Assessment of Ocular Deformation in Pathologic Myopia Using 3-Dimensional Magnetic Resonance Imaging

This study attempts to assess global ocular deformation in pathologic myopia.

Key Points

Question

What are some morphological patterns of ocular deformation in the equatorial and posterior regions of eyes with pathologic myopia and how are they associated with myopic fundus changes and visual impairment?

Findings

In this cross-sectional study of 180 eyes, 5 equatorial and 6 posterior shapes were identified with pyriform in the equatorial region and bulb-shaped in the posterior region being the predominant types of deformation. Ocular shapes differed in axial length, visual acuity, morphological indexes, and myopic maculopathy.

Meaning

These results suggest that ocular deformation is globally involved, diverse, and associated with some ocular parameters and myopic fundus manifestations.

Abstract

IMPORTANCE

Ocular deformation in pathologic myopia can affect the entire globe. However, few studies have investigated the equatorial pattern of ocular shape. In addition, the correlation between equatorial and posterior morphology needs to be further explored.

OBJECTIVE

To assess global ocular deformation in pathologic myopia.

DESIGN, SETTING, AND PARTICIPANTS

This hospital-based, cross-sectional study included 180 pathologic myopic eyes with atrophic maculopathy grading C2 (diffuse chorioretinal atrophy) or more from 180 participants who underwent comprehensive ophthalmic examination, including high-resolution 3-dimensional magnetic resonance imaging. In addition, 10 nonpathologic myopic eyes of 10 participants were set as the control group.

Main Outcomes and Measures

According to the cross-sectional view of equator, equatorial shape was classified as round, rectangular, pyriform (noncircular and more protruded in 1 direction), vertical-elliptical, or horizontal-elliptical; according to the nasal and inferior views, the posterior shape was categorized as spheroidal, conical, bulb-shaped, ellipsoidal, multidistorted, and barrel-shaped. Equatorial circularity and ocular sphericity were used to quantitatively assess the morphological variability of the equatorial and posterior regions, respectively. The association between ocular morphology and ocular parameters and myopic maculopathy was also investigated.

Results

The mean (SD) age of 180 participants with pathologic myopia was 55.14 (10.74) years, 127 were female (70.6%), and the mean (SD) axial length of studied eyes was 30.22 (2.25) mm. The predominant equatorial shape was pyriform (66 eyes [36.7%]), followed by round (45 eyes [25.0%]). The predominant posterior shape was bulb-shaped (97 eyes [52.2%]), followed by multidistorted (46 eyes [24.7%]). Equatorial circularity and equatorial shapes were correlated (r = −0.469; 95% CI, −0.584 to −0.346; P < .001) and ocular sphericity was correlated with posterior shapes (r = −0.533; 95% CI, −0.627 to −0.427; P < .001). In eyes with a vertical-elliptical equator, equatorial circularity and ocular sphericity were positively linearly correlated (R² = 0.246; 95% CI, 0.050-0.496; P = .002) and the prevalence of inferior staphyloma was higher (27.8%; P = .04). Eyes with a horizontal-elliptical equator have the most horizontally oriented axis of corneal flat keratometry (median, 43.55 [interquartile range, 43.84] degrees; P = .01) and tended to present with multidistorted posterior shape (21.7%; P = .04).

Conclusions and Relevance

These findings suggest ocular deformation is common in pathologic myopia and can affect the entire eye, including the equatorial and posterior regions. The morphological classification may enhance the understanding of the diverse patterns of ocular shape in pathologic myopia.

Introduction

Ocular deformation, especially posterior staphyloma, is a unique feature of pathologic myopia.^1,2 Three-dimensional (3-D) magnetic resonance imaging (MRI) can reconstruct the entire eye and allow noncontact in vivo morphology studies.^3,4,5,6 The posterior curvature changes in high myopia have been found to be associated with vision-threatening complications in several studies.^4,5,6,7 Moriyama et al⁵ classified the posterior shape into nasally distorted, temporally distorted, cylinder-shaped, and barrel-shape based on ocular elongation and symmetry in the inferior view and found that eyes with multiple protrusions tended to have severe fundus atrophy and that visual-field defects were more prevalent in temporally distorted eyes. Then in a modified classification by Guo et al,⁴ spheroidal, conical, and ellipsoidal shapes were added. They found that spheroidal shape, which was considered to be undistorted, was the most common type in high myopia with relatively young age and short axial length (AL), whereas ocular deformity was associated with fundus degeneration and vision loss.

Considering the continuity of the entire ocular tissues, deformation in pathologic myopia may not occur only in the posterior region. Although it has been previously reported that the equator in high myopia is not always regular,³ to the best of our knowledge, no studies have further investigated the morphological classification of the equator and its association with posterior shape. In addition, previous studies evaluated posterior morphology only from the inferior view.^4,5 However, pathologic myopia has a more complex posterior shape in 3 dimensions,^3,8 where the shape may appear inconsistent in the horizontal and sagittal planes.^3,5,9 Therefore, it is difficult to provide a complete description of ocular morphology based on ocular symmetry from a single viewpoint.

Therefore, this study aimed to propose a modified classification describing both the equatorial and posterior ocular shapes based on high-resolution 3-D MRI. In addition, we investigated the association between morphological indexes, ocular parameters, fundus features, and ocular shapes.

Methods

This cross-sectional study was conducted following the tenets of the Declaration of Helsinki at the Zhongshan Ophthalmic Center, Sun Yat-sen University, and was approved by the relevant ethical review committee. This study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Participants

Clinical data of participants with high myopia who registered at our center between May 2018 and May 2021 were reviewed. Registered participants were voluntary and received no stipends or incentives. All participants underwent a comprehensive ophthalmic examination, including MRI, after written informed consent was obtained. One eye of each participant was randomly included to determine eligibility. According to the inclusion and exclusion criteria (see the eMethods in Supplement 1), only highly myopic eyes with pathological fundus change (myopic atrophic maculopathy [MAM] grading C2 [diffuse chorioretinal atrophy] or more¹⁰) were included. Because this study was conducted at a tertiary specialist hospital and MRI is not a routine examination in healthy individuals, the control group of nonpathologic myopia (AL less than 26.0 mm and MAM less than C2) was screened from the excluded eyes.

Ocular Shape Classification System and Morphological Indexes

A 3-D model of the eye is formed by volume rendering based on MRI localization sequences and intraocular-fluid imaging sequences (details are described in eTable 1 and eMethods in Supplement 1). The classification system was based on the 3-D MRI results and consists of 2 parts describing the equatorial and posterior regions, respectively (eMethods in Supplement 1). Five types of equatorial ocular shapes and 6 types of posterior shapes were identified (Figure 1 and Figure 2) with detailed definitions provided in eTable 2 in Supplement 1. The posterior shape classification part was a modification of previous classifications described by Moriyama et al⁵ and Guo et al⁴ (eTable 2 in Supplement 1 for a comparison of definitions). Included morphological indexes were ocular volume, surface area, and sphericity,¹¹ as well as equatorial area, perimeter, maximum diameter, and circularity^12,13 were measured (eMethods in Supplement 1). Circularity and sphericity describe the deviation of an equator from a perfect circle and of an eye from a perfect sphere, respectively. Both indexes take values from 0 to 1; the closer to 1, the smaller the deviation.

Figure 1. Illustration of Equatorial Shape Classification, Including Typical Examples and the Corresponding Diagrams.

Equatorial shapes are shown in the frontal, bottom, and side views. Some eye models have been resized for demonstration purposes without affecting the shape.

Figure 2. Illustration of Posterior Shape Classification, Including Typical Examples and the Corresponding Diagrams.

Posterior shapes are shown in the inferior, nasal, and back views. The reconstructed optic nerve is presented to demonstrate the attachment site. The back-view diagrams are aimed at showing possible posterior protrusion(s) with dashed lines, rather than the true outline of the eye. Some eye models have been resized for demonstration purposes without affecting the shape.

Ophthalmic Assessment

All participants underwent a comprehensive ophthalmic examination including assessment of refraction error, AL, best-corrected visual acuity (BCVA), corneal parameters, etc, the details of which can be found in the eMethods in Supplement 1. MAM was classified according to the Meta Analysis for Pathologic Myopia classification¹⁰ and myopic tractional maculopathy (MTM) was classified following the atrophy-traction-neovascularization classification.¹⁴ The presence of posterior staphyloma,^1,2,15 myopic neovascularization (MNV),¹⁶ dome-shaped macula,¹⁷ and vitreoretinal interface abnormalities including epiretinal membrane, posterior vitreoschisis, and vitreomacular traction were also documented. Details of fundus evaluation are presented in the eMethods in Supplement 1.

Statistical Analysis

One-way analysis of variance or Kruskal-Wallis H test were performed to evaluate differences in continuous data among subgroups. Pearson χ² test was used to detect differences in categorical data among subgroups with Cramer V measuring the strength of association. Spearman rank correlation and linear regression analysis were used to measure the strength of the association between morphological indexes and the classification system and ocular parameters, respectively. Cohen κ coefficient was used to estimate the strength of agreement between the 2 graders. 95% CIs for coefficients were provided. All reported P values were 2-sided and not adjusted for multiple comparisons and significance was set at P < .05. SPSS version 21.0 (IBM, SPSS) was used for statistical analysis.

Results

A total of 180 pathologic myopic eyes of 180 participants were included after the screening procedure (eFigure 1 in Supplement 1). In addition, 10 nonpathologic myopic eyes of 10 participants were used as the control group. The comparisons of demographic and clinical characteristics are summarized in eTable 3 in Supplement 1.

Ocular Shape of Control Group

Of the 10 control eyes, 8 had a round equator, 1 had a pyriform equator protruding toward the superior-nasal quadrant, and 1 had a vertical-elliptical equator. All eyes showed a spheroidal posterior region.

Equatorial Shape of Pathologic Myopia

Pyriform shape was the predominant equatorial shape, which was found in 66 of 180 eyes (36.7%). The most frequent protruded direction was superior temporal (32 of 66 [48.5%]), followed by superior nasal (25 of 66 [37.9%]), inferior nasal (7 of 66 [10.6%]), and inferior temporal (2 of 66 [3.0%]). The second most common equatorial shape was round (45 of 180 [25.0%]), followed by vertical-elliptical (36 of 180 [20.0%]), horizontal-elliptical (17 of 180 [9.4%]), and rectangular (16 of 180 [8.9%]). The κ coefficient was 0.811 (95% CI, 0.742-0.880; P < .001), indicating a substantial classification agreement.

Equatorial shapes differed in equatorial area, perimeter, maximum diameter, and circularity (all P < .001), and circularity was correlated with equatorial shape by the classification order (r = −0.469; 95% CI, −0.584 to −0.346; P < .001) (Table 1). Circularity of round equator was similar to that of the control (mean [SD]; 0.9945 [0.002] vs 0.9944 [0.002]; P = .94), but its area, perimeter and maximum diameter were larger (mean [SD]; area, 467.77 [43.09] vs 418.39 [36.44] mm²; P < .001; and perimeter, 76.80 [3.53] vs 72.65 [3.09] mm; P < .001; and maximum diameter, 24.92 [1.19] vs 23.68 [0.95] mm; P = .001), and round equator was the closet to the ideal circle compared with the other shapes (circularity, mean [SD], 0.9945 [0.002]; all P < .001)(Table 1; eTable 4 in Supplement 1. In addition, horizontal-elliptical equator had the largest diameter than other shapes (mean, 26.18 [SD, 1.33], mm; all P < .001) (Table 1; eTable 4 in Supplement 1). Equatorial area (R² = 0.412; 95% CI, 0.304-0.596; P < .001), perimeter (R² = 0.415; 95% CI, 0.291-0.526; P < .001), and maximum diameter (R² = 0.396; 95% CI, 0.275-0.513; P < .001) showed a linear correlation with AL, whereas no correlation was found between equatorial circularity and AL (eFigure 2 in Supplement 1). In addition, vision impairment and circularity showed a negative correlation (r = −0.165; 95% CI, −0.289 to −0.040; P = .02).

Table 1. Morphological Indexes of Equatorial Shape Classification.

Equatorial shape	Control	Round	Rectangular	Pyriform	Vertical-elliptical	Horizontal-elliptical	P value 1a	P value 2b	r (95% CI)c
No. of eyes	10	45	16	66	36	17	NA	NA	NA
Area, mean (SD), mm2	418.39 (36.44)	467.77 (43.09)	448.04 (35.44)	461.40 (35.84)	450.08 (33.47)	494.20 (48.73)	<.001d	.13e	0.110 (−0.049 to 0.265)
Perimeter, mean (SD), mm	72.65 (3.09)	76.80 (3.53)	75.28 (3.04)	76.46 (2.97)	75.49 (2.79)	79.07 (3.93)	<.001d	.09e	0.123 (−0.037 to 0.259)
Maximum diameter, mean (SD), mm	23.68 (0.95)	24.92 (1.19)	24.67 (0.99)	25.02 (1.00)	24.88 (0.93)	26.18 (1.33)	<.001d	<.001e	0.250 (0.106-0.386)
Circularity, mean (SD)	0.9944 (0.002)	0.9945 (0.002)	0.9919 (0.003)	0.9903 (0.003)	0.9912 (0.003)	0.9911 (0.002)	<.001f	<.001e	−0.469 (−0.584 to −0.346)

Abbreviation: NA, not applicable.

^a P value for intergroup comparison of the variables.

^b P value for correlation measurement between the variables and the classification by order.

^c Correlation coefficient for measuring the correlation between the variables and the classification by order, followed by the 95% CIs.

^d One-way analysis of variance.

^e Spearman rank correlation.

^f Kruskal-Wallis H test.

Equatorial shapes differed in AL but not in BCVA or age (eTable 5 in Supplement 1). Eyes with a horizontal-elliptical equator showed the longest AL the longest AL (mean, 32.14 [SD, 3.11] mm; P < .001) compared with eyes with a round pyriform, vertical-elliptical and rectangular equator (eTable 6 in Supplement 1). In terms of corneal parameters in phakic eyes, the flat keratometry axis was most horizontally oriented in eyes with a horizontal-elliptical equator (median, 43.55 [interquartile range, 43.84] degrees; P = .01) compared with that of eyes with a round pyriform vertical-elliptical or rectangular equator (eTable 6 and eTable 7 in Supplement 1). In addition, equatorial shape was not associated with a history of keratorefractive surgery (eTable 8 in Supplement 1).

Posterior Shape of Pathologic Myopia

The predominant posterior shape was bulb-shaped (97 of 180 [53.9%]), followed by multidistorted (46 of 180 [25.6%]), conical (12 of 180 [9.4%]), ellipsoidal (9 of 180 [5.0%]), barrel-shaped (7 of 180 [3.9%]), and spheroidal (4 of 180 [2.2%]). The κ coefficient was 0.829 (95% CI, 0.762-0.896; P < .001), indicating a substantial agreement for the posterior classification.

Ocular volume, surface area, and sphericity were different between posterior shapes (all P < .001), correlated with the classification order (r = 0.538; 95% CI, 0.430-0.625; r = 0.562; 95% CI, 0.462-0.647; and r = −0.533; 95% CI, −0.627 to −0.427, respectively; all P < .001) (Table 2) and showed linear correlations with AL (R² = 0.750; 95% CI, 0.684-0.803; r = 0.786; 95% CI, 0.728-0.834; and r = 0.220; 95% CI, 0.132-0.314, respectively; all P < .001) (eFigure 3 in Supplement 1). Multiple comparisons of morphological indexes between posterior shapes were summarized in eTable 9 in Supplement 1. Age (r = 0.233; 95% CI, 0.086-0.369; P = .002), AL (r = 0.630; 95% CI, 0.529-0.715; P < .001), and BCVA (r = 0.279; 95% CI, 0.142-0.418; P < .001) were associated with posterior shapes by the classification order (eTable 10 in Supplement 1). Vision impairment and sphericity showed a negative correlation (r = −0.266; 95% CI, −0.395 to –0.137; P < .001). In addition, posterior shape was not associated with a history of keratorefractive surgery (eTable 8 in Supplement 1).

Table 2. Morphological Indexes of Posterior Shape Classification.

Posterior shape	Control	Spheroidal	Conical	Bulb-shaped	Ellipsoidal	Multi-distorted	Barrel-shaped	P value 1a	P value 2b	r (95% CI)c
No. of eyes	10	4	17	97	9	46	7	NA	NA	NA
Volume, mean (SD), mm3	6671.99 (790.60)	8112.41 (689.40)	8086.45 (719.02)	8331.18 (1228.05)	9402.72 (852.12)	9607.42 (1518.45)	11 178.82 (1172.60)	<.001d	<.001e	0.538 (0.430-0.625)
Surface area, mean (SD), mm2	1703.28 (134.70)	1978.37 (112.12)	1979.53 (116.27)	2026.94 (197.98)	2195.57 (130.92)	2238.93 (233.24)	2486.73 (172.31)	<.001d	<.001e	0.562 (0.462-0.647)
Sphericity, mean (SD)	0.9844 (0.004)	0.9871 (0.003)	0.9834 (0.003)	0.9781 (0.006)	0.9804 (0.004)	0.9737 (0.006)	0.9729 (0.004)	<.001d	<.001e	−0.533 (−0.627 to −0.427)

Abbreviation: NA, not applicable.

^a P value for intergroup comparison of the variables.

^b P value for correlation measurement between the variables and the classification by order.

^c Correlation coefficient for measuring the correlation between the variables and the classification by order, followed by the 95% CIs.

^d Kruskal-Wallis H test.

^e Spearman rank correlation.

Staphyloma was found in 158 of 180 eyes (87.0%) and the staphyloma location was summarized in eTable 11 in Supplement 1. MAM (Cramer V = 0.344; 95% CI, 0.300-0.455; P < .001) and MTM (Cramer V = 0.253; 95% CI, 0.229-0.366; P < .001) had different distributions among the posterior shapes (eFigure 4 in Supplement 1). The proportion of advanced MAM rose according to the classification order, whereas the prevalence of MTM increased and then decreased. In addition, more dome-shaped macular were identified in the barrel-shaped eyes (Cramer V = 0.312; 95% CI, 0.181-0.536; P = .004), whereas no difference was found in the MNV or vitreoretinal interface abnormalities prevalence (eFigure 5 in Supplement 1).

Association Between Equatorial and Posterior Shape

Equatorial circularity was not correlated with ocular sphericity in general (eFigure 6 in Supplement 1). However, in the subgroup linear analysis, the circularity showed a positive correlation with sphericity in the control (R² = 0.591; 95% CI, 0.014-0.968; P = .01) and the eyes with a vertical-elliptical equator (R² = 0.246; 95% CI, 0.050-0.496; P = .002), whereas no correlation was found in the other pairs.

The distributions of equatorial shapes were different across the types of posterior shapes (Cramer V = 0.213; 95% CI, 0.188-0.337; P = .04) (eFigure 6 in Supplement 1). Pyriform was the predominant equatorial shape in conical (8 of 17 [47.1%]), bulb-shaped (35 of 97 [36.1%]), ellipsoidal (5 of 9 [55.6%]), and multi-distorted (16 of 46 [34.8%]) eyes. The pyriform equator also has a different distribution of the protruded direction across different posterior shapes (Cramer V = 0.466; 95% CI, 0.235-0.660; P < .001). Similar to the overall results, superior-temporal was the dominant direction in conical (6 of 8 [75.0%]) and bulb-shaped (19 of 35 [54.3%]) eyes, whereas superior-nasal was the dominant direction in ellipsoidal (3 of 5 [60.0%]) and multi-distorted (9 of 16 [56.3%]) eyes. In addition, the dominant equatorial shape in barrel-shaped eyes was different, being the horizontal-elliptical equator (3 of 7 [42.9%]). The horizontal-elliptical equator appeared frequently in multi-distorted eyes (10 of 46 [21.7%]), where the ocular deformation was marked.

The prevalence of inferior staphyloma was different in equatorial shapes (Cramer V = 0.232; 95% CI, 0.124-0.420; P = .04) (eFigure 6 in Supplement 1). Eyes with a vertical-elliptical equator had the highest prevalence at 27.8% (10 of 36), where the inferior staphyloma appeared as a bulb-shaped posterior region with an inferior edge that was smoothly contiguous with a vertically expanded equator.

Discussion

In this study, we proposed a classification system to investigate the various patterns of ocular shape in pathologic myopia which contributed to the previous systems.^4,5 Our classification system may have several advantages. First, the interobserver agreement was excellent. In particular, eyes with multiple posterior protrusions can be quickly classified without judging the symmetry. Second, we investigated the equatorial morphology in detail. We found diverse equatorial patterns in pathologic myopia and their association with ocular parameters and global morphology. Third, morphological indexes further confirmed the ability of our classification to elaborate on the degree of ocular deformation. In addition, our classification system can validate the differences in myopic fundus manifestations and ocular parameters across ocular shapes like the previous systems.^4,5,6,18

The equatorial shape in pathologic myopia is not always round. Ohno-Matsui³ reported a noncircular shape viewed from the posterior, which presented as pyramidal or horizontal-elliptical in 22 highly myopic eyes. In our study, various patterns of equatorial deformation were identified in pathologic myopia and the equator of 2 nonpathologic myopic eyes in the control group was not round.

The mechanism underlying equatorial deformation is unclear. Equatorial morphology may be influenced by extraocular anatomical factors.³ In the Ohno-Matsui’s study,³ irregular equators protruded only superior nasally. However, we found that superior-temporal was the predominant direction in similar eyes, which may be a result of degeneration. The superior-temporal quadrant lacks the envelopment of extraocular elements and the intermuscular membrane degeneration was marked in high myopia.¹⁹ Subsequently, thin sclera may protrude toward this weakest quadrant. However, not all quadrants can be explained by this degeneration hypothesis.

Another possible explanation is that these shapes are driven by optical mechanisms. In infant monkeys, peripheral defocus and form deprivation were shown to play an important role in myopia development and to act in a region-selective manner.^20,21 In addition to causing central axial myopia, unbalanced signals due to asymmetric peripheral vision may contribute to the irregular development of equator. In turn, the deformed equator may further alter peripheral vision, leading to myopia progression. A recent study in children²² found that wearing keratoplasty lenses resulted in less nasal hyperopic defocus and a more symmetrical posterior shape. However, we did not find differences in ocular shape with a history of corneal surgery in adulthood. This may be due to the relatively small sample size or to the effects of peripheral vision on ocular shape that may occur primarily in childhood.

Equatorial circularity and ocular sphericity were correlated in the eyes with a vertical-elliptical equator, indicating the contribution of equator to the global deformation. Vertical-elliptical equator may be due to the expansion of the inferior part of eye, which is considered to be weak and susceptible to pathological expansion.^3,8,23,24 This may also explain the higher prevalence of coexisting inferior staphyloma (Curtin type V). In addition, the eye position is closer to the orbital roof,^25,26 which allows more space below and may contribute to the vertical expansion of equator. An association between the regional shapes was also observed in eyes with a horizontal-elliptical equator. Horizontal-elliptical equator often coexists with a multidistorted or barrel posterior shape where the posterior sclera was greatly horizontally stretched. In addition, eyes with a horizontal-elliptical equator had the most horizontally oriented flat keratometry axis, suggesting that due to extensive scleral remodeling, ocular shape may be reflected as a measurable difference in some corneal parameters.²⁷

Identification and classification of equatorial morphology is clinically relevant. First, different posterior shapes may indicate that a particular equatorial morphology is more likely to accompany them. Predicting equatorial shape based on posterior morphology may help protect these potentially vulnerable regions during surgery involving the equator and alert the examiner to possible peripheral retinal and choroidal degeneration. Second, we speculate that equatorial morphology may influence the force pathways of extraocular muscles during contraction and may affect eye position and movement in myopia. Third, as discussed above, equatorial morphology may be associated with myopia development and may contribute to the design of myopia control systems based on peripheral vision.

In pathologic myopia, spheroidal and ellipsoidal shapes may be the result of uniform global expansion and equatorial elongation, respectively.²⁶ However, the increase in ocular volume and area is sufficient to cause marked macular degeneration in the absence of local protrusion, which was different from the relatively young population with high myopia.⁴ We also separated conical shape from bulb shape and multi-distorted shape from barrel shape as distinct entities with different irregularities and severity of myopic fundus complications, which were confounded in previous studies.^4,5,28 Although the conical shape does not meet the definition of staphyloma,^1,2 it may represent the early stage of staphyloma where the deformation may be sufficient to cause myopic maculopathy. In addition, posterior expansion may also increase the possibility of Bruch membrane rupture and retinal pigment epithelium exposure, facilitating the development of MNV.¹⁴ Miyake et al¹⁸ reported that MNV tended to occur in a flat and smooth fundus.¹⁸ However, Ohno-Matsui et al⁶ observed a higher incidence of MNV in eyes with irregular curvature. In our study, we found that the prevalence of MNV was not associated with posterior shape, emphasizing that clinicians should be aware of the risk of MNV in general.

Limitations

The study has limitations. First, this study was conducted at a tertiary referral specialist hospital and there may be selection bias when we recruited the participants. Therefore, the proposed classification should be carefully extended. Second, although we proposed a classification of equatorial morphology in pathologic myopia, due to the sample size of control group, the exact definition of pathological equatorial shape needs further studies sand the lack of longitudinal data prevents us from determining the natural alteration of these shapes in the myopic progression in this study. Third, although ocular sphericity was found to correlate well with posterior shape, it may include regularities of the anterior, equatorial, and posterior regions. The morphological indexes need further validation and modification to represent the ocular deformation more accurately and perhaps be used as a clinical indicator.

Conclusion

In conclusion, this cross-sectional study suggests that ocular deformation in pathologic myopia is global and diverse and that these changes are associated with ocular parameters, such as visual acuity and myopic fundus manifestations. The morphological classification may enhance the understanding of the diverse patterns of ocular shape in pathologic myopia.

Supplement 1.

eMethods

eFigure 1. Flowchart of the screening of study participants and eyes

eFigure 2. Linear regression analysis of equatorial morphological indexes and axial length

eFigure 3. Linear regression analysis of ocular morphological indexes and axial length

eFigure 4. Macular atrophic and tractional maculopathy in different posterior shapes

eFigure 5. Myopic fundus features in different posterior shapes

eFigure 6. Association between equatorial and posterior shape

eTable 1. Three-dimensional magnetic resonance imaging sequences

eTable 2. Definitions of ocular shapes in the proposed classification and comparison with previously described classifications

eTable 3. The demographic and characteristics of the participants and eyes.

eTable 4. Multiple comparisons of morphologic indexes between control and different equatorial shapes

eTable 5. Age and ocular parameters for equatorial shapes

eTable 6. Multiple comparisons of axial length and the axis of flat keratometry between different equatorial shapes

eTable 7. Astigmatism and corneal parameters for equatorial shapes

eTable 8. Equatorial and posterior shapes in eyes with and without a history of keratorefractive surgery

eTable 9. Multiple comparisons of morphological indexes between control and posterior shapes

eTable 10. Age, axial length, and best-corrected visual acuity for different posterior shapes

eTable 11. Location of staphyloma involvement in different posterior shapes

Supplement 2.

Data sharing statement