All OCT studies have shown that mean peripapillary RNFL thickness is decreased in MS patients compared to normal subjects. Greater RNFL thinning occurs in MS patients who are not treated with immunomodulatory agents [11]; this is unsurprising, since disease-modifying agents reduce the number of MS relapses and, therefore, would be expected to reduce axonal damage. The magnitude of RNFL atrophy is greater in the eyes with a history of prior acute MSON; on average, the RNFL thins by about 20 % following an attack of acute MSON; additional episodes of MSON result in more severe RNFL loss [12, 13].

In the immediate aftermath of acute MSON, the RNFL appears thicker due to inflammatory edema; this swelling normally resolves within about 6 weeks [14, 15]. This early reduction in RNFL thickness may be mistaken for atrophy and has therefore been termed “pseudoatrophy”; current technology does not permit differentiation between axonal loss (i.e., true atrophy) and inflammation-related pseudoatrophy. The majority of patients experience significant RNFL loss within 1 year after an attack of acute MSON; most of this loss occurs within 3–6 months following the event and stabilizes at between 6 and 12 months depending on the area analyzed [15, 16].

Consistent with neuropathological studies that demonstrate optic nerve demyelination in virtually all MS patients, the RNFL atrophies by approximately 2-μm per year in most patients, regardless of history of prior acute MSON [17]. The earliest and most prominent RNFL loss occurs in the temporal quadrant [18–20], corresponding with the fibers of the papillomacular bundle, which mediates central vision. RNFL loss is the result of retrograde retinal ganglion cell axon degeneration following inflammatory demyelination of the optic nerve [21]. Alternately, damage to the retrochiasmatic visual pathways can also cause RNFL thinning as a consequence of retrograde transsynaptic degeneration of the retinal ganglion cells and their axons [22–27].

A robust relationship between RNFL atrophy and visual function is well established. RNFL loss correlates with reduced Snellen (high contrast) visual acuity in most studies [14, 16, 28–34], but not all [35, 36]. Reduced visual acuity logMAR scores are also associated with RNFL atrophy [18, 37–40]. Interestingly, RNFL atrophy may be present despite 20/20 Snellen visual acuity, suggesting that RNFL thickness is a more sensitive marker of MS disease activity, compared to high-contrast visual acuity. Alternately, low-contrast letter acuity (a more sensitive marker of visual dysfunction than high-contrast visual acuity) correlates better with RNFL loss [33–35, 38, 39, 41]—for every one line of low-contrast letter acuity lost, the mean RNFL thickness decreased by 4 μm [35]. Additionally, loss of RNFL correlates with other measures of visual function, including impaired color vision [42] and visual field loss [14, 16, 27, 32, 38, 40]. Once the RNFL thickness falls below 75 μm, every 10 μm drop in RNFL is associated with a 6.46 dB decrease in visual field mean deviation (a value that gives the average of differences from normal expected value for the patient’s age) [16]. However, visual fields may be normal despite the presence of RNFL atrophy [28], confirming that visual field testing may not be sensitive enough to detect retinal ganglion cell damage until it is at an advanced stage. Furthermore, from the glaucoma literature, it is well known that about 25–35 % of retinal ganglion cells need to be lost prior to detection of visual field loss [76–78].

RNFL loss correlates with perturbations in visual electrophysiology as well. Visual evoked potential (VEP) amplitudes are reduced in patients with RNFL atrophy, the consequence of axonal loss in MS [32, 37, 43–46]. Several studies have also demonstrated an association between prolonged VEP latencies (the electrophysiologic hallmark of visual pathway demyelination) and RNFL loss [6, 32, 45, 47]. Similar findings have been found using multifocal VEP [45, 48]. Furthermore, RNFL atrophy correlates with abnormalities of the pattern electroretinography P50 latency as well as P50–P95 amplitude [6], the likely consequence of inner retinal degeneration resulting from optic nerve demyelination. However, electrophysiologic dysfunction may precede any detectable RNFL atrophy [31, 49]. Therefore, it remains possible that early pathologic changes that begin at a cellular level may escape the detection of even the most sensitive, high-resolution, technologically advanced OCT.

MS-related changes in the visual pathway mirror the pathological damage sustained by other anatomically distinct neural structures; as such, RNFL atrophy has been shown to correlate with a host of other markers of neurological impairment in MS and provides a window into disease processes. For instance, RNFL thinning correlates with the severity of neurologic disability and loss of quality of life in MS patients [18–20, 36–38, 50–53]. In the clinically isolated syndrome, RNFL atrophy correlates with cognitive impairment [54]. In addition, RNFL loss correlates with changes in other radiologic markers of disease in MS, particularly those related to brain atrophy [36, 51, 52]. Specifically, RNFL loss has been correlated with T1 lesion volume [51], T2 lesion volume [20, 50, 51], normalized brain volume [20, 50], brain parenchymal fraction [51, 52], gray matter volume [20, 51], white matter volume [20], magnetization transfer ratio [50, 55], bicaudate ratio [56], fractional anisotropy, and diffusion tensor imaging [30, 50], as well as an increase in CSF volume [52]. The correlation between RNFL atrophy and radiologic markers of the disease is tantalizing; for monitoring disease progression, it is possible that OCT metrics may one day be used in the place of many of these time-consuming, technically demanding, and costly radiologic studies in the clinical setting.

While RNFL loss represents axonal damage, total macular volume (TMV) loss indicates neuronal death since retinal ganglion cells constitute a large population of the macula. TMV loss correlates with RNFL thinning in MS patients; when compared to healthy controls, TMV is reduced in MS patients, especially in the eyes affected by prior acute MSON [37]. A 10-μm difference in RNFL thickness corresponds to a 0.20-mm³ reduction in TMV [57]. TMV loss has been correlated with impaired ambulatory capabilities in MS patients, as measured by the timed 25-ft and the 6-min walk [58]. On the other hand, a subset of MS patients with more severe disease demonstrate evidence of primary retinal neuronopathy (termed “macular thinning predominant phenotype”), evidenced by loss of TMV that predominantly affects the outer retina instead of the inner retina [59, 60]. This fascinating observation suggests that the inflammatory changes associated with MS optic neuropathy may begin in the retina in some patients; similar findings have been shown in murine models of MS [61].

Advancements in segmentation technology have improved the resolution of OCT devices and enabled measurements of each layer of the retina, in addition to the RNFL. This has led to the discovery that ganglion cell–inner plexiform layer (GCIPL) thickness correlates better with visual function and radiologic markers of MS disease activity, compared to RNFL thinning [62–67]. Indeed, GCIPL atrophy may be a more accurate measurement of the damage sustained from attacks of acute MSON, since unlike the RNFL, GCIPL thickness is unaffected by inflammatory edema in the acute stages of acute MSON [65]. Furthermore, GCIPL atrophy occurs rapidly following acute MSON (even in a matter of days) and is marked at 1 month following the ictus; in fact, GCIPL atrophy precedes macular volume and macular RNFL loss [80]. Indeed, GCIPL loss in the first month following acute MSON is a good predictor of short-term visual dysfunction. A loss of less than 4.5 μm predicts low-contrast acuity recovery; if the loss does not exceed 7 μm, good visual field and color vision recovery will usually take place [80]. Additionally, improved OCT resolution has enabled the identification of small vacuolar inclusions in the inner nuclear layer, the so-called macular microcysts, which appear to be confined to regions with corresponding RNFL and GCIPL atrophy and are associated with a characteristic perifoveal crescent of punctuate changes that spares the fovea centralis on fundoscopic scanning laser ophthalmoscopy [68, 69]. These changes were first reported in MS patients and neuromyelitis optica (NMO) patients with prior acute MSON [70, 71] and attributed to retinal inflammation. Since then, however, macular microcysts have been identified in optic neuropathies from many other etiologies, including ischemia, compression, glaucoma, trauma, hereditary causes (dominant optic atrophy, Leber’s hereditary optic neuropathy), non-demyelinating inflammation (chronic relapsing inflammatory optic neuropathy), toxic–metabolic etiologies, and hydrocephalus [68, 69, 72–74], as well as in optic tract lesions [68]. Therefore, macular microcysts are not specific for inflammatory pathologies but may occur in conditions that result in retrograde degeneration of the retinal ganglion cells.

The OCT may be used as a clinical tool for evaluating suspected ocular pathology in MS patients with visual complaints. We have found OCT to be very useful in monitoring patients on fingolimod for evidence of macular edema—an infrequent but serious adverse effect of the drug. Alternatively, OCT is very useful for evaluating for non-MS-related ocular or retinal disorders that may affect patients (e.g., epiretinal membranes, posterior vitreous detachment, neuroretinitis). We have also detected incidental ocular diseases in MS patients, including a few cases of ocular bartonellosis. The ability to detect such conditions in the MS clinic would allow timely ophthalmologic referral and institution of interventions or therapies that could potentially save the patient from serious visual loss.

As discussed in the preceding text, there is a robust correlation between OCT metrics and visual function in MS both cross-sectionally and longitudinally; OCT is a promising tool for screening and studying neuroprotective and repair-promoting therapies [79]. Furthermore, sample sizes needed to demonstrate the neuroprotective effects of novel therapies may be smaller if OCT metrics were used as outcome measures. Using serial RNFL measurements from their study, Henderson et al. [15] generated sample sizes for putative clinical trials of acute neuroprotective agents for acute MSON and suggested that such trials could be limited to 6 months since RNFL thinning appears to stabilize beyond this point. Overall sample sizes for detecting a moderate 50 % neuroprotective effect at 6 months—inferred from a 50 % reduction in the amount of RNFL loss compared to the fellow eye in a trial comparing active and placebo arms—are ∼35 per arm (95 % CI ∼20, 70) with 80 % power and ∼50 per arm (CI ∼30, 110) with 90 % power. The equivalent sample sizes for a lesser 30 % neuroprotective effect at 6 months are ∼100 (CI ∼55, 200) and ∼130 (CI ∼70, 300) per arm [15]. For neuroprotective trials using GCIPL thickness as an outcome (based on an end point and optimal power as a loss of GCIPL thickness of less than 4.5 μm 1 month following an attack of optic neuritis), Gabilondo et al. [80] calculated that 25 patients with acute MSON per group would suffice to detect differences between the groups.

Indeed, OCT is beginning to be used as the primary or secondary outcome measure for several therapeutic trials related to MS. For example, anti-IL-17 antibodies were shown to prevent RNFL and GCIPL loss in experimental autoimmune encephalomyelitis mouse studies [81]. In MS clinical trials, OCT metrics are currently being used as paraclinical outcomes for ongoing studies of erythropoietin in optic neuritis, as well as two potential neuroprotective agents (phenytoin and amiloride) [82]. A randomized placebo-controlled phase II trial of autologous mesenchymal stem cell use in MS demonstrated a trend toward fewer gadolinium-enhancing lesions but failed to show any significant changes in OCT data (peripapillary RNFL thickness and macular volume) [83].