Progressive Choriocapillaris Changes on Optical Coherence Tomography Angiography Correlate With Stage Progression in AMD

Abstract

Purpose: We investigated the association between inner choroid flow deficit percentage (IC-FD%) using swept-source optical coherence tomography angiography (SS-OCTA) and progression of AMD.

Methods: Retrospective, observational study including 64 eyes (42 participants) with early or intermediate AMD at baseline. Participants had two or more consecutive swept-source optical coherence tomography angiography covering a period of at least 18 months. Demographics, visual acuity, and AMD staging based on Beckman classification were reviewed. OCT was analyzed for hyperreflective foci, subretinal drusenoid deposits, hyporeflective drusen cores, and subfoveal choroidal thickness. IC-FD% was measured within the central 3- and 6-mm using a 16-µm slab, after compensation and binarization (Phansalkar method). Mixed-effects Cox regression models assessed the association between imaging biomarkers and AMD progression.

Results: During follow-up (37 ± 9 months), 4 eyes with early AMD (31%) progressed to intermediate AMD and 30 (59%) eyes with intermediate AMD developed late AMD (19 geographic atrophy; 11 wet AMD). Baseline hyporeflective drusen core was associated with geographic atrophy development (P < 0.01), whereas greater IC-FD% (3-mm) was associated with wet AMD (P = 0.03). Time-varying analysis showed that faster subfoveal choroidal thickness reduction and IC-FD% (6-mm) increase were associated with geographic atrophy onset (P < 0.05), whereas IC-FD% (3-mm) increase was associated with wet AMD (P = 0.03). Notably, greater IC-FD% increases in the 3 mm (area under the curve = 0.72) and 6 mm (area under the curve = 0.89) were better predictive of wet AMD and geographic atrophy development, respectively.

Conclusions: Our longitudinal IC-FD% assessment emphasizes the role of progressive choriocapillaris changes as a biomarker for AMD progression. Our findings support that widespread choriocapillaris alterations (6 mm) may precede progression to geographic atrophy, whereas more central choriocapillaris loss (3 mm) may provide an ischemic stimulus for wet AMD.

AMD is the leading cause of vision loss in patients over 50 years of age in developed countries, with its prevalence expected to increase by 2040 with global population aging.^1,2 The multifactorial pathogenesis of AMD involves various molecular and cellular pathways resulting in the degeneration of photoreceptors, retinal pigment epithelium (RPE), and choriocapillaris.^3,4 Although the initial course of AMD is characterized by minimal visual symptoms, advanced manifestations tend to impair patients’ vision and quality of life severely. As such, there is a critical need for better predictors of disease progression to facilitate early interventions. 

Numerous studies have explored potential associations between spectral domain-optical coherence tomography (OCT) retinal biomarkers and progression to late AMD. Historically, structural imaging biomarkers such as central drusen volume, intraretinal hyperreflective foci (iHRF), subretinal drusenoid deposits (SDDs), and hyporeflective drusen cores (hDCs) have been identified.^5–7 In addition to these biomarkers, substantial evidence suggests that choriocapillaris might play a significant role in the development and progression of AMD.³ However, in vivo assessment of choriocapillaris alterations has been challenging historically owing to the limitations of traditional dye-based angiography, which provides insufficient depth-resolved information and low digital resolution.⁸ 

The advent of swept-source OCT angiography (SS-OCTA) has markedly enhanced our understanding of AMD pathogenesis markedly by elucidating the role of inner choroidal alterations.^9–11 SS-OCTA, with its deeper penetration and advanced segmentation algorithms, has enabled high-resolution visualization and quantification of the choriocapillaris, effectively addressing many issues associated with drusen shadowing effect.¹¹ Previous studies indicated that inner choroid flow deficit percentage (IC-FD%) represents the most accurate metric for quantifying the extent of choriocapillaris impairment. Notably, a greater IC-FD% correlates with AMD severity,¹² is associated with greater risk of drusen formation or enlargement,⁷ and may predict AMD progression ultimately.^13,14 

Despite these advances, longitudinal SS-OCTA data assessing choriocapillaris in AMD with extended follow-up periods remain limited. Therefore, our study aims to investigate the association between IC-FD%, various OCT biomarkers, and AMD progression longitudinally in patients with dry nonadvanced AMD. 

Methods

Design and Patient Selection

This retrospective, longitudinal, observational study was conducted at the Retina Service of Massachusetts Eye and Ear (MEE). The study adhered to the tenets of the Declaration of Helsinki and complied with Health Insurance Portability and Accountability requirements. Approval was obtained from the Massachusetts General Brigham Institutional Review Board (protocol no: 2019P001863), and written informed consent was obtained from all participants. 

Clinical records and imaging studies of patients diagnosed with early or intermediate AMD seeking treatment at the Retina Service of MEE between September 2018 and April 2023 were reviewed systematically. Inclusion criteria were as follows: (1) clinical diagnosis of early or intermediate (dry nonadvanced) AMD based on the Beckman classification¹⁵; (2) age of 55 years or older at the time of diagnosis; (3) at least two same-day spectral-domain OCT (Spectralis HRA+OCT2; Heidelberg Engineering, Heidelberg, Germany) and SS-OCTA (PLEX Elite 9000; Carl Zeiss Meditec Inc, Dublin, CA) examinations (4a) before progression or (4b) covering a minimum follow-up of 18 months; (5) absence of nonexudative macular neovascularization (MNV) or complete RPE and outer retinal atrophy (cRORA) at baseline.^16–18 Exclusion criteria included (1) the presence of motion artifacts in SS-OCTA data or signal strength index of less than 7, (2) significant media opacities, (3) the use of systemic medications (e.g., hydroxychloroquine), ocular (e.g., glaucoma), or systemic conditions (e.g., diabetes) that could potentially affect the analysis, (4) a positive family history of inherited retinal dystrophies, (5) refractive errors of greater than ±6 diopters (spherical equivalents), (6) a history of intraocular inflammation, and (7) ocular treatment during the follow-up period, except for cataract extraction or anti-vascular endothelial growth in case of MNV development. Both eyes of the same patient were included if they met these criteria. 

Data Collection

Baseline and follow-up visits were reviewed for the following clinical and demographic data: age, sex, race/ethnicity, smoking status (never, former, or current), AREDS2 supplementation, statin use, lens status (phakic or pseudophakic), visual acuity (converted to ETDRS letters), AMD stage, stage of the fellow eye (earlier/same or more advanced), and, when applicable, the date of progression to a more advanced stage. AMD staging was determined according to the Beckman classification¹⁵ by four experienced retina specialists (J.B.M., D.G.V., J.W.M., and D.H.) based on clinical examination and confirmed by assessing multimodal imaging studies. The development of late dry AMD was defined as the appearance of at least one area of cRORA on SD-OCT following criteria established by the CAM group.¹⁷ Conversion to wet AMD was defined based on specific features on OCT, including (1) elevation of the RPE with heterogeneous reflectivity and signs of exudation (type 1 MNV), (2) a neovascular complex located in the subretinal space with possible subretinal hyperreflective material (type 2), or (3) a hyperreflective extension from the middle retina to the RPE with intraretinal fluid (type 3), as recommended by the CONAN study group.¹⁹ Fluorescein angiograms were also reviewed for confirmation when available. 

Imaging Analysis

Imaging studies from same-day OCT and SS-OCTA visits were independently reviewed by two masked ophthalmologists with medical retina training (F.R. and X.D.). The mean values were used for quantitative variables, and discordant qualitative findings were adjudicated by a senior retina specialist (J.B.M.). OCT grading was conducted using spectral-domain OCT scans (20° × 15°, 37 lines with ≥9 automatic real-time tracking frames per scan) owing to their higher axial resolution and improved signal-to-noise ratio.^9,10 Specifically, macular OCT volumes were reviewed with the Heidelberg Eye Explorer 2 software (HEYEX2, version 2.5.7; Heidelberg Engineering GmbH) for (1) iHRF, well-circumscribed round lesion located in the neurosensory retina with reflectivity equal to or higher than the RPE and a size of at least 30 µm,²⁰ (2) SDD, subretinal accumulation of cone- or mound-shaped lesions with intermediate to high reflectivity,²¹ (3) hDC, exhibiting “heterogeneous internal reflectivity” within drusenoid lesions,²² and (4) subfoveal choroidal thickness, measured as the distance between the outer RPE border and the hyporeflective sclero-choroidal interface.²³ 

SS-OCTA volume scans covering a 6 × 6 mm area (100,000 A-scans per second) were used to analyze drusen volume in the central 3 mm (DV3) and IC-FD% on the PLEX Elite 9000 review software (version 2.1.0.55513; Carl Zeiss Meditec, Dublin, CA). Only scans with satisfactory image quality (signal strength index of >7) and devoid of motion artifacts were included in the study, after manually correcting the RPE, RPE fit, and Bruch's membrane references for segmentation errors. The Advanced RPE Analysis software (version 6.0), which has been approved by the US Food and Drug Administration, provided by the manufacturer was used to generate drusen maps (color-coded elevation maps between the RPE and the RPE fit lines²⁴) and to measure DV3. A customized inner choroidal slab with a thickness of 16 µm, starting 4 µm beneath the Bruch's membrane, was created following recent imaging guidelines suggested by Chu et al.²⁵ The resulting 6 × 6 mm angiograms and structure en face images were exported as .tiff files, visually inspected for degradative artifacts by the two graders, and imported to FIJI (ImageJ2 version 2.9.0; NIH, Bethesda, MD). After inverting the en face slab, the two images were multiplied to compensate for signal loss beneath drusen and other hyperreflective structures.²⁶ Images were then binarized using a low-contrast local thresholding method (Phansalkar) with two different radii (4–15 pixels; only the latter is reported in the results).^25,27 The IC-FD% was calculated using the “Analyze Particles” command after creating a mask to eliminate large vessels from superficial capillary plexus slab and drusen from the drusen map.^28,29 The large vessel masks were also used to align follow-up slabs using the Data Science for Health Image Alignment tool.³⁰ The IC-FD% was measured within the central 3-mm and 6-mm circles, excluding flow deficits smaller than 24 µm, considered to fall within the normal range of intercapillary distance (Fig. 1).³¹ 

Figure 1.

 

Schematic representation of the inner choroid flow deficit percentage (IC-FD%) analysis. A customized angiography choriocapillaris slab (offset, 4–20 µm) placed underneath the Bruch's membrane and the corresponding inverted structural en face slab were multiplied to obtain a compensated image. The resulting image was binarized using a low-contrast local thresholding technique (Phansalkar method; radius = 4 and 15 pixels). Two masks were generated from the superficial capillary plexus and the drusen map to highlight the structures that might exert a shadowing effect on the underlying choriocapillaris. The IC-FD% was then measured in the central 3- and 6-mm regions using the ‘Analyze Particle’ tool after excluding the areas detected on vessel and drusen masks.

Statistical Analysis

Statistical analyses were conducted using SPSS software version 28.0 (IBM Corporation; Armonk, NY) and R software version 4.3.1 (The R Foundation for Statistical Computing; Vienna, Austria; https://cran.rproject.org) with the EpiR, coxme, and ggplot2 packages. Statistical significance was set at a P value of less than 0.05, and all tests were two sided. Descriptive statistics are presented as mean ± standard deviation (range), median (quartiles), or frequency (percentage, %), as appropriate. Interrater agreement between the two ophthalmologists was assessed using intraclass correlation coefficient (95% confidence intervals) and Cohen's k factor (κ, 95% confidence intervals). 

Kaplan–Meier estimates were used to calculate and visualize the cumulative incidence of AMD progression and visual loss over time. Mixed-effects Cox regression analysis was performed to assess the effect of the analyzed OCT and SS-OCTA biomarkers on (1) any AMD progression, (2) cRORA, and (3) wet AMD development. All models accounted for nesting of eyes within patients and included age, AREDS2 supplementation, and the status of the fellow eye as covariates.³² Two distinct models were generated for each outcome: one including baseline status and another with time-varying characteristics as fixed effects. Results were reported through hazard ratios (HR with 95% confidence intervals and P values). Last, receiver operating characteristic (ROC) analysis was conducted to determine the area under the curve (AUC) for each imaging variable, thereby predicting the probability of progressing to a more advanced stage. 

Results

We included 64 eyes from 42 patients, with the majority being women (31 [73.8%]). The mean age at baseline was 71.4 ± 6.8 years (range, 55-83 years), and the mean follow-up period was of 37.4 ± 9.0 months (range, 18.0–51.1 months). The mean baseline visual acuity was 78.5 ± 7.8 ETDRS letters (20/25 Snellen) and decreased to 74.0 + 12.1 ETDRS letters at the last follow-up visit (20/32), with 11 eyes experiencing a loss of three or more ETDRS lines (17.2%). 

At baseline, 13 eyes were classified as having early AMD (20.3%) and 51 as intermediate AMD (79.7%). Based on OCT examination, iHRFs were identified in 55 eyes (85.9%), SDD, in 20 (31.3%), and hDCs in 11 (17.2%) at baseline. The mean IC-FD% was 16.4% ± 3.2% (range, 9.8%–25.0%) within the central 3-mm and 16.5% ± 3.4% in the 6-mm circle (range, 9.6%–25.8%). The complete demographic and clinical features of our cohort are summarized in Table 1. 

Table 1.

 

Demographic and Clinical Features of the Studied Cohort

 

Table 2.

 

Clinical and Imaging Differences Between Intermediate AMD Patients Remaining Stable and Progressing to Either Complete RPE and Outer Retinal Atrophy or Wet AMD

 

Baseline Characteristics and AMD Progression

Over the follow-up, 34 eyes demonstrated progression to a more advanced AMD stage (Table 2). Specifically, four eyes diagnosed initially with early AMD progressed to intermediate (30.8%), 19 eyes with intermediate AMD developed cRORA (37.3%), and 11 eyes with intermediate AMD converted to wet AMD (21.6%). The estimated median times to any progression, cRORA and wet AMD development were 23.4 months (14.7–37.7 months), 30.4 months (17.9–41.6 months), and 15.2 months (5.5–20.2 months), respectively. 

In our univariate analysis, baseline characteristics associated with AMD progression included older age (73.8 ± 6.5 vs 69.5 ± 6.7 years; ß = 2.62; P = 0.01), less frequent AREDS2 supplementation (76.5% vs 93.3%; ß = −1.54; P = 0.05), and worse baseline visual acuity (76.6 ± 9. ETDRS letters 1 vs 80.6 ± 5.5 ETDRS letters; ß = 2.10; P = 0.04). The Kaplan–Meier estimates for AMD progression and the loss of three or more ETDRS lines are presented in Figure 2. 

Figure 2.

 

Kaplan-Meier estimates for the incidence of AMD stage progression and visual acuity (VA) loss during the follow-up. Incidence rates for any AMD progression (green lines), geographic atrophy development (red lines), and wet AMD conversion (blue lines) were estimated to be 33.5, 20.4, and 11.4, person-years, respectively. The estimated incidence rate for VA loss of at least three Early Treatment for Diabetic Retinopathy Study (ETDRS) lines was of 13.9 person-years; this characteristic was particularly evident in eyes converting to wet AMD.

Imaging Biomarkers and AMD Progression

Our multivariate Cox regression model, adjusted for age, AREDS2 supplementation and fellow eye status, revealed that the baseline presence of hDC (HR, 4.43; 95% CI, 1.59–12.37; P = 0.004), SDD (HR, 2.83; 95% CI, 1.07–7.50; P = 0.04), and larger IC-FD% within the central 3 mm (HR, 1.26; 95% CI, 1.07–1.50; P = 0.007) were associated with an increased risk of any stage progression. Specifically, baseline hDC was significant for the development of cRORA (HR, 8.94; 95% CI, 2.02–39.61; P = 0.004), whereas a larger IC-FD% in the central 3 mm was associated with conversion to wet AMD (HR, 1.36; 95% CI, 1.03–1.80; P = 0.03) (Fig. 3). 

Figure 3.

 

Mixed-effect Cox regression models to estimate AMD progression from baseline (top) and time-dependent (bottom) imaging characteristics. The occurrence of any AMD progression was associated with the presence of hDC (HR, 4.43; P = 0.004), subretinal drusenoid deposits (SDDs; HR, 2.83; P = 0.04), and greater IC-FD% in the central 3 mm (HR, 1.26; P = 0.007) at baseline and with faster increase of IC-FD% in the central 3-mm circles (HR, 1.56; P = 0.02) and 6-mm circles (HR, 2.02; P = 0.01) over the follow-up. Progression to geographic atrophy was associated with the presence of hDC at baseline (HR, 8.94; P = 0.004) and with faster subfoveal choroidal thinning (HR, 1.25; P = 0.04) and IC-FD% increase in the 6-mm circle (HR, 3.32; P = 0.01). Conversion to wet AMD was associated with greater IC-FD% in the 3-mm circle at baseline (HR, 1.36; P = 0.03) and faster increase during the follow-up (HR, 1.87; P = 0.03).

Our Cox regression model for time-varying biomarkers revealed that an increased risk of any AMD progression was associated with a faster increase in IC-FD% in the central 3 mm (HR, 1.56; 95% CI, 1.07–2.26; P = 0.02) and in the 6-mm region (HR, 2.09; 95% CI, 1.17–3.75; P = 0.01). For progression to late AMD, cRORA development was associated with faster subfoveal choroidal thickness thinning (HR, 1.25; 95% CI, 1.02–1.55; P = 0.048) and a faster IC-FD% increase in the 6-mm region (HR, 3.32; 95% CI, 1.32–8.35; P = 0.01), whereas wet AMD conversion was associated independently with faster IC-FD% increase in the central 3 mm (HR, 1.87; 95% CI, 1.07–3.24; P = 0.03) (Fig. 3). Changes in IC-FD% over time in the different sub-groups are presented in Figure 4 and Supplementary Figure S1. 

Figure 4.

 

Expected values and 95% confidence intervals for IC-FD% in our AMD cohort over the follow-up. (Top) Eyes progressing to more advanced AMD stages exhibited a faster increase in IC-FD% both in the central 3- and 6-mm circles (red line), reaching a plateau after approximately 24 months of follow-up. (Bottom) When assessing intermediate AMD, both eyes progressing to geographic atrophy (red line) and those converting to wet AMD (blue line) showed a greater increase in IC-FD% in the two analyzed circles.

ROC Analysis

An ROC analysis of individual baseline imaging characteristics revealed that a larger IC-FD% had the highest predictive value for any AMD progression (AUC = 0.72). The identification of hDC had a predictive value for cRORA development (AUC = 0.68), and more extensive DV3 for wet AMD conversion (AUC = 0.69). Similarly, ROC analysis for time-varying characteristics showed greater diagnostic accuracy than for baseline characteristics. A faster IC-FD% increase in the central 3 mm was predictive of any AMD progression (AUC = 0.84) and of wet AMD conversion (AUC = 0.72), whereas a faster IC-FD% in the 6-mm region was highly predictive of cRORA development (AUC = 0.89) (Supplementary Fig. S2). 

Discussion

In this longitudinal study, we analyzed the association between IC-FD%, various OCT biomarkers, and AMD progression over a mean follow-up period of 36 months. Our findings suggest that the development of cRORA was associated with the presence of hDC at baseline, as well as with greater subfoveal choroidal thinning and IC-FD% increase during the follow-up. Conversely, wet AMD conversion was linked strongly with more extensive IC-FD% changes within the central 3 mm both at baseline and longitudinally. 

The quest for accurate imaging biomarkers predicting progression to late AMD holds significant importance for several reasons, including enhancing patient counselling, establishing tailored follow-up plans, and providing standardized endpoints for clinical trials.^33,34 Traditionally, soft drusen and pigmentary abnormalities have been considered primary predictors of progression,³⁵ and only recently have SDDs been recognized as a third macular risk feature.^36,37 Recent advances in retinal imaging have expanded the range of novel imaging biomarkers for detailed exploration in clinical and research settings.³⁸ 

In our study, we observed AMD progression in 34 eyes (53.1%) over a mean of 36 months, aligning with the diverse reported 5-year progression rates reported in the literature, which vary depending on severity (0.4% to 53%).³⁵ We identified significant associations between the presence of hDC, SDD, and greater IC-FD% at baseline and overall AMD progression. Notably, eyes presenting with hDC at baseline were strongly associated with the onset of cRORA, whereas those with greater IC-FD% in the central 3 mm were more prone to develop wet AMD. These findings bolster previous research indicating that SDD presence is linked to a 2- to 6-fold increased risk of late AMD.^21,39 Furthermore, we corroborate the notion that hDC serve as precursors to cRORA formation, indicative of substantial RPE impairment preceding drusen collapse.^22,40 

Our findings on the associations between longitudinal IC-FD% changes and AMD progression are of particular interest for gaining insights on AMD pathogenesis. We found that a faster rate of choriocapillaris loss in the central 3-mm circle over time was a significant risk factor for wet AMD conversion. This observation aligns with a previous study by Corvi et al.⁴¹ on the topographic distribution of choriocapillaris alterations, suggesting that eyes progressing to wet AMD have a predilection for greater central choriocapillaris loss. Histopathological studies on postmortem eyes have also supported this hypothesis, indicating that choriocapillaris drop-out is an early AMD characteristic that increases with age and in the presence of drusen.^42–44 In addition, Bhutto and Lutty⁴⁵ proposed that choriocapillaris loss with a relatively intact RPE may serve as the initial insult for the development of MNV owing to macular ischemia deriving from stenosis of large choroidal vessels. 

Our research identified a faster IC-FD% increase in the central 6 mm and subfoveal choroidal thinning over time as predictors of dry AMD progression. This finding is consistent with a growing body of literature linking choriocapillaris impairment on OCTA with cRORA development and enlargement.^46–48 These results are also in line with previous studies indicating that more peripheral or generalized choriocapillaris changes may predict progression to cRORA.^41,49 Furthermore, Corradetti et al.¹³ demonstrated that greater baseline IC-FD% on SS-OCTA may predict the formation of nascent geographic atrophy over 24 months of follow-up. However, their study had a limited sample size (28 eyes) and focused only on baseline features, failing to assess choriocapillaris alterations over time. 

Numerous imaging and histopathological studies have also evaluated the role of choroidal thickness in the pathogenesis of AMD.^43,50–52 The temporal relationship between the analyzed choroidal variables and progression to cRORA might indicate a secondary involvement of the choroid in the pathogenesis of dry AMD. In fact, Bhutto and Lutty⁴⁵ hypothesized that RPE dysfunction may serve as the primary trigger for geographic atrophy, followed by a secondary diffuse choriocapillaris loss, which might be insufficient to support an angiogenic response. 

Last, in our cohort, we did not find an association between iHRF and AMD progression,⁵³ which may be due to the high prevalence of iHRF at baseline and the lack of a quantitative, spatially resolved approach in our analysis. This finding highlights the potential of emerging machine learning-based approaches, combined with larger datasets, to refine new biomarkers.^54,55 

We acknowledge that our research has certain limitations. First, the retrospective design may introduce some selection bias. Second, the sample size is modest and likely underpowered to detect minor differences impacting AMD progression. Other constraints include the lack of averaging of multiple OCTA volumes and the variability in RPE and choroidal involvement across different AMD phenotypes. The frequency of same-day OCT and SS-OCTA examinations varied per patient, being influenced by the fellow eye stage. Last, our choriocapillaris findings are specific to the PLEX Elite 9000 device and thus may not translate directly to other OCTA systems. The strengths of this research include its relatively longer follow-up, the longitudinal assessment of IC-FD% in our AMD cohort, and the robust statistical analysis. 

In conclusion, our study provides the longest longitudinal assessment of choriocapillaris alterations’ impact on AMD progression using SS-OCTA. The conversion to wet AMD was strongly associated with greater IC-FD% in the central 3 mm at baseline and over time, whereas the development of cRORA seemed to be predicted better by the identification of hDC at baseline and a higher rate of IC-FD% increase across the macular area. Although our findings may not be definitive owing to potential underpowering, they provide deeper insights into the pathophysiology of AMD, supporting the strong relationship between progressive choriocapillaris insufficiency and increased AMD severity. This work emphasizes the importance of choriocapillaris alterations as potential biomarker for risk stratification and monitoring in future clinical trials, paving the way for more targeted and effective interventions in AMD. 

Acknowledgments

Disclosure: F. Romano, None; X. Ding, None; M. Yuan, None; F. Vingopoulos, None; I. Garg, None; H. Choi, None; R. Alvarez, None; J.H. Tracy, None; M. Finn, None; P. Ravazi, None; I.V.M. Stettler, None; I. Laìns, None; D.G. Vavvas, Drusolv (A); patents on neuroprotection held by Mass Eye and Ear (F); D. Husain, Allergan (C), Genetech (C), Omeicos Therapeutics (C), National Eye Institute (F), Lions VisionGift (F), Commonwealth Grant (F), Lions International (F), Syneos LLC (F), Macular Society (F); J.W. Miller, Genetech/Roche (C), Sunovion (C), KalVista Pharmaceuticals (C), ONL Therapeutics (C), Heidelberg Engineering (F), Lowy Medical Research Institute, Ltd. (F), US 7,811,832 (with royalties paid by ONL Therapeutics to Massachusetts Eye and Ear), US 5,798,349, US 6,225,303, US 6,610,679, CA 2,185,644, CA 2,536,069 (with royalties paid by Valeant Pharmaceuticals to Massachusetts Eye and Ear) patent; J.B. Miller, Alcon (C), Allergan (C), Carl Zeiss (C), Sunovion (C), Topcon (C), Genentech (C)