In the present study, a detailed microstructural evaluation of retinal layers was performed on en face imaging in eyes with macular holes. A comprehensive assessment showed the presence of intraretinal cystic changes in the inner and outer retinal layers, namely the INL and OPL, respectively. Although the area of these cystic changes was significantly larger in the IPL as compared to the OPL, the percentage of cysts was more in the OPL. At the same time there was a positively correlation between the area of the cysts in both the layers, implying a simultaneous congruous enlargement of cystoid spaces in both layers. Interestingly, an increase in the basal diameter of the hole was associated with a simultaneous enlargement in the area of the cyst in the IPL and the OPL respectively and not related to percentage of cysts within these layers. The inlet diameter correlation with cyst area or percentage of cysts was however not so straightforward.
Presence of intraretinal cystoid cavities have been previously reported with various nomenclatures such as ‘intraretinal cavities’, ‘cystoid changes', and ‘cystoid spaces’.15–17 However, the pathogenesis of intraretinal cysts remains a mystery. In 2015, Matet et al evaluated the cystic changes around the macular holes on en face imaging.13 They reported a dual profile of cystic cavities in the retina, with elongated stellate pattern in OLP and Henle fiber layer (HFL) and a rounded vertical cylindrical pattern in INL, respectively. This arrangement of cystic spaces corresponded to the “Z-shaped” configuration of Müller cells.13, 18 They proposed that these cystic spaces result from either fluid collection in between the retinal framework of Müller cells and Henle fibers or from intracellular fluid accumulation within the Müller cells.13 The retinal pigment epithelium (RPE) pump is vital for keeping the overlying neurosensory retina (NSR) attached and maintaining the intraretinal and subretinal spaces dry.19 In the presence of MH, there is loss of contact between the RPE and the NSR, thus compromising the normal intraretinal physiology. Failure of the RPE pump to act upon the overlying detached retinal tissue can conceivably lead to collection of intraretinal fluid in the form of IPL and OPL cysts. In our study, we observed a positive concordance between the area of the cysts in IPL and OPL along with their synchronized enlargement with expanding basal diameter of the macular hole. In other words, wider the basal diameter of the MH, i.e., wider the area of RPE contact loss, more was the associated intraretinal cystic fluid collection. Thus, we propose that the intraretinal cysts probably arise due to altered retinal physiology due to RPE separation which we label as the “RPE contact loss” theory. Additionally, these cysts are known to disappear rapidly during the early post-operative period, thus supporting the hypothesis of underlying RPE contact loss as the etiopathogenesis for intraretinal cystoids. Moreover, the micromorphic location of these cysts is presumably extracellular, as an intracellular location within the Müller cells would unlikely result in their speedy disappearance post-operatively.
Our study demonstrated a greater area of cysts in the INL as compared to the OPL. Thus, the INL layer contained cysts that were significantly more widespread in nature. Conversely, Matet et al found the mean area of cysts significantly lower in the INL as compared to the OPL and the Henle’s fiber layers.13 However, they did not evaluate the percentage of cysts whereas in our study, we observed that the percentage of cysts in the OPL was significantly more than the INL, i.e., in the OPL the cysts were significantly larger in dimensions as compared to the INL cysts. Rizzo et al have shown similar results of small round cysts in the INL and larger stellate cysts in the OPL.20 Their group also demonstrated that OCTA slabs underestimate INL cyst area as compared to OPL cyst area, especially on automated segmentation and proposed that en face imaging is superior for identification of INL cysts as compared to optical coherence tomography angiography (OCTA).20 To overcome these limitations, in our study we utilized manual segmentation on en face OCT for accurate imaging analysis using two independent graders. The larger percentage of cysts observed in the OPL is suggestive of either early involvement of this layer with / without associated structural weakness in this layer. Early involvement of OPL is supported by the “RPE contact loss” hypothesis which we propose for the pathogenesis of cystoid changes. Additionally, the configuration of the Müller cells which are arranged obliquely in the OPL / Henle’s layer of the foveal region, makes this layer highly vulnerable for both MH formation secondary to tractional forces and development of larger cysts.13, 21, 22 At the same time, the Müller cells are vertically oriented in the inner retinal layers,21, 22 making them structurally more stable and accounting for smaller size cysts in IPL noted in our study.
On stratification of eyes based on visual outcomes, we noted a similar finding of cystic involvement in eyes with worse visual outcomes, namely IPL showing a larger area of cyst involvement while the OPL demonstrating a greater percentage of cystic spaces. Although the proportion of cystic spaces were greater in OPL in eyes with good visual outcomes, we did not note any difference in the area of cyst involvement amongst the INL and OPL in these eyes. These cystic changes in the INL form the crux of the hydration theory and the associated degenerative inner retinal changes.23 In the late 1980s, Gass proposed that hole formation was secondary to tangential traction exerted by the vitreous cortex at the foveal edges.24 This tangential traction is associated with the splitting of glial material in the inner part of the retina leading to hole formation. In 2003, Tornambe et al proposed that macular hole progression occurs due to destabilization of the underlying retina with progressive hydration of retinal layers.23 Shahlaee et al also provided supplementary evidence for "hydration theory" in their study related to en face based quantitative assessment of intraretinal cystic changes in macular holes.4 A greater area of cystic involvement of inner layers in our study is suggestive of supplementary degenerative changes and is compatible with the “hydration theory”.23 Nevertheless, we did not observe any significant difference in the cyst area between INL and OPL in eyes with good visual outcomes. This is indicative of lesser degenerative damage to the inner retina amongst these eyes, and subsequently a better visual outcome after the surgery.
Previous reports have shown multiple factors such as age, pre-operative visual acuity, size of the MH, stage of MH, ELM integrity, and EZ integrity which may affect the post-operative visual recovery in MH surgery.6, 7, 25, 26 The ELM is located at the junction between the cell bodies and the inner segment of the photoreceptors and represents the junctional complexes between the Müller cells and the photoreceptors.27 Its integrity plays a critical role in the functioning of photoreceptors and thus visual restoration.6 The gliotic proliferation of Müller cells is critical not only for the closure of the macular hole but also to restore the integrity of ELM.28, 29 Madreperla et al have illustrated the closure of ELM defect by Müller cell processes in an eye operated for stage III macular hole.30 Studies have shown that early postoperative restoration ELM is associated with better visual outcomes and is critical for the restoration of the photoreceptor layer.31 Similar to these previous reports, where the integrity of ELM and EZ formed a critical part in visual recovery post macular hole surgery,6, 7, 31 our findings suggest that the ELM and EZ defects were a significant predictor of final visual outcomes. Additional, other parameters including the pre-operative visual acuity, the basal diameter of MH, the minimal inlet of the MH, the area of cysts in INL and OPL were observed to be significant predictors of final visual outcomes in our study.
Our study has few limitations including retrospective design and shorter follow-up. Also, a lack of comparison with post-operative en face imaging is a shortcoming. Further long-term studies with a comparison of foveal micromorphic structures on en face imaging at follow-up visits to evaluate the changes in intraretinal cysts and restoration of retinal layer integrity would be needed to corroborate our findings. Nonetheless, the strengths of our study include the adequate sample size and performing a detailed morphometric evaluation of retinal structural changes in MH eyes. Also, to the best of our knowledge, this is the first study to perform the en face imaging analysis based on manual segmentation. This is vital as it eliminates any potential segmentation errors arising from automated segmentation slabs. Moreover, all imaging analysis was performed by two graders independently with a very good intergrader agreement of 0.81.
To conclude, in our study, we characterize the intraretinal cystic changes in various retinal layers and correlate the en face imaging features with visual outcomes in macular holes. Correlation of the cystic area dimension in the OPL and IPL along with their relationship to the basal diameter of MH supports the theory of “RPE contact loss” for the origin of cystic changes. Additionally, the role of Müller cells is critical in pathogenesis and recovery of macular holes as evident by en face features such as the greater percentage of cysts in the OPL, larger area of cyst involvement in IPL, and ELM and EZ defects being critical predictors of final visual outcome. The other factors affecting the final visual acuity include the pre-operative visual acuity, the basal diameter of MH, the minimal inlet of the MH, and the area of cysts in IPL and OPL, respectively. Further imaging studies and histopathological corroboration is needed to elucidate these architectural retinal changes and their long-term functional impact.