Retinal and Choriocapillaris Thickness Changes in Spontaneously Diabetic Macaques

Abstract

Purpose: Beyond the clinically apparent damage to the retinal vasculature, diabetes affects the neuroretina and choroid. Nonhuman primates (NHPs) serve as valuable models for human retinal diseases, including diabetic retinopathy. This study aimed to investigate changes in the thickness of the retina, RPE, and choriocapillaris in spontaneously diabetic cynomolgus macaques.

Methods: Optical coherence tomography (OCT) images were obtained from 25 diabetic macaques and 26 age-matched nondiabetic macaques. The thickness of individual retinal layers in the macula, along with RPE and choriocapillaris, was measured manually. Choriocapillaris thickness was assessed in postmortem human globes from diabetic and nondiabetic individuals.

Results: Diabetes predominantly affected the inner retina more than the outer retina in macaques. Notably, the nasal retina showed greater thinning compared to the temporal retina. A slight but significant thickening of the inner nuclear layer was observed. No changes were detected in the thickness of RPE/Bruch's membrane complex via OCT. However, a significant increase in choriocapillaris thickness was noted in the diabetic macaques. Postmortem human specimens from patients with nonproliferative diabetic retinopathy (NPDR) also demonstrated thickening of choriocapillaris/Bruch's membrane.

Conclusions: Consistent with humans, early diabetes in cynomolgus macaques results in notable alterations in retinal thickness, particularly affecting the nasal inner retina. The observed increase in choriocapillaris thickness in both diabetic macaques and human patients with NPDR likely indicates pathologic changes and remodeling due to diabetes. Cynomolgus macaque presents a valuable NHP model for studying diabetic retinopathy.

Diabetic retinopathy (DR) has traditionally been viewed as a microvascular disorder due to the clinically visible manifestations, including hemorrhages and lipid exudation, which result from retinal vascular damage. However, there is now much greater appreciation and understanding of pathogenic involvement of other retinal components beyond the retinal vasculature. This has led to the emergence of the concept of DR as a disease of the neurovascular unit, encompassing neuronal elements, glial elements, and microglia in addition to vascular cells.^1,2 Furthermore, although DR primarily affects the inner retina, it also alters the outer retina, specifically photoreceptors, as well as the retinal pigment epithelium and choroid.³ Emerging evidence strongly indicates pathogenic involvement of cellular elements beyond the retinal vascular cells. 

Recent studies highlight retinal neuronal alterations early in diabetes. Some researchers propose that DR could be a primary neurodegenerative disorder⁴ and suggest that neuronal changes occur in the retina before clinically evident DR and possibly even before vascular changes, at least with current research methodologies.⁵ Diabetes causes neuronal apoptosis, especially in the inner retina, in both humans and rats,⁶ leading to an irreversible loss of neurons that begins relatively early in diabetes.⁷ Imaging technologies, especially optical coherence tomography (OCT), have revealed thinning of the retinal nerve fiber layer (RNFL) and ganglion cell layer (GCL) in diabetes, even before clinically evident retinopathy.⁸ Beyond these more evident changes in the inner retina in DR, the outer retina, RPE, and choroid are also affected.^2,3,9 Studies in mice demonstrate diabetes-induced oxidative stress in photoreceptors, implicating these cells as the major source of reactive oxygen species in diabetic retinas.¹⁰ In addition, dysfunction of the RPE can compromise the outer blood–retinal barrier, with leakage from the choriocapillaris that contributes to retinal edema.^9,11 Histopathologic studies of postmortem human specimens have demonstrated a “diabetic choroidopathy” involving neutrophil-mediated damage and vascular insufficiency.^9,12 Together, these studies indicate the importance of studying changes in multiple cell types to improve understanding of DR. Interdependence between vascular, glial, and neuronal elements highlights the potential value of neuroprotective treatments for DR.¹³ 

Animal models have allowed tremendous insights into DR and the effects of diabetes on the retina. These studies, combined with human studies, have helped reformulate our understanding of diabetic retinopathy, with a more comprehensive view of this condition, including its neurodegenerative aspects.^1,4,7 However, most animal studies have focused on rodent models. Important differences between rodents and humans, including lack of a macula in rodents, may limit the ability to extrapolate changes in these animals to humans. Given the strong insights provided by animal models, an important gap in knowledge is the relative paucity of research on DR with respect to nonhuman primates (NHPs), particularly concerning changes in neuronal elements, RPE, and choroid. 

NHPs, including macaques, are excellent animal models of human retinal diseases.¹⁴ NHPs are particularly unique as an animal model given their cone-rich macula,^15,16 a critical site of damage in diabetic retinopathy. Importantly, a systematic study of rhesus macaques found great anatomic and functional similarities in the eye and retina between NHP and humans, including macular structure.¹⁷ In DR research, multiple NHP models have been investigated. As with humans, long durations of diabetes are required before vascular pathology, including capillary dropout and blood–retinal barrier dysfunction.¹⁸ 

With the increasing recognition that earlier changes occur in the retina in diabetes in humans and rodents, it is of interest to identify changes in NHP in early diabetes, which could expand our insights for DR. We previously examined a population of cynomolgus macaques (Macaca fascicularis).¹⁹ From a population of more than 800 macaques, 28 had spontaneous diabetes, with no clinically evident signs of DR. Histopathologic studies of a small subset did reveal acellular capillaries and pericyte ghosts, and half the subjects exhibited sub-RPE drusenoid deposits around the vascular arcades with minimal involvement of the macular area. In the current study, we performed a systematic analysis of their macular OCT images, examining the thickness of the individual retinal layers as well as RPE and choriocapillaris, including their spatial relationship within the macula. We also conducted a complementary analysis of the choriocapillaris in postmortem human specimens comparing nondiabetic individuals, those with diabetes without retinopathy, and patients with DR. We found diabetes-related changes in retinal thickness, especially in the inner retina. Additionally, thickening of the choriocapillaris was noted in both macaques and human specimens. These findings indicate diabetes-induced changes in the neuroretina and choroid that may have implications for DR. 

Materials and Methods

Human Ocular Specimens

Postmortem human globes from autopsy specimens of individuals at Johns Hopkins Hospital were collected from 16 nondiabetic controls without ocular disease, 14 diabetic individuals without diabetic retinopathy, and 10 with DR. The tissues were fixed in 10% formalin and embedded in paraffin as described.²⁰ Study protocols were approved by the institutional review board at the Johns Hopkins School of Medicine (IRB00045080 approved August 21, 2014). 

Animals

All animal procedures in this study adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved and monitored by the Institutional Animal Care and Use Committee of Zhongshan Ophthalmic Center (No. W2021023). The animals were maintained at ∼26°C and 40% to 70% room humidity on a 12-hour/12-hour light/dark cycle. Diabetes was first screened among macaques over 15 years old from two animal facilities in southern China by measuring random blood glucose followed by a confirmatory test of positive fast glucose and an increased HbA1c level as previously described.¹⁹ A diagnosis of diabetes was made when a macaque had both a fasting glucose of 150 mg/dL or higher and an HbA1C of more than 6.5%. Animals were monitored by a trained technician and a veterinarian at all times. The demographic and basic clinical information are listed in Table 1. 

Table 1.

 

Demographic and Clinical Information of Macaques

 

OCT Examination and Retinal Thickness Measurement in Macaques

Spectral domain OCT (SD-OCT) imaging was performed as previously described.¹⁹ Macaques were anesthetized with Zoletil 50 (VIRBAC S.A.) (4 mg/kg), followed by topical application of proparacaine HCl (Alcaine, 0.5%; Alcon Laboratories, Geneva, Switzerland) and 0.5% tropicamide. SD-OCT images were captured with a Cirrus HD-OCT 5000 (Carl Zeiss Meditec, Dublin, CA, USA). The macular cube 512 × 128 protocol (512 A-scans per B-scans with a 5-µm axial resolution and a scanning depth of 2 mm) that covered an area of 6  ×  6 mm centered on the fovea was used for the scanning. The animals were monitored by a trained technician and a veterinarian at all times. 

Measurement of each retinal layer was performed as previously described.²¹ The section on a horizontal line scan through the foveal center was used for the analysis. The thickness of each retinal layer was measured at 1.5 mm nasal to the foveal center, at the foveal center, and at 1.5 mm temporal to the foveal center using the caliper tool with ImageJ (National Institutes of Health, Bethesda MD, USA). As shown in Figure 1, the layers included in the measurement include the RNFL, GCL, inner plexiform layer, inner nuclear layer (INL), outer plexiform layer, outer nuclear layer (ONL), photoreceptor inner segments (ISs), photoreceptor outer segments, RPE (also known as RPE–Bruch's membrane complex), and choriocapillaris (CC). All measurements were performed in a masked fashion by one (L.Z.) grader. 

Figure 1.

 

Retinal thickness measurements with OCT. Red arrows indicate individual retinal layers. CC, choriocapillaris; GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; IS, inner segments; ONL, outer nuclear layer; OPL, outer plexiform layer; OS, outer segments; RNFL, retinal nerve fiber layer; RPE, retinal pigment epithelial.

Human Sample Hematoxylin and Eosin Staining and Quantification

Hematoxylin and eosin (H&E) staining of 5-µm-thick formalin-fixed, paraffin-embedded human retinal sections was used to evaluate the CC morphology as previously reported.²⁰ The section of the macular region with the thickest ganglion cell layer was assessed. To evaluate the thickness of CC/Bruch's membrane in human specimens, four consecutive fields of images were taken. The CC layer remained horizontal in every image. Choriocapillaris/Bruch's membrane was measured from the top of Bruch's membrane to the bottom of the CC, selected using the magic lasso tool in Adobe Photoshop (Adobe, San Jose, CA, USA). The selection of the CC area for each image was done in a masked fashion. The average thickness of the CC was calculated by dividing the area of the CC by the length of the measured area. 

Statistical Analysis

All bar graphs represent mean ± SD. Linear mixed-effects models (LMEMs) were used to model each outcome (thickness of a retinal layer) as a function of diabetes status and sex while accounting for potential between-eye correlation using a random intercept. To fully examine the potential effect of sex, model building started with a full model including the main effects and the interaction between diabetes status and sex. For all the outcome variables, the interaction term was not significant. Thus, the final model for each outcome variable was the LMEM including the main effects of diabetes status and sex. A similar model-building strategy was used for the analysis of human data. Because there were few human samples where data for both eyes were eligible, the random intercept effect was unestimable, and the final model included diabetes status group (three levels) and sex as fixed effects. For the monkey data, because there were 27 outcome variables being examined at the same time, to adjust for multiplicity from the multiple comparisons, the Hochberg method was used to obtain multiplicity adjusted P values using the raw P values from the LMEMs. All statistical analyses were conducted in R (R Project for Statistical Computing, Vienna, Austria). 

Results

Retinal Thickness Changes in Diabetic Macaques

It has been reported that neuronal changes occur early in diabetes.^4,5 In our previous study, we performed clinical examination and OCT testing of 28 spontaneously diabetic macaques identified from our previous study,¹⁹ comparing them with a matched group of 26 nondiabetic macaques. The diabetic macaques did not exhibit any clinical evidence of DR, including microaneurysms, microhemorrhages, or lipid deposits.¹⁹ We investigated the effect of diabetes on the thickness of individual retinal layers using OCT. Scans with poor quality were excluded, resulting in the inclusion of 45 eyes from 25 diabetic macaques and 46 eyes from 26 age-matched nondiabetic macaques in our analysis. 

We examined the focal thickness of each retinal layer at 1.5 mm nasal and temporal to the foveal center as well as at the foveal center (Fig. 1). As shown in Figure 2 and Table 2, we observed a significantly decreased thickness of the inner retina, including the RNFL and the GCL, at both the nasal and temporal side of the fovea in diabetic macaques. The reduction of RNFL and GCL was 20.73% and 12.31% on the nasal side, respectively, and 19.72% and 9.20%, on the temporal side, respectively, in the diabetic macaques compared to age-matched nondiabetic macaques. 

Figure 2.

 

Thickness of each layer of retina in diabetic and age-matched nondiabetic macaques. (A) Thickness of retinal layers at 1.5 mm nasal to the foveal center. (B) Retinal thickness at the foveal center. (C) Thickness of retinal layers at 1.5 mm temporal to the foveal center; n = 46 eyes from 26 macaques for nondiabetic group; n = 45 eyes from 25 macaques for diabetic group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, no significant difference with linear mixed-effects models.

Table 2.

 

Retinal Thickness of Diabetic and Nondiabetic Macaques

 

Compared to the RNFL and GCL, the effect of diabetes on the remaining retinal layers was relatively mild. We observed a significant increase (9.61%) in the INL and a reduction in the ONL (7.46%) in the nasal retina. On the temporal side, a similar trend of retinal thickness change was observed, although it was not statistically significant. We also observed that the foveal IS was significantly thicker (5.50%) in diabetic macaques compared to nondiabetic macaques, and both temporal and nasal IS showed a trend toward thinning that was not statistically significant. There was a minor (2.14%) but significant reduction of total retinal thickness on the nasal side in the diabetic macaques. No significant difference in the total retinal thickness was observed at the temporal side and foveal center between the diabetic and nondiabetic macaques. 

RPE and Choriocapillaris Thickness Changes in Diabetic Macaque

We also investigated the thickness of RPE–Bruch's membrane CC in diabetic and nondiabetic macaques. As shown in Figure 3 and Table 3, no significant differences were observed in RPE–Bruch's membrane between the diabetic and nondiabetic macaques. SD-OCT in macaques exhibits a hyporeflective band corresponding to the CC, with a well-defined posterior limit with a peak of reflectivity from Sattler/Haller's layer due to the presence of large and darkly pigmented uveal melanocytes between the medium and large vessels of the choroid. At the same time, these melanocytes also obscure the visualization of the choroidal–scleral junction.²² In our cynomolgus macaque population, there was clear visualization of the CC, as expected (Figs. 1 and 3). We measured the thickness of the CC layer in diabetic and age-matched nondiabetic macaques. As shown in Figure 3 and Table 3, we observed a significant increase in CC thickness in the nasal and temporal macula but not at the fovea center in the diabetic macaques. Compared to the nondiabetic macaques, the CC thickness was increased by 20.23% and 22.96%, respectively, in the nasal and temporal side of the diabetic macaque. 

Figure 3.

 

Thickness of RPE/Bruch's membrane and CC in diabetic and age-matched nondiabetic macaques. (A) Representative OCT images from diabetic and nondiabetic macaques. Lower panels show the enlarged field of the CC; red arrowheads point to RPE/Bruch's membrane; yellow arrowheads point to CC. (B) RPE/Bruch's membrane thickness at 1.5 mm nasal to the foveal center, at the foveal center, and at 1.5 mm temporal to the foveal center. (C) Choriocapillaris thickness at 1.5 mm nasal to the foveal center, at the foveal center, and at 1.5 mm temporal to the foveal center. *P < 0.05; ns, no significant difference with linear mixed-effects models.

Figure 4.

 

H&E staining of postmortem human eye specimens. (A) Representative images of H&E staining on CC/Bruch's membrane in nondiabetic individuals, patients with diabetes without DR, and patients with NPDR. Lower panels show CC/Bruch's membrane in blue. (B) Quantitation of the average CC/Bruch's membrane thickness for each group. n = 23 eyes from 16 subjects for nondiabetic control; n = 19 eyes from 14 subjects for diabetes without DR; n = 13 eyes from 10 subjects for NPDR. *P < 0.05 with linear mixed-effects models.

Table 3.

 

RPE and CC Thickness of Diabetic and Nondiabetic Macaques

 

Choriocapillaris Thickness Changes in Human Diabetes

In contrast to macaques, SD-OCT in humans exhibits highly variable reflectivity levels between the RPE and sclera, and importantly, there is no distinct single hyporeflective CC to enable delineation of its boundary due to the loose distribution of uveal melanocytes, which are small and more lightly pigmented, in the choroid.²² SD-OCT–based measures of the CC in humans are therefore generally avoided. To investigate the alteration of CC in humans, we instead examined the thickness of CC/Bruch's membrane in patients with diabetes using H&E staining of human postmortem retinal specimens. We included 23 eyes from 16 nondiabetic patients, 19 eyes from 14 diabetic patients without retinopathy, and 13 eyes from 10 patients with nonproliferative diabetic retinopathy (NPDR). Four consecutive macular fields were used to quantify the thickness of CC/Bruch's membrane. Patient demographic information is described in Table 4. We observed an increase in average thickness of the CC/Bruch's membrane in the NPDR specimens compared to the nondiabetic and diabetes without retinopathy specimens. 

Table 4.

 

Demographic Information for Postmortem Eyes

 

Discussion

Nonhuman primates are recognized to be among the best models for human retinal diseases, in part because of the presence of a macula.¹⁴ Advances in research methodologies, including OCT imaging, have enabled recognition of very early changes in the retina in diabetes in both humans and rodents, including neuronal changes that may precede vascular changes.^4,5,8 NHP research has the potential to provide unique insights into the impact of diabetes on the retina, given the presence of a macula, a major and critical site of pathology in human DR. In this study, we used OCT imaging to examine retinal thickness in spontaneously diabetic cynomolgus macaques and their age-matched controls. We observed reductions in the thickness of the inner retina in diabetic macaques compared to their age-matched controls. Additionally, we found an increased thickness of the CC, but not RPE, in diabetic macaques. H&E staining of postmortem eye specimens from diabetic patients also showed increased thickness of the CC in NPDR compared with the age-matched nondiabetic patients and diabetic patients without DR. These macular findings in diabetic macaques further highlight the impact of diabetes on cellular elements beyond vascular cells in the inner retina that could play a role in the development and progression of DR. 

Macaques share a similar retinal anatomy with humans, including the presence of a macula.^14,17 The distinct characteristics of the macula, including its cone-rich nature, may be important in confirming and uncovering important diabetes-induced changes in other species that inform human retinal disease. In addition, cynomolgus macaques have a smaller body size, making them a more accessible nonhuman primate for research investigation. Compared to macaque models of induced diabetes, the spontaneous diabetes macaque model shares the greatest similarity to humans with respect to clinical features and disease pathology of diabetes.²³ Previous studies of spontaneously diabetic macaques mainly focused on vascular changes in the retina. Microaneurysms and nonperfused retina were observed in the late stage of diabetes in rhesus macaques.^24,25 

Our current study is the first that focuses on neuroretinal changes at an early stage of diabetes with a large number of cynomolgus monkeys with spontaneous diabetes. It should be noted that the macaques included in this study showed no significant clinical signs of diabetic retinopathy, such as microaneurysms, microhemorrhages, or lipid deposits. We found that thinning of the RGC, RNFL, and potentially ONL are observed at earlier stages of diabetes in macaques before significant vascular changes, such as microaneurysms, leakage, and neovascularization, were observed, although the macaques did exhibit histopathologic evidence of pericyte ghosts and acellular capillaries.¹⁹ This is similar to observations in human diabetic patients. Notably, retinal thinning in our diabetic macaque population was more significant on the nasal side, consistent with findings in human diabetes studies.^26,27 There was a higher degree of thinning of RNFL and RGC in the nasal versus temporal macula, and ONL was thinned in the nasal but not the temporal macula. The more pronounced thinning of the inner retina could potentially be related to the dense arrangement of RGC axons nasally. Thinning of the inner neuroretina is recognized as an early change in diabetes in both humans and rodents. Our current similar findings of inner retinal thinning in the macula of cynomolgus monkeys in early diabetes strongly reinforce an effect of diabetes on retinal neurodegeneration, particularly in retinal ganglion cells. 

Interestingly, we also observed a slight but significant increase in thickness of the INL in the nasal retina. The INL consists of bipolar cells, amacrine cells, and Müller cells. A study in humans has reported an increase in INL thickness in patients with early diabetes.²⁸ This was postulated to be from reactive hypertrophy observed of Müller glia in INL,^29,30 which is known to be associated with damage to regional neurons. 

We did not observe a difference in thickness of RPE/Bruch's membrane between diabetic and nondiabetic macaques with OCT, consistent with previous human studies.⁹ An increase in RPE height was found in a diet-induced rhesus macaque model.³¹ The difference in our findings could be due to the different species of macaques or differences in methodology. It is worth mentioning that compared with immunohistochemistry staining, OCT provides a more precise approach for comparing thickness measurements at the same location across different animals. Of note, recent OCT studies of RPE/Bruch's membrane in humans have demonstrated either no effect of diabetes³² or diabetes-associated thinning.³³ 

In humans, investigations of choroidal thickness in diabetes have yielded varying results. In particular, the pattern of loose distribution of the uveal melanocytes in the human choroid makes delineation of the CC difficult on SD-OCT,²² and such measurements are therefore generally not reported in humans. In comparison, the CC layer of macaques is much better visualized with OCT due to the distribution of dense pigmented uveal melanocytes that are only located between the medium- to large-caliber vessels in Sattler's and Haller's layers but not the CC layer in the choroid.²² Strikingly, we found an increase in CC thickness in diabetic macaques compared to nondiabetic control macaques. In postmortem human specimens, we investigated the height of the CC/Bruch's membrane in the macular region microscopically and observed thickening in patients with NPDR compared to those with DM without DR and nondiabetic individuals. At first glance, these findings are somewhat surprising, since studies have indicated capillary dropout and loss of capillaries in diabetic choroid.¹² Flow impairments in the CC have been reported in patients with NPDR using swept-source OCT angiography.³⁴ We surmise, however, that the observed thickening of the CC in diabetes might be attributable to remodeling or damage to this region. Contributing factors could include vascular basement membrane thickening, a hallmark of diabetes,³⁵ as well as vessel dilation³⁴ or damage related to leukocyte-induced damage.¹² Thickening of the CC could therefore be a marker of diabetic eye disease as imaging technologies advance. Of note, the increase in CC thickness in diabetic macaques was found in the nasal and temporal macula (1.5 mm from the foveal center) but not at the foveal center. The reasons for this regional change are currently unclear. However, it is noteworthy that regional differences have also been found in human diabetes and DR with respect to the CC. One study of nondiabetic and diabetic humans indicated parafoveal CC flow deficits that were greater than foveal flow deficits in NPDR but not DR or proliferative diabetic retinopathy (PDR).³⁶ Another study similarly found reduction of parafoveal but not foveal vessel density of the CC in humans with NPDR.³⁷ An additional point is that the diabetic macaques included in our study had no significant clinical signs of DR. Therefore, these macaques were more similar to the human subjects with diabetes but no DR and yet already manifested thickening of the CC. Future studies will be helpful to clarify the reasons for this distinction. For instance, it is possible that the course of changes in the CC is more rapid in diabetes in macaques relative to humans. 

Although changes in CC height/thickness have not been well studied in macaques in diabetes, it has been better studied in aging. A previous study also found that there was a correlation between increased CC thickness and aging in rhesus macaques.²¹ There have been very few OCT-based studies of CC thickness with aging in humans. However, one report found an increase in thickness of CC in patients with early-stage AMD.³⁸ This was postulated to be from intrachoroidal neovascularization or accumulation of ghost vessels. Although studies of diabetes have understandably focused much more on the retina than the choroid, our work further highlights the potential importance of future studies of the choroid, which could potentially affect the retina in diabetes, given the role of the choroid in regulating both the outer retina and RPE.⁹ Of note, diabetes is known to cause dysfunction of photoreceptors¹⁰ and breakdown of the outer blood–retinal barrier constituted by the RPE.^39,40 These initial changes can potentially affect the inner retina as well.^3,9,10 

Nonhuman primates represent the closest animal model to human retinal diseases, including DR, because of the presence of a macula and the very close anatomic similarities.^14,17 Our findings of macular thickness changes in our diabetic cynomolgus macaque population enhance the concept of a diabetes-induced primary retinal neurodegeneration, particularly of the retinal ganglion cells. In addition, our observations of CC thickness in this macaque population and in human postmortem retinas indicate a heretofore unappreciated diabetes-induced thickening of the CC. These findings further highlight the involvement of other cellular elements altered by diabetes beyond the retinal vasculature. It also highlights changes in the inner retina and CC as contributors to the pathogenesis of DR and as markers for progression. 

Acknowledgments

Supported by research grants from the National Institutes of Health (EY035549 and EY035897, EJD), the Altsheler-Durell Foundation, P30EY001765 to Wilmer Imaging and Microscopy Core and Wilmer Biostatistics Core Modules, unrestricted funds from the Research to Prevent Blindness, grants from the Guangdong Provincial Key Area R&D Program (Grant No. 2023B1111050004, XL and WY), and Guangzhou Science and Technology Plan Project (Grant No. 2024B01J1121, XL and WY). 

Disclosure: L. Zhou, None; S. Fan, None; H. Lu, None; W. Zhang, None; K. Ri, None; Z. Xu, None; X. Kong, None; A.H. Kashani, None; X. Liu, None; C. Eberhart, None; W. Yi, None; E.J. Duh, None