Idiopathic epiretinal membrane (iERM) is a common macular disease with unknown etiology. It is characterized by abnormal fibroblast proliferation on the inner macular surface; moreover, it is commonly observed in older individuals (1, 2). The contraction characteristics of iERM can cause anatomical destruction and stratified changes, which results in visual impairment, metamorphopsia, and aniseikonia as the main symptoms (3, 4). Surgical pars plana vitrectomy (PPV) with iERM peeling is the standardized procedure for releasing retinal traction and enhancing retinal repair. However, there are some cases of insufficient anatomical and functional recovery even after successful surgery without significant complications (5–7). Therefore, it is important to predict the probability of postoperative visual recovery and to determine the timing of surgery.

There has been increasing interest in central macular thickness (CMT) and ectopic inner foveal layer (EIFL), which primarily react to retinal alterations caused by vertical iERM traction (6, 8–11). However, iERM contraction or shrinkage can cause foveal displacement. This can result in the loss of light dispersion due to vessel displacement (12), and light propagation anomalies due to muller cell deformation (13–15). Daiki et al. (16) proposed tangential displacement as a parameter for characterizing tangential morphological alterations in the outer retina. They reported that the horizontal and vertical metamorphopsia scores were positively correlated with the absolute values of the vertical and horizontal tangential displacement, respectively.

Given that foveal thickening and deviating under iERM traction is a three-dimensional problem, analysis through appropriate geometry methods is necessary. To further elucidate the pathogenesis of visual impairment, we calculated the horizontal distance (HD) and total distance (TD) in iERM using horizontal and vertical B-scan images from the bottom of the central fovea to the innermost vertex of the central fovea. Further, we analyzed the correlations of HD and TD with baseline and postoperative visual acuities.

This study had several principal findings. First, TD was an independent predictor of the baseline and postoperative BCVA in the multiple linear regression analysis. Second, the HD and TD were significantly correlated with the baseline and last follow-up BCVA, with TD showing the highest correlation coefficient among all the retinal parameters. Third, the preoperative HD and TD values significantly increased with advanced EIFL staging. There were no significant differences between the vertical and horizontal components.

Therefore, the HD and TD may accurately reflect the degree of iERM stereo traction and may be predictive factors for the degree of preoperative visual impairment and postoperative visual prognoses.

The formation of iERM shares similar characteristics as fibrosis. Specifically, the transdifferentiation of various precursor cells into myofibroblasts and the massive production of extracellular matrix proteins containing extracellular fibrils results in fibrous contraction and distortion of the normal tissue structure (4). Retinal thickening, disorganization, and altered foveal morphology can be caused by the vertical traction and horizontal tangential forces of iERM (11, 27–29). The area and depth of iERM traction were correlated with the foveal alteration and BCVA (30). Therefore, the iERM-generated forces and the resulting macular structure damage can be three-dimensionally considered. Stereo quantitative analysis of the central macular displacement might allow a better description of iERM tractions.

Although previous studies have used fundus autofluorescence imaging to evaluate retinal displacement based on relative changes in the position of blood vessels (31), this is confined to the inner retinal layer. Our study showed that the HD and TD significantly increased with staging, indicating a significant positive correlation between retinal stereo deviations with iERM severity in the whole retina. The lack of significant differences between the horizontal and vertical components of both indexes could be attributed to the indeterminate traction location, which is consistent with the results of Sakai et al. (16).

Sakai et al. (16) reported no significant correlation of the fovea’s horizontal and vertical tangential distortion with BCVA. Hsiao-Fan Tung et al. (32) reported that postoperative horizontal movement of the central fovea was positively correlated with the severity of BCVA worsening; however, this parameter cannot be preoperatively obtained. We found that both HD and TD were significantly associated with baseline BCVA, with TD showing the highest correlation index. Moreover, TD was an independent predictor of preoperative BCVA.

Ichikawa et al. (33) reported that the ratio of baseline to postoperative measurements of the distances between the intersections of retinal vessels on near-infrared spectroscopy images was correlated with the postoperative M-score. Another recent study reported that the postoperative change in the projection distance relative to the PRE layer at the vascular bifurcation was associated with BCVA at 6 postoperative months (34). However, both aforementioned parameters only reflected changes in the horizontal plane.

In our study, the last follow-up BCVA was significantly correlated with the baseline HD and TD, with TD having the highest correlation coefficient. Moreover, TD was a significant predictor of the last follow-up BCVA. Thus, serious stereo heterotopia of the retina was an adverse factor for visual prognosis. Furthermore, it is important to consider the TD when determining the timing of surgery.

The FAZ area, perimeter, and AI significantly correlated with staging. Previous studies have reported that FAZ-related parameters differed significantly compared with healthy controls (28), and changed significantly with disease severity (29, 35, 36). The FAZ area and perimeter were significantly correlated with the baseline and postoperative BCVA. The correlation between FAZ area and baseline BCVA was also proven in previous studies (28, 35, 37–39) and BCVA recovery (28, 37, 40, 41). Due to the differences in sample size, follow-up time, scanning depth range of different OCTA devices, and software vascular detection algorithm, many studies have controversial results on the correlation between FAZ-related indicators and visual acuity before (29, 35, 42, 43) and after surgery (35, 38, 43, 44). Although FD was not significantly correlated with the baseline staging, it was significantly associated with and was an independent predictor of postoperative visual acuity. Consistent with our findings, Kim et al. (41) reported that the parafoveal capillary density was significantly associated with postoperative visual acuity. This suggested that vessel displacement could alter the related tissue supply, with independent prognostic implications. Previous studies on FD-300 reported that the recovery at 6 months postoperatively did not reach statistical significance. The absence of a correlation with postoperative visual acuity was probably attributed to the limited sample size and insufficient follow-up period (8). CB was not significantly correlated with visual function, which is consistent with previous reports (22). The incidence of MME was positively correlated with EIFL staging progression; further, it was an independent predictor of the baseline and postoperative BCVA. Consistent with our findings, previous studies (45–47) have shown that MME is a nonnegligible prognostic index.

This study has several limitations. First, estimating the HD and TD through horizontal and vertical B-scan images may involve errors due to its retrospective nature. Theoretically, high-density imaging, with the 360° rotating B-scans centered on the bottom of the central fovea, can simultaneously obtain the image from the start to the endpoint. Accordingly, this may be used to accurately measure the HD and TD; however, future studies are warranted to test this hypothesis. Second, although the dose and exposure time of indocyanine green was minimized and it is relatively safe, its potential retinal toxicity still requires attention (48, 49). Third, the postoperative follow-up period was not exactly 12 months due to patient compliance issues. Forth, this was a small-scale retrospective study.

In conclusion, the baseline HD and TD were positively correlated with EIFL stage progression; moreover, they were predictors of the baseline and postoperative BCVA. TD showed the highest correlation indexes and was an independent predictor of baseline and postoperative visual acuity. FD-300 and MME, which are markers related to microvascular changes and microstructural damage, also showed significant predictive utility for postoperative visual acuity. Therefore, the HD and TD may accurately reflect the degree of iERM stereo traction and allow prediction of baseline and postoperative visual acuity.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by 2019168K160. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

LS, JM, and ZZ contributed to the conception of the study. LS, JM, ZZ, JL, NC, XD, and YijC performed the experiment. ZZ, JL, and YijC contributed significantly to analysis and manuscript preparation. ZZ, JT, and YiqC performed the data analyses and wrote the manuscript. ZZ and JM helped perform the analysis with constructive discussions. All authors contributed to the article and approved the submitted version.

The Zhejiang Medical Health Science and Technology Project (2023KY490) and the Natural Science Foundation of Zhejiang Province (LTGY23H120005) supported this study. Nonetheless, funding bodies were not involved in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.