction, irreversible axonal and neuronal degeneration are also final common pathways to permanent visual loss (12–14, 24, 64). While the relationship between visual loss and brain MRI lesion burden discussed above is a relatively new finding, the extremely high predilection for MS to involve the optic nerves is well-documented and was initially observed long before the advent of advanced optical and neuroimaging. Autopsy studies have shown that up to 94–99% of MS patients have detectable optic nerve lesions (65, 66). Sub-clinical changes related to visual loss may involve the optic nerves or chiasm, or post-chiasmal regions of the optic tract (13, 67–69).
OCT provides a non-invasive technique to measure axonal and neuronal loss in the anterior visual pathways (24, 28–32). As the optical analog of ultrasound B-mode imaging, OCT allows us to image the retinal nerve fiber layer (RNFL) (15, 20–32, 47, 59, 70–82). Since, within the retina, these axons are non-myelinated, the RNFL is an ideal structure with which to visualize the processes of neurodegeneration, neuroprotection, and potentially even neuro-repair. In contrast to the peripapillary RNFL, which contains axons, the macula contains a large proportion of retinal ganglion cell neurons (about 34% of total macular volume) (81). The development of Fourier-domain (or spectral-domain [SD]) detection has in particular enhanced ophthalmic OCT technology. While many of the studies cited in this review utilized time-domain (TD) OCT, some have incorporated the newer SD OCT technologies. Table 1 presents some representative group data for OCT measures in patients with MS, ON, and disease-free controls. Since OCT measures have been shown to correlate with clincially meaningful changes in visual function and QOL (see above), the differences in mean values between groups shown in Table 1 are likely to have clinical significance. Larger studies and additional meta-analyses will allow us to further refine the precision of these representative average values.
The earliest application of OCT technology to the study of MS was reported by Parisi et al. in 1999 (15). In this study, which utilized first-generation OCT technology, 14 patients with MS who had completely recovered from a previous event of acute optic neuritis (ON) were analyzed. The thickness of the RNFL was shown to be reduced by 46% in the affected eyes of the patients with MS versus the control eyes (P <0.01), and by 28% when affected eyes were compared with the ‘unaffected’ eyes of the same patient (P <0.01). Even in the clinically unaffected eyes of patients, however, there was a 26% reduction in RNFL thickness when compared with control eyes (P <0.01).
In 2005, Trip and colleagues (21) reported their observations with OCT in 11 patients with MS and 14 patients with clinically isolated syndrome (CIS), all of which individuals had a history of a single episode of ON. The study was a cross-sectional analysis with follow-up ranging from 1 to 9 years after the optic neuritis event. Corroborating the previous findings by Parisi et al. (15), the investigators found a 33% reduction in RNFL thickness in the eyes of the patients when compared with the eyes of matched controls, and a 27% reduction when the affected and unaffected eyes of the same patient were compared (P <0.001). Trip et al. (21) extended the utility of OCT by also showing that the macular volume (a reflection of retinal ganglion cell neuronal integrity) was reduced by 11% in the eyes of patients with a history of ON when compared with control eyes (P <0.001), and by 9% in the affected versus the unaffected eye of the same patient (P <0.001).
In 2006, Costello and colleagues (20) reported that the majority of patients with MS who have ON (about 75%) will sustain 10–40 µm of RNFL loss within a period of approximately 3–6 months. This finding is striking given that the RNFL is only about 110–120 µm thick by the age of 15 years, and that most individuals (without a history of glaucoma or macular degeneration) will lose only about 0.017% per year in retinal thickness, which equates to approximately 10–20 µm over 60 years (25, 83). Costello et al. (20) also provided compelling evidence identifying an injury threshold within the RNFL of about 75 µm; thinning of the RNFL below this level led to a corresponding decline in visual function, as measured by automated perimetry. Using low-contrast letter acuity in MS eyes with a history of acute ON, we have found a similar threshold for abnormal visual function and axonal loss at 80 µm (Fig. 3).
The trajectory and time course of RNFL axonal loss following an episode of acute ON is an important for determining the “window of opportunity” within which a neuroprotective or repair agent should be administered in a treatment trial. Sample size calculations for a proposed ON trial are also likely to be based on such data. Henderson et al. (84) recently addressed both of these questions in a study of 23 patients with acute clinically isolated unilateral ON. Patients underwent time-domain (TD) OCT assessment of RNFL thickness and total macular volume; visual function testing with high-contrast VA, low-contrast acuity (Sloan charts at 1.25%), Farnsworth-Munsell 100 color testing, and automated visual fields; and pattern VEP testing. The mean time to 90% loss from the initial RNFL thickness values for affected eyes (initial median 116.5 µm---likely reflecting slight edema) was 2.38 months (95% CI 1.95, 2.77; P<0.05). Ninety-nine percent of the loss from the initial value to the asymptotic value (lowest point where loss leveled off) occurred by a mean of 4.75 months (95% CI 3.90, 5.54; P<0.05). The time of first detectable RNFL atrophy compared to the baseline fellow eye value was 1.64 months (95% CI 0.96, 2.32; P<0.05). Eyes with poor recovery (n=5), defined as logMAR VA>0.1, had a significantly greater decline of RNFL from baseline to 3 months (P=0.002). Macular volumes also declined significantly to the time of last follow-up. Sample size estimates using their models showed that, assuming 80% power, approximately 100–150 patients would be needed to detect a 30% decrease in RNFL loss between the treatment and placebo groups at a 3–6 month endpoint. These findings are critical in setting the stage for use of acute ON as a model for neuroprotection and repair trials (84).
One of the most important findings that has resulted from the use of OCT MS studies is the correlation between RNFL thinning and visual loss, as measured by low-contrast letter acuity. In 2005, Fisher et al. (22) conducted a cross-sectional study that compared RNFL thickness among MS eyes with a history of ON (MS ON eyes), MS eyes without a history of ON (MS non-ON eyes) and disease-free controls. In addition to OCT measurement of OCT, they conducted low-contrast visual assessments with low-contrast letter acuity, contrast sensitivity (Pelli-Robson charts), high contrast VA (ETDRS charts). The authors found that RNFL thickness was reduced significantly among MS patients (92 µm) versus controls (105 µm, P<0.001), and particularly reduced in MS ON eyes (85 µm, P<0.001). Furthermore, lower visual function scores were associated with reduced average overall RNFL thickness in MS eyes; for every 1-line decrease in low-contrast letter acuity or contrast sensitivity score, the mean RNFL thickness decreased by 4 µm. These findings not only supported the validity of low-contrast visual assessment as a secondary clinical outcome measure in MS trials, they also suggested a potential role for OCT in trials that examine neuroprotective and other disease-modifying therapies. Several other investigations have demonstrated correlations between RNFL thinning and visual loss (21, 25, 77, 85–87).
While the above studies used cross-sectional studies to examine relationships between biomarkers measured with OCT and visual function, the first longitudinal study to study the relationship between RNFL thickness and visual loss was carried out by Talman et al in 2010 (47). These investigators used OCT to measure RNFL thickness at baseline and at 6-month intervals during a mean follow-up of 18 months at 3 centers. Low-contrast letter acuity (2.5% and 1.25%) and high contrast visual acuity (VA) were tested. The results indicated that among 299 patients (593 eyes) with at least 6 month follow-up, eyes with visual loss showed greater RNFL thinning compared to eyes with stable vision (low-contrast acuity, 2.5%: P<0.001; VA: P= 0.005). RNFL thinning increased over time, with average losses of 2.9 µm at 2 to 3 years and 6.1 µm at 3–4.5 years (P<0.001 vs. 0.5–1 year follow-up). The authors concluded that progressive RNFL thinning occurs as a function of time in some patients with MS, even in the absence of ON (47). Studies such as this draw the important connection between structure (RNFL thickness and macular volume, as measured by OCT) and function (visual performance as measured by low-contrast acuity), and are likely to be instrumental in improving our understanding of the MS disease process.
Recently, OCT has also been used to show the capacity of RNFL thinning to distinguish MS disease subtypes. Costello and colleagues (85) found that RNFL comparisons involving eyes without ON yielded greater differences between MS subtypes than ON-affected eyes. Overall RNFL values in non-affected eyes were reduced in Secondary Progressive MS patients (SPMS) (83.4 µm), relative to ON as a clinically isolated syndrome (CIS) (101.2 µm) (P=0.0009), and relapsing-remitting MS (RRMS) patients (103.7 µm) (P=0.001); and temporal RNFL atrophy was greater in RRMS (64.4 µm) eyes as compared to CIS eyes (73.2 µm, P=0.02). In ON-affected eyes, RNFL atrophy was greater in SPMS patients (39.5 µm) than CIS patients (58.1 µm, P=0.03), and in RRMS patients (48.2 µm) relative to CIS patients (P=0.05). These authors concluded that RNFL thickness might represent an important MS structural marker because RNFL thinning reflects disease progression.
While RNFL thinning is most marked in patients with SPMS, those with benign MS, traditionally defined as EDSS ≤3 and ≥15 years disease duration, are thought to follow a milder course (88). We recently conducted an analysis of a longitudinal MS cohort to determine the extent of visual pathway axonal loss by OCT RNFL thickness. At three academic centers, a subset of patients with EDSS scores, visual function, OCT, and QOL assessments was analyzed. Low- and high-contrast letter acuity were performed to assess visual function. RNFL thickness was determined using OCT-3. QOL scales included the NEI-VFQ-25 and SF-36. Among 68 patients (135 eyes) studied longitudinally, 13 (26 eyes) had benign MS using criteria of EDSS ≤3 and ≥15 years disease duration. Benign MS eyes had as much RNFL thinning (−3.6 µm, P=0.0008 vs. baseline, paired t-test) as typical MS eyes (−3.3 µm, P<0.0001). Both groups had significant low-contrast acuity loss over time. Prior history of optic neuritis (ON) was more frequent in benign MS (69% vs. 33% of eyes). History of ON distinguished benign vs. typical MS (P=0.002) and correlated with RNFL thickness at baseline (P=0.002) and disease duration (P=0.03), but not EDSS (P=0.32, logistic regression). NEI-VFQ-25 scores were also worse for benign MS, accounting for age (75±21 vs. 88±11, P=0.005). These findings demonstrated that patients with benign MS have RNFL axonal loss that is as marked as that of typical MS, and have reduced vision and QOL. While overall neurologic impairment is mild, visual dysfunction, not well-captured by the EDSS, accounts for a substantial degree of disability in benign MS.
In addition to axonal degeneration, neuronal loss is increasingly recognized as a correlate of disability in MS (16–19, 89–91). While investigations of time-domain OCT have shown reductions in total macular volume as a potential marker for neuronal loss in MS eyes, more specific measurement of the retinal ganglion cell layer (GCL) and other layers by segmentation has only recently emerged following the introduction of high-resolution and high-speed -domain (SD) OCT techniques to study of MS (Table 1) (25–27, 82, 92–100). Earlier studies of total macular volume using TD- OCT suggested that retinal ganglion cell neuronal loss occurs in MS eyes and that this correlates with visual function (81). These observations were necessarily based on the structural assumption that ~34% of the total macular volume is comprised of ganglion cells.
Using SD- OCT, some of the first observations of specific retinal ganglion cell layer (GCL) loss in MS eyes were generated by a study that involved manual delineation of the GCL on a Spectralis OCT platform (Fig. 4) (97). In this pilot investigation, eyes of patients with MS (n=16) had significantly lower GCL volumes than control eyes (P<0.001, accounting for age and within-patient, inter-eye correlations). Lower volumes were noted among MS eyes with a prior history of ON (n=4) compared to MS non-ON eyes (P<0.001). Reduced GCL volumes were not associated with worse high-contrast VA (P=0.14), but did predict visual loss by low-contrast acuity (P=0.003).
While manual OCT retinal segmentation demonstrated important findings in these studies, the labor-intensive nature of this technique (~2 hours per eye) limits its large scale use in MS cohorts. Computerized segmentation algorithms have been used successfully in studies of glaucoma (99, 100) and have provided the basis for two recent investigations in MS cohorts. Saidha et al. (96) used a segmentation method developed commercially to study patients with a macular thinning predominant (MTP) MS phenotype. These patients have peripapillary RNFL thickness that is within the normal ranges, but exhibit severe <5th percentile thinning of the macular region on spectral-domain OCT. This group of patient eyes demonstrated thinning of the outer retinal layers (P<0.001 for inner and outer nuclear layers in MTP vs. MS eyes), with relative sparing of the GCL. Patients with the MTP phenotype also had relatively greater degrees of overall neurologic disability, suggesting that these individuals may have a distinct or primary process underlying their neuronal loss.
Recent pathologic studies of post-mortem eyes from heterogeneous MS cohorts (n=82) demonstrated that the GCL is a site of neuronal dropout in 79% of eyes (101). Retinal ganglion cell loss has also been shown to occur in vivo in mouse models of relapsing-remitting experimental autoimmune encephalomyelitis (EAE) with optic nerve involvement, and to be reduced by neuroprotective therapies such as resveratrol in these models (102). To examine GCL loss in MS in vivo in patients with MS, our collaborative MS vision research group has used a novel computerized segmentation algorithm developed at the University of Pittsburgh (103). In this study, patients with MS (n=122 subjects, 239 eyes) and disease-free controls (n=31 subjects, 61 eyes) underwent Cirrus SD-OCT (Carl Zeiss-Meditec, Dublin, CA). Images were captured using Macular Cube (200×200 or 512×128) and ONH Cube 200x200 scanning protocols. Retinal layer segmentation was performed using algorithms originally designed for studies of glaucoma. Thicknesses of the ganglion cell layer/inner plexiform layer (GCL+IPL), RNFL, outer plexiform/inner nuclear layer (OPL+INL), and outer nuclear/photoreceptor layer (ONL+PRL) were measured and compared in MS vs. control eyes as well as MS ON vs. non-ON eyes. Since the IPL is a thin layer that is currently inseparable from the GCL using OCT, the combination GCL+IPL is used to estimate GCL thickness. Macular RNFL and GCL+IPL were significantly decreased in MS subjects vs. controls (P<0.001, P=0.001) and in MS ON eyes vs. non-ON eyes (P<0.001 for both measures). Peripapillary RNFL, macular RNFL, and GCL+IPL were all significantly correlated with VA (P≤0.001), 2.5% low-contrast acuity (P<0.001), and 1.25% low-contrast (P≤0.001). Among OCT measurements, reductions in GCL+IPL (P<0.001) and macular RNFL (P=0.006) were the most strongly associated with lower (worse) NEI-VFQ-25 and 10-Item Supplement composite scores for QOL; GCL+IPL thinning was significant even accounting simultaneously for macular RNFL thickness (P=0.03 for GCL+IPL and P=0.39 for macular RNFL). These data demonstrate that GCL+IPL thinning is most significantly correlated with both visual function and vision-specific QOL in MS. GCL thickness is likely to emerge as a useful structural marker of disease. These findings parallel those of MRI studies that show gray mater disease is a marker of neurologic disability in MS.
The incorporation of OCT and visual outcome measures into MS clinical trials has benefited from the presence of OCT reading centers. The University of California (UC) Davis Reading Center recently published the results of Stratus (time-domain) OCT quality control in two multi-center MS clinical trials (104). The authors evaluated 19,961 OCT scans from 981 patients with the goal of determining the influence of OCT quality control procedures on error rate. In Trial 1 (design and therapeutic agent not specified in publication), there was no ophthalmic technician certification and data were obtained by the Reading Center retrospectively. However, in Trial 2, technicians were certified and submitted data prospectively according to the study protocol. OCT scans in Trial 2 had higher signal strengths, fewer errors, and more usable data compared to Trial 1 scans. This study showed that certified technicians and prompt transmission of data for ongoing quality control monitoring provide higher data quality; these factors and the use of Reading Centers should be considered in the design of clinical trials for MS and other neuro-ophthalmologic disorders.
Visual dysfunction is not only an important contributor to impairment and disability in MS, but represents a unique opportunity for studying disease mechanisms and for testing new therapies that involve neuroprotection and repair. The advent of ocular imaging with OCT has allowed investigators to examine in vivo the morphological changes that accompany visual loss. Sensitive visual function tests, including low-contrast letter acuity, have been shown to correlate with OCT measures of axonal and neuronal loss as well as with patient-reported assessments of QOL. These observations have been instrumental in the establishment of a structure-function paradigm for using the anterior visual pathway as a model in MS. Low-contrast letter acuity, vision-specific QOL measures (NEI-VFQ-25 and 10-Item Neuro-Ophthalmic Supplement), and OCT measures have been incorporated into recent MS clinical trials. It is likely that these emerging data will yield important findings for therapeutics in MS, ON, and other neuro-ophthalmologic causes of visual loss.
Funding for Studies of Vision and OCT in MS: National MS Society PP1115, RG 3208-A-1, RG 3428-A/2, National MS Society Translational Research Partnership TR 3760-A-3, NIH/ NEI K24 EY 014136, R01 EY 019473 (Dr. Balcer), McNeill Foundation, and DADs Foundation.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest: Dr. Galetta has received speaking and consulting honoraria from Biogen-Idec, Novartis, and Teva. Dr. Balcer has received speaking and consulting honoraria from Biogen-Idec, Bayer, and Novartis.