Deep Optic Nerve Head Morphology in Tilted Disc Syndrome and Its Clinical Implication on Visual Damage

Abstract

Purpose: To study deep optic nerve head (ONH) morphology in tilted disc syndrome (TDS) and identify factors associated with retinal nerve fiber layer (RNFL) defect.

Methods: In patients with TDS, we evaluated the optic disc shape using the Bruch's membrane opening (BMO)–anterior scleral canal opening (ASCO) offset and measured the border tissue (BT) length, depth, and angle in the direction of the tilt, using radial ONH optical coherence tomography (OCT). We compared the parameters between the TDS groups with and without RNFL defects.

Results: Twenty-one eyes had no glaucomatous RNFL defect, and 38 eyes had a glaucomatous RNFL defect. The group with RNFL defects had a higher baseline IOP, larger tilt axis of BMO-ASCO optic disc margin (76.4° ± 14.5° vs. 87.9° ± 15.4°, P = 0.012), larger BMO–lamina cribrosa insertion (LCI) angle (25.6° ± 9.3° vs. 43.6° ± 15.2°, P < 0.001), and more lamina cribrosa (LC) defects (4.3% vs. 30.6%, P = 0.028) than without RNFL defects. The tilt axis and BMO-LCI angle were significant factors after adjusting for baseline IOP and LC defect. The BMO-LCI angle had excellent diagnostic power for glaucomatous RNFL defect in TDS, similar to the visual field mean deviation.

Conclusions: OCT-based large deep ONH BT angle and tilt axis were factors associated with the presence of RNFL defects in TDS. The results suggest a mechanism of RNFL defect associated with structural ONH deformation. Further investigations are warranted to understand the role of ONH structures in a general population with and without optic disc tilt.

Tilted disc syndrome (TDS) is characterized by a collection of findings, including oblique orientation of the optic disc axis, inferonasal tilting of the optic disc with an inferonasal crescent, a thinning of the retinal pigment epithelium (RPE) and choroid in the inferonasal fundus, situs inversus of the retinal vessels, and a posterior staphyloma.¹ Visual field (VF) defects may be present in some cases.^2–8 

TDS has not received much attention, but recent rigorous research showed the importance of optic nerve head (ONH) deformation in the pathogenesis of the retinal nerve fiber layer (RNFL) and VF damage using optical coherence tomography (OCT).^9–13 Being an extreme form of ONH deformation, observed in 1.6% to 3.5% of the population,^6,14–16 the significance of structural alterations in TDS may be better focused regarding the mechanisms of tilt-related optic nerve damage. However, studies on the deep ONH structure using OCT in TDS have been rare. 

Therefore, we aimed to investigate the deep ONH morphology of TDS eyes with OCT by performing qualitative and quantitative analyses. We also tried to identify the structural factors associated with the presence of RNFL loss. To the best of our knowledge, this is the first quantitative report on the structure of the deep ONH and the factors associated with vision loss in TDS. 

Methods

This was a cross-sectional study, and the medical records of patients who visited Samsung Medical Center (Seoul, South Korea) between September 2018 and December 2021 were reviewed. This study followed all guidelines for experimental investigation in human subjects, was approved by the Samsung Medical Center Institutional Review Board, and adhered to the tenets of the Declaration of Helsinki. 

Patient Inclusion/Exclusion

The inclusion criteria for TDS, based on conventional descriptions using color fundus photographs, were as follows: (1) inferior or inferonasal tilt of the optic disc,^1,6,14 (2) ratio of shortest to longest disc diameter (tilt ratio) <0.8,¹⁷ (3) elevation of the superior neuroretinal rim,¹ and (4) inferior or inferonasal conus.^1,6,14 

Each participant underwent a comprehensive ophthalmic examination, including slit-lamp biomicroscopy, Goldmann applanation tonometry, refraction, gonioscopic examination, dilated stereoscopic examination of the ONH, color- and red-free fundus photography (TRC-50DX model; Topcon Medical System, Inc., Oakland, NJ, USA), automated perimetry using a central 30-2 Humphrey field analyzer (HFA model 640; Humphrey Instruments, Inc., San Leandro, CA, USA) with the Swedish interactive threshold algorithm standard, axial length measurement (ARGOS; Movu Inc., Santa Clara, CA, USA), and ultrasonographic pachymetry (Tomey SP-3000; Tomey Ltd., Nagoya, Japan). The baseline IOPs were measured at the first and second visits without IOP-lowering medication, and the average values were used in the analysis. 

Inferior RNFL defects were identified using both red-free fundus photography and OCT. An RNFL defect appears as a dark, wedge-shaped area with the apex touching the optic disc border.¹⁸ A thickness more than 1% below the lower limit of normal (red areas) in the OCT deviation map and a dark blue area in the thickness map were also used as criteria. Shifting of the RNFL thickness distribution (e.g., temporalization) was considered. 

Glaucomatous VF defect was defined as a second VF defect using more than one reliable test for at least two of the following three criteria: (1) a cluster of three points with a probability of less than 5% on the pattern deviation map, including at least one point with a probability of less than 1% or a cluster of two points with a probability of less than 1%; (2) a glaucoma hemifield test result outside normal limits; or (3) a pattern standard deviation of 95% outside the normal limits. 

The exclusion criteria were as follows: (1) eyes with any episode of IOP >21 mm Hg; (2) presence of systemic or ocular disease that could affect VF test results; (3) eyes with noncorresponding or equivocal optic disc, RNFL, OCT, and VF tests; and (4) eyes with poor-quality OCT images that did not offer interpretable information with respect to Bruch's membrane opening (BMO), border tissue (BT), and lamina cribrosa (LC) insertion due to prelaminar tissue or vessels, defined as a scan with <70% of the anterior LC visible due to prelaminar tissue, retinal pigment epithelium, or overlying vessels in ≥3 of the 24 radial line scans.^13,19 

Evaluation of TDS Using Conventional Color Fundus Photographs

For qualitative evaluation, we recorded the presence of characteristic features of TDS: situs inversus of retinal vessels,²⁰ inferior fundus pallor (depigmentation or tesselation),²¹ inferior coloboma, and staphyloma.¹ 

Optic Disc Shape Evaluation Using Conventional Color Fundus Photographs

For quantitative evaluation, we measured tilt ratio as the ratio between the longest and shortest optic disc diameter²² and tilt axis as the angle of the short axis perpendicular to the long axis of the optic disc from the fovea–disc center line (Fig. 1).²² The clinical disc margin, as determined in color fundus photographs, was defined as the inner border of the reflective tissue that was internal to any pigmented tissue and within which only neural tissue was present. If there was no clear reflective tissue present, then the disc margin was defined as the innermost termination of pigmented tissue.²³ The peripapillary atrophy (PPA) ratio was defined as the ratio of the PPA area to the optic disc area. 

Figure 1.

 

Measurement of tilt ratio and tilt axis, as well as BT length, depth, and angle. Measurement of tilt ratio and tilt axis from color fundus photograph-based clinical optic disc margin (A) and OCT-based BMO-ASCO optic disc margin (B). Tilt ratio was defined as the ratio between the longest and shortest diameter. Tilt axis was defined as angle of the short axis perpendicular to the long axis of the optic disc from the fovea–disc center line. (C–E) Fundus photograph, infrared image, and OCT image at the tilt axis. (E) The BMO-choroid (blue dot), BMO-BT (superficial BT endpoint, red dot), and BMO-LCI (green dot) points are shown. We set the superficial BT endpoint as the most protruded point (red dot). We set the LCI point as the meeting point of the anterior LC surface and border tissue. We set the choroidal end as the point where the peripapillary choroidal tissue ended (blue dots). The yellow line indicates the BMO connecting line. (F–H) Measurements of depth, length, and angle from BMO end to LCI. Other landmark point measurements were performed identically.

Performance of Radial ONH OCT

Enhanced depth imaging spectral-domain OCT (Heidelberg Engineering, Heidelberg, Germany) with built-in software was used to investigate deep ONH structures, as reported previously.¹⁰ Scans were obtained using 24 radial-line B-scan modes (each at an angle of 7.5°) centered on the clinical optic disc margin demonstrated in the infrared images. The scans were not centered on the BMO centroid due to severely tilted optic discs, where the BMO center was located outside the optic disc in the middle portion of the externally oblique border tissue (EOBT). The scans centered on the visible optic discs provided more adequate views of the radial distribution of EOBT. 

Qualitative Evaluation of Deep ONH Using OCT

For qualitative evaluation, we recorded the presence of a posteriorly sloping LC, a protrusion of the upper edge of Bruch's membrane and choroid, and an elevation of the nerve tissue at the upper margin of the optic disc, as in a previous study.²⁴ We also recorded the presence of peripapillary intrachoroidal cavitation (PIC)^25,26 and the presence of a peripheral LC defect.²⁷ 

Optic Disc Shape Evaluation Using OCT

We defined the optic disc margin based on OCT findings using the concept of BMO–anterior scleral canal opening (ASCO) offset.²⁸ BMO-ASCO offset was reported to identify the location and extent of oblique border tissue regions.²⁹ The neural canal opening aligned to the clinical optic disc margin in eyes with EOBT, and the termination of BT at the ASCO was the disc margin using OCT-based optic disc margin anatomy.³⁰ We superimposed the BMO and ASCO on infrared fundus photographs provided by OCT, creating a BMO margin continuum on the superior side³⁰ and the ASCO margin on the inferior side.²⁹ From this OCT-based BMO-ASCO optic disc margin, we measured the tilt ratio and tilt axis as performed for photograph-based measurements (Fig. 1). Additionally, the direction of maximum EOBT among the 24 radial scans, which was used for quantitative measurements, was measured. We also measured the circularity of the BMO margin, as 4π (area/perimeter²): a value of 1.0 indicated a perfect circle. 

Quantitative Evaluation of Deep ONH Using OCT

We analyzed the OCT image in the direction of maximum EOBT along the tilt axis, as previously performed in other studies.^13,19 It was because we speculated that the scans obtained in the direction of maximum EOBT formation would most closely represent the most severe ONH structural deformation and associated biomechanical ONH microenvironment for each given eye. 

For quantitative evaluation, we measured the length, depth, and angle of the BT against the BMO (Fig. 1). We set three anatomic landmarks for measurement: (1) the superficial BT endpoint, defined as the farthest protruding point of the BT from the BMO end. This landmark followed previous methods.^27,31 (2) The LC insertion (LCI), defined as the point where the peripapillary sclera (PPS) and the anterior surface of the LC meet. The LCI point was chosen because it can be an objective point of measurement and because of its importance as a biomechanical anchor connecting the scleral border tissue, scleral flange, and peripheral LC.^32,33 (3) The choroidal endpoint, defined as the point where the peripapillary choroidal tissue ended, is gauged by tracing the BT of the choroid.³⁴ The choroidal endpoint was chosen because of the reported significance of the anterior scleral opening and BMO offset.^28,35,36 The distance was measured with the built-in program of Heidelberg OCT, and the angle was measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA). In peripheral LC defect cases in the analyzed OCT scan, the LCI was measured from the anterior LC surface extension line, as previously performed in eyes with an optic disc pit.³³ Two independent observers (EJL and JCH), who were masked to the clinical information, reviewed the images, and discrepancies were resolved by consensus. 

Data Analysis

The interobserver reproducibility of BMO-BT, BMO-LCI, and BMO–choroidal landmark identification was assessed by calculating the intraclass correlation coefficient (ICC) of the length measurement. We used a generalized estimation equation to compare the parameters between eyes with and without RNFL defects to consider the possible relationship between the two eyes of each patient. Correlation was assessed by calculating Pearson's correlation coefficient. The diagnostic accuracy of significant factors was determined by analyzing the area under the receiver operating characteristic (ROC) curve. Statistical significance was set at P < 0.05. All statistical analyses were performed using SPSS (version 15.0; SPSS, Inc., Chicago, IL, USA). 

Results

From the 79 eyes of 66 patients enrolled initially, we excluded 4 eyes of 4 patients with a tilt ratio ≥0.8, 5 eyes of 3 patients with episodes of high IOP, 1 eye of 1 patient with Parkinson disease with unreliable VF results, 2 eyes of 1 patient with concomitant superior segmental optic nerve hypoplasia, 3 eyes of 3 patients with equivocal results for the presence of RNFL defect, and 5 eyes of 5 patients with poor-quality OCT images. 

Finally, 59 eyes of 49 patients were included in the analysis; 23 eyes of 21 patients had no RNFL defects, and 36 eyes of 29 patients had RNFL defects. 

Clinical Characteristics

Table 1 shows the clinical characteristics of patients. In the overall TDS group, the spherical equivalent refractive error, cylindrical refractive error, and the axial length were not different between the two groups. The untreated baseline IOP was higher in the group with RNFL defects than in the group without (15.6 ± 1.7 vs. 13.5 ± 2.6 mm Hg, P = 0.006), with less difference after adjustment for central corneal thickness (P = 0.038). However, there was no abnormally high IOP (all IOPs ≤21 mm Hg). 

Table 1.

 

Clinical, Photograph, and OCT-Based Characteristics of Patients With Tilted Disc Syndrome With and Without RNFL Defect

 

Qualitative and Quantitative Photograph-Based Characteristics

The two groups had no significant difference in the distribution of situs inversus of retinal vessels, inferior fundus pallor (depigmentation), inferior coloboma, and staphyloma (Table 1). Situs inversus was the most commonly observed associated finding (47 of 59 eyes, 79.7%). 

The tilt ratio was not significantly different between the two groups (0.65 ± 0.09 vs. 0.68 ± 0.09, P = 0.276), but the tilt axis was larger (more deviated from the disc–fovea line, 85.6° ± 13.0° vs. 76.6° ± 9.5°, P = 0.009) in the group with RNFL defects than in the group without RNFL defects (Table 1 and Fig. 2). 

Figure 2.

 

Difference in clinical, photographic, and OCT-based parameters between the two TDS groups and their diagnostic power. (A) Distribution of IOP, photographic parameters, and OCT-based parameters in the two groups, respectively. (B) ROC curves for the analyzed parameters. Note that the BMO-LCI angle has a markedly distinct between-group distribution compared with the other parameters and shows the best diagnostic power, similar to visual field MD. Asterisks indicate statistical significance. White dots indicate eyes with TDS but no RNFL defect, and black dots indicate eyes with RNFL defect.

OCT-Based Evaluation of Optic Disc Shape

Despite considerable optic disc tilt, as depicted in fundus photographs, we observed that the BMO maintained a relatively round shape in both groups (circularity 0.891 ± 0.037 vs. 0.893 ± 0.038, P = 0.836). The overall morphologies of BMO and ASCO were similar to those reported for myopic eyes.³⁶ The calculated tilt ratio derived from the BMO-ASCO–delineated disc margin did not differ between the two groups. Similarly, the tilt axis calculated from the BMO-ASCO–delineated disc margin was significantly smaller in the group without RNFL defects than in the group with RNFL defects (Table 1). Additionally, the maximum EOBT formation located closely along the optic disc tilt axis, as delineated by the BMO-ASCO margin. Accordingly, the tilt axis measured in the maximum EOBT direction exhibited similar results (Table 1 and Fig. 2). 

The calculated tilt axes from the photographic disc margin, BMO-ASCO disc margin, and the axis of maximum EOBT manifestation were highly correlated (Pearson coefficient 0.778 between the photographic disc margin and BMO-ASCO disc margin, 0.799 between the BMO-ASCO disc margin and maximum EOBT, and 0.964 between the photographic disc margin and maximum EOBT; all P < 0.001). 

Qualitative OCT-Based Deep ONH Morphology

In the qualitative description of ONH structure as done by Shinohara et al.,²⁴ there was no difference in any feature between the two groups. We observed significantly more frequent presence of peripheral LC defects in the group with RNFL defects than that in the group without RNFL defects (11/25 vs. 1/22, P = 0.028). The OCT B-scan images were not morphologically distinguishable from that of myopic tilted optic discs. In 5 of the 12 eyes with peripheral LC defects, the defects were present in the analyzed OCT section. 

Quantitative OCT-Based Deep ONH Morphology

The interobserver reproducibility of the BMO-BT, BMO-LCI, and BMO-choroid lengths was excellent (ICC = 0.901, 95% confidence interval [CI], 0.834–0.941; ICC = 0.964, 95% CI, 0.940–0.979; and ICC = 0.923, 95% CI, 0.870–0.954, respectively, all P < 0.001). Table 1 and Figure 2 show the differences in deep ONH morphology between the two groups. 

As for BT lengths, the BMO-BT length (the BT length measured by the superficially most protruded point) was greater in the group without RNFL defects compared to the group with RNFL defects, but with marginal statistical significance (832.8 ± 362.0 µm vs. 599.0 ± 295.9 µm, P = 0.050). BT lengths measured by the BMO-choroid and BMO-LCI showed no difference between the two groups. None of the BT depths showed any difference between the two groups. 

In contrast, the BMO-based BT angles were all significantly smaller in the group without RNFL defects than in the group with RNFL defects. The BMO-LCI angle (25.6° ± 9.3° vs. 43.6° ± 15.2°, P = 0.001) had the largest difference among the three landmark-based measurements. It is noteworthy that the distribution of BMO-LCI angle was more distinct between the groups compared to other parameters that showed a wide overlap between the groups (Fig. 2). 

The tilt axes in the three different methods and BMO-LCI angle were significantly different between the two groups after adjusting for baseline IOP and the presence of LC defect (Supplementary Table S1). IOP was not a significant factor. We selected the most significant parameter from the same category to be included in the model because of possible redundancy. 

ROC Curve and Diagnostic Accuracy

The areas under the ROC (AUROC) curves are shown in Figure 2 and Table 2. The BMO-LCI angle showed the best discrimination among the structural parameters and was similar to that of the mean deviation (MD) value and RNFL thickness (AUROC curve 0.853, 0.852, and 0.909, respectively, all P < 0.001). The BMO-LCI angle had excellent diagnostic accuracy, as AUROC curves of 0.7 to 0.8 are considered acceptable, 0.8 to 0.9 excellent, and more than 0.9 outstanding.³⁷ 

Table 2.

 

Areas Under ROC Curves of ONH Morphologic Parameters to Discriminate Between TDS Eyes With and Without RNFL Defects

 

Representative Cases

Figure 3 shows representative cases. It is noteworthy that despite the similar photographic appearances of obvious disc tilt, OCT shows different deep ONH morphology between the TDS groups with and without RNFL defects. 

Figure 3.

 

Representative cases of TDS with and without RNFL defect and the illustrative diagram from the hypothetical significance. A case with TDS without RNFL defect (A). Inferior optic disc tilt and inferior conus are obvious in the fundus photograph. OCT RNFL/ganglion cell-inner plexiform layer (GC-IPL) thickness maps and visual field tests show no RNFL damage. Deep ONH OCT shows that BT has a shallow slope with a BMO-LCI angle of 23.4°. Another TDS case with RNFL and VF defect (B). The radial deep ONH OCT shows steep posterior angulation of BT and steep BMO-LCI angle of 44.1°. (C) With steeper slope of BT to LCI (BMO-LCI angle), the change in the plane of the altered LCI zone may be greater in TDS with the RNFL defect than in TDS without the RNFL defect. Greater changes in three-dimensional geometry and disruption of the physiologic protective function of the PPS-LCI zone may result in greater LC strain and damage to axons that pass through it. The analyzed OCT scans were in the direction of maximal EOBT formation along the tilt axis.

Discussion

VF defects in TDS are considered congenital,³⁸ do not progress, and do not require treatment.^{1,5,16,39–41} However, the mechanism of visual loss, or how TDS and glaucoma are different, is not fully understood. 

OCT has not been widely used in studying the ONH of the eyes with TDS. In TDS, OCT was mostly used to evaluate peripapillary RNFL thickness⁴² or changes in the macula⁴³ but not in the ONH. One study examined the ONH in TDS eyes with swept-source OCT, but their description was qualitative and did not compare cases with RNFL damage and those without. Therefore, our study would be the first to report both qualitative and detailed quantitative evaluation of the deep ONH using OCT, comparing eyes with and without nerve damage in TDS. 

In addition, previous reports have highlighted that the clinical optic disc margin and OCT-based structures comprising the margin are not identical. Furthermore, the disc margin is not a single anatomic structure.^28–30 Particularly for tilted discs, Strouthidis et al.³⁰ suggested that when the border tissue is externally oblique, this tissue or the anterior scleral canal opening may represent the disc margin. It is also reported that where an overhang of Bruch's membrane extends beyond the innermost termination of border tissue (whether internally or externally oblique), the termination of the BMO is detected as the disc margin, and where the innermost termination of an externally oblique border tissue is internal to the termination of Bruch's membrane, the border tissue is visible clinically and comprises the disc margin.⁴⁴ Therefore, we also evaluated optic disc margin using OCT-based parameters. 

To date, TDS has only been described according to photographic findings. Using OCT, we observed that ONH in TDS had (1) significant superior-side BMO-ASCO offset resulting in tilted BMO-ASCO optic disc margins, (2) BMO protrusion on the superior side and EOBT formation on the inferior side, (3) neural tissue elevation on the superior side, (4) maximal EOBT formation around the BMO-ASCO disc tilt axis direction, and (5) relatively round BMO. 

On the basis of these findings, we suggest an OCT-based definition of TDS, as (1) superior-side BMO-ASCO offset forming a superior disc margin by BMO and an inferior disc margin by ASCO at locations where the EOBT is formed, (2) elevated nerve fiber tissue and Bruch's membrane/choroid protrusion in the superior disc, and (3) the tilt axis as evaluated by BMO-ASCO disc margin in the inferior to the inferonasal side. Our OCT findings conform to the classic photograph-based features of tilted disc syndrome: (1) the oval optic disc shape, as demonstrated by the oval BMO-ASCO disc margin; (2) situs inversus, as the BMO-ASCO offset is often accompanied by significant vessel trunk displacement; (3) inferior crescent, as the EOBT forms in the inferior side; and (4) partly the staphyloma or fundus ectasia, where occasional presence of PIC may be the minor phenotype. 

The OCT findings of EOBT and β- and γ-zone PPA in this study are morphologically similar to the reported deep ONH OCT findings in myopic tilted discs.^45–47 Therefore, our OCT results confirm that the ONH deformation in TDS is morphologically indifferent from that of myopia except in the direction of tilt, as previously reported.^{7,19,20,41,48} It corroborates the microscopic findings that the PPA in myopic tilted optic discs and the inferior conus in TDS are histologically indistinguishable, both lacking RPE, BM, and choroid.^7,20,48 

From these structural similarities, we considered that TDS and glaucomatous visual damage may share similarities. We also noted several clinical similarities between TDS and glaucoma developed in eyes with an optic disc tilt. First, myopic normal-tension glaucoma (NTG) is also characterized by optic disc tilt, VF defect, and normal IOP.^1,20,21,41 The pattern of the VF defect is also strikingly similar. While superior and superotemporal defects are most common in TDS, altitudinal and arcuate defects have all been reported.^2–8 Cohen et al.²¹ discussed that these VF defects look fascicular, related to the blind spot. Moreover, the bitemporal VF defect in TDS that mimics a chiasmal lesion is actually incomplete, is limited to the superior quadrants, and does not respect the vertical meridian, and the vertex is directed toward the blind spot.³ Moreover, all the RNFL defects in our study were wedge-shaped, indistinguishable from that of glaucoma. 

Second, while the progression of the VF defect has been considered the only clue that indicates glaucomatous damage from TDS,^1,20 there is no longitudinal study to prove the stability of VF defects in TDS.^20,21 In addition, the progression is also very slow in myopic NTG, even without treatment,^49–51 which may seem stable depending on the follow-up period. For example, Han et al.⁵¹ reported that in some cases of myopic NTG, it took almost 10 years until progression of VF defects was confirmed. Therefore, the difference in prognosis is not completely distinguishing the two diseases. 

Third, the deep BT structure was also a significant factor for the presence of glaucoma,^9–13,52 with a topographical relationship of visual damage.⁵³ For example, the greater angle of BT and LC and tilt axis were associated with glaucoma,^9–13 and the PPS had a more pronounced V-shaped configuration in eyes with glaucoma than in controls.⁵² Moreover, the LCI shows posterior migration in glaucoma,^32,33 which may also be associated with a steep BT angle. 

On the basis of these similarities, we speculated that the large BMO-LCI angle may also have significant roles in patients with glaucoma and tilted optic discs. This assumption would be in line with that in myopic NTG where optic disc tilt is common and IOP remains normal, and the optic disc tilt has been considered a cause of glaucomatous damage.^{11,51,54–56} However, it is speculative, and further investigations are warranted. 

A more prevalent peripheral LC defect may have affected the BMO-LCI angle. However, as we measured the LCI from the anterior LC surface extension, the LC defect may not have resulted in a falsely aggravated BMO-LCI angle or deepened LCI position. Moreover, the number of eyes with concomitant LC defects and LCI measurements was small (five eyes). 

Our results suggest the importance of the biomechanical environment of the ONH in tilted discs. The special circular collagen geometry in the PPS has a protective function to reduce the LC strain,⁵⁷ which is critical for axonal health.⁵⁸ Accordingly, the LCI zone is reinforced with crisscrossing circumferential collagen fibers in the PPS, radial fibers in BT, sagittal fibers of the scleral flange, and elastin fibers that are absent elsewhere.^34,59–62 The destroyed PPS collagen arrangement had a significantly altered distribution of anisotropy in highly myopic human eyes.⁶³ In this regard, a steeper slope of BT to LCI (BMO-LCI angle) may indicate greater deformation of protective architecture (the PPS-LCI zone) and the function. The LC strain is critical for axonal damage,⁵⁸ and the changes may result in higher LC strain and subsequent RNFL damage in TDS, even under normal IOP (Fig. 3). Grytz et al.⁶⁴ also suggested that myopic connective tissue changes in the posterior sclera and the ONH may alter the principal stress/strain direction, which may lead to secondary load-driven remodeling, potentially increasing the risk of glaucoma. Large tilt axis may also be interpreted similarly. 

Nevertheless, the clinical application of BMO-LCI angles may yet be limited in several aspects. First, angular diversity was compared only within eyes and locations where EOBT formation was significant. Therefore, applying the findings to a more generalized population with diverse ONH structures (e.g., internally oblique border tissue) is difficult. Second, many other factors determine the BMO-LCI angle, including peripheral LC defects, posterior LCI migration, and overall neural canal obliqueness. It has been demonstrated that the LCI is posteriorly displaced in glaucoma in humans and animals.^33,65 Therefore, any glaucomatous eye may have a larger BMO-LCI angle compared to healthy status, whether the eyes have TDS or not. However, the LCI depth did not differ between the two groups, indicating that different BMO-LCI angles may not be the sole result of glaucomatous damage and represent meaningful geometric structural variation. Third, the angle is sensitive to the scan position. Subtle location and direction differences in the scan may result in large variability and may significantly impair an easy use in clinical settings. 

Indeed, the neural canal obliqueness evaluated in OCT may be a more robust and intuitive parameter for evaluating ONH tilt.^28,36 In addition, the deep ONH structure in eyes with TDS is highly complex, especially around the location of LCI, where important biomechanical relationships are established. In this regard, we speculate that the BMO-LCI angle may represent a small but significant local variation in the PPS-EOBT-LC continuum in addition to the overall neural canal axis. Further studies are warranted to better understand the significance of the structural deformations. 

One interesting but underrated question is how most myopic populations remain healthy without developing glaucoma despite a significant optic disc tilt,⁶⁶ especially myopic NTG. It would be interesting to investigate whether the BMO-LCI angle in the myopic eyes may be related to the development of glaucoma as in this study. However, studies are needed to understand whether the two conditions share pathogenic similarities or if they only resemble morphologic features. Glaucoma is a leading cause of blindness worldwide; thus, it would be meaningful to investigate whether the BT angle is a biomechanical parameter that modulates the overall health of ONH in eyes with significant deep ONH deformation. 

This study had several limitations. First, the sample size is small; thus, the size might have not been large enough to evaluate the relative influence of each parameter altogether.⁶⁷ Additionally, the small number of patients might have prevented us from determining significant differences between the two groups for factors that were not statistically significant. However, recruiting a set of patients with and without visual loss in a rare disease is difficult. Nevertheless, studies with a larger number of eyes are warranted. Second, in the group with RNFL defect, IOP was 2.1 mm Hg higher, which requires further studies. However, it was not significant in multivariate analysis, with a wide overlap between the groups. We speculated that because we included patients with normal IOP, the impact of IOP may not be as great as that of high-IOP glaucoma. Lastly, we could not demonstrate whether the defect was congenital or due to coexistent glaucoma. Further investigation through longitudinal studies is warranted. 

In conclusion, we used OCT to show that the deep ONH morphology in TDS is characterized by the BMO-ASCO offset to the superior side, the formation of EOBT to the inferior and inferonasal side, and elevation of superior nerve fiber layer tissue; the OCT findings of deep ONH structure in TDS resembled those of myopic optic disc tilt in the given direction of tilt. The OCT-based steep BT angle and large tilt axis of BMO-ASCO optic disc margin were associated with the presence of RNFL defects in eyes with TDS. Further studies are warranted to better understand the clinical significance of the deep ONH morphology in the overall visual prognosis of eyes with tilted optic discs. 

Acknowledgments

Disclosure: E.J. Lee, None; J.C. Han, None; C. Kee, None