Molecular Vision 2021; 27:466-479
Received 31 August 2020 | Accepted 15 July 2021 | Published 17 July 2021
1Department of Ophthalmology, Shanghai General Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China; 2Shanghai Key Laboratory of Fundus Disease, Shanghai, China; 3Shanghai Engineering Center for Visual Science and Photomedicine, Shanghai, China; 4Shanghai Engineering Center for Precise Diagnosis and Treatment of Eye Diseases, Shanghai, China; 5National Clinical Research Center for Eye Diseases, Shanghai, China
Correspondence to: Bilian Ke, Department of Ophthalmology, Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, 100 Haining Road, Shanghai 200080, China; Phone: ??; FAX: ??; email: email@example.com
Purpose: To evaluate changes in the retinal microvascular network after posterior chamber phakic implantable collamer lens (ICL) surgery using optical coherence tomography angiography (OCTA) in patients with high myopia.
Methods: Patients with high myopia who underwent ICL surgery were enrolled in this study. All patients underwent comprehensive ophthalmologic exams preoperatively and 1 week, 1 month, and 3 months postoperatively. The vascular densities (VDs) in the retina and the superficial and deep capillary plexuses of different annular and quadrantal areas were evaluated from OCTA images (Zeiss Cirrus 5000). Correlations between the variations in microvascular density and possible factors were further analyzed.
Results: The study comprised 32 eyes of 32 patients. The mean age of the patients was 26.91±7.610 years (15 men and 17 women). A statistically significant reduction in microvascular density in the retina and the superficial plexus was found 1 week and 1 month postoperatively (p<0.05, repeated-measures ANOVA). Further fractal analysis found that the VD of the outer ring declined statistically significantly (p<0.05). A statistically significant decrease was also found in the inferior nasal sector 1 week postoperatively, with an overall decrease in all four quadrants 1 month postoperatively. The microvascular density recovered toward the baseline level 3 months postoperatively. No correlations were observed between the variation in microvascular density and the spherical equivalent (SE), axial length (AL), intraocular pressure (IOP), amplitude of accommodation (AA), or contrast sensitivity.
Conclusions: Retinal microvascular density was decreased postoperatively and then recovered toward the baseline level after 3 months. ICL surgery may have a transient influence on the retinal microvascular network without affecting visual function.
It is estimated that 50% of the world’s population will be affected by myopia by 2050 as prevalence is increasing globally, especially in East Asia . High myopia is becoming the leading cause of vision impairment, resulting in a higher risk for cataracts, glaucoma, and vitreoretinal complications . Changes in the macular and optic disc, such as posterior staphyloma, chorioretinal atrophy, lacquer cracks, and peripapillary atrophy, are frequently found in eyes with high myopia and may lead to irreversible visual loss [3,4].
Implantable posterior chamber phakic intraocular lens (ICL) surgery has been introduced for ametropic correction, especially for patients with high myopia . A 3-year prospective clinical trial conducted by the U.S. Food and Drug Administration (FDA) supported the safety, efficacy, and predictability of ICL surgery for moderate and high myopia . Ninety-three percent of the study eyes were within ±1.0 diopters (D) after a 4-year observation period . Although the safety and predictability of ICL surgery have been confirmed [8,9], it has been reported that complications, such as anterior capsular opacification [5,10,11], endothelial cell loss , and pigmentary glaucoma [13,14], might still occur after ICL implantation. Although ICL surgery is an intraocular procedure that might further affect the stability of the whole eye, little is known about the effect of ICL implantation surgery on the posterior segment.
Optical coherence tomography angiography (OCTA) is a noninvasive technique used for imaging microvasculature in different layers of the retina and the choroid, with advantages such as a short acquisition time and no need for an injectable dye such as fluorescein. OCTA has been widely used for screening and diagnosing various ocular diseases, such as diabetic retinopathy, macular degeneration, vascular occlusions, central serous chorioretinopathy, and glaucoma . OCTA can also produce a highly sensitive and specific characterization of myopia-related changes in the retina . In previous studies, use of fractal analysis of OCTA images revealed that the density of the superficial and deep microvascular plexuses was statistically significantly lower in the high myopia group than in the control group [17,18], indicating the possible application of OCTA for assessing changes in the microvasculature and predicting the possibility of retinal complications induced by ICL surgery. In the present study, the OCTA technique was used to investigate changes in the microvasculature induced by ICL implantation surgery.
A total of 32 patients (15 male, 17 female) who underwent ICL surgery in the Department of Ophthalmology at the Shanghai General Hospital from December 2016 to December 2018 were enrolled in this study. The mean age of the patients was 26.91±7.61 years. Patients who underwent ICL surgery in the Department of Ophthalmology at the Shanghai General Hospital from December 2016 to December 2018 were enrolled in this study. This prospective study was approved by the Institutional Review Board of Shanghai General Hospital, Shanghai Jiaotong University. The study adhered to the tenets of the Declaration of Helsinki and the ARVO statement on human subjects. Informed consent was obtained from all patients before enrollment.
The inclusion criteria of this study were as follows: patients with nonpathologic high myopia (spherical equivalent ≤ −5.0 D) who were willing to undergo ICL surgery for refractive correction and were older than 18 years old. Patients with intraocular pressure (IOP) higher than 21 mmHg, anterior chamber depth (the distance from the endothelium to the anterior surface of the crystalline lens) of less than 2.8 mm, any vitreoretinal disease, presence of media opacity, presence of staphyloma, history of intraocular surgery, and any systemic disease or any other condition that was ineligible for ICL surgery were excluded from this study.
Clinical information, including age, sex, spherical equivalent (SE), corrected distance visual acuity (CDVA), axial length (AL), IOP, and amplitude of accommodation (AA), was collected before surgery and at each follow-up visit. Patients also underwent comprehensive ophthalmologic examinations and funduscopic examinations to exclude possible vitreoretinal diseases. Refraction was measured with an autorefractometer and described as the SE (spherical dioptric power plus half of the cylindrical dioptric power). CDVA (logMAR) was measured manually by an experienced optometrist. The axial length was measured with an intraocular lens master (IOL Master; Master 500, Carl Zeiss, Jena, Germany). IOP was obtained with a noncontact tonometer (Full Auto Tonometer TX-F, Canon, Utsunomiyashi, Japan). AA was measured with the push-up test. The contrast threshold (CT) was measured with a CGT-1000 Contrast Glare Tester (Takagi, Nakano, Japan). The CT was automatically determined with six sizes of annular stimuli ranging from 6.3° to 0.7° of the visual angle, with and without glare conditions. Postoperative follow-up visits were conducted 1 week, 1 month, and 3 months postoperatively. All ophthalmic parameters were independently measured three times by an experienced doctor and an experienced optometrist.
The STAAR sizing formula was applied to determine the size of ICLs used for patients, according to the horizontal corneal diameter (HCD) and the anterior chamber depth (ACD). Standard ICL implantation surgery was performed on all patients by the same doctor (BL Ke). Briefly, on the day of surgery, patients were administered 0.5% tropicamide to fully dilate the pupils, and then topical anesthesia was applied with 0.4% dibucaine hydrochloride half an hour preoperatively. After the placement of hyaluronic acid, the surgeon slowly inserted the ICL into the anterior chamber through a 3 mm corneal incision. After gently removing the remaining viscoelastic substances, the surgeon then irrigated and aspirated the eyes with normal saline solution. Topical corticosteroids and antibiotics were administered postoperatively to prevent infection and inflammation.
Optical coherence tomography angiography (OCTA) was obtained using a Zeiss Cirrus 5000 with an Angioplex OCTA device (Carl Zeiss Meditec, Dublin, CA). (Figure 1) The technique was described previously . Briefly, 6 × 6 mm scans were acquired for the right eye of each patient at a scan rate of 68,000 Hz. Images of the retinal vascular network (RVN), superficial vascular plexus (SVP), and deep vascular plexus (DVP) were exported for further processing and fractal analysis. The SVP indicated the vessel network between the internal limiting membrane (ILM) and the inner plexiform layer (IPL), while the DVP was defined as the vascular network located on the surface of the outer nuclear layer (ONL) between the inner nuclear layer (INL) and the outer plexiform layer (OPL) .
Magnification effects produced by variations in the axial length might affect the ocular biometric results measured with OCT as well as OCT angiograms; thus, we used Bennett’s formula (scaling factor = 0.013062 × [AL – 1.82]) to resize the images according to the patients’ axial length  . The details of the image processing were described previously . The OCTA image acquired was adjusted according to the axial length and then cropped to 1024 × 1024 pixels (Figure 2). The processed image was used for further analysis, which met the image requirements of the software program used for vessel segmentation. The segmentation software was developed and run in the MATLAB environment (MathWorks, Inc., Natick, MA) and processed the images by inverting, equalizing, and removing background noise and nonvessel structures, and then creating a binary image (Figure 3). Vessels with diameters larger than 25 µm were considered large vessels and were separated from the images. The remaining vessels were small, which were used for analysis of microvascular density. The removal of large vessels helped to eliminate projection artifacts from the superficial vascular plexus in the deep vascular plexus. The microvascular network was then skeletonized and partitioned. The center of the foveal avascular zone (FAZ) was detected and delineated by the software. The FAZ (diameter = 0.6 mm) centered on the fovea was removed. The annulus from 0.6 mm to 5.0 mm was defined as the annular zone. The annulus was further subdivided into four quadrants as follows: superior nasal (SN), inferior nasal (IN), superior temporal (ST), and inferior temporal (IT). Moreover, the annulus was subdivided into six thin annuli with a bandwidth of 0.37 mm (Figure 4). ImageJ software (v1.48; National Institutes of Health, Bethesda, MD) was used to outline the FAZ of the superficial microvascular plexus and draw a contour of the area. The fractal analysis toolbox (TruSoft BENOIT Pro 2.0; TruSoft International Inc., St. Peterburg, FL) was used to perform fractal analysis in each partition with the box-counting method. The fractal dimension (Dbox) was used to represent the vascular density in each zone.
The data analysis was performed with SPSS Statistics 22 (SPSS Inc., Chicago, IL). Data characteristics were recorded as the mean ± standard deviation (SD). The Shapiro-Wilk test was used to assess the normal distribution of the data. Repeated-measures ANOVA (Re-ANOVA) tests with post hoc Bonferroni corrections were performed to compare the data from the preoperative and postoperative visits. The Pearson coefficient was calculated to determine the relationships between the magnitude of the change in the vascular density and risk factors. A p value of less than 0.05 was considered statistically significant.
A total of 32 eyes of 32 patients (15 men, 17 women) were included in this study. The patients’ age was 26.91±7.610 years. The preoperative SE was −10.58±3.310 D, ranging from −5.000 D to −16.75 D. The preoperative clinical information is listed in Table 1. All ICL implantations were successful, and no adverse events or complications occurred during the surgery or postoperative follow-up visits.
The refraction of all patients was corrected and statistically significantly improved after surgery, with a mean standard error of the mean (SEM) of −0.26±0.29 D after 1 month (p<0.05), and the refraction remained stable 3 months after surgery. An improving trend of CDVA (logMAR) was observed, but it did not reach statistical significance over the course of the preoperative visit and three follow-up visits (Re-ANOVA, p=0.1033). The pre- and postoperative IOPs were also not statistically significantly different during the observation period, and all IOPs were within a normal range (Re-ANOVA, p=0.3211). Moreover, a statistically significant postoperative increase in the amplitude of accommodation was found (Re-ANOVA, p<0.001; Figure 5). The pre- and postoperative CT data of six different degrees of visual angle are shown in Figure 6. No statistically significant difference was found during the course of the preoperative visit and three follow-up visits in CT with or without glare (p>0.05).
The microvascular density of the annular zone was found to be statistically significantly changed preoperatively and at the three follow-up visits in the retinal vascular network (RVN) and the SVP (Re-ANOVA, p=0.017* for the RVN, p=0.03* for SVP, respectively). However, the macrovascular density did not change statistically significantly (p>0.05 for the RVN, SVP, and DVP) (Table 2, Figure 7). Compared with the preoperative VD, the microvascular density was first decreased after 1 week in the SVP (p=0.046*). A statistically significant decrease in microvascular density was then observed after 1 month in the RVN (p=0.001) and the SVP (p=0.002). The decreased percentage of microvascular density (Dbox) was 0.701% and 0.698% in the RVN and the SVP, respectively. The average density also declined postoperatively in the DVP, but the change did not reach statistical significance in this study (Re-ANOVA, p=0.065). The microvascular density had recovered toward the baseline with no statistically significant difference 3 months postoperatively (p>0.05).
We then further investigated the microvascular density in six thin annuli from the center to the periphery (C1–C6; Figure 8). C1, C4, and C6 were found to be statistically significantly lower postoperatively in the RVN (C1: p=0.037*, C4: p=0.026*, C6: p=0.024*)(Table 3). Moreover, the density of C1, C4, and C6 statistically significantly decreased 1 month postoperatively and then recovered to the preoperative level (p>0.05). Similar changes were observed in the SVP, with C4 and C6 statistically significantly decreasing after 1 month and recovering to preoperative levels 3 months postoperatively (Table 4).
Quadrantal partitioning was also performed to demonstrate different distributions in the four quadrants of the OCTA images (Figure 9, Table 5). A reduction in VD was first observed in the inferior nasal sector 1 week postoperatively. Then, all four quadrants showed a statistically significant decrease in microvascular density 1 month after surgery, which recovered toward the baseline with no statistically significant difference 3 months postoperatively.
No statistically significant correlation was found between the variation in the microvascular density of the RVN or the SVP and the refractive errors or the axial length. No statistically significant correlation between VD and IOP, AA, or CT was found (Table 6). Taken together, the evidence presented here indicates that ICL implantation surgery does not have a long-lasting impact or an adverse effect on the retinal microvasculature.
In this study, we demonstrated that retinal microvascular network densities transiently decreased following ICL surgery. Changes in the retinal microvasculature might be related to visual function, and patients with eyes with high myopia are typically at higher risk for retinal vascular dysfunctions, which might further affect visual acuity (VA) . Recent studies have found that the retinal vasculature might change in patients with diabetes, glaucoma, and high myopia [23-26]. Clinical evidence also suggests that surgery might cause changes in the retinal vasculature  [28,29]. Therefore, it is crucial to evaluate changes in the retina following ICL surgery in a population with high myopia. To the best of our knowledge, this was the first study to perform a quantitative assessment of retinal vascular density with partitioning analysis using OCTA in patients who underwent ICL implantation.
We found that the microvascular density was decreased following ICL surgery, with 0.701% and 0.698% reductions in the RVN and the SVP, respectively, 1 month postoperatively. Comparatively, the macrovascular density was not affected by the surgery, indicating that the changes were limited to the microvascular network, which has not been previously reported. Moreover, the superficial retina showed a reduction in vascular density, while the deep retina showed suggestive changes that did not reach statistical significance. The underlying reason for the changes observed in the retinal microvasculature remains unclear. Several factors might have contributed to these changes. First, previous studies have demonstrated that the microvascular density of highly myopic eyes is significantly lower than that of healthy eyes [17,18]. ICL surgery is considered a minimally invasive surgery, but a fluctuation of intraocular pressure might still occur, which would further destabilize the intraocular microenvironment [30,31,32]. It was reported that the IOP decreased 3 h following ICL surgery and gradually recovered to the preoperative level 24 h postoperatively . In the present study, the postoperative IOP levels were all within the normal range and showed no statistically significant difference postoperatively. The pulsatile ocular blood flow also increased after cataract surgery , which might also induce changes in macular thickness as well as vascular density . Second, Chen et al. found that the diameter of retinal vessels and the oxygen saturation of retinal venules underwent a transient decrease and returned to preoperative levels after ICL implantation , which followed the pattern of the change in microvascular density in the present study. Therefore, it was speculated that the changes in the microvasculature occurred without other changes in vascular diameter as a result of a higher rate of oxygen metabolism in the human retina following ICL surgery. Third, the superficial capillary plexuses were found to be more readily affected by surgery. Lorenzo et al. reported that in eyes with idiopathic vitreomacular traction, the perfusion density in the superficial capillary plexus was reduced 1 month after injection compared to that of the deep capillary plexus and choriocapillaris . It was also reported that the reduction in the microvascular density of the superficial peripapillary retina was associated with axial elongation in patients with myopia, whereas the deep retina showed no association [38,39]. However, the underlying causes of the progressive decline in microvascular density in the SVP until 1 month postoperatively might be multifactorial, which warrants additional investigation. Compared with the deep retina, the superficial retina might be more vulnerable and sensitive to external stimulation, particularly in patients with high myopia. Thus, we suggest that VD changes in the SVP should be emphasized clinically in ICL follow-up visits, especially for patients with high myopia.
An uneven decrease in VD was observed through the further analysis of the annular and quadrantal compartments. The outer ring (C4 and C6) of the RVN and the SVP showed a lower microvascular density than the inner ring in the annular analysis. The retinal vascular system in different parts of the fundus presents various responses to external stimulation . Excessive axial elongation results in mechanical stretching and thinning of the retina. Myopic retinas have thinner perifoveal fields than nonmyopic retinas . Moreover, the average retinal thickness of the outer ring is less than that of the inner ring in highly myopic eyes . Thus, the variation in retinal thickness in highly myopic eyes results in different topographic patterns at the macula ; in particular, the outer ring of the retina is more likely to be affected. Another reason might be autoregulation of the retinal blood supply. Autoregulation of the retinal blood supply maintains a relatively constant blood flow if the perfusion pressure changes . Autoregulation of retinal vessels also varies within different regions of the fundus. Zong et al. found that the parafoveal region had a greater response to the Valsalva maneuver than the peripapillary region . Because ICL surgery might disturb the retinal microcirculation and blood perfusion , we speculated that the retina has a lower capability to regulate blood perfusion in the outer ring, resulting in a significant decrease in VD in the outer retinal zone. The inferior nasal sector was found to decline first, 1 week postoperatively, in the quadrantal analysis in the present study. It was reported that VD changes in the inferior nasal sector were largest in patients with high myopia, which might be related to the loss of the nerve fiber layer . Thus, the inferior nasal sector might be more sensitive and show an earlier response than the other three quadrants.
However, in the present study, the microvascular density variation was not correlated with age, SE, AL, changes in IOP, AA, or contrast sensitivity. The postoperative fluctuation of the retinal microvascular density did not affect the patients’ visual function. All recruited patients in this study presented a trend of improving CDVA and remained stable 3 months postoperatively. The contrast sensitivity did not show a statistically significant change, suggesting that the VD reduction found in this study was transient, and ICL surgery is considered a safe surgical technique for refractive correction.
This study has several limitations. First, the participants enrolled were mostly young, and the response to surgical stimulation might differ in older patients. Thus, we could not evaluate the effect of age on the retinal vascular network. Second, most patients preferred postoperative follow-up visits within 3 months. Moreover, patients recruited showed improved and stable refraction after surgery, and the changes in the retinal microvasculature soon recovered back to the baseline level. Nevertheless, these limitations do not affect the conclusions of this study.
In conclusion, we found that the microvascular density was transiently decreased in macular areas 1 month after ICL surgery and then recovered toward the baseline without adverse effects on visual function. Moreover, the pre- and postoperative application of OCTA can help ophthalmologists to better understand microvascular responses to ICL surgery and prevent fundus complications following ICL surgery in patients with high myopia.
The authors acknowledge Dr. Jianhua Wang for kindly supplying us with customized software for the assessment of vascular density from OCTA images. Funding details: This work was supported by Grant 2020YFC2003904 from National Key Research & Development Program,Grant 81770953 from National Natural Science Foundation, Grant 2018ZHYL0222 from intelligent medical project of Shanghai, Grant 17411950204 from the Science and Technology Commission of Shanghai Municipality, Grant CTCCR-2018B01 from Clinical Research Innovation Plan of Shanghai General Hospital, Grant 82070992 from National Natural Science Foundation,Grant YG2021ZD18 from Shanghai Jiaotong University Medical Engineering Cross Research.