Optic nerve magnetisation transfer ratio after acute optic neuritis predicts axonal and visual outcomes

Abstract

Magnetisation transfer ratio (MTR) can reveal the degree of proton exchange between free water and macromolecules and was suggested to be pathological informative. We aimed to investigate changes in optic nerve MTR over 12 months following acute optic neuritis (ON) and to determine whether MTR measurements can predict clinical and paraclinical outcomes at 6 and 12 months. Thirty-seven patients with acute ON were studied within 2 weeks of presentation and at 1, 3, 6 and 12 months. Assessments included optic nerve MTR, retinal nerve fibre layer (RNFL) thickness, multifocal visual evoked potential (mfVEP) amplitude and latency and high (100%) and low (2.5%) contrast letter acuity. Eleven healthy controls were scanned twice four weeks apart for comparison with patients. Patient unaffected optic nerve MTR did not significantly differ from controls at any time-point. Compared to the unaffected nerve, affected optic nerve MTR was significantly reduced at 3 months (mean percentage interocular difference = -9.24%, p = 0.01), 6 months (mean = -12.48%, p<0.0001) and 12 months (mean = -7.61%, p = 0.003). Greater reduction in MTR at 3 months in patients was associated with subsequent loss of high contrast letter acuity at 6 (ρ = 0.60, p = 0.0003) and 12 (ρ = 0.44, p = 0.009) months, low contrast letter acuity at 6 (ρ = 0.35, p = 0.047) months, and RNFL thinning at 12 (ρ = 0.35, p = 0.044) months. Stratification of individual patient MTR time courses based on flux over 12 months (stable, putative remyelination and putative degeneration) predicted RNFL thinning at 12 months (F(2,32) = 3.59, p = 0.02). In conclusion, these findings indicate that MTR flux after acute ON is predictive of axonal degeneration and visual disability outcomes.

Figure 1. Serial measurements of MTR in patient affected and unaffected optic nerves during the 12 months following acute optic neuritis.

Significant differences were observed at three, six and 12 months (** indicates p<0.01 for paired t-test). The control distribution is indicated on the left for comparison. Errorbars indicate one standard error of the mean (SEM).

Figure 2. Individual patient MTR asymmetry time courses (grey lines) and mean fitted time courses (black lines) for three subgroups defined based on the pattern of MTR flux.

*p<0.05.

Figure 3. Scatterplots showing distributions for RNFL thinning (upper) and mfVEP amplitude loss (lower) based on subgroups.

*p<0.05.