Relationship Between Optical Coherence Tomography Parameter and Visual Function in Eyes With Epiretinal Membrane

Abstract

Purpose: To investigate the associations between visual function and the optical coherence tomography (OCT) parameters in eyes with idiopathic epiretinal membrane (ERM).

Methods: Thirty-nine consecutive eyes with ERM were enrolled. In addition to OCT parameters, such as central retinal thickness (CRT), the area of gap between the ERM and the retinal surface (SUKIMA) was newly defined and calculated from the vertical and horizontal OCT images (SUKIMAv and SUKIMAh). The average of SUKIMAv and SUKIMAh (SUKIMAave) was used for the statistical analysis. The vertical and horizontal metamorphopsia scores (MV, MH) and the average of MV and MH (Mave) were also used for the analysis.

Results: The Mave was not significantly associated with logMAR visual acuity (VA) (P = 0.57, linear regression analysis). Analysis using second-order bias-corrected Akaike information criterion model selection identified the age, CRT, and SUKIMAave as being associated with logMAR VA. On the other hand, among the OCT parameters, SUKIMAave and CRT were associated with the Mave. In addition, there was a significant relationship between SUKIMAh and MV (P = 0.011) and between SUKIMAv and MH (P = 0.0014).

Conclusions: We identified SUKIMA as a novel OCT parameter that is useful to predict both VA and metamorphopsia in patients with ERM.

Figure 1.

Measurement of OCT parameters in eyes with ERMs. (A) Vertical and horizontal scans 5.8 mm in length (scan angle 20°) were used for the analysis. (B) Each OCT image through the fovea (1024 × 496 pixel) was obtained in eyes with idiopathic ERMs. (C) The following OCT parameters were calculated using ImageJ software: CRT (white arrow), maximum retinal fold depth (yellow bar), and SUKIMA (blue area).

Figure 2.

Correlation between SUKIMAave and visual functions. There were significant correlations between (A) SUKIMAave and logMAR VA (r = 0.397; 95% CI, 0.0936–0.633; P = 0.012) and (B) SUKIMAave and degree of metamorphopsia (Mave) (r = 0.578; 95% CI, 0.322–0.756; P = 0.00012, linear regression analysis).

Figure 3.

Relationships between SUKIMA and metamorphopsia score. There were significant correlations between (A) SUKIMAv and MH score (r = 0.494; 95% CI, 0.211–0.700; P = 0.0014), (B) SUKIMAh and MV score (r = 0.405; 95% CI, 0.103–0.639; P = 0.011), (C) SUKIMAh and MH score (r = 0.599; 95% CI, 0.349–0.769; P < 0.0001), and (D) SUKIMAv and MV score (r = 0.486; 95% CI, 0.202–0.695; P = 0.0017, linear regression analysis).

Figure 4.

Relationships between SUKIMA and OCT parameters. (A) There was a significant relationship between SUKIMAave and the maximum retinal fold depth (r = 0.600; 95% CI, 0.350–0.769; P < 0.0001). (B) A significant correlation was also observed between SUKIMAave and CRT (r = 0.385; 95% CI, 0.0792–0.625; P = 0.016, linear regression analysis).

Figure 5.

Representative cases with good and poor correlation between SUKIMA and metamorphopsia score. (A) Vertical and horizontal scans of 85-year-old female with a good correlation between SUKIMA and metamorphopsia score. SUKIMAv, SUKIMAh, and SUKIMAave were 0.062, 0.033, and 0.047 mm², respectively. MH, MV, Mave score, and logMAR VA were 0.8, 0.5, 0.65, and 0.523, respectively. (B) This 71-year-old female showed a good correlation between SUKIMA and metamorphopsia score. SUKIMAv, SUKIMAh, and SUKIMAave were 0.137, 0.107, and 0.122 mm², and MH, MV, Mave score, and logMAR VA were 0.9, 0.7, 0.8, and 0.155, respectively. (C) This 72-year-old female showed a poor correlation between SUKIMA and metamorphopsia score. SUKIMAv, SUKIMAh, and SUKIMAave were 0.061, 0.0083, and 0.035 mm², and MH, MV, Mave score, and logMAR VA were 1.4, 1.7, 1.55, and 0.155, respectively. Of note, ERM was not fully identified in both vertical and horizontal scans, whereas retinal folds were observed (white arrows). (D) This 84-year-old male showed a poor correlation between SUKIMA and metamorphopsia score. SUKIMAv, SUKIMAh, and SUKIMAave were 0.120, 0.067, and 0.093 mm², and MH, MV, Mave score, and logMAR VA were 1.7, 2.0, 1.85, and 0.398, respectively. ERM at the temporal region was not identified, and it is possible that the SUKIMA score might have been underestimated in this case.