Based on our clinical observations, the closure rate of large macular holes remains unsatisfactory. Given the significant role of EZ recovery in visual improvement, it is possible that EZ-related parameters have the potential to predict the early outcome of surgery. However, we found a lack of studies examining the relationship between EZ-related parameters and the early evaluation of the success of primary macular hole surgery. In our previous study, we discovered that the recovery of the EZ itself, rather than EZ-related angles, was significantly associated with postoperative visual acuity in patients with successfully closed macular holes [10]. In this study, we aimed to further explore the predictive value of EZ-related angle parameters in the successful closure of large macular holes. Our results revealed that preoperative EZ-NFL, EZ-GCL, EZ-INL, EZ-OPL, and EZ-ONL angles were significantly larger in patients with closed macular holes compared to those with unclosed macular holes after surgery. Additionally, we introduced a new parameter called angle regularity, which represented the degree of irregularity in the aforementioned angles across four directions. It was observed that the AR of unclosed patients was significantly larger than that of closed patients, particularly in the nasal direction. This may suggest varying degrees of retinal damage in the nasal, inferior, temporal, and superior directions. Comparison of preoperative and postoperative parameters in unclosed patients revealed some degree of recovery in structure-related parameters, despite the lack of significant improvement in postoperative visual acuity.

Our findings indicate that the MLD, BD, diameter of ELM and EZ disruption were significantly larger in patients with unclosed macular holes compared to those with closed holes. However, no significant difference was observed for H. These results align with previous studies that have demonstrated the predictive value of BD as the strongest indicator for anatomical and functional prognosis in patients after macular hole surgery [8, 9, 13]. Other studies demonstrated that the angle between the upward edge of the retinal surface and RPE did not significantly correlate with the type of MH closure [14], our research also showed no significant difference between the two groups of patients before the surgery. Thus, further investigation is needed to identify more valuable angle parameters that accurately predict surgical outcomes.

In our study, we introduced EZ-related angle parameters to explore their predictive value for successful macular hole (MH) surgery. We found that the EZ-NFL, EZ-GCL, EZ-INL, EZ-OPL, and EZ-ONL angles in patients with closed MHs were significantly larger than those in patients with unclosed MHs before surgery. Since H was comparable between the two groups, the size of the EZ-related angle was largely dependent on the diameter of EZ disruption. Because unclosed MH patients tended to have a larger size of EZ disruption, the EZ-related angles in the unclosed group were significantly larger than those in the closed group. These findings indicated that only EZ-related angles, rather than classic parameters related angle (Basal or Basal-NFL angles), were capable of predicting prognosis. We propose that classical parameters such as MLD, BD, and the diameter of EZ disruption are one-dimensional parameters that reflect length, while angles are two-dimensional parameters that encompass these length measurements. Therefore, angles can provide a more comprehensive description of the anatomical status of the macula compared to length parameters alone. Consequently, EZ-related angle parameters may be more accurate in predicting the outcome of primary MH surgery compared to classical length parameters. We propose that classical parameters such as MLD, BD, and the diameter of EZ disruption are one-dimensional parameters that reflect length, while angles are two-dimensional parameters that encompass these length measurements. Therefore, angles can provide a more comprehensive description of the anatomical status of the macula compared to length parameters alone. Consequently, EZ-related angle parameters may be more accurate in predicting the outcome of primary MH surgery compared to classical length parameters. Importantly, the EZ represents the mitochondria-rich region of the photoreceptors [15], and it reflects the functional status of these photoreceptors. Therefore, damage to the EZ, comparing with other structures, was more indicative of the extent of macular function affected by the hole. EZ disruption appears to be a more valid predictor of functional recovery after surgery, which may explain why the basal angle or basal-NFL angle alone is not able to predict the prognosis of a macular hole.

In our study, we introduced the concept of angle regularity (AR) to assess the uniformity of angle distribution in closed and unclosed macular holes (MHs), which may have an impact on the prognosis of MH surgery. Our findings revealed that the AR of most angles in unclosed patients was significantly higher compared to closed patients, indicating that the distribution pattern of angles could affect the success rate of surgery in patients with large MHs. Furthermore, we observed that the main differences in angles between the two groups were primarily in the nasal and inferior directions, with the nasal direction showing the most significant disparity. It is widely accepted that posterior vitreous detachment (PVD) plays a critical role in the formation of MHs [16]. During the early stages of PVD, the perifoveal vitreous detaches, while the vitreofoveal adhesion and optic disc adhesion persist [17]. Subsequently, the vitreofoveal vitreous detaches before the optic disc, leading to traction forces that contribute to the formation of a macular hole. Based on our speculation, during this stage, the vitreous traction on the retina mainly remains at the optic disc, resulting in a persistent traction on the nasal side of the macula. This may explain the more significant differences observed in the angles in the nasal direction. Additionally, the papillomacular bundle, which contains a higher fiber density and shorter nerve fibers compared to other regions, is located on the nasal side of the macula [18]. This region presumably produces greater tension than in other directions and more likely results in significant differences. Therefore, optimizing surgical techniques to release nasal traction as much as possible may increase the likelihood of successful closure of the macular hole.

In our study, we conducted a comparison of parameters before and after surgery in patients with unclosed macular holes. The results revealed that after surgery, the MLD, BD and H of the MH significantly decreased. However, there was no statistical significance observed in the diameter of the EZ and ELM. Previous studies by Ikuno et al. [19] and John et al. [20] reported an increase in the size of the macular hole after unsuccessful surgery. They noted an enlargement of the hole from 480 μm to 610 μm and from 594 μm to 675 μm, respectively. However, our study demonstrated a dramatic reduction in the size of the hole after surgery. This difference could be attributed to variations in the follow-up period, as the previous studies had a minimum follow-up period of three months, while our study focused on early postoperative changes. It is possible that the hole may become larger over time in patients who did not achieve successful closure in the primary surgery. Consistent with previous findings, our study also indicated that unclosed patients did not experience significant functional improvement in BCVA [20]. We observed that almost all EZ-related angles significantly decreased after surgery. This could be attributed to the peeling of the ILM, as the release of macular traction allowed each retinal layer to return to its normal position. Additionally, we noticed the alleviation of cystoid macular edema (CME), which facilitated the recovery of angle parameters and contributed to the relief of extra tension on the macula. Our results demonstrate that surgery can release the traction on the macula, prevent the enlargement of the macular hole, and allow each retinal layer to return to its normal position, even in cases where the closure of the MH was not achieved.

Our study had some limitations. Firstly, we only analyzed the short-term risk factors of patients with large MHs. To fully understand the long-term risk factors, a more extended follow-up duration is necessary. Secondly, our sample size was small, as it was limited to matching patients with similar characteristics in terms of BD, MLD and diameter of EZ disruption. Therefore, subsequent studies are needed to expand the sample size to evaluate the predictive value of the angle parameters when matching these classical parameters. Furthermore, our study only included patients with stage IV macular holes. However, we found that the diameter of EZ disruption was significantly longer than the size of the hole itself, so further study may construct the model of EZ damage in patients with different stages of MH and modify grading system combining EZ parameters to reflect functional damage of MH patients. This would provide valuable evidence for determining the optimal timing and appropriate surgical intervention procedures. Due to the nature of retrospective studies, patients all underwent standard ILM peeling without modified procedure, and in the future, we will enroll patients with modified ILM peeling techniques to gain a better prognosis.

In conclusion, our findings support the notion that large macular holes, characterized by large EZ-related angle and angle AR, are more likely to have unfavorable outcomes following primary MH surgery. The EZ-related angles demonstrate potential as valuable parameters in predicting the prognosis of surgical interventions. Notably, the angles in the nasal direction were found to be particularly influential. Consequently, it is crucial to release nasal traction during surgery to improve the likelihood of a favorable prognosis.