Underlying Microstructure of the Lamina Cribrosa at the Site of Microvasculature Dropout

Abstract

Purpose: To determine the microstructure of the lamina cribrosa (LC) associated with microvasculature dropout (MvD) of the deep optic nerve head (ONH) in primary open-angle glaucoma (POAG) and to identify factors related to the presence of MvD.

Methods: POAG eyes that exhibited MvD in the LC (MvD-LC) or MvD in the peripapillary choroid (MvD-PC) underwent optical coherence tomography and optical coherence tomography angiography (OCTA) to evaluate the structure and microvasculature of the deep ONH, respectively. The presence of MvD-LC or MvD-PC was determined using en face OCTA images of the deep ONH. The sectoral LC thickness (LCT) and LC curvature index (LCCI) (at MvD-LC site, when applicable), the mean LCT and LCCI of the global ONH, and other clinical characteristics were measured and compared between eyes with and without MvD-LC.

Results: The study included 93 eyes with and 51 without MvD-LC. The presence of MvD-LC was associated with lower sectoral LCT (odds ratio [OR] = 0.96, P < 0.001) and mean LCT (OR = 0.97, P = 0.032), larger visual field pattern standard deviation (PSD; OR = 1.20, P = 0.038), and higher pretreatment intraocular pressure (IOP; OR = 1.22, P = 0.012). Fifteen percent of the eyes with MvD-LC (14/93) did not present MvD-PC. Those eyes had younger age (P = 0.043), thicker juxtapapillary choroid (P = 0.018), larger sectoral LCCI (P = 0.040), thicker retinal nerve fiber layer (P = 0.024), smaller PSD (P = 0.008), and higher pretreatment IOP (P = 0.006) than those with both MvD-LC and MvD-PC.

Conclusions: MvD-LC was associated with a localized morphologic alteration of the LC, and eyes with MvD-LC tended to have a higher pretreatment IOP. The clinical implications of MvD-LC should differ from those of MvD-PC in eyes with POAG.

Understanding the perfusion of the deep optic nerve head (ONH) is crucial for elucidating the pathophysiology of glaucomatous optic neuropathy. Studies using optical coherence tomography angiography (OCTA) have identified vascular impairments in the tissues of the ONH and peripapillary area.^1–15 Localized microvasculature dropout (MvD) in the peripapillary choroid (MvD-PC) observed in the deep layer en face OCTA image is of particular interest and is often observed at the site of glaucomatous damage.^5–13 Deep ONH perfusion is believed to be largely dependent on circulation in the adjacent parapapillary choroidal, which emits fine centripetal branches to the prelaminar region.¹⁶ MvD-PC may therefore indicate localized perfusion impairment of the ONH in glaucoma. MvD-PC corresponds with a perfusion defect in the choroid, as revealed by indocyanine green angiography (ICGA),⁸ and is frequently accompanied by paracentral visual field (VF) defects, indicating vascular insufficiency.⁶ Recent studies also found that the MvD-PC was significantly associated with faster glaucoma progression, indicating its pathogenic importance in glaucomatous optic neuropathy.^17,18 

MvD can also be found in the deep tissues of the ONH. MvD in the optic disc, identified using OCTA, was found to match the filling defect within the ONH as revealed by ICGA,⁸ suggesting that the optic-disc MvD indicates a true perfusion defect in the ONH. The optic-disc MvD was observed at the MvD-PC site in eyes with glaucoma^2,19 and was associated with worse VF mean deviation (MD).² Progressive retinal nerve fiber layer (RNFL) thinning occurred more rapidly when MvD-PC was accompanied by optic-disc MvD.¹⁰ However, the pathogenic mechanisms of optic-disc MvD and its relationship with MvD-PC remain poorly understood. 

Recent advancements in OCTA technology, including projection artifact removal and multiple image averaging, have enabled more accurate imaging of the microvasculature in the deep tissues of the ONH, including the lamina cribrosa (LC).^20,21 A previous study found significant improvement in vessel density in the deep ONH with reversal of the LC curvature following surgery to reduce intraocular pressure (IOP).²² This finding suggests that the microvasculature within the deep ONH may be influenced by morphological changes of the LC. 

The purpose of the present study was to determine the structural characteristics of the ONH underlying MvD in the LC (MvD-LC) and to determine the clinical characteristics associated with MvD-LC in eyes with primary open-angle glaucoma (POAG). 

Methods

This study enrolled patients with POAG who participated in the Investigating Glaucoma Progression Study, which is an ongoing prospective study of patients with glaucoma at the glaucoma clinic of Seoul National University Bundang Hospital. All subjects provided written informed consents for their participation. The study protocol was approved by the Institutional Review Board of Seoul National University Bundang Hospital and followed the tenets of the Declaration of Helsinki. 

All subjects underwent comprehensive ophthalmic examinations, which included best-corrected visual acuity assessments, Goldmann applanation tonometry, refraction tests, slit-lamp biomicroscopy, gonioscopy, stereo disc photography, red-free fundus photography (using an EOS-D60 digital camera; Canon, Tokyo Japan), measurement of peripapillary RNFL thickness, enhanced depth-imaging scanning of the ONH using spectral-domain OCT (SPECTRALIS; Heidelberg Engineering, Heidelberg, Germany), swept-source OCT and OCTA of the ONH and peripapillary area (DRI OCT Triton; Topcon, Tokyo, Japan), and standard automated perimetry (Humphrey Field Analyzer II 750, 24-2 Swedish Interactive Threshold Algorithm; Carl Zeiss Meditec, Dublin, CA, USA). Other ophthalmic examinations included measurements of corneal curvature (KR-800; Topcon), central corneal thickness, and axial length (IOL Master 5; Carl Zeiss Meditec). 

Untreated IOP was defined as that measured before the initiation of ocular hypotensive treatment or as identified in the referral notes. Diurnal variation in patients with an untreated IOP of <21 mm Hg was measured every 2 hours from 9 AM to 5 PM, with the mean of the five measurements considered the untreated IOP for that patient. In patients with an untreated IOP of >21 mm Hg, IOP was measured twice before IOP-lowering medication was administered, with the mean of the two measurements considered the untreated IOP for that patient. Patients who were undergoing treatment with ocular hypotensive medication at the time of the initial visit received the diurnal variation measurements after a 4-week washout period. 

Systolic and diastolic blood pressures (BPs) were measured using a digital automatic BP monitor (Omron HEM-770A; Omron, Kyoto, Japan). The mean arterial pressure (MAP) was calculated as diastolic BP + 1/3(systolic BP – diastolic BP), and the mean ocular perfusion pressure was calculated as 2/3(MAP – IOP) at the time of OCTA being performed. 

POAG was defined as the presence of an open iridocorneal angle and a glaucomatous VF defect, with signs of glaucomatous optic nerve damage (i.e., neuroretinal rim thinning or notching or an RNFL defect). A glaucomatous VF defect was defined as a defect that met one or more of the following criteria: (1) outside normal limits on a Glaucoma Hemifield Test; (2) three abnormal points with a <5% probability of being normal and one abnormal point with a <1% probability of being normal according to pattern deviation; or (3) a pattern standard deviation (PSD) of probability < 5% confirmed on two consecutive reliable tests. Reliability in these tests was defined as fixation loss, false-positive, and false-negative rates of 20%, 15%, and 25%, respectively. 

Considering that the normal morphological characteristics of the ONH^23–25 and thickness profiles of the PC^26,27 differed between the superior and inferior sectors, and that most occurrences of MvD-PC and MvD-LC were located in the inferior sectors,^2,10,19 only eyes with ONH damage to the inferior sectors were included in the present study. 

Eyes were excluded if they had a best-corrected visual acuity worse than 20/40, a spherical equivalent of <−9.0 D or >+3.0 D, a cylinder correction of <−3.0 D or >+3.0 D, a history of intraocular surgery with the exception of uneventful cataract surgery, retinal diseases (e.g., diabetic retinopathy, retinal vessel occlusion, retinoschisis), or neurological diseases (e.g., pituitary tumor). Eyes with γ-zone parapapillary atrophy (PPA)²⁸ were also excluded because the microstructure underlying MvD-PC differs between PPA in the β- and γ-zones, with MvD-PC in the latter consisting only of border tissue, not choroid, which indicates a different pathomechanism.²⁹ When a subject had two eligible eyes, one was randomly selected for inclusion in the study. 

OCTA Assessments of MvD

The optic nerve and peripapillary area were imaged using the commercially available swept-source DRI OCT Triton OCTA device operating at a central wavelength of 1050 nm, an acquisition speed of 100,000 A-scans per second, and axial and transverse resolutions in the tissue of 7 mm and 20 mm, respectively. Scans were taken from 4.5-mm × 4.5-mm cubes, with each cube consisting of 320 clusters of four repeated B-scans centered on the optic disc. The values of corneal curvature and axial length were entered into the DRI OCT system before the scan to eliminate magnification errors. En face projections of volumetric scans allowed visualization of the structural and vascular details within segmented layers. An eye was excluded from the analysis when the OCTA images were of poor quality (e.g., due to blurring) or when the vascular signal was blocked by artifacts (e.g., due to blinking or masking).³⁰ 

The DRI OCT Triton device allows the microvasculature in regions of interest to be evaluated in a customized manner. Using manual segmentation, en face OCTA images were first produced from the segmented layers of the LC (from the anterior to posterior borders of the LC) and PC (below the retinal pigment epithelium, including the choroid and anterior sclera). 

MvD-PC was defined as a focal sectoral capillary dropout with no visible microvascular network in the PC in the en face images,^7,8 and MvD-LC was defined as a complete loss of all OCTA signals within the ONH at the LC.^2,19 MvD-PC or MvD-LC was considered to be present when the dropout had a maximum diameter of >200 µm.^2,13,19 Only the temporal side of the vertical line passing through the centroid of the optic disc was used in the analysis because the ability to analyze the deep-layer microvasculature at the nasal portion of the ONH is restricted by the overlying neural rim and large vessels.^2,10,19,31 Two observers (YJR, DKH) independently identified MvDs while masked to the clinical information of the subjects. Disagreements between these observers were resolved by a third adjudicator (EJL). 

All included OCT B-scan images had image quality scores > 30, which were considered acceptable based on the manufacturer's recommendation. Eyes with poor-quality OCTA images (e.g., blurred images that hampered MvD delineation) were excluded from the analysis. 

Measuring the MvD Areas

The MvD-PC and MvD-LC areas were measured in the en face OCTA images using the caliper tool of the IMAGEnet software provided with the DRI OCT Triton device. Because the absolute areas of MvD-PC and MvD-LC may differ with the areas of the PPA and optic disc, respectively, the MvD-PC and MvD-LC areas were calculated relative to the β-PPA and optic disc areas, respectively (Fig. 1). The measurements were made by two observers (YJR, DKH) who were masked to clinical information. The mean values of the measurements made by these two observers were used in the main analysis. 

Figure 1.

 

Measurements of the areas of MvD-PC (red area) and MvD-LC (yellow area) relative to the β-PPA (blue area) and optic disc (light-blue area) areas. Upper (A–D) and lower (E–H) images show disc photographs (A, B, E, F) and en face OCTA images of the deep ONH (C, D, G, H) of the eyes having only MvD-PC and having both MvD-PC and MvD-LC, respectively.

Investigations of LC Morphology

The LC morphology was assessed by measuring the LC curvature and thickness on the volume and radial scan images of the ONH obtained by enhanced depth-imaging OCT (Fig. 2). The general morphology of the LC was determined by obtaining the mean LC curvature and mean LC thickness (LCT) from the horizontal B-scans of the whole ONH. The LC curvature was evaluated using the LC curvature index (LCCI), defined as the degree of inflection of the curve representing a section of the LC.³² The LCCI was determined by measuring the width of the anterior LC surface reference plane (W) and then by measuring the LC curve depth (LCCD) within the anterior LC surface plane in each B-scan, with LCCI calculated as follows: (LCCD/W) × 100. Because the curvature was normalized to the LC width, the LCCI is an indicator of the shape of the LC independent of the actual size of the ONH. Only the LC within the measured width was considered, because the LC was often not clearly visible outside this region. In eyes with LC defects, the LCCI was measured using a presumed anterior LC surface that represented the best fit to the curvature of the remaining part of the LC or by excluding the area of the LC defect. The LCCI was obtained from the three selected B-scan images (midhorizontal, superior midperipheral, and inferior midperipheral regions of the ONH) to calculate the mean LCCI of the eye. The LCT was assessed in the same three scans that were used to measure the LCCI and measured as the distance between the anterior and posterior borders at the central three points (100 µm from each point) in the direction perpendicular to the anterior LC surface at the measurement point.³³ The measurements made on the three images were used to calculate the mean LCT of the eye. 

Figure 2.

 

Measurements of the global and sectoral LCCI and LCT and the sectoral JPCT. (A) Disc photograph showing three horizontal planes (green lines) where the global LCCI and LCT were measured. (B) B-scan image obtained in the inferior midperipheral plane (A, solid green line). The LCCI was measured by dividing the LCCD by the width of the anterior LC surface reference line (W) connecting the two points meeting the perpendicular lines from the Bruch's membrane opening plane, then multiplying by 100. The LCT was measured as the distance between the anterior and posterior borders (red dots) at the most central three points (100 µm from each point) in the direction perpendicular to the anterior LC surface at the measurement point. (C) En face OCTA image of the deep ONH showing the location where the sectoral LCCI, LCT, and JPCT were measured (green line). (D) Radial B-scan image obtained at the sites of MvD-PC and MvD-LC (C, green line). The sectoral LCCI and LCT were measured in the same manner as that used to measure the global LCCI and LCT, but the measurements were performed within the inferotemporal sector, between the central ridge and the peripheral LC. The area of the choroidal tissue within 500 µm of the border tissue of Elschnig was measured (light-blue area), and was divided by 500 µm to obtain the sectoral JPCT.

Because the purpose of the study was to characterize the ONH structure associated with MvD-LC, the LC morphology was also assessed locally in the inferotemporal quadrant of the ONH. The measurements were made on the radial B-scan images passing through the inferotemporal sector of the ONH. The sectoral LCCI of the inferotemporal sector of the ONH was measured by dividing the sectoral LCCD (LCCDsec) by the width of the reference plane of the inferotemporal ONH quadrant (Wsec). LCCDsec was determined as the maximally depressed point of the anterior LC surface from the inferotemporal sectoral reference plane, which was defined as the line connecting the central ridge of the LC and the inferotemporal LC insertion point. Wsec was measured as the distance between the point of the central LC ridge and the inferotemporal LC insertion point. The sectoral LCCI was calculated as (LCCDsec/Wsec) × 100. The sectoral LCT was assessed in the same radial scan, and measured in the same manner as that used to measure the global LCT. 

The manual caliper tool in the Heidelberg Eye Explorer software (version 1.10.4) was used to measure the LCCI and LCT. Measurements were made by two observers (YJR, DKH) who were masked to clinical information. The mean values of the measurements made by these two observers were used in the main analysis. 

A focal LC defect was defined as an anterior laminar surface irregularity violating the normal smooth curvilinear U- or W-shaped contour.³⁴ To avoid false positives, defects had to be >100 µm in diameter and >30 µm in depth.³⁵ The radial B-scan images were independently reviewed by two observers (YJR, DKH) who were masked to all other clinical information, and the presence of focal LC defects was determined. The B-scan locations were subsequently compared with the stereo disc photographs to ensure that any identified focal LC defects were not artifacts caused by vascular shadowing. The presence of a focal LC defect was double checked on horizontal B-scans. 

Measurement of Juxtapapillary Choroidal Thickness

The juxtapapillary choroidal thickness (JPCT) of the inferotemporal sector was measured on the same radial B-scan image that was used for evaluating the sectoral LC morphology (Fig. 2). The area of the choroidal tissue within 500 µm of the border tissue of Elschnig was measured using the built-in drawing tool of the SPECTRALIS viewer software. The sectoral JPCT was obtained by dividing the measured choroidal area by 500 µm. Measurements were performed by two masked observers (YJR, DKH), and the mean of the measurements was used in the analysis. 

Although subfoveal choroidal thickness shows diurnal variation,^36–38 little is known about whether JPCT is also influenced by circadian changes. Assuming that any impact of diurnal variation would be randomly distributed among the groups, we did not consider the time of OCT scanning used for measuring JPCT in the inferotemporal sector. 

Data Analysis

Except where stated otherwise, all data are presented as mean and standard deviation values. Interobserver agreement regarding the presence of MvD-LC or MvD-PC was assessed using kappa statistics (κ values). The interobserver reproducibilities in determining the relative areas of MvD-LC and MvD-PC and LCT, LCCI, and JPCT were assessed by calculating intraclass correlation coefficients (ICCs). Comparisons between groups were performed using independent-samples t-tests for parametric data and Mann-Whitney U tests for nonparametric data or comparison between groups with small sample size (n < 30). Factors associated with the presence of MvD-LC were assessed using logistic regression analysis, first with a univariable model and subsequently with a multivariable model that included variables in the univariable model for which P < 0.10. All statistical analyses were performed using the SPSS Statistics 22.0 (IBM, Chicago, IL, USA), with P < 0.05 considered significant. 

Results

Of the 212 eyes with POAG that were initially enrolled, 27, 15, and 12 were excluded because of poor-quality OCTA images or B-scan images that did not allow clear delineation of MvD, the LC, or the choroid, respectively. An additional 14 eyes were excluded due to the presence of glaucomatous damage in the superior sector. MvD-LC was identified in 93 of the remaining 144 eyes with POAG that initially met our inclusion criteria: 79 MvD-LCs were accompanied by an adjacent MvD-PC, and 14 presented as isolated MvD-LC without an accompanying MvD-PC. Fifty-one eyes without MvD-LC had an MvD-PC in the inferotemporal sector. The interobserver agreements for determining the presence of MvD-LC and MvD-PC as assessed based on κ values were 0.777 and 0.970, respectively. The interobserver reproducibilities in determining the relative areas of MvD-LC and MvD-PC and the LCT, LCCI, and JPCT as assessed based on ICC values were 0.937 (95% confidence interval [CI], 0.907–0.958), 0.954 (95% CI, 0.935–0.967), 0.913 (95% CI, 0.880–0.937), 0.952 (95% CI, 0.934–0.966), and 0.976 (95% CI, 0.967–0.983), respectively. 

Table 1 lists the clinical characteristics of the study subjects. Eyes with MvD-LC were more frequently accompanied by focal LC defects (P = 0.028) and had a smaller sectoral LCT in the inferotemporal sector (P < 0.001), smaller mean LCT (P = 0.001), and larger sectoral LCCI in the inferotemporal sector (P = 0.004) than those without MvD-LC. They also had worse VF MD (P = 0.023) and PSD (P < 0.001) and higher pretreatment IOP (P = 0.005) relative to the eyes without MvD-LC. 

Table 1.

 

Comparison of the Characteristics Between the POAG Eyes Having MvD-LC and Those Without

 

The univariable logistic regression analysis showed that the significant factors associated with MvD-LC were the presence of an LC defect (odds ratio [OR] = 4.77, P = 0.044), smaller sectoral LCT in the inferotemporal sector (OR = 0.96, P < 0.001), smaller mean LCT (OR = 0.97, P = 0.001), larger sectoral LCCI in the inferotemporal sector (OR = 1.14, P = 0.005), lower MD (OR = 0.94, P = 0.025), higher PSD (OR = 1.18, P = 0.001), and higher pretreatment IOP (OR = 1.15, P = 0.005) (Table 2). The significant factors in the multivariable analysis were a smaller sectoral LCT in the inferotemporal sector (OR = 0.96, P < 0.001), smaller mean LCT (OR = 0.97, P = 0.032), higher PSD (OR = 1.20, P = 0.038), and higher pretreatment IOP (OR = 1.22, P = 0.012) (Table 2). Figures 3A to 3D and Figures 3E to 3H show representative cases with and without MvD-LC, respectively. 

Table 2

 

Factors Associated With the Presence of MvD-LC

 

Figure 3.

 

Representative cases of eyes with POAG with (A–D) and without (E–H) MvD-LC. (A, E) Color disc photographs. (B, F) En face OCTA images of the deep ONH layer indicating the radial planes where the radial B-scan images were obtained (dashed lines). (C, G) Radial B-scan images where the LC morphology was examined. (D, H) Grayscale plots from VF examinations. The eye in the upper panel shows both MvD-LC and MvD-PC (B), but only MvD-PC is evident in the eye in the lower panel (F). Note that the LCT is smaller and the VF damage is more advanced in the eye with MvD-LC than in the eye without MvD-LC (C, G; red glyphs).

The 14 of the 93 eyes with isolated MvD-LC without accompanying MvD-PC had a younger age (P = 0.043), thicker juxtapapillary choroid (P = 0.018), larger sectoral LCCI (P = 0.040), larger sectoral RNFL thickness (RNFLT; P = 0.024) in the inferotemporal sector, lower VF PSD (P = 0.008), and larger pretreatment IOP (P = 0.006) than those with both MvD-LC and MvD-PC (Table 3). In the comparison with the eyes with isolated MvD-PC without MvD-LC, younger age (P = 0.013), thinner sectoral (P < 0.001) and mean (P = 0.007) LCT, larger sectoral (P = 0.001) and mean (P = 0.046) LCCI, and larger pretreatment IOP (P < 0.001) were associated with the presence of only MvD-LC (Table 3). Figure 4 shows a POAG eye showing only MvD-LC without accompanying MvD-PC. 

Table 3.

 

Comparison of the Characteristics Between POAG Eyes Having Both MvD-LC and MvD-PC and Those Having Only MvD-LC But Not MvD-PC

 

Figure 4.

 

A POAG eye showing an MvD-LC not accompanied by an MvD-PC. (A) Color disc photograph. (B) En face OCTA image of the deep ONH layer indicating the radial planes where the radial B-scan image was obtained (dashed line). (C) Radial B-scan image where the LC morphology was examined. (D) Grayscale plots from a VF examination. A focal LC defect was detected at the MvD-LC site (C, arrow) in this eye.

Discussion

This study has elucidated the structural characteristics at the MvD-LC site and associated features in eyes with POAG, in comparison with those without MvD-LC. MvD-LC sites exhibited greater structural deformation of the LC, characterized by a thinner and more curved LC, and an increased frequency of focal LC defects. Eyes with MvD-LC tended to have a higher pretreatment IOP and more severe glaucomatous damage. 

Previous studies investigating MvD in the optic disc have utilized the whole-signal mode of OCTA to visualize the intrapapillary MvD.^2,10,19 The optic-disc MvD was accompanied by a MvD-PC in all cases, which has led to suggestions of a pathogenic link between the optic-disc MvD and MvD-PC.^2,10 

The present study focused on MvD in the deeper ONH tissue (i.e., MvD-LC), which means that the results are not directly comparable with those of previous studies. We found that some cases of MvD-LC appeared as isolated MvD-LC—that is, not accompanied by MvD-PC. Moreover, the clinical characteristics of eyes with MvD-LC differed from those with only MvD-PC. These findings indicate that the pathogenic mechanisms of MvD-LC may be independent of those of MvD-PC, suggesting that the clinical significance differs between these two conditions. Although both MvD-LC and MvD-PC are downstream of the short posterior ciliary arteries (SPCAs), there are no connections between the microvessels in the PC and LC. This makes it less likely that MvD-LC and MvD-PC are causally related to each other. 

The LC morphology was primarily assessed using LCT and LCCI, which have been suggested as indicators of increased mechanical stress^39–41 and associated with rapid glaucoma progression.^17,33,42 Focal LC defects have also been considered to be significant pathogenic sites^43–45 and crucial factors in rapid glaucoma progression.^46,47 In the present study, eyes with MvD-LC exhibited smaller sectoral LCT and mean LCT and larger sectoral LCCI relative to eyes without MvD-LC, with focal LC defects being observed more frequently in the former group. The eyes with MvD-LC also displayed more advanced VF damage and higher IOP than those without MvD-LC. The multivariable analysis identified smaller LCT, higher VF PSD, and higher IOP as significant risk factors for the presence of MvD-LC. These findings together suggest that MvD-LC represents focal mechanical disruption of the LC that is associated with increased IOP-related stress, leading to localized severe glaucomatous damage. Supporting our results, Akagi et al.² found that optic-disc MvD was associated with more advanced VF MD and focal LC defects. The inferotemporal peripapillary region is known to be susceptible to mechanical damage and aligns with the macular vulnerability zone affecting central vision.⁴⁸ This means that LC disruption in this area may result in more severe damage involving the central VF. 

We previously found that MvD-PC was not associated with structural changes in the choroid.⁷ This suggested that MvD-PC does not arise from mechanical compression of the PC vessels. Instead, it might result from partial or complete obliteration of the vessels supplying or draining the choriocapillaris, leading to a perfusion defect in a choriocapillaris lobule. In the present study, the eyes exhibiting only MvD-PC (i.e., without MvD-LC) presented a lower pretreatment IOP and smaller degree of LC deformation, implying that these eyes were less affected by IOP-related stress. MvD-PC has been considered to be an indicator of decreased ocular perfusion in glaucoma.^6,8,17 Because the PC nourishes the prelaminar tissue, a perfusion defect in the choroid could compromise the viability of the retinal ganglion cell axons. However, we found no evidence of decreased ocular perfusion—such as low systemic BP or reduced ocular perfusion pressure—in the eyes showing only MvD-PC, which may be at least partly attributable to most of the included eyes having MvD-PC. Whether different pathomechanisms underlie eyes with only MvD-PC and those with both MvD-PC and MvD-LC remains unclear. Given the less severe glaucomatous VF damage in eyes with only MvD-PC, it would be intriguing to investigate whether MvD-LC develops as glaucoma progresses in these eyes. 

It has previously been suggested that MvD-PC is caused by disruption of the laminar beam, because the microvasculature in both the PC and LC are downstream from the SPCA¹³; however, this is not supported by the present findings. LC disruption could directly affect the PC microvasculature if there were anastomoses between the LC and choroid, which is less likely. It could be hypothesized that reduced blood flow in the mainstream SPCAs led to hypoperfusion of both the PC and the LC, resulting in the simultaneous development of MvD-PC and MvD-LC. However, our findings suggest that a more plausible scenario is focal LC disruption damaging the microvessels within the LC, leading to MvD-LC and subsequent damage to nearby axons. As this axonal damage progresses, the adjacent choroidal microvessels supplying the axons might also undergo decreased perfusion, ultimately resulting in apparent MvD-PC. Long-term observations would be necessary to confirm this hypothesis. 

Some of the POAG eyes with MvD-LC in this study did not exhibit MvD-PC. These eyes were characterized by a thicker juxtapapillary choroid, smaller VF PSD, and higher pretreatment IOP compared with eyes having both MvD-PC and MvD-LC. The sectoral LCCI was significantly larger, but both the sectoral and mean LCT were comparable to those in eyes showing both MvD-PC and MvD-LC. These findings could be explained by a relatively recent elevation in IOP causing deformation of the LC that led to the development of MvD-LC but not yet MvD-PC. Alternatively, it is also possible that MvD-PC was present but undetected in eyes with a small β-PPA. These findings must therefore be validated in further studies because the present study could not determine the clinical significance of only MvD-LC being present due to the small number of included samples. 

A previous study found that the coexistence of optic-disc MvD and MvD-PC was associated with faster RNFL thinning.¹⁰ The LC and PC contain the microvessels perfusing the ONH axons in the laminar and prelaminar regions, respectively, and they are both crucial for axon survival. Thus, impairment of perfusion in both the LC and PC may exert more severe detrimental effects on the retinal ganglion cells. Furthermore, the MvD-LC site may be indicative of mechanical stress, suggesting an additional burden imposed on the ONH axons. 

This study had several limitations. First, eyes without any types of MvD were not included because a small MvD might not be detectable and so could erroneously be considered to be absent; consequently, only eyes with definite MvD-PC or MvD-LC were included. Second, eyes with a well-visualized LC were included to determine the presence of MvD-LC, which may have led to the preferential selection of those with large discs and thin neuroretinal rims. Interobserver agreement was lower for the detection of MvD-LC than for MvD-PC, indicating that determining MvD-LC could be more challenging. This limitation should be addressed by further advancements in OCTA technology to improve the visualization of the deep ONH vasculature. Third, we included only eyes with glaucomatous damage affecting the inferotemporal sector, based on the anatomical differences between the inferior and superior sectors and the challenges of visualizing the deeper structure of the nasal optic disc due to large retinal vessels and thick neuroretinal rims. Fourth, the measurements of sectoral LC morphology may not be robust, as they can be affected by the angle at which they are taken. Fifth, despite our strict definition of the presence of MvD, we might not have been able to detect microvasculature signals that were smaller than the threshold in areas with complete loss of OCTA signals. Finally, the cross-sectional design of this study meant that it could not reveal causal or spatial relationships between MvD-PC and MvD-LC. 

In conclusion, MvD-LC in eyes with POAG was associated with structural deformation of the LC and higher pretreatment IOP. The pathogenic significance of MvD-LC probably differs from that of MvD-PC. Long-term studies are needed to clarify the clinical implications of MvD-LC and MvD-PC, as well as their causal relationships. 

Acknowledgments

Supported by a Seoul National University Bundang Hospital Research Fund (06-2022-0031) and by grants from the Patient-Centered Clinical Research Coordinating Center funded by the Ministry of Health & Welfare, Republic of Korea (HI19C0481, HC19C0276). The funding organizations played no role in the design or conduct of this research. 

Disclosure: E.J. Lee, None; D.K. Han, None; Y.J. Roh, None; T.-W. Kim, None