The present study showed that the ONL layer around the macular area is naturally horizontally long, and that it is more prominent in ERM patients than in normal controls.
The photoreceptor layer, the first set of neurons involved in light perception, consists of the retinal pigment epithelium (RPE), outer segment,(OS), inner segment (IS), cell bodies, and long axons. SD-OCT enables the visualization of these structures in vivo via the white and black intensity images. A recent immunohistogical study has shown that the hypo-reflective wide layer over the ELM line is considered to be the cell body and the long axons of the photoreceptor.15 In that study, the ONL-B was around 150 µm from the RPE line, which indicated that the ONL-B was evaluated at the level of the cell body and the Henle nerve fiber layer of the photoreceptor.15
The present study showed that the ONL-B is naturally horizontally long, and it is more evident in ERM patients (Fig.3). This horizontal long structure is consistent with the form of the nerve fiber layer, the temporal fibers are dispersed widely, and a small percentage of arcuate fibers are dispersed horizontally.16 In the present study, the ONL-B is more evident in ERM patients, the average axis was 2 degrees from the horizontal line, and 90% of ERM patients showed an axis within ±45 degrees, which may be due to the ERM contraction direction. Our colleague recently reported that the vertical retinal displacement was greater than the horizontal retinal displacement, as assessed by measuring the position of the retinal vascular in ERM patients17, which indicates that retinal morphology is likely to deform to be horizontally long with ERM vertical contraction. The results of the present study are consistent with this observation.
Classically, the origin of metamorphopsia was considered to be the displacement of the cones, which in turn, causes a false localization of the image.18 Macropsia and micropsia, special subtypes of metamorphopsia, show that the retinal displacement is correctly logically associated with the symptom, outward retinal displacement-derived macropsia,19 and inward retinal displacement-derived micropsia.20 Moreover, Amsler suggested that outer retinal and choroidal changes cause metamorphopsia and scotoma.1
Many clinical studies using retinal images have investigated the cause of metamorphopsia,21 and several previous studies support that changes of the photoreceptor layer are the cause of metamorphopsia. Okamoto et al. reported that preoperative MCHART score was significantly correlated with the ONL + OPL thickness using SD-OCT iamges.9 Ooto et al. reported that cone morphologic feature patterns in the foveal photoreceptor layer was associated with metamorphopsia, as assessed using adaptive optics scanning laser ophthalmoscopy images.22
In the present study, circularity and area ratio of the ONL-B was significantly correlated with MH/MV. In the MCHART examination, subjects recognize the misalignment of dots in each examination line. Therefore, we hypothesized that as the vertical contraction of the retina only affects the dot spacing in MV, and does not result in a lateral slip of the dot, and thus, it does not increase MV. On the other hand, it is assumed that the vertical contraction of the retina increases MH because it induces lateral displacement of the dot of MH. Conversely, contraction in the horizontal direction affects MV score mainly. We assumed that the MH/MV value is high and that the longitudinal deformation of the retina is greater than the lateral deformation of the retina.
Circularity is the parameter indicating how the ONL is close to a perfect circle, and area ratio is the parameter indicating whether the contour line of the ONL-B is irregular. MH/MV is negatively correlated with circularity, which indicates that an increase in MH influences the deformation of the ONL in the vertical direction. Area ratio is higher in ERM patients, which indicates that the ONL in ERM patients has a more complicated deformation with contraction.
We found that the correlation between the metamorphopsia parameters and morphological characteristic parameters is more significant in cases where the long axis is close to vertical (which is ±10°). This result may indicate that metamorphopsia contains different direction components of distortion, and that the MCHART would be affected by the component orthogonal to their axis. Arimura et al, who developed the MCHART test, reported a significant correlation between horizontal contraction of the retina and the MV score and between vertical contraction of the retina and the MH score by measuring the superficial retinal position 23, and our result is consistent with their report.
We observed the characteristics of the outer retina and their effect on metamorphopsia. However, several recent reports documented that the inner nuclear layer was significantly correlated with metamorphopsia.9 19 Moreover, we observed metamorphopsia in patients with diseases of the superficial retinal layer, such as small branch vein occlusion and superficial small ERM. Therefore, we think that the mechanism of metamorphopsia involves multiple factors, including the conscious effect,24 and the pathology of metamorphopsia is different in each disease.
ERM could be classified on the basis of macular morphology into retinal thickening, pseudo hole, and lamellar hole.25 In the subtype of retinal thickening, which is the most frequently observed type of ERM, membrane-attached macular and tangential traction destroyed the macular morphology. In contrast, fovea-sparing ERM types, such as the pseudo hole type, the membrane does not cover the foveal area and has minimal influence on macular morphology; it has also been reported to show well-preserved foveal function. We included only patients with the retinal thickening type in the present study to focus on the influence of deformation direction on metamorphopsia direction.
There are limitations in this study. This was a retrospective study on a limited number of participants. The IOWA Reference Algorithm does not work well automatically for a region where the ONL steeply rises. Therefore, we had to perform manual segmentation for these parts. New algorithms may be needed to enable segmentation of OCT images in diseased eyes.