High-Resolution Imaging of Photoreceptors in Macular Microholes

Abstract

Purpose.: To assess photoreceptor structure in macular microholes by using adaptive optics scanning laser ophthalmoscopy (AO-SLO) and spectral-domain optical coherence tomography (SD-OCT) and compare with visual acuity.

Methods.: Fourteen eyes from 12 patients with macular microholes underwent a full ophthalmologic examination and imaging with a fundus camera, SD-OCT, and an original prototype AO-SLO system at each visit.

Results.: All eyes had a cone outer segment tip line disruption and a normal retinal pigment epithelium line on SD-OCT images. Adaptive optics scanning laser ophthalmoscopy revealed foveal cone disruption (13 eyes, round or oval; 1 eye, T-shaped) in all eyes. Cone disruption area (mean = 14,805 ± 9120 μm²; range, 3495–35,901 μm²) positively correlated with logMAR visual acuity at the first visit (P = 0.015, r_s = 0.679). During the follow-up period, cone disruption area increased in two eyes, was stable in seven eyes, and decreased in five eyes. At the last visit, cone disruption area (mean = 8717 ± 7432 μm²; range, 0–25,746 μm²) also positively correlated with logMAR visual acuity (P = 0.035, r_s = 0.610). In one patient with bilateral microholes and no apparent vitreous traction, lesion size gradually increased. Cone disruption area decreased and visual acuity improved following oral prednisone therapy.

Conclusions.: Cone disruption occurs in eyes with macular microholes and a larger cone disruption area translates into a poorer visual acuity. Macular microholes, which are commonly observed as foveal cone inner and outer segment disruptions, may occur in eyes with or without vitreofoveal traction.

Introduction

A macular microhole is a little known condition that causes mildly reduced visual acuity, central scotoma, and metamorphopsia. Cairns and McCombe ¹ first described this condition in 1988 as “microholes of the fovea centralis” in a series of 17 patients with red, well-defined, foveal defects near the center of the capillary free zone. In 1996, Reddy et al. ² described 14 patients with small, reddish macular holes, and termed these holes, “microholes of the macula.” Using time-domain optical coherence tomography (OCT), Douglas et al. ³ reported on 13 patients with a single foveal or juxtafoveal, red, irregular lesion with sharply defined borders. They named these defects, “foveal spots.” In 2005, Zambarakji et al. ⁴ reported OCT and retinal thickness analyzer findings in a series of 22 patients with a well-demarcated, red, intraretinal, foveal, or juxtafoveal defects. They also named the defects, “macular microholes.” Optical coherence tomography images of eyes with macular microholes have revealed macular structural abnormalities in the outer retina, including disruption of the photoreceptor inner and outer segment junction (IS/OS) line and/or the retinal RPE. ^3–5 However, the current understanding of macular microhole pathology is limited and further investigation is needed. 

Optical coherence tomography and other imaging modalities (e.g., scanning laser ophthalmoscopy [SLO]) fail to provide sufficient detail of the photoreceptor microstructure, primarily because of ocular optical aberrations. Aberrations can be compensated for by using imaging systems with adaptive optics (AO). These include wavefront sensors that measure optical aberrations and deformable mirrors or spatial light modulators, which compensate for aberrations. ^6–10 Adding AO to imaging systems, including flood-illuminated ophthalmoscopes, SLO equipment, and OCT, has allowed researchers to identify abnormalities in individual cone photoreceptors in patients with various retinal disorders. ^11–34  

Here, we used AO-SLO and spectral-domain OCT (SD-OCT) to assess pathologic changes in photoreceptors in eyes with macular microholes. Cone abnormalities on AO-SLO and retinal and vitreous findings on SD-OCT images are discussed, along with the relationship between structural abnormalities and visual function. 

Methods

Study conduct adhered to the tenets of the Declaration of Helsinki. This study and the use of the prototype AO-SLO system were approved by the Institutional Review Board and the Ethics Committee of Kyoto University Graduate School of Medicine. The nature of this study, participation in it, and possible risks and benefits were explained to study candidates, after which written informed consent was provided by all participants. 

Participants

There were a total of 14 eyes from 12 participants (five men, seven women) in this observational case series. Average patient age was 59.4 ± 11.5 years (range, 36–73 years) and all patients had macular microholes with no other macular abnormalities, glaucoma, or inherited color blindness. All patients visited the Kyoto University Hospital in Kyoto, Japan, between February 2008 and May 2012. All patients were followed for at least 12 months. 

All patients were diagnosed with macular microholes, based on the presence of a well-demarcated, red, intraretinal, foveal, or juxtafoveal defect. Patients with a full-thickness macular hole (MH), solar retinopathy (including history of visual disturbance after sun gazing), laser pointer burn, any toxic maculopathy, history of trauma, amyl nitrate abuse, welding arc maculopathy, lightning maculopathy, cone dystrophy, and occult macular dystrophy were excluded from participation. Eyes with high myopia (axial length > 26.5 mm) were also excluded. 

Ophthalmologic Examinations

All participants underwent comprehensive ophthalmologic examinations at baseline, which included measurement of best-corrected visual acuity (BCVA), intraocular pressure (IOP), and axial length (IOLMaster; Carl Zeiss Meditec, Dublin, CA, USA). Measurement of BCVA was performed using the Landolt chart and expressed as logMAR. In addition, indirect ophthalmoscopy, slit-lamp biomicroscopy with a noncontact lens, color fundus photography, SD-OCT, and AO-SLO were performed. If necessary, fluorescein angiography, fundus autofluorescence, and multifocal electroretinography were performed to obtain a differential diagnosis. At each follow-up visit, all participants underwent BCVA measurement, color fundus photography, SD-OCT, and AO-SLO. 

Spectral-Domain Optical Coherence Tomography

Examinations with SD-OCT were performed in all eyes using a commercial device (Spectralis HRA+OCT; Heidelberg Engineering, Dossenheim, Germany). Horizontal and vertical line scans through the center of the fovea were obtained at a 30° angle, after which we performed 12 radial scans centered on the fovea. We then obtained volume scans (horizontal B-scans at 10° × 30°). At each location of interest on the retina, 50 SD-OCT images were acquired and averaged to reduce speckle noise. 

The reflectivity of the inner segment/outer segment (IS/OS) or cone outer segment tip (COST) was measured from within a 40-μm slab, and the IS/OS or COST line was quantified by using the plot profile function of ImageJ software (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA) with a 6-pixel fixed-width line. The border of the IS/OS or COST disruption was defined as the line on the grayscale image along which the IS/OS or COST reflectivity diminished by 2 SD from the reflectivity of the IS/OS or COST line in the unaffected retina. ³⁵  

Outer nuclear layer (ONL) thickness was measured at the center of the fovea. Foveal ONL thickness was defined as the distance between the vitreoretinal interface and the external limiting membrane. We have also measured the ONL + outer plexiform layer (OPL) thickness 0.5 and 1.0 mm from the center of the fovea. 

Prototype AO-SLO

Our AO-SLO system has been previously described in full. ^28,33 Briefly, the AO-SLO system used in this study was designed and constructed in our laboratory, based on previous reports of how incorporating a wide-field SLO is useful. ^36,37 The AO-SLO system is confocal, allowing it to create high-contrast “en face” images in any plane. The system is composed of four primary optical subsystems: the AO subsystem (wavefront sensor, spatial light modulator), the high-resolution confocal SLO imaging subsystem, the wide-field imaging subsystem, and the pupil observation subsystem. The pupil subsystem facilitates initial pupil alignment with the AO-SLO system's optical axis through chin rest position adjustment. The wavefront sensor measures aberrations in the eye as a whole, and the spatial light modulator compensates for these aberrations. The total acquisition time was 10 minutes for one eye. 

Cone Mosaic Imaging With the AO-SLO

A series of AO-SLO images was acquired at each of several macular locations by shifting the focal point from the retinal nerve fiber layer to the RPE, with particular attention paid to acquiring images showing the cone mosaic. We automatically created a montage of AO-SLO images by using MosaicJ (National Institutes of Health, Bethesda, MD, USA). If necessary, we made manual corrections by selecting both the area of interest and each image to be included in the montage. Proper reconstructions of AO-SLO images were verified by comparing the final montage to the corresponding wide-field images for that eye. The postprocessing time was 30 minutes for each image. 

The size of the dark lesion in each eye was quantified in AO-SLO images by two independent, experienced examiners using ImageJ software, a Java-based image-processing software program (Fig. 1). In ImageJ, the command path of Image > Adjust > Threshold was used to differentiate the dark area from the nondark area. Brightness was adjusted manually by the grader. To measure the dark area, the command path of Analyze > Measure was used. To obtain accurate scan lengths, the magnification effect in each eye was corrected using the adjusted axial length method, previously devised by Bennett et al. ³⁸  

Figure 1

 

Measurement of foveal dark area seen on AO-SLO. (A) An adaptive optics–SLO image of the area that includes the center of the fovea of a normal eye. (B) An adaptive optics-SLO image of the area that includes the center of the fovea of an eye with macular microhole (case 10, left eye). Note that the macular microhole is seen as a hyporeflective dark area. (C) The size of the dark lesion was quantified using ImageJ (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA). In ImageJ, the command path of Image > Adjust > Threshold was used to differentiate the dark area (indicated by red) from the nondark area. Brightness was adjusted manually by each grader. To measure the dark area, the command path of Analyze > Measure was used.

To evaluate cone density, we applied the automated cone labeling process of Li and Roorda. ³⁹ After automated cone labeling, an experienced observer examined each image. The area for quantification (80 × 80 μm) was identified manually by selecting the area without vascular shadows and by placing the computer cursor on the area to be quantified. This method has been used in other studies. ^{19,24,25,26,28,33} As has been reported for similar systems, ^6–28 we found that our system did not always allow clear visualization of individual cones within much of the central fovea. However, we could clearly distinguish individual cones ≥ 0.5 mm from the center of the fovea. Therefore, we obtained an estimate of cone density in areas 0.5 mm from the foveal center by dividing the number of cones in each imaging area by the size of the area. We measured cone density in each of four directions (superior, lower, nasal, and temporal), and the mean density was calculated from the densities in all four directions. 

Statistical Analyses

Statistical analysis was done using only one eye per patient. We used the data of the right eye in bilaterally affected cases. For intraobserver measurements, one-way random, average measure intraclass correlation coefficients (ICC) were obtained. Wilcoxon signed tests were used to compare parameters between the first and last clinic visits. The Spearman rank correlation coefficient was calculated to examine the association between cone disruption area and logMAR BCVA. All statistical calculations were performed using a commercially available statistical software program (SPSS, version 17; SPSS, Inc., Chicago, IL, USA). Statistical significance was defined as P < 0.05. 

Results

Participant clinical characteristics are summarized in Table 1. Fourteen eyes from 12 patients (five men, seven women) were included in this observational case series. Mean participant age was 59.4 ± 11.5 years (range, 36–73 years) and mean logMAR BCVA was 0.082 ± 0.087 (range, 20/12 to 20/25) at the first visit. At the last visit, on average 26.9 ± 11.4 months (range, 12–44 months) later, mean logMAR BCVA was −0.041 ± 0.081 (range, 20/12 to 20/30), which was no different from the first visit (P = 0.097). 

Table 1

 

Clinical Data of Patients With Macular Microholes

Case	Age/Sex	Eye	BCVA First Visit	BCVA Last Visit	Vitreous Condition	OPL/ONL, μm*	OPL/ONL, μm†
1	73/M	L	20/20	20/20	VFS	98	91
2	54/F	L	20/16	20/20	VFS	102	101
3	58/M	L	20/20	20/16	VFS	88	108
4	69/F	R	20/20	20/16	VFS	103	119
5	60/F	R	20/25	20/16	VFS	83	90
6	66/F	R	20/20	20/20	VFS	114	137
7	68/F	L	20/16	20/12.5	Complete PVD	98	101
8	61/F	R	20/25	20/16	Complete PVD	98	98
9	73/M	R	20/12	20/16	Complete PVD	103	116
10	46/F	R	20/20	20/16	No VFS or PVD	114	127
		L	20/12	20/16	No VFS or PVD	106	127
11	36/M	R	20/25	20/20	No VFS or PVD	119	124
		L	20/20	20/16	No VFS or PVD	119	124
12	49/M	L	20/25	20/32	No VFS or PVD	106	114

 OPL/ONL, thickness of outer plexiform layer and outer nuclear layer.

*  1.0 mm temporal from the center of the fovea.

†  0.5 mm temporal from the center of the fovea.

‡  0.5 mm nasal from the center of the fovea.

§  1.0 mm nasal from the center of the fovea.

Table 1

 

Extended

ONL Fovea, μm	OPL/ONL, μm‡	OPL/ONL, μm§	COST Disruption	Hyperreflective Lesion	Cone Loss Area First Visit, μm2	Cone Loss Area Last Visit, μm2	Follow-up, mo
84	98	93	+	+	11,523	11,901	12
93	93	111	+	–	3,495	4,090	44
57	119	108	+	+	8,549	8,570	33
104	116	103	+	+	10,427	0	32
72	96	93	+	–	35,901	5,768	31
57	114	103	+	–	16,742	12,089	44
93	114	104	+	–	13,039	11,123	34
83	90	95	+	–	25,218	0	40
100	108	111	+	–	7,524	0	16
116	119	106	+	–	11,924	11,879	23
120	119	111	+	–	11,620	11,760	23
95	119	124	+	–	9,980	13,435	15
102	124	119	+	–	8,147	10,534	15
80	116	111	+	–	23,346	25,746	15

A small, red, well-demarcated, intraretinal, foveal defect was seen in all eyes (Figs. 2 1552 1552 1552 1552 1552 1552–9, Supplementary Figs. S1, S2). Three eyes (21%) had a complete posterior vitreous detachment (PVD) and 6 eyes (43%) had a vitreofoveal separation (VFS, Fig. 2). The two participants with bilateral macular microholes and one patient with a unilateral microhole did not have a PVD, a VFS, or vitreous traction (Figs. 4, 7). 

Figure 2

 

Unilateral macular microhole with vitreofoveal separation. Images from the left eye of a 58-year-old man with a macular microhole (case 3). Best-corrected visual acuity was 20/20. (A) Fundus photograph showing a small, faint, irregular, red lesion at the foveal center (arrow). (B) High-magnification view of the foveal center. (C) Infrared image. (D) Spectral-domain optical coherence tomography image. Horizontal line scan through the foveal center, taken in the direction of the arrow in (C). Small outer retinal defects, selective thinning of the outer nuclear layer in the fovea, and a vitreofoveal separation (green arrowheads) are visible.

Figure 3

 

Magnified SD-OCT and AO-SLO images of the fovea (case 3). (A) The SD-OCT revealed photoreceptor IS/OS junction line and COST line defects at the fovea (blue arrowhead). The retinal pigment epithelial line was intact. The adaptive optics–SLO images show dark regions, representing cone disruption. (B) Three years later, dark area was almost stable. Scale bar: 100 μm.

Figure 4

 

Nonprogressive bilateral macular microholes without apparent vitreous traction. Images from the right eye of a 46-year-old woman with bilateral macular microholes (case 10); BCVA was 20/20. (A) Fundus photographs showed small, localized, round, red lesions at the foveal center. (B) High-magnification view of the foveal center. (C) Multifocal electroretinography showed a near-normal response. (D, E) SD-OCT images. Horizontal (D) and vertical (E) line scans through the foveal center showed small outer retinal defects. The posterior vitreous membrane (green arrowheads) is attached to the macula and neither eye had evident vitreoretinal traction.

Figure 5

 

Magnified SD-OCT and AO-SLO images of the fovea (case 10, right eye). (A) The SD-OCT image showed COST line defects and an irregular IS/OS line at the foveal center (blue arrowhead). The retinal pigment epithelial line was intact. The adaptive optics–SLO images show dark, round regions, representing central foveal cone disruption. (B) Two years later, the size and shape dark regions was almost stable. Scale bar: 100 μm.

Figure 6

 

Montage AO-SLO images of the fovea and parafovea (case 10, right eye). Right: High-magnification views of the areas outlined by the white boxes. Note that cone mosaic is normal except for the center of the fovea. Scale bar: 100 μm.

Figure 7

 

Progressive bilateral macular microholes without apparent vitreous traction. Images of the right (A–D) and left (E–H) eyes from a 36-year-old man with bilateral macular microholes (case 11). Best corrected visual acuity was 20/25 in the right eye and 20/20 in the left. (A, E) Fundus photographs showing small, localized, round, red lesions at the foveal center. (B, F) High-magnification view of the central fovea. (C, G) Fluorescein angiograms were normal in both eyes. (D, H) Spectral-domain optical coherence tomography images. A horizontal line scan through the foveal center showed small COST line defects and IS/OS line irregularities (blue arrowheads) in both eyes. The retinal pigment epithelial line was intact. The posterior vitreous membrane (green arrowheads) was attached to the macula, but neither eye had evident vitreoretinal traction.

Figure 8

 

Progressive bilateral macular microholes without apparent vitreous traction. Magnified SD-OCT and AO-SLO images from case 11. The AO-SLO images showed dark, round regions in the foveal center, representing cone disruption. The lesion size and shape was similar between the right and left eyes. The dark area gradually increased gradually, corresponding to an increase in scotoma size. RE, right eye; LE, left eye. Scale bar: 100 μm.

Figure 9

 

Resolving bilateral macular microholes after oral prednisone. Magnified SD-OCT and AO-SLO images from case 11. Treatment with oral prednisone (40 mg/d) 16 weeks after the initial visit prevented further vision loss; (BCVA: 20/32 in the right eye, 20/25 in the left eye). By the last visit, the cone disruption region size had decreased and visual acuity had subsequently improved (BCVA: 20/20 in the right eye, 20/16 in the left eye). Note that the dynamic lesion changes are more apparent on AO-SLO images than in SD-OCT images. Scale bar: 100 μm.

All 14 eyes had a disruption of the IS/OS and COST line on SD-OCT images and a normal RPE line (Figs. 2 1552 1552–5, 7–9, Supplementary Figs. S1, S2). Additionally, three eyes with VFS also had a moderately reflective lesion in the outer nuclear layer (ONL, Fig. 3). The mean ONL thickness of the fovea was 86.2 ± 17.9 μm (Table 2). 

Table 2

 

Photoreceptor Abnormalities Detected by SD-OCT and AO-SLO in Macular Microholes

	First Visit	Last Visit	P*
SD-OCT
ONL thickness, μm	86.2 ± 17.9	89.7 ± 36.2	0.410
IS/OS disruption  size, μm	105.8 ± 61.8	62.8 ± 63.5	0.084
COST disruption  size, μm	136.3 ± 78.9	74.3 ± 74.3	0.012
AO-SLO
Dark area, μm2	14,805 ± 9,120	8,717 ± 7,432	0.158
Cone density,  cones/mm2†	31,812 ± 2,715	31,664 ± 2,515	0.583

n  , 12 eyes of 12 patients.

*  Wilcoxon rank test.

†  0.5 mm from the central fovea.

Foveal dark lesion was observed in all 14 eyes on AO-SLO. In 13 eyes, the lesion was round or oval (Figs. 1, 5, 6, 8, 9, Supplementary Fig. S2) and in 1 eye it was T-shaped (Fig. 3). Cone abnormalities were limited to the fovea in all eyes (Fig. 6). The reproducibility of the dark area measurements was evaluated through an interobserver ICC. The ICC was 0.944 for measurement at the first visit and 0.952 at the last visit. The 95% confidential intervals for ICC values were 0.834 to 0.982 at the first visit, and 0.858 to 0.984 at the last visit. 

The IS/OS and COST disruption size on SD-OCT correlated with the dark area on AO-SLO (P = 0.006, r_s = 0.741 and P = 0.003, r_s = 0.776, respectively) at the first visit. The IS/OS and COST disruption size correlated with the dark area on AO-SLO (P = 0.005, r_s = 0.749 and P < 0.001, r_s = 0.858, respectively) at the last visit as well. 

Cone disruption area (mean = 14,805 ± 9120 μm², range, 3495–35,901 μm²) positively correlated with logMAR visual acuity at the first visit (P = 0.015, r_s = 0.679). The dark area increased by >20% in two eyes (14%), was stable in seven eyes (50%), and decreased by >20% in five eyes (36%). Three eyes showed complete resolution of cone disruption (Supplementary Figs. S1, S2). Of the 5 eyes where cone disruption decreased during the follow-up period, three had a VFS and two had a complete PVD. Interestingly, the dark area was either stable or decreased in all eyes with a VFS or a complete PVD. At the last visit, cone disruption area (mean = 8717 ± 7432 μm²; range, 0–25,746 μm²) also positively correlated with logMAR visual acuity (P = 0.035, r_s = 0.610). The mean cone density 0.5 mm from the center of the fovea (i.e., the area around the lesion) was 31,812 ± 2715 at the first visit and 31,664 ± 2515 at the last visit (P = 0.583; Table 2), which was within normal range. ^19,25,26,28  

Five eyes (36%) had no apparent signs of either past or present vitreous traction. In one 36-year-old participant with bilateral microholes and no evident vitreous traction, retinal lesion size gradually increased. Unfortunately, the patient also experienced a gradual increase in scotoma size. Treatment with oral prednisone (40 mg/d) was initiated 16 weeks after the initial visit to prevent further cone disruption, and subsequent vision loss. With systemic steroid treatment, dark area decreased and visual acuity improved (Figs. 8, 9). 

Discussion

Macular microholes are diagnosed based on the presence of a red, well-demarcated, intraretinal, foveal, or juxtafoveal defect on fundoscopic examinations or fundus photography. ^1–5 However, the lesion is usually very small, and the disease may be easily overlooked during standard fundoscopic examinations. Although SD-OCT is more sensitive than AO-SLO generally, macular microholes may even be overlooked on SD-OCT images without dense B-scans. The intervals of each B-scan are more than 20 μm, even using a raster scan, with lateral resolution of ∼20 μm, whereas the AO-SLO, with lateral resolution of 3 μm, can detect ∼20-μm–wide abnormalities. In fact, the dynamic lesion changes were more apparent on AO-SLO images than on SD-OCT images (Figs. 8, 9). Thus, the combination of fundus examination, SD-OCT, and AO-SLO may be useful in accurately diagnosing macular microholes and for monitoring patients with the disease. Recently Flatter et al. ⁴⁰ reported the SD-OCT and AO-SLO findings of photoreceptor damage after blunt trauma, which are similar to those of macular microholes. In both conditions, the lesions were very small, and AO-SLO is useful for evaluating and monitoring patients. 

We propose that the “dark area” on AO-SLO represents abnormalities at the photoreceptor level. This proposal is based on several findings. First, foveal ONL thickness was smaller in eyes with macular microholes (86 ± 18 μm), compared with normal eyes (122 ± 23 μm). ²⁸ Second, a comparison of AO-SLO images with wide-field SLO images or fundus photographs allowed us to rule out the possibility that the dark areas represented the shadows of blood vessels. Moderately reflective foveal lesions (Fig. 3) seem to have little effect on the penetration of light reflected from the deeper layers. In fact, on SD-OCT, which uses a light source with a wavelength (840 nm) identical to that of our AO-SLO system, no shadows were observed in the photoreceptor layer or RPE. Third, the dark areas on AO-SLO positively correlated with the areas of disruption in the IS/OS and COST line on SD-OCT images. 

Recently, Spaide and Curcio ⁴¹ hypothesized that the IS/OS lines and COST lines correspond to the ellipsoid zone of the photoreceptors and the contact cylinder of the cones. The ellipsoid section is a part of the photoreceptor inner segments and is densely packed with mitochondria. The OS continues to the RPE, whereupon it is enveloped in specialized apical processes, forming a contact cylinder. Thus, it is possible that the appearance of the presumed IS/OS line reflects the function of the photoreceptor inner segments, and the appearance of the presumed COST line reflects the function of the photoreceptor outer segments. The high reflectance of the cone mosaic in AO-SLO is thought to be caused by reflectance from both the IS/OS and the COST in the normal retina. ⁴² In fact, the current study showed that the dark area was correlated with larger decreases in the reflectivity sizes of the IS/OS and COST. However, dark areas were also observed on AO-SLO images in the areas where the IS/OS line was continuous (though it was irregular), but the COST line was disrupted, on SD-OCT (Figs. 7 1552–9). This finding is consistent with the results of a study by Kitaguchi et al., ¹⁶ in which the dark area observed on AO fundus camera images corresponded with the areas where the COST line, rather than the IS/OS, was disrupted on SD-OCT images. Thus, the OS probably plays a more important role in the reflectance of the photoreceptor mosaic on AO-imaging devices. This is also supported by the recent split detector AO-SLO imaging performed by Scoles et al. ⁴³  

Previous studies examining OCT images of macular microholes have revealed retinal structural abnormalities. ^3–5 Using time-domain OCT, Zambarakji et al. ⁴ found an outer retinal abnormality and/or RPE defect on many OCT series. Using SD-OCT, Gella et al. ⁵ also reported the presence of photoreceptor layer abnormalities in all OCT series, with some eyes also having RPE abnormalities. However, both studies used single, and not averaged, images so speckle noise likely limited detailed layer analyses. In the current study, the speckle-noise-reduction capabilities of the SD-OCT device (eye tracking combined with multiple B-scan averaging, Spectralis; Heidelberg Engineering) allowed us to obtain highly detailed images of all retinal layers. Disruptions of the IS/OS and COST line were found in all eyes examined, and the RPE line was normal. In addition, AO-SLO revealed dark area in the fovea in all eyes. Thus, the current study shows that foveal cones, especially the inner and outer segments, are mainly affected in eyes with macular microholes. 

Several researchers have proposed that macular microhole pathogenesis may be associated with anteroposterior vitreous traction at the macula. ^2,4,5 In the current study, nine eyes (64%) had evidence of acute or prior vitreous traction (i.e., VFS, complete PVD) on the fovea. This led us to hypothesize that, when a certain kind of macular microhole forms, cone photoreceptors are pulled away from the RPE, as a result of anteroposterior traction on the photoreceptor layer, which is mostly made up of cones in the fovea. We believe that this traction is caused by perifoveal posterior vitreous traction or acute detachment of vitreous. 

Lai et al. ⁴⁴ examined spontaneous resolution of macular microholes and found that a type of macular microhole can form while a full-thickness macular hole is resolving. Histopathologic studies of repaired full-thickness macular hole have shown that photoreceptor cells are replaced by Müller cells and/or astrocytes at the site of the macular hole. ^45–47 These cellular changes may explain the moderately reflective ONL lesions seen on OCT that were first reported by Ko et al. ⁴⁸ on ultrahigh-resolution OCT images. In the current study, three eyes with macular microholes, all of which had VFS, had these moderately reflective ONL lesions in the macula. Together, this evidence suggests that these eyes may have had a resolving full-thickness macular hole. 

During the follow-up period (26.9 months on average), the dark area was stable or decreased in all eyes with VFS or a complete PVD, both of which are evidence of past anteroposterior vitreal traction on the macula. This finding implies that macular microholes, originally caused by acute anteroposterior vitreous traction on the macula, may spontaneously decrease in size once the traction is no longer present. Additionally, if the photoreceptor cell body and the inner segment remain intact, the cell will regenerate its outer segment and visual acuity improves. This phenomenon may be consistent with the gradual recovery of the IS/OS line in closed full-thickness macular hole after vitreous surgery. ^49–51  

Five eyes (36%) had no apparent signs of either past or present vitreous traction. Thus, macular microholes may be divided into at least two subtypes, based on the involvement of anteroposterior vitreous traction. The pathology of macular microholes without the involvement of anteroposterior vitreous traction may be similar to that of acute zonal occult outer retinopathy (AZOOR) complex disease, in which disruptions of the IS/OS and the COST lines are the characteristic findings on OCT as well. ^52–55 Indeed, in one middle-aged patient with bilateral microholes, lesion size gradually increased with no apparent vitreous traction. When the size of cone disruption regions decreased following oral steroid therapy, visual acuity also improved, which is similar to the report of Spaide et al. ⁵³ that oral steroid and immunosuppressants reconstituted IS/OS line defects in eyes with AZOOR complex disease. Although macular microholes caused by acute anteroposterior vitreous traction are photoreceptor defects secondary to vitreofoveal traction, this microhole subtype may be caused by the primary damage of the photoreceptor outer segments. 

In the current study, a larger dark area in the fovea on AO-SLO coincided with a worse visual acuity at both the first and list clinic visit. This pattern is consistent with a previous report using SD-OCT and microperimetry techniques, which found correlations between the macular microhole size and the retinal sensitivity reduction. ⁵ Cumulatively, these findings suggest that macular functional impairment is closely associated with foveal cone photoreceptor changes in eyes with macular microholes. 

Our study has several limitations. First, this study examined a relatively few eyes due to practical limitations associated with the rarity of macular microholes. Second, although it has better lateral resolution than commercially available SD-OCT, our AO imaging equipment was unable to clearly show individual cone photoreceptors near the foveal center. However, each dark area, representing cone disruption, was larger than the diameter of a single central foveal cone. In fact, cone disruption area was detectable near the foveal center, even though individual cones could not be imaged in the same location. Despite these limitations, our study shows that the combination of fundus examination, SD-OCT, and AO-SLO may be useful in diagnosing and monitoring macular microholes. Using these images, we were able to associate photoreceptor and visual acuity changes in eyes with macular microholes. Macular microholes, which are commonly observed as inner and outer segment disruptions in the fovea, may occur in cases with and without vitreofoveal traction. 

Supplementary Materials

pdf S1 

Acknowledgments

Supported in part by the New Energy and Industrial Technology Development Organization (NEDO; P05002), Kawasaki, Japan. 

Disclosure: S. Ooto, None; M. Hangai, NIDEK, Co., Ltd. (C); K. Takayama, None; N. Ueda-Arakawa, None; Y. Makiyama, None; M. Hanebuchi, NIDEK Co., Ltd. (E); N. Yoshimura, NIDEK, Co., Ltd. (C)