Multiple sclerosis: is it all black and white in optical coherence tomography?

This scientific commentary refers to ‘Brain and retinal atrophy in African-Americans versus Caucasian-Americans with multiple sclerosis: a longitudinal study’, by Gonzalez Caldito et al. (doi:10.1093/brain/awy245).

Multiple sclerosis is an inflammatory demyelinating disease of the CNS. Histopathologically, degenerative processes have also been identified early in the disease course and independent of the demyelinating white matter disease. MRI of brain atrophy is currently the gold standard to assess neurodegeneration in vivo. However, longitudinal MRI of brain atrophy is difficult at the individual level owing to great interscan variability. Thus, alternative proxies for neurodegenerative processes that can predict future clinical disability in patients with multiple sclerosis need to be identified. With the advent of optical coherence tomography (OCT), retinal layers can be assessed with high accuracy and low variability, and the ganglion cell/inner plexiform layer (GCIP) in particular appears to reflect brain atrophy in patients. With this tool, a variety of major issues in multiple sclerosis research can be addressed including mechanistic cascades of neurodegeneration remote from inflammatory foci and—even more important from a clinical perspective—the establishment of reliable prognostic markers that may help to stratify patients for distinct therapies. Other prognostic factors previously suggested to predict a milder or more severe disease course include age at onset, sex, first symptom, relapse rate and lesion load (Degenhardt et al., 2009; Tintore et al., 2015). Ethnic background has also been considered a determinant of the disease course (Shapira et al., 2010): although the incidence of multiple sclerosis is lower in African Americans than in Caucasian Americans, African Americans develop more aggressive multiple sclerosis as compared to Caucasian Americans, including faster pyramidal system involvement and more severe visual impairment (Phillips et al., 1998), consistent with a more pronounced lesion accumulation on MRI. In this issue of Brain, Gonzalez Caldito and Saidha and their co-workers use OCT to substantially add to our understanding of retinal and cerebral neurodegeneration in African American as compared to Caucasian American patients with multiple sclerosis (Gonzalez Caldito et al., 2018).

OCT is a non-invasive, safe and highly reproducible imaging tool for the measurement of retinal layers. Since its introduction in 1991, OCT has become an important tool for the assessment of retinal pathology. In patients with multiple sclerosis, the retinal nerve fibre layer (RNFL) loses thickness at a greater rate compared to control individuals even in the absence of clinically manifest optic neuritis. Low RNFL thickness is an indicator of disability worsening in patients with multiple sclerosis (Martínez-Lapiscina et al., 2016). In addition, GCIP atrophy correlates well with brain substructure atrophy on MRI (Saidha et al., 2015), which is probably a consequence of retrograde trans-synaptic axonal degeneration, and clinically mirrors disability progression. A negative correlation between GCIP thickness and intrathecal B cell immunity has been demonstrated (Knier et al., 2017), laying a foundation for building hypotheses of potential pathophysiological events resulting in neuronal atrophy. The inner nuclear layer (INL) on the other hand, mostly reflects inflammatory disease activity. Thickening of the INL correlates with MRI parameters that indicate inflammatory activity such as new or enlarging T₂ lesions. The occurrence of microcystoid macular pathology (MMP) in the INL is also associated with worsening of Expanded Disability Status Scale (EDSS) scores (Gelfand et al., 2012). Consequently, reduction and normalization of INL volumes might increase the likelihood of achieving the status ‘no evidence of disease activity’ in patients with multiple sclerosis under appropriate treatment (Knier et al., 2016).

Figure 1

Visual pathway and changes upon inflammation and degeneration in MRI and OCT. (A) Schematic diagram of the visual pathway including retinal ganglion cells and their axons assembling the optic nerve. From here, the signal is transmitted to neurons in the lateral geniculate nucleus (LGN) and is further conducted via the optic radiation to neurons in the visual cortex. Retrograde trans-synaptic axonal degeneration is displayed as a model for secondary damage and consecutive atrophy of inner retinal layers after axonal injury in the optic pathway. Dashed lines indicate indirect effects. (B) Inflammatory changes in multiple sclerosis result in hyperintense fluid-attenuated inversion recovery (FLAIR) lesions on MRI and thickening of the INL, which is occasionally accompanied by microcystoid macular pathology (MMP, between arrow heads in the INL). MMP appeared more frequently in African American (AA) patients with multiple sclerosis than in Caucasian American (CA) patients. (C) Degenerative changes in multiple sclerosis displayed by whole brain volume decrease in T₁-weighted multiplanar reconstructed (MPR) MRI and by atrophy of inner retinal layers, particularly the common layer of the ganglion cells and inner plexiform layer (GCIP).

Open in new tabDownload slide

At present, MRI is the most widely used method to visualize inflammatory activity in the brain and spinal cord of patients with multiple sclerosis through the detection of new and enlarging T₂ lesions and contrast-enhancing lesions. As such, MRI has been established as the gold standard with which to monitor treatment response to immunomodulatory therapies in all pivotal trials. However, worsening of disability and underlying neurodegenerative processes are still difficult to capture by conventional MRI. Thus, retinal OCT could be a suitable tool to fill the niche of monitoring neurodegeneration in the course of multiple sclerosis.

Here, Gonzalez Caldito et al. applied a multimodal approach to assess longitudinal regional (grey matter, white matter and nuclear thalamic) brain volume loss by MRI and retinal layer changes in patients with multiple sclerosis. They ingeniously chose to investigate a cohort of patients, i.e. African Americans, in whom epidemiological data suggest a particularly rapid disease course, in order to generate prospective longitudinal data with large effect sizes. They also analysed whether distinct patterns of brain or retinal atrophy are detectable in African American patients as compared with Caucasian American patients. Longitudinal MRI data were retrieved from 22 African American and 60 Caucasian American patients. In the OCT arm of the study, 116 patients were enrolled in each of the African American and the Caucasian American cohorts and followed for a mean of 4.0 (African American patients) and 5.0 years (Caucasian American patients), respectively.

African Americans with multiple sclerosis showed increased rates of atrophy on MRI compared to Caucasian American patients, with approximately doubled atrophy rates in cortical grey matter, white matter and nuclear thalamic volumes. At the same time, the T₂ lesion volume also increased twice as fast in patients of African American versus Caucasian American ancestry. In OCT, Gonzalez Caldito et al. observed thicker peripapillary RNFL (pRNFL) values in healthy African Americans than in healthy Caucasian Americans. However, the difference between African Americans and Caucasian Americans was reduced and in fact almost inverted in the multiple sclerosis patient cohorts, suggesting a more pronounced thinning of pRNFL in African American individuals with multiple sclerosis than in Caucasian American patients. In accordance with greater pRNFL thinning in African American patients, GCIP thickness was lower in African American patients than in Caucasian American patients at baseline and diminished at about twice the rate in African American patients as compared to Caucasian American patients. In contrast, the INL—the thickness of which is positively correlated with inflammatory activity in multiple sclerosis—increased in a more pronounced manner in African American patients as compared with Caucasian American patients. At the same time, other changes in INL, namely MMP, were detected in 12.1% of African American patients and only 0.9% of Caucasian American patients. This is an intriguing finding and might be at least partly responsible for the more pronounced thickening of the INL in African American patients with multiple sclerosis in this study.

Overall therefore, Gonzalez Caldito et al. provide evidence that various MRI and OCT parameters that have been associated with neurodegenerative processes and with inflammation, respectively, deteriorate faster in African American than in Caucasian American patients. Together, these findings might help to explain the hazard ratio of 1.96 in time to walking aid dependence of African American versus Caucasian American patients with multiple sclerosis (Khan et al., 2015).

The study also has some limitations. The classification of African American and Caucasian American status by self-reporting without accounting for mixed ancestries might dilute relevant characteristics between the two groups. Additionally, the inclusion of different and non-matched multiple sclerosis phenotypes, lack of knowledge of the heterogeneous geographical provenance of the African American individuals, and the lack of correction for socio-economic status as an individual risk factor for poorer disease course in multiple sclerosis, might result in bias and should be considered in future studies.

These limitations aside, the authors present a comprehensive longitudinal study of cerebral and retinal changes measured by MRI and OCT, respectively, with a long follow-up duration in African American versus Caucasian American patients with multiple sclerosis. We can learn exciting lessons from these data: first, the pattern of atrophy in African American versus Caucasian American patients with multiple sclerosis is similar (only faster in African Americans as compared to Caucasian Americans) and globally affects grey matter, white matter and thalamic nuclei in both ancestries. Second, the higher rate of brain volume and GCIP loss in African American patients is associated with variables (both in MRI and OCT) that indicate more pronounced inflammatory activity of the disease. Thus, neurodegenerative processes do not appear disconnected from inflammation but depend on it both in African American and Caucasian American patients. Third, the accelerated progression of the disease in African American patients is not due to a greater susceptibility of the target tissue (the CNS) to neurodegeneration in African Americans, but appears to be driven by a more vigorous inflammatory process in African American patients as compared to Caucasian American patients with multiple sclerosis. As a next step, the integration of biological markers of CNS tissue atrophy, such as neurofilament light chain levels in serum (or CSF), with MRI and OCT data might be a missing link that could bring us closer to a mechanistic understanding of how inflammation relates to neurodegeneration in human CNS autoimmunity. Finally, this study emphasizes the usefulness of OCT for the measurement of atrophy and the monitoring of both inflammatory and neurodegenerative processes in multiple sclerosis. If OCT is an appropriate method to reliably detect differences as subtle as in these cohorts with different racial background, it might also help us to assess how variables that indicate inflammation and neurodegeneration behave in later stages of multiple sclerosis, in which the link between inflammation and neurodegeneration is less firmly established than at disease onset.

Competing interests

The authors report no competing interests.

References