Stereopsis and retinal microstructures following macular hole surgery

The aim of this prospective study was to evaluate the changes in stereopsis in patients who underwent vitrectomy for macular hole (MH) and assess the relationship between stereopsis and retinal microstructures. Fifty-two patients who underwent successful vitrectomy for unilateral MH and 20 control participants were recruited. We examined stereopsis using the Titmus Stereo Test (TST) and TNO stereotest (TNO), optical coherence tomography, and best-corrected visual acuity measurements, preoperatively, and at 3, 6, and 12 months postoperatively. As a result, preoperative and postoperative 3, 6, and 12-month values of stereopsis assessed by TST (log) were 2.7, 2.2, 2.2, and 2.2, respectively. TNO (log) were 2.8, 2.5, 2.4, and 2.4, respectively. Stereopsis in MH after surgery was significantly worse than that in normal participants (p < 0.001). Preoperative TST significantly correlated with MH size and defect length of external limiting membrane (ELM). Postoperative TST demonstrated significant correlation with the preoperative ELM defect length, and postoperative TNO was associated with the preoperative interdigitation zone defect length. Vitrectomy for MH significantly improved stereopsis, although not to normal levels. The ELM defect lengths, which approximately correspond to TST circles, are prognostic factors for postoperative stereopsis by TST. The interdigitation zone defect length, similar in size to the TNO index, is a prognostic factor for postoperative stereopsis by TNO.

traumatic MH, ophthalmic diseases except mild cataract and refractive errors, and any systemic disease that influenced ocular motility.
The examinations included measurements of stereopsis, best-corrected visual acuity (BCVA), and retinal microstructure before surgery and at 3, 6, and 12 months after surgery. Stereopsis was measured using the Titmus Stereo Test (TST) and TNO stereotest (TNO). These tests were performed at a distance of 40 cm with appropriate spectacle correction. The results of stereopsis tests were expressed as 'seconds of arc' . We converted these values to logarithms for statistical evaluation.
The retinal microstructure was measured with spectral-domain OCT (Cirrus high-definition OCT; Carl Zeiss, Dublin, CA, USA). We used 5-line raster scans for each eye using an analytical software package (Cirrus analysis software, version 3.0; Carl Zeiss) with a signal strength of more than 7/10. We quantified the following parameters before surgery based on OCT images: minimum diameters of MH, base diameters of MH, macular thickness, and external limiting membrane (ELM), ellipsoid zone (EZ), and interdigitation zone (IZ) defect lengths ( Fig. 1) 8 . Based on the images of the 5-line raster scans, we quantified the parameters with an image processing software (ImageJ bundled with Java 1.8.0_172, developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available from https ://rsbwe b.nih.gov/ij/index .html) 8 . The defect length of each line was determined by agreement between two blinded, well-trained observers (Y.S. and Y.M.), and the mean value of the length of each line was used for further analysis.
All surgeries were performed by two surgeons (F.O. and Y.O.) under sub-Tenon's local anaesthesia. Clinically significant cataracts were simultaneously operated. We induced posterior vitreous detachment and performed core vitrectomy. Subsequently, we injected 0.2 mL of 0.025% brilliant blue G solution over the macula for 10-20 s and thereafter, washed it with an irrigation solution. Internal limiting membrane (ILM) peeling and fluid/gas exchange were performed for all cases. The patients were instructed to be in a face-down position for 1-3 days postoperatively.
The mean scores were compared, and standard deviations were calculated for each parameter of visual function and OCT measurement. BCVA was measured with the Landolt Chart and expressed as a logarithm of the minimum angle of resolution. The Mann-Whitney U test was used to compare stereopsis between patients with MH and normal participants. The Wilcoxon signed rank test was used to compare differences in stereopsis before and after surgery in patients with MH. Repeated-measures analysis of variance was used to clarify the changes in visual acuity and stereopsis. When a significant difference was detected, we conducted the Dunnett post hoc test for multiple comparisons to reveal the time point that demonstrated a significant difference from the baseline value. The associations between stereopsis and OCT parameters and visual acuity were examined by the Spearman rank correlation test. All tests for evaluating associations between parameters were considered statistically significant if P was < 0.05. All analyses were performed using StatView (version 5.0, SAS Inc., Cary, NC, USA). Table 1 reveals the visual functions and OCT parameters of patients with MH before surgery. Preoperatively, 40 eyes were phakic and 12 eyes were pseudophakic with posterior chamber intraocular lenses. All phakic eyes were performed with cataract surgery combined with vitrectomy.

Relationship between stereopsis and visual acuity in patients with MH. Preoperative stereop-
sis assessed by TST showed a significant correlation with preoperative BCVA (r = 0.32, P < 0.05), but stereopsis  Relationship between stereopsis and OCT parameters before surgery. The preoperative TST values showed significant correlation with the minimum and base diameters of MH (P < 0.01 and P < 0.01, respectively) and with the ELM defect length (P < 0.01). TNO values were significantly associated with the base diameter of MH (P < 0.05; Table 2).
Preoperative OCT parameters affecting postoperative stereopsis. Table 3 shows the relationship between the postoperative 12-month values of stereopsis and preoperative OCT parameters in patients with MH. The postoperative 12-month TST values significantly correlated with preoperative ELM defect length (P < 0.05, Fig. 3A). Postoperative TNO values showed a significant correlation with preoperative IZ defect length (P < 0.05, Fig. 3B).

Discussion
The postoperative visual acuity in patients with MH tended to improve gradually over 1 year, while stereopsis improved immediately within 3 months after surgery, after which it remained unchanged. Stereopsis is a more advanced visual function than visual acuity, and there may be a limit to the extent of surgical improvement of stereopsis. It has already been reported that stereopsis improves after surgery for treating ERM and MH; [12][13][14][15] ; however, these studies did not report detailed temporal changes in stereopsis. This is the first study to report temporal changes in stereopsis in patients with MH. Stereopsis was compromised in patients with MH and was worse in patients with MH than that in normal controls in this study. Previous studies have investigated disturbances in stereopsis in retinal disorders, including unilateral ERM and RD, and compared them with stereopsis assessed in normal controls 15,16 . Stereopsis was affected in patients with MH similar to that in patients with other unilateral retinal diseases 14 . Moreover, Stereopsis assessed by TST was associated with visual acuity in patients with MH. The relationship between visual acuity and stereopsis has been investigated in normal participants 26,27 and in patients who underwent surgery for MH 12,13 and RD 28 . Lam et al. 29 reported a difference in visual acuity between the eyes as a cause of disturbance of stereopsis, while Burian 30 reported that stereopsis deteriorated with worsening visual acuity in one eye. Judging from these findings, stereopsis, which is binocular vision, is impaired even if visual acuity in one eye is impaired. We included MH patients with one eye and normal vision in the other eye in this study. It is consistent with the results of these articles that worsening visual acuity in one eye impairs stereopsis.
In contrast, preoperative stereopsis assessed by TNO was not significantly associated with visual acuity in patients with MH. This difference may be attributed to the different index sizes used in the two stereotests. The stimulus used for assessing fine stereopsis in TNO was much larger than that used in the TST circles. The TST circles subtend a visual angle of 0.7°, while those in TNO subtend an angle of 8.5°. The diameters of the fovea and foveola were 1500 and 350 μm, respectively, with visual angles of approximately 5° and 1.2°, respectively. Therefore, TST circles are more relevant to visual acuity than TNO. Generally, MH patients do not complain of stereoscopic vision. This may be because the central stereopsis deficit indicated by TST is supplemented by paracentral stereopsis indicated by TNO in the visual field pathway in the brain.
Preoperative stereopsis had a significant correlation with various OCT parameters, including the minimum diameter of MH, base diameter of MH, and ELM defect length. Several studies have investigated the relationship between visual functions and OCT findings in patients with MH. Visual acuity was significantly associated with the minimum and base diameters of MH 19 , and the EZ 20-23 , and IZ defect lengths 24,25 . Moreover, metamorphopsia was associated with the minimum and base diameters of MH 6,7 , asymmetric elongation of foveal tissue 9 , and intraretinal cysts within the fluid cuff 8 . Stereopsis was related to the size of the MH, outer retinal layers, visual acuity, and metamorphopsia.
The prognostic factors for the assessment of postoperative stereopsis in patients with MH were ELM defect length by TST and IZ defect length by TNO. The difference in the prognostic factors between TST and TNO is probably due to the different index sizes used in the two tests as mentioned previously. Generally, the defect length of IZ is longer than that of ELM because of the shape of the MH. In this study, the mean defect length of ELM was 755 µm and that of IZ was 2126 µm. Interestingly, when the indices and results of defect lengths of the two tests are superimposed on the fundus photograph, the indices and defect lengths of the two tests coincide (Fig. 4). Therefore, the extent of impairment of retinal structures was reflected in the results of stereopsis.
Our study has several limitations, especially the small sample size and short follow-up duration. Visual acuity and stereopsis in patients with MH may improve over the course of a longer follow-up period. Although factors such as eye dominance 31,32 , pupil size 33,34 , and accommodation 26,35 , are known to affect stereopsis, we did not evaluate these factors in this study. In addition, our study population was relatively small MH patients (Minimum diameters of MH was 394 µm in average). The present results may not be totally suitable for larger macular holes. Future studies that include a larger sample size, longer follow-up duration, and evaluate other factors will further improve our understanding of stereopsis and other visual functions in patients with MH.