Optical coherence tomography (OCT) angiography

Transcription

1 CLINICAL TRIAL ENDPOINTS FOR OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY IN NEOVASCULAR AGE- RELATED MACULAR DEGENERATION EMILY D. COLE, BS,* DANIELA FERRARA, MD, PHD,* EDUARDO A. NOVAIS, MD,* RICARDO N. LOUZADA, MD,* NADIA K. WAHEED, MD, MPH* Purpose: To describe qualitative and quantitative optical coherence tomography (OCT) angiography (OCTA) parameters for choroidal neovascularization (CNV) secondary to agerelated macular degeneration (AMD) and their applicability as potential clinical trial endpoints. Methods: A review of current literature related to the topic of OCTA and AMD. Results: There are a number of promising OCTA parameters that can be used to diagnose the presence of CNV and to monitor the activity and progression of the lesion, pre- and post-treatment morphological characteristics, CNV dimensions, and automated quantitative parameters such as vessel density. Conclusion: The OCTA parameters described in this review have promise for the future development of clinical trial endpoints, but require further validation before they can be widely used. RETINA 36:S83 S92, 2016 Optical coherence tomography (OCT) angiography (OCTA) has been recently used to characterize the vascular morphology of choroidal neovascularization in a variety of pathologic conditions, including age-related macular degeneration (AMD) and its response to anti vascular endothelium growth factor (anti-vegf) agents. 1 4 The expanded use of this imaging modality may limit the use of invasive fluorescein angiography (FA) and indocyanine green angiography (ICGA) for the assessment of neovascular AMD (namd). The OCTA is a comfortable, From the *New England Eye Center, Tufts Medical Center, Boston, Massachusetts; Department of Electrical Engineering and Computer Science, and Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Massachusetts; Federal University of Goiás, Goiânia, Brazil; and School of Medicine, Federal University of São Paulo, São Paulo, Brazil. Supported in part by a grant from the Macula Vision Research Foundation, New York. E. A. Novais and R. N. Louzada are researchers supported by CAPES Foundation, Ministry of Education of Brazil, Brasília, DF, Brazil. D. Ferrara is an employee at Genentech, Inc and has Roche stock/stock options. N. K. Waheed has received research support and speaker s fees from Carl Zeiss Meditec and Optovue. The remaining authors have no conflicting interests to disclose. Reprint requests: Nadia K. Waheed, MD, MPH, New England Eye Center, Tufts Medical Center, 260 Tremont Street, Biewend Building, 9 11th Floor, Boston, MA 02116; nadiakwaheed@gmail.com fast, and noninvasive imaging strategy that has promise in future clinical trials for namd. In clinical trials, a primary endpoint characterizes the outcome of interest, may be clinically meaningful, and may represent how a patient feels or functions. In most studies of retinal pathology, visual acuity has been historically adopted as the most common functional endpoint, although it has recognized limitations in many retinal diseases. 5 Anatomical endpoints are often described as surrogate endpoints, in that they may act as a surrogate for a clinically meaningful endpoint or for long-term visual prognosis. These are supported by the therapeutic pathway or mechanism of action from drug to outcome, 6 8 and usually demonstrate the ability to predict a clinically relevant outcome as well as a treatment effect on this outcome. 9 Criteria for a good surrogate endpoint include being objective, reproducible, sensitive, specific, unbiased, easy to interpret, free of errors of ascertainment or measurement, stable, observable, and independent of treatment assignment. Moreover, clinical trial endpoints should ideally be chosen a priori at the time of design of the clinical trial. In clinical trials for namd, OCT biomarkers such as central retinal thickness have often been used as surrogate endpoints S83

2 S84 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2016 VOLUME 36 NUMBER 12 Potential clinical endpoints of namd on OCTA still require further development and validation with independent studies. When evaluating OCTA-based clinical endpoints, some of the criteria listed above for a good surrogate endpoint may not be achievable with the currently available technology. While OCTA imaging will undergo significant advances in both hardware and software, at a minimum, OCTA endpoints should be reproducible, vary in the presence of disease activity, and remain stable in patients without disease. Furthermore, these endpoints need to be reproducible between and within scans on different commercially available OCTA devices. Many of these features are related to the nature of the assessment, but others are fundamentally related to the interpretation of the multidimensional OCTA data set. Therefore, the validity of OCTA relies on experienced assessors. Ideal candidates for a clinical endpoint in namd should reliably identify the presence or absence of CNV and identify features that correlate to lesion activity. It should also provide additional information beyond what is already available on structural OCT scans, or obviate the need for more invasive studies such as FA or ICGA. Anti-VEGF agents have been extensively demonstrated in multiple clinical trials to improve visual acuity in patients with namd Imaging modalities such as FA and OCT have been used as surrogate endpoints as well as part of retreatment criteria in clinical trials. The MARINA and ANCHOR trials used FA leakage as part of the criteria for retreatment The CATT, PrONTO, and VIEW 1 and 2 trials used OCT to assess central retinal thickness and the presence of subretinal or intraretinal fluid for both recruitment and retreatment The aim of this article is to review the current literature describing the qualitative and quantitative features of namd on OCTA. Proposed OCTA clinical endpoints for namd can be divided into two categories: qualitative and quantitative features of CNV. Qualitative Features of CNV Qualitative features of CNV include the presence or absence of CNV, CNV type, morphology, pre- and post anti-vegf treatment characteristics, and choriocapillaris analysis. The first and the most important characteristic is accurately grading for the presence or absence of CNV. To do so, the OCTA must be evaluated in a multidimensional fashion alongside the corresponding cross-sectional scan with visible segmentation Fig. 1. Examples of areas of geographic atrophy (GA) that can artifactually give the appearance of CNV due to increased signal penetration and visualization of underlying choroidal vessels. All images were performed on an OptoVue Avanti RTVue XR with Angio- Vue software (Fremont, CA). A. Structural en face OCT scan segmented at the level of the choriocapillaris showing a small, focal area of GA. A larger choroidal vessel can be seen on the en face image and the OCTA (yellow arrow). B. A 3 mm 3 mm en face OCT angiogram segmented at the level of the choriocapillaris showing a smaller focal area of GA that may be misinterpreted as CNV because of the appearance of branching vascular structures on OCTA segmented at the level of the choriocapillaris. The corresponding cross-sectional OCT scan shows a focal area of hypertransmission in the choroid.

3 OCTA ENDPOINTS FOR NEOVASCULAR AMD COLE ET AL S85 lines, and the en face and cross-sectional structural OCT scans. This allows for evaluation of imaging artifacts, such as shadowing, that may obscure the appearance of CNV; and allows better assessment of the integrity of the tissues, such as in small, focal areas of geographic atrophy, which can give a false-positive appearance of CNV (Figure 1). The corresponding cross-sectional structural OCT scan, with or without angiographic-pixel overlay, can also be used to grade the type of CNV based on its location relative to the retinal pigment epithelium in CNV Types 1, 2, or 3. 2,24 The pixels corresponding to the area of OCTA decorrelation signal may more obviously delineate the location of CNV than standard cross-sectional OCT. A variety of CNV morphologies have been described, including the medusa shape, seafan shape, glomerular shape, and the dead tree appearance. 25,26 However, not all CNV follow these patterns, suggesting that there is further variation in vascular branching and morphology in these lesions (Figure 2). In addition, the clinical relevance of these morphologic variations is still debatable. Although there is much anecdotal evidence, these morphologies have not yet been correlated to CNV activity, or the appearance of treated versus treatment-naive lesions and the prognostic significance of these morphologies is yet to be determined. In current schemes, the CNV is described by the appearance of the vascular core and the appearance of the surrounding vascular network. 27 Coscas et al have described a series of microvascular features common to both treatment-naive and previously treated lesions, and correlated the presence of these features to the treatment outcome, either prompt anti-vegf treatment or no need for treatment. 28 Microvascular features associated with prompt treatment in this study included a lacy wheel or seafan shape, numerous branching tiny capillaries, the presence of anastomoses or loops, a peripheral arcade at the vessel termini, and the presence of a perilesional hypointense halo surrounding the lesion. 28 However, further studies are needed to replicate and validate these findings, and currently there are no established treatment paradigms for CNV based on OCTA. It is hypothesized that anti-vegf treatment may prune smaller vessels in the CNV lesion. 29 OCTA features can be evaluated for pruning of the peripheral arcade and areas within the lesion. 29,30 Although these small vessels may be pruned, persistence of a trunk vessel can be evaluated on OCTA by assessing the complete tridimensional data set through the en face angiographic cube and tracing the location of these large vessels. 29 Figure 3 shows an example of a 3 3 and6 6 mm OCTA image of a CNV lesion pre- and post- anti- VEGF treatment, with the 3 3 scan being the highest resolution scan currently commercially available for appreciation of microvascular changes. The lesion has reduced in size after anti-vegf treatment, and shows increased intercapillary spaces within the lesion and fewer smaller vessels at the periphery of the lesions. It has been also hypothesized that the choriocapillaris is compromised in eyes with namd, either directly as result of the disease itself or after prolonged treatment with anti-vegf agents. 3,31 36 The OCTA can be used to visualize the integrity of the choriocapillaris in those patients. The ability to image the choriocapillaris structure and flow impairment may be useful for detecting and monitoring the progression of namd, and for monitoring treatment responses to therapies that target disease progression. 37 Fig. 2. Select CNV morphologies as visualized on OCTA. The yellow arrows in (A) and (B) correspond to the trunk vessel of the lesion from which smaller vessels emanate. A. Medusa morphology, which shows vessels radiating in all directions from a central vessel. B. Seafan morphology, which shows a rake-like pattern of vessels radiating from one side of the lesion. C. Glomerular morphology, which shows globular structures of entwined vessels separated by hypodense areas. Image (A) was captured on Zeiss Cirrus HD-OCT with AngioPlex (Carl Zeiss Meditec, Inc, Dublin, CA), while (B) and (C) were imaged on an OptoVue Avanti RTVue XR with AngioVue software. Both devices are spectral-domain OCT devices that operate at a wavelength of 840 nm and 68,000 A-scans per second (Avanti) or 7000 A-scans per second (Zeiss).

4 S86 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2016 VOLUME 36 NUMBER 12 Fig. 3. The OCTA of a treatment-naive CNV at presentation and 4 weeks following intravitreal anti-vegf treatment. A. A3mm 3 mm OCTA of the CNV shows dense, numerous small capillaries with small vessels at the periphery of the lesion. The corresponding crosssectional OCT scan shows subretinal fluid; (B) 3mm 3mm OCTA 4 weeks after intravitreal anti-vegf treatment. There is loss of small vessels within the lesion and fewer small vessels at the periphery of the lesions. C. The 6 mm 6 mm OCTA pretreatment highlights the full extent of the irregularly shaped CNV. D. The 6 mm 6mm OCTA posttreatment highlights the reduced size of the CNV lesion. In the 6 mm 6 mm OCTA scans, a perilesional hypointense halo can be clearly seen surrounding both the preand posttreatment lesions. However, with the current available technology, the assessment of the choriocapillaris must be done with extreme care, since image artifacts or anatomical abnormalities such as shadowing from the CNV and associated fluid, blood, or fibrosis can give a false appearance of choriocapillaris loss where there is none. Figure 4 shows an example of the OCTA image segmented at the level of the choriocapillaris pre- and post anti-vegf treatment showing a circular area of shadowing from the overlying CNV and fluid, which can be described as a perilesional hypointense halo. The posttreatment image shows a slight reduction in shadowing, which may be

5 OCTA ENDPOINTS FOR NEOVASCULAR AMD COLE ET AL S87 Fig. 4. Measurements of GLD and CNV area On OCTA of a treatment-naive CNV at presentation and 4 weeks after intravitreal anti-vegf treatment. Images were performed on an OptoVue Avanti RTVue XR with AngioVue software. (A) and (B) are 6 mm 6mmOCTA images segmented at the level of the outer retina pre and posttreatment, respectively. The GLD measurements were made on ImageJ (National Institutes of Health, Bethesda, MD; available at html). The pretreatment GLD was mm and the posttreatment GLD was mm, as indicated by the white dotted lines. The CNV area was made by manually tracing the border of the CNV lesions and was mm 2 pretreatment and mm 2. D. The 6 mm 6mm OCTA segmented at the level of the choriocapillaris showing a circular area of shadowing due to the CNV and fluid (white arrows on corresponding crosssectional OCT scan). The arrows on the OCTA image highlight an area of shadowing that appears dark black on the pretreatment image and appears more gray in the posttreatment image (E). In the posttreatment image (E), this reduction in shadowing may be due to regression of fluid, reduction in CNV size, choriocapillaris reperfusion, or slightly different segmentation. F and G. are same day cross-sectional OCT scans showing areas of subretinal and intraretinal fluid. misinterpreted as reperfusion. There are significant challenges associated with accurate imaging and interpretation of the choriocapillaris especially on spectral-domain OCTA, mostly related to segmentation, signal penetration, wavelength of the OCTA device used, and the effect of artifacts. 38 These must be overcome before choriocapillaris analysis can be more widely used, and may be achieved with better software tools as well as commercialization of swept-source OCTA systems. Quantitative Features of CNV Quantitative features of CNV that can be evaluated on OCTA include measurement of CNV dimensions, automated flow analysis, and other automated algorithms for OCTA image analysis. The area and greatest linear diameter (GLD) of CNV can be quantified with digital tracing tools. Such measurements require experienced assessors, as reproducibility is challenging. Reductions in CNV area and GLD have been shown following anti-vegf treatment in various studies using different imaging modalities, including standard FA as well as OCTA. 1,39 41 Figure 4 shows an example of GLD measurements made with an automated digital tracing tool that showed a reduction in this parameter that was correlated with reduction in CNV dimension. Automated algorithms for OCTA image analysis have been validated in diabetic retinopathy, and may become more commonly used as OCTA use

6 S88 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2016 VOLUME 36 NUMBER 12 expands and quantitative analysis of these images is more widely performed. 42,43 There are also several automated options for analyzing areas of flow and nonflow on commercially available OCTA devices, which can include the assessment of peri- and interlesional areas of nonflow(figure5)orvessel density analysis (Figures 6 and 7). However, these algorithms have not been adequately validated and are dependent on automated segmentation that may fail in the presence of disrupted outer retina or image artifacts (Figure 5). For use in clinical trials, therefore, the segmentations may have to be manually adjusted to encompass the region of interest before applying the algorithms. Again, expert Fig. 5. Areas of vascular flow and nonflow highlighted with a pseudoautomated imaging analysis tool on the OptoVue Avanti RTVue XR with prototype AngioVue software. Flow area in a circular field of the 6 mm 6 mm OCTA with a fixed radius before (A) and after (B) intravitreal anti-vegf treatment. In the pretreatment image, the choroidal neovascularization can be clearly delineated; however, in the posttreatment image, the surrounding areas of projection artifact make it more difficult to clearly visualize the borders of the lesion. C. A perilesional area of nonflow in the pretreatment 6mm 6 mm image, as highlighted with a pseudoautomated imaging algorithm. D. Interlesion areas of reduced flow highlighted in the 3 mm 3mm posttreatment image.

7 OCTA ENDPOINTS FOR NEOVASCULAR AMD COLE ET AL S89 Fig. 6. Vessel density analysis before and after anti-vegf treatment. A 6 mm 6 mm OCTA images in both the superficial (A and B) and deep (C and D) plexus. This is an automated tool on the OptoVue AngioVue software. Overall, the vessel density is greater in the deep plexus because of the morphology of numerous small, interconnected vessels compared with the larger vessels of the superficial plexus. In the superficial plexus, the vessel density values are smaller in the posttreatment OCTA image. The deep plexus does not show the same trend. assessors need to be very cognizant of the variability that this can introduce in the data, as well as concerns regarding reproducibility. At present, therefore, the automated segmentation algorithms currently available are not optimal for use in clinical trials, since they may lead to inconclusive results. Limitations of OCTA OCTA has distinct advantages and disadvantages when compared to FA and indocyanine green, current standard imaging modalities, when being used to generate surrogate endpoints for clinical trials. 44

8 S90 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2016 VOLUME 36 NUMBER 12 Fig. 7. Specific microvascular features of CNV which have been previously described by Coscas et al. The CNV indicated by the yellow arrow in (A)is an example of a lacy-wheel shape with condensed, looping vessels. The neovascular membrane in (B) has large, mature vessels with increased intercapillary spaces between the abnormal vessels. The vessel termini in the CNV in (C) shows a peripheral arcade of fine vessels, which is highlighted with yellow arrows. The CNV in (D) has a branching network of tiny capillaries, and a portion of the membrane, indicated by the yellow arrow, is bordered by a perilesional hypointense area. The OCTA images are generated through a decorrelation signal, which is the difference in sequential OCT scans at the same retinal location. Because the OCT scans are rapidly acquired, the movement between the scans represents erythrocyte movement, assuming all other anatomical elements remain unchanged OCTA is a static image of the vasculature in which the microvascular details that may be obscured by leakage or pooling on dye-based standard angiography (FA or ICGA) can be clearly visualized. However, commercially available OCTA devices have a smaller field of view compared to standard angiography (either 3 mm 3mmor6 6 mm), so they cannot visualize CNV that extends beyond these regions unless the photographer can direct OCT scanning to the area of interest and capture several scans. Artifacts present in OCTA images can affect the accurate interpretation and visualization of CNV. 44,49,50 These artifacts include segmentation error, signal blockage, motion artifact, and projection artifact. The presence of subretinal or intraretinal fluid can affect automated segmentation algorithms and cause segmentation errors, which may affect quantitative and qualitative assessment of CNV. These segmentation errors can be identified by manual assessment and careful examination of the en face OCTA volumetric cube as well as the corresponding cross-sectional OCT image showing the segmentation lines, in a multidimensional fashion. 49 However, in some cases, altered retinal contour obscures visualization of the neovascular membrane on OCTA. On OCTA, areas of signal blockage are characterized by dark, hyporeflective regions that appear as discrete areas of the same shape on both OCTA angiogram and structural cross-sectional or en face OCT. These dark areas of signal blockage may be misinterpreted as flow impairment in the choriocapillaris or deeper segmentations; therefore, it is vital to analyze OCTA angiograms alongside their corresponding structural cross-sectional or en face OCT scans. Motion artifact can disrupt the continuity of the vessel images, making it difficult to visualize the microvascular details of the lesion. Finally, projection artifacts from vessels in the superficial and deep retinal plexus can occur in the choriocapillaris segmentation slab, particularly if it includes the retinal pigment epithelium, and can be mistaken for abnormal choroidal vasculature. 44,49 51 Conclusion OCTA is still a field in its infancy and surrogate endpoints on OCTA are currently being evaluated. With further validation, these proposed endpoints can be more widely adopted for use in clinical trials on namd. However, at present, these endpoints are still exploratory, and their significance in terms of disease prognosis and treatment response is still being evaluated. Longitudinal studies of CNV are warranted to determine the reliability, repeatability, and clinical significance of these findings. The current limitations of OCTA are mostly related to imaging artifacts, which can compromise the optical reconstruction and the performance of algorithms, as well as to the complexity of image interpretation. In this sense, evaluation by expert graders that can identify and

9 OCTA ENDPOINTS FOR NEOVASCULAR AMD COLE ET AL S91 adjust for image artifacts and inaccurate segmentation is key. Future studies should also determine the reproducibility of OCTA findings across different devices and algorithms. There are currently several spectraldomain OCTA devices available, and further expansion of commercial swept-source OCTA devices is expected in the near future. Variations in the hardware and software of OCTA, including automated segmentation algorithms, can affect the appearance of CNV and therefore both its quantitative and qualitative evaluation. These parameters will need to be further validated before they can be more widely and consistently used as standardized endpoints. In conclusion, further studies are needed to evaluate qualitative and quantitative features of CNV on OCTA, and to determine which characteristics should be considered clinically meaningful for the diagnosis, treatment, and monitoring of namd. Biomarkers for lesion activity may include lesion shape, branching structure, anastomoses, features of the vessels at the periphery of the lesion, and the appearance of the choriocapillaris surrounding the lesion. Future studies should investigate the features of CNV morphology that are associated with lesion activity, to establish new treatment paradigms for namd, so these features can be used to identify highrisk lesions and treat them at an earlier stage before the development of irreversible vision loss. Key words: optical coherence tomography, optical coherence tomography angiography, neovascular agerelated macular degeneration. References 1. Muakkassa NW, Chin AT, de Carlo T, et al. Characterizing the effect of anti-vascular endothelial growth factor therapy on treatment-naive choroidal neovascularization using optical coherence tomography angiography. Retina 2015;35: Novais EA, Adhi M, Moult EM, et al. Choroidal neovascularization analyzed on ultra-high speed swept source optical coherence tomography angiography compared to spectral domain optical coherence tomography angiography. Am J Ophthalmol 2016;164: Moult E, Choi W, Waheed NK, et al. Ultrahigh-speed sweptsource OCT angiography in exudative AMD. Ophthalmic Surg Lasers Imaging Retina 2014;45: Coscas G, Lupidi M, Coscas F, et al. Optical coherence tomography angiography during follow-up: qualitative and quantitative analysis of mixed type I and II choroidal neovascularization after vascular endothelial growth factor trap therapy. Ophthalmic Res 2015;54: Mehta S. Age-related macular degeneration. Prim Care 2015; 42: Gobburu JV. Biomarkers in clinical drug development. Clin Pharmacol Ther 2009;86: Lassere MN, Johnson KR, Boers M, et al. Definitions and validation criteria for biomarkers and surrogate endpoints: development and testing of a quantitative hierarchical levels of evidence schema. J Rheumatol 2007;34: Prentice RL. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med 1989;8: Medeiros FA. Biomarkers and surrogate endpoints in glaucoma clinical trials. Br J Ophthalmol 2015;99: Fung AE, Lalwani GA, Rosenfeld PJ, et al. An optical coherence tomography-guided, variable dosing regimen with intravitreal ranibizumab (Lucentis) for neovascular age-related macular degeneration. Am J Ophthalmol 2007;143: Holz FG, Amoaku W, Donate J, et al. Safety and efficacy of a flexible dosing regimen of ranibizumab in neovascular agerelated macular degeneration: the SUSTAIN study. Ophthalmology 2011;118: Lalwani GA, Rosenfeld PJ, Fung AE, et al. A variable-dosing regimen with intravitreal ranibizumab for neovascular agerelated macular degeneration: year 2 of the PrONTO Study. Am J Ophthalmol 2009;148:43 58.e Comparison of Age-related Macular Degeneration Treatments Trials Research Group, Martin DF, Maguire MG, et al. Ranibizumab and bevacizumab for treatment of neovascular agerelated macular degeneration: two-year results. Ophthalmology 2012;119: Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology 2012;119: Chakravarthy U, Harding SP, Rogers CA, et al. Alternative treatments to inhibit VEGF in age-related choroidal neovascularisation: 2-year findings of the IVAN randomised controlled trial. Lancet 2013;382: Miller JW, Le Couter J, Strauss EC, Ferrara N. Vascular endothelial growth factor a in intraocular vascular disease. Ophthalmology 2013;120: Cruess AF, Berger A, Colleaux K, et al. Canadian expert consensus: optimal treatment of neovascular age-related macular degeneration. Can J Ophthalmol 2012;47: Veritti D, Sarao V, Lanzetta P. Neovascular age-related macular degeneration. Ophthalmologica 2012;227(suppl 1): Dhoot DS, Kaiser PK. Ranibizumab for age-related macular degeneration. Expert Opin Biol Ther 2012;12: Mitchell P. A systematic review of the efficacy and safety outcomes of anti-vegf agents used for treating neovascular age-related macular degeneration: comparison of ranibizumab and bevacizumab. Curr Med Res Opin 2011;27: Brown DM, Michels M, Kaiser PK, et al. Ranibizumab versus verteporfin photodynamic therapy for neovascular age-related macular degeneration: two-year results of the ANCHOR study. Ophthalmology 2009;116:57 65.e Brown DM, Kaiser PK, Michels M, et al. Ranibizumab versus verteporfin for neovascular age-related macular degeneration. N Engl J Med 2006;355: Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 2006;355: Jung JJ, Chen CY, Mrejen S, et al. The incidence of neovascular subtypes in newly diagnosed neovascular age-related macular degeneration. Am J Ophthalmol 2014;158: e El Ameen A, Cohen SY, Semoun O, et al. Type 2 neovascularization secondary to age-related macular degeneration imaged by optical coherence tomography angiography. Retina 2015; 35: Kuehlewein L, Bansal M, Lenis TL, et al. Optical coherence tomography angiography of type 1 neovascularization in age-

10 S92 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2016 VOLUME 36 NUMBER 12 related macular degeneration. Am J Ophthalmol 2015;160: e Sadda S. OCTA Consensus group on neovascular AMD. Macula Society; Coscas GJ, Lupidi M, Coscas F, et al. Optical coherence tomography angiography versus traditional multimodal imaging in assessing the activity of exudative age-related macular degeneration: a new diagnostic challenge. Retina 2015;35: Spaide RF. Optical coherence tomography angiography signs of vascular abnormalization with antiangiogenic therapy for choroidal neovascularization. Am J Ophthalmol 2015;160: Huang D, Jia Y, Rispoli M, et al. Optical coherence tomography angiography of time course of choroidal neovascularization in response to anti-angiogenic treatment. Retina 2015;35: Bhutto I, Lutty G. Understanding age-related macular degeneration (AMD): relationships between the photoreceptor/retinal pigment epithelium/bruch s membrane/choriocapillaris complex. Mol Aspects Med 2012;33: McLeod DS, Grebe R, Bhutto I, et al. Relationship between RPE and choriocapillaris in age-related macular degeneration. Invest Ophthalmol Vis Sci 2009;50: Lutty G, Grunwald J, Majji AB, et al. Changes in choriocapillaris and retinal pigment epithelium in age-related macular degeneration. Mol Vis 1999;5: Curcio CA, Messinger JD, Sloan KR, et al. Subretinal drusenoid deposits in non-neovascular age-related macular degeneration: morphology, prevalence, topography, and biogenesis model. Retina 2013;33: Jia Y, Bailey ST, Wilson DJ, et al. Quantitative optical coherence tomography angiography of choroidal neovascularization in age-related macular degeneration. Ophthalmology 2014;35: Choi W, Moult EM, Waheed NK, et al. Ultrahigh-Speed, swept-source optical coherence tomography angiography in Nonexudative age-related macular degeneration with geographic atrophy. Ophthalmology 2015;122: Waheed NK, Moult EM, Fujimoto JG, Rosenfeld PJ. Optical coherence tomography angiography of dry age-related macular degeneration. Dev Ophthalmol 2016;56: Choi W, Mohler KJ, Potsaid B, et al. Choriocapillaris and choroidal microvasculature imaging with ultrahigh speed OCT angiography. PLoS One 2013;8:e Kuehlewein L, Sadda SR, Sarraf D. OCT angiography and sequential quantitative analysis of type 2 neovascularization after ranibizumab therapy. Eye (Lond) 2015;29: Jia Y, Bailey ST, Hwang TS, et al. Quantitative optical coherence tomography angiography of vascular abnormalities in the living human eye. Proc Natl Acad Sci U S A 2015;112:E2395 E Lumbroso B, Rispoli M, Savastano MC. Longitudinal optical coherence tomography-angiography study of type 2 naive choroidal neovascularization early response after treatment. Retina 2015;35: Hwang TS, Gao SS, Liu L, et al. Automated quantification of capillary nonperfusion using optical coherence tomography angiography in diabetic retinopathy. JAMA Ophthalmol 2016: Agemy SA, Scripsema NK, Shah CM, et al. Retinal vascular perfusion density mapping using optical coherence tomography angiography in normals and diabetic retinopathy patients. Retina 2015;35: Cunningham ET, Jr., Ferrara D, Mrejen S, et al. Imaging the choroid and choroidal neovascularization in eyes with Inflammation. Ocul Immunol Inflamm 2016;24: Zhi Z, Chao JR, Wietecha T, et al. Noninvasive imaging of retinal morphology and microvasculature in obese mice using optical coherence tomography and optical microangiography. Invest Ophthalmol Vis Sci 2014;55: Matsunaga D, Yi J, Puliafito CA, Kashani AH. OCT angiography in healthy human subjects. Ophthalmic Surg Lasers Imaging Retina 2014;45: Jia Y, Wei E, Wang X, et al. Optical coherence tomography angiography of optic disc perfusion in glaucoma. Ophthalmology 2014;121: Nagiel A, Sadda SR, Sarraf D. A Promising future for optical coherence tomography angiography. JAMA Ophthalmol 2015; 133: Ferrara D. Image artifacts in optical coherence tomography angiography. Clin Exp Ophthalmol 2016;44: Ferrara D, Waheed NK, Duker JS. Investigating the choriocapillaris and choroidal vasculature with new optical coherence tomography technologies. Prog Retinal Eye Res 2016; 52: Spaide RF, Fujimoto JG, Waheed NK. Image artifacts in optical coherence tomography angiography. Retina 2015; 35: