MS Minute: Retinal Optical Coherence Tomography for MS

Sensitive, pragmatic biomarkers are needed to aid diagnosis, prognosis, and monitoring in demyelinating diseases (eg, multiple sclerosis [MS] and optic neuritis [ON]). Retinal optical coherence tomography (OCT) is a useful research tool increasingly being incorporated into routine clinical care for MS. OCT is noninvasive, inexpensive, rapid, highly reproducible, and able to quantify, on the scale of microns, discrete retinal layers. OCT has shown value for evaluating acute inflammatory events as well as quantifying insidious neural atrophy resulting from local or distant contiguous axon pathology in MS, as well as many other conditions with neurodegeneration.^1-3

OCT measures correlate with key MS outcomes, including but not limited to low-contrast visual acuity (VA) and disability;^4-6 response to disease-modifying therapy (DMT);⁷ whole brain, thalamic, and gray matter atrophy;⁸ future relapses;⁹ disability progression; and the likelihood of eventually meeting clinically definite MS criteria.⁵

This review, focused on practical applications of OCT in MS and ON, summarizes some of the existing evidence and is informed by experience at our center. Although not yet completely validated to guide clinical decision making in MS, OCT offers useful information in the context of other clinical factors. OCT measures with the most robust data for clinical use in demyelinating diseases are the peripapillary retinal nerve fiber layer (RNFL) and the macular ganglion cell-inner plexiform layer (GCIPL) thicknesses.^10,11

Retinal Anatomy and Physiology

OCT is analogous to high-resolution, light-based “ultrasound.” A light beam is directed at the retina and each retinal layer reflects that light differently. The resulting interference pattern is processed to render cross-sectional fundoscopic-like images (called a deviation map, Figure 1A), a 3-dimensional map, and retinal layer measurements (Figure 1B).

Key regions are the macula (central retinal area, with the fovea at its center) for GCIPL, and the peripapillary region (around the optic nerve head) for RNFL measurement. The RNFL is innermost and contains axons from retinal ganglion cells (RGCs) that meet at the optic disc, form the optic nerve, and then become myelinated. RNFL axons originate from the RGC soma that comprise the GCIPL. The inner plexiform layer contains dendrites and other structures that are technically difficult to distinguish from the RGC body layer. So, dendrites and soma are combined to make up the GCIPL, but this layer is thought of largely as representing the RGC soma (Figure 1C).

Why Study the Retina in MS?

The retina is the most accessible part of the central nervous system (CNS) and a frequent site of local inflammation, blood-retina barrier breakdown, and neuron (soma and axons) loss in MS and ON. In addition, the retina is affected by other conditions (eg, increased intracranial pressure [ICP] or global cerebral degeneration). Approximately half of people with MS experience clinically apparent ON over their lifetime,¹² but some optic nerve demyelination is virtually ubiquitous in MS.¹³ Although degeneration from optic nerve inflammatory demyelination is reflected in the retina, the retina itself is unmyelinated. Thus, it is an appealing target for assessing neurodegeneration without the confound of myelin or local inflammation. In those without a history of ON, retinal thinning tends to be less pronounced but correlates more closely with brain atrophy—highlighting the importance of measuring subclinical optic neuropathy.¹⁴ This is especially critical in progressive MS where axon loss is a poorly understood but key factor in the development of irreversible disability.¹⁵

RNFL and GCIPL measurements may provide information for diagnosis (eg, the presence of current or prior ON), monitoring (eg, response to DMT), and prognosis (eg, evidence of recovery after ON vs ongoing axon loss consistent with progressive disease).¹⁶ GCIPL thinning mirrors cerebral gray matter thinning and GCIPL changes vary by DMT.^7,8 GCIPL thickness appears to have a stronger correlation with disability compared with RNFL thickness and other measures in MS,^6,17 possibly owing to better GCIPL interscan reproducibility, RNFL gliosis masking atrophy, and RNFL edema during acute ON/retinal inflammation.¹⁸

Potential Uses of OCT in Clinical Neurology

Diagnosis

Differential Diagnosis of Neuroinflammatory Sequelae. In addition to use for diagnosing and monitoring acute ON as discussed later in this article, certain patterns raise concern for diagnoses other than MS or ON:

• more profound and diffuse atrophy seen in neuromyelitis optica spectrum disorders,¹⁹
• rapid RNFL swelling to >119 μ>m and thinning in myelinoligodendrocyte glycoprotein-antibody disease (MOGAD),²⁰
• patchy inner retinal thinning in Susac syndrome,²¹ and
• unusual findings suggesting other rare conditions (eg, marked symmetric diffuse thinning in genetic ON).²²

Diagnosis of Retrospective ON. Certain patterns (eg, predominantly temporal RNFL thinning)⁴ may suggest a history of ON or subclinical ON, although more data is needed before OCT can be relied upon as objective evidence of prior ON in the case of an unclear clinical history.²³ OCT also has potential to assist in risk stratification for conversion of radiologically isolated syndrome (RIS) to MS. The GCIPL in MS, even without a history of clinical ON, is significantly thinner than in normative data from age-matched people.¹⁷

Prognosis

OCT measures correlate with outcomes and future disability risk and may aid initial DMT selection as well as provide evidence for “aggressive” or progressive disease.

Baseline RNFL thickness predicts disability accumulation at 3 and 5 years. Baseline GCIPL thickness better predicts outcomes at 10 years than baseline RNFL thickness.^24,25

Monitoring

Evidence of Ongoing Degeneration. Because patients may compensate for ongoing degeneration and appear clinically stable, decreases in GCIPL thickness over time may help provide objective evidence of ongoing degeneration. This could be a datapoint to include when considering DMT changes.²⁶

Evidence of Progressive Disease. An accelerated rate of RNFL and GCIPL thinning may be seen in progressive MS.²⁷ This is being used as an outcome measure in clinical trials for association with biomarkers (eg, neurofilament and genes associated with disease progression). A consensus recommendation for reporting quantitative OCT results in clinical trials, the Advised Protocol for OCT Study Terminology and Elements (APOSTEL), has been published.²⁸

A Practical and Accessible Clinical Tool

OCT might be of use in locations or situations in which the expense and burden of MRI are not feasible and for those who cannot readily get MRI (eg, individuals with metal implants or devices, claustrophobia, prohibitive body habitus, or children with possibly raised ICP).

Expected Evolution of OCT Measures in ON

In acute ON, edema results from impaired transport in inflamed axons; this edema is reflected in the RNFL, with thickness peaking in the first month often higher than 110 μ>m (Figure 2). “Pseudonormalization” occurs as the edema resolves over the next few weeks. RNFL thinning usually becomes detectable 3 to 6 months after an acute attack of ON,^17,18 with approximately 90% of RNFL atrophy visible by 6 months.

Because the GCIPL does not swell significantly in acute ON, GCIPL thinning is apparent sooner, often within the first 5 weeks after acute ON¹⁷ and usually within the first 1 to 3 months.^19,30,31 GCIPL thinning at 1 to 2 months after acute ON may predict visual function at 6 months.¹⁸

Practical Implementation of OCT for MS Care

Staff must be trained to acquire OCT images, and, ideally, to also measure VA. The OCT technician should assess scan quality immediately to determine if repeat scanning is needed. In our center, these team members, often coordinators and postdoctoral research fellows, also screen the OCT and 3-dimensional map in concert with the ordering clinician for incidental ophthalmologic findings, which can affect reliability. There should be a process in place for rapid referrals or emergent evaluation by ophthalmology and neuro-ophthalmology in the case of urgent findings (eg, papilledema or retinal detachment). Several excellent resources delineate incidental findings²⁹ and comorbid and technical issues.³¹

Ideally, the device used for an individual should be from the same manufacturer for maximally consistent longitudinal comparison. Patients must be able to focus their gaze on a red dot for at least 2 seconds. OCT is likely unfeasible for those with very poor visual acuity, profound nystagmus, or an inability to cooperate with the evaluation (eg, a young child).

Expert consensus statements²⁸ note that pupillary dilation is not needed, particularly if the room is moderately dark; dilation may slightly decrease scan quality. For those having VA testing, we perform both 100% and 2.5% contrast VA and collect brief ophthalmologic history, including use of corrective lenses and indication.

We obtain OCT annually on most of our patients with MS. More frequent checks may be useful after acute ON, if there are clinical concerns for retinal edema (eg, with sphingosine-1-phosphate receptor [S1PR] modulator DMTs), or if a dramatic change was seen on recent prior OCT.

Challenges in OCT

Mitigating test-retest variation related to hardware and protocols as well as physiologic variation remains challenging. There are several potentially confounding ophthalmologic and systemic comorbidities (eg, myopia or diabetes). Scarred cells may impede accurate capture of functional atrophy. The RNFL and GCIPL are heterogeneous in the general population, and pediatric healthy control data is limited. In some cases, symmetric, diffusely high or low RNFL or GCIPL measures may simply represent a normal variant.

Interpretation of OCT in MS Care

OCT processing software automatically calculates the average thickness and thickness by sector of the RNFL, GCIPL, and macular cube with comparison to age-matched controls as seen in Figure 3 where all components of an OCT report are detailed. An online-only Table and Checklist outline our typical process for interpretation and generation of a clinical report.

OCT Quality Assessment

Staff should use the “OSCAR-IB” and recently expanded “OSCAR-AI”³² consensus approaches to quality assessment with OCT.³³ Training on these approaches is currently available online at no cost. At a minimum, 1 trained team member must review the entire 3-dimensional map. At least half of incidental findings are in the periphery and therefore unlikely to be captured in the representative centrally located B image generated for the report.²⁹ Some findings best seen on review of the 3-dimensional map may also be clinically relevant. For example, microcystic macular edema (MME), or microcystoid macular pathology (MMP), are hyporeflective cyst-like findings mostly seen in the inner nuclear layer (INL); these occur in approximately 5% of people with MS and may be associated with more severe disease course.³⁴

Conclusion

With its practical benefits and compelling evidence for use in the diagnosis, prognosis, and monitoring in MS, ON, and other conditions, routine clinical use of OCT in neurology is increasing. At present, OCT is a useful adjunct—akin to a quantitative ophthalmoscope. OCT measures offer additional data to help guide decision-making in the complex field of neuroimmunology and other cerebral neurologic conditions.