IN VIVO CHARACTERIZATION OF RETINAL VASCULARIZATION MORPHOLOGY USING OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY MARIA CRISTINA SAVASTANO, MD, PHD, BRUNO LUMBROSO, MD, MARCO RISPOLI, MD Purpose: To evaluate retinal vessel morphology using split-spectrum amplitudedecorrelation angiography with optical coherence tomography in healthy eyes. Methods: Fifty-two eyes of 26 healthy volunteers (age range from 35 to 48 years; mean age 41.94 years; SD: ±4.13) were evaluated by optical coherence tomography angiography in the macular region. The protocol acquisition consisted of a 216 · 216 A-scan that was repeated 5 times in the same position, in 3 · 3 mm centered into the fovea. Results: All 52 eyes showed 2 separate vascular networks in the inner retina: the superficial network, located in the nerve fiber layer and in the ganglion cell layer, and the deep network, detected in the outer plexiform layer. The superficial and deep networks showed interconnections of vertical vessels. The reference planes to observe the 2 networks were defined at 60 mm, with an inner limiting membrane reference (6 mm offset), and 30 mm, with an inner plexiform layer reference (60 mm offset), respectively. Conclusion: Optical coherence tomography angiography can separately detect the superficial vascular and the deep vascular networks. These networks are overlaid and seem to be fused when seen with standard angiographies. Furthermore, optical coherence tomography angiography technology allows for the visualization of abnormal blood column and vessel wall details. RETINA 35:2196–2203, 2015
T
nous dye injection can cause some side effects.3 Furthermore, in a diseased eye, the fluorescein leaks radially through the damaged fenestrations of the vessels, not allowing an evaluation of anatomical details of the vascular morphology. The main problem with fluorescein angiography is that it cannot dissociate the two (superficial and deep) retinal networks. With the introduction of optical coherence tomography (OCT) in 1991, clinical practice saw remarkable development.4 The high-resolution spectral domain OCT provides information comparable with a histological examination. Despite the rapid evolution of imaging, even the most modern OCT devices have not been able to provide an adequate microvasculature visualization of chorioretinal diseases. This limitation frequently requires the diagnosis to be completed with fluorescein angiography or indocyanine green angiography to investigate retinovascular disorders, such as diabetic
he present body of knowledge on retinal vascularization morphology is derived from the works of many investigators, dating from 1930 to 1960. These studies show that there are two (superficial and deep) retinal networks located in the inner retina. After the 1960s, more interest was shown in fluorescein angiography. Since the pioneer studies by Novotny and Alvis more than 50 years ago, fluorescein angiography of the retina has been considered to be the best imaging modality to assess and study the retinal vascular network.1,2 Although fluorescein angiography is able to
detect significant microvascular details, the intraveFrom the Italian Macula Center, Rome, Italy. None of the authors have any financial/conflicting interests to disclose. M. C. Savastano and B. Lumbroso contributed equally to this study. Reprint requests: Maria Cristina Savastano, MD, PhD, via Angelo Brofferio, 7 00195 Rome- Italy; e-mail: [email protected]
2196
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
retinopathy, age-related macular degeneration, and vascular occlusions. Innovative and promising OCT angiography is able to assess retinal vessels without intravenous dye injection. The first report of OCT angiography was made by Makita et al.5 They used Doppler shifts of the bulk motion for axial displacement between adjacent A-lines and fast correlation-based algorithm for axial shifts between neighboring images. Fingler et al6 and Kim et al7 described the retinal microvasculature for the first time using phasevariance OCT. In this new approach, the identification of the vessels is based on areas of motion between consecutive B-scans that are compared with static areas. In the retina, the regions showing motion correspond to the vasculature because the vessels are distinct from other retinal tissues that are static. The variance of motion-corrected phase captured with multiple B-scans at the same position, and repeated over the entire volumetric scan, generated a threedimensional phase-variance OCT image of the retinal vasculature. Recently, a new analytical algorithm, split-spectrum amplitude-decorrelation angiography with optical coherence tomography, was described by Jia et al.8 Split-spectrum amplitude-decorrelation angiography with optical coherence tomography is able to improve the signal-to-noise ratio of flow detection using the spectral bands separately and then averaged, showing the microvascular network and allowing the automated removal of motion errors. The images observed with the split-spectrum amplitude-decorrelation angiography with optical coherence tomography represent the blood flow. The classical histological and anatomical studies of retinal vessels report the clear distinction between the superficial and deep vascularization. These important studies were performed before fluorescein angiography allowed a fast and unique in vivo study. All classical histology publications highlight that there are two parallel vascular networks at the inner retinal level.9 The superficial network consists of about 75 mm caliber vessels in a network, lying in the nerve fiber layer. The deep network is represented by a dense and complex system of smaller vessels (about 20 mm), lying in the outer plexiform layer. The superficial network is interconnected to the deep network through small vessels in vertical course.10–13 Druault10 and Redslob11 report that the retinal capillary system is formed by two superposed vascular networks that are in communication with each other. The first has a wide mesh (75 mm) and is located in the nerve fiber layer and in the ganglion cell layer. It originates directly from arterioles and, itself, is the source of venule orig-
2197
ination. The second, the outer network, has a small mesh (20 mm) and is located in the ganglion layer. Its farthest loops reach the outer plexiform layer. The outer network originates from the first net or is connected to it by vertical vessels (arterioles). In the same year, Redslob12 reported that retinal vessels, arteries, and veins have a sinuous course inside the nerve fiber layer under the inner limiting membrane. Thin nerve fibers innervate the vessel walls. A first vascular network is located in the inner plexiform layer and a second in the outer plexiform, at the edge of the outer nuclear layer. The two networks are connected by vertical arterioles. The inner vascular net is composed of a wide mesh. This mesh is much tighter in the outer network; the two networks merge in venules that follow the course of the arterioles. Anatomical studies by Duke-Elder13 described the retinal artery emerging from the main branches of the central artery in the superficial parts of the nerve fiber layer, interdigitating in their arrangement with the corresponding veins. From these vessels, two capillary plexuses are given off: 1) the superficial (inner) capillary network, lying in the more superficial parts of the nerve fiber layer, arranged in a two-dimensional pattern and 2) the deep (outer) capillary network, which forms a denser and more complex pattern than the other, lying in the boundary plane between the inner nuclear layer and the outer plexiform layer. Retinal vessel assessment by fluorescein angiography delivers vascular anatomy information, merging the retinal vascular structures, the superficial and deep retinal networks. Fluorescein angiography cannot dissociate the different networks that form the complex retinal vascularization. Because fluorescein imaging cannot dissociate the two parallel vascular networks at an inner retina level, most authors ceased stressing the existence of this important anatomical feature after the 1960s. The aim of our study was to evaluate the retinal vascular anatomy by splitspectrum amplitude-decorrelation angiography with optical coherence tomography (OCT angiography) in healthy eyes by analyzing the superficial network and deep network morphologic features.
Methods This study adhered to the tenets of the Declaration of Helsinki, and written informed consent was obtained from all participants after a detailed description of our study and its goals. A total of 52 eyes from 26 healthy volunteers (from 35 to 48 years; mean age: 41.94 years; SD: ±4.13) were included in the study. Inclusion criteria were no other eye pathologies (i.e.,
2198 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
Fig. 1. B-scan of healthy eye showing the superficial vascular network distribution (120 mm caliper) and deep vascular network arrangement (60 mm caliper).
uveitis, glaucoma, and so forth), emmetropic eyes (spherical equivalent ±1.00 diopter [D]), and no diopter opacities, to obtain good image quality. The exclusion criterion was inadequate cooperation to obtain satisfactory images. All recruited eyes underwent ocular fundus imaging evaluation (spectral domain OCT RTVue XR AVANTI, version 2014.2.0.13; Optovue Inc, Fremont, CA), producing 5 consecutive B-scan acquisitions of 3 · 3 mm. The OCT angiography protocol acquisition consists of a 216 · 216 a-scan
Fig. 2. The superficial network was defined at a 60 mm thickness with an inner limiting membrane reference (6 mm offset), and the deep network has been identified at 30 mm with an inner plexiform layer reference (60 mm offset). Scan projections on white paper were used to highlight major vessels and secondary vessels and to draw them, simplifying the network to its main feature.
2015 VOLUME 35 NUMBER 11
repeated 5 times on the same position. The acquisition is divided into 2 steps: the first consists of 2 horizontal acquisitions (the operator may choose the best one), followed by 2 vertical acquisitions. The aim of the double acquisition is to reduce the motion artifacts. The grid area may be customized between 3 · 3 mm and 6 · 6 mm, although the pixel number in the analysis area is the same. The immediate evidence is the lack of resolution in the larger grid, whereas the 3 · 3 and 4 · 4 grids have more pixels · mm. Results According to the anatomical features, OCT angiography was able to detect the superficial network in the ganglion cell layer and the deep network in the outer plexiform layer in all assessed eyes. The anastomotic vessels run from the superficial to the deeper network. To define the reference plane for the different networks visualization by OCT angiography, selected parameters were chosen. The requirement to establish
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
references was to compare different operator analyses. The considered parameters were outline (inner limiting membrane, inner plexiform layer, retinal pigment epithelium, retinal pigment epithelium reference), thickness (layer analysis selected), and offset (distance from the outline). Figure 1 shows the B-scan OCT visualization of the superficial network with a mean representative caliper measurement (white arrowhead) of 120 mm. The evaluation of this network by OCT angiography was selected at a 60 mm thickness from ILM, to define all superficial vessels (Figure 2). The deep network distribution area, lying in the outer plexiform layer, was detected by the caliper (black arrowhead). The reference was set at a 30 mm thickness scan to obtain
2199
the entire visualization of the deep network (Figures 3 and 4). As reported by Hogan et al,9 the deep network has small capillaries with a mean diameter of 20 mm. The references outline are set at 30 mm to incorporate all the capillaries of the deep network as they described. Fluorescein angiography cannot differentiate between the superficial and deep network, merging vascular anatomy information of all retinal vascular structures (Figure 5). Superficial Network Features In healthy eyes, the superficial network showed a defined silhouette morphology with a linear and continuous white shape against a black background. The wall thickness was homogenous along the entire
Fig. 3. Four different samples of the superficial network OCT angiography representation and the respective references on B-scan. In all eyes, OCT angiography shows the superficial network harmonic distribution around the foveal avascular zone with a spider web feature. B-scans show depth and shape of the layer under study.
2200 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 11
Fig. 4. Same eyes described in Figure 3, with scans taken at deep network level. Optical coherence tomography angiography shows small complex fans with multiple interconnections with irregular features around the foveal avascular zone. B-scans show depth and shape of the layer under study.
scan; the vessel course was regularly distributed, originating from the arched branches, superior and inferior, around the foveal avascular zone. The vessel flow representation (decorrelation) was homogenous for the visualized scan. The superficial network may be compared with a spider web. Around the foveal avascular zone, the vessel design shows continuous regularly spaced macular web branches, centered on the fovea, and homogeneously distributed (Figure 6). Deep Network Features The deep network in healthy eyes had a regular distribution around the foveal avascular zone, with more complex multiple tiny radial and horizontal
interconnections. The vessel flow was dense throughout the entire visualized scan, and the vasculature texture was homogeneous. The vascular network showed multiple vertical anastomoses connecting the superficial to the deeper network. From the inferior end of each vertical anastomosis, originate horizontal fan-like capillaries that interconnect to form a continuous web, parallel to the superficial network. Figure 7 shows a stack of parallel images captured in sequence through the superficial and deep network. In all the images, the persistence of the flow between the superficial and deep network can be followed showing anatomical capillary interconnections. Moreover, it is possible to follow the capillary flow from the superficial network to the deep network.
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
2201
Fig. 5. Fluorescein angiography and OCT angiography superficial and deep network representations of the same healthy eye. The fluorescein angiography shows the overlapping information of the superficial and deep networks. The OCT angiography revealed the possibility to individually view the two networks.
Discussion Optical coherence tomography has become an essential tool in the management of retinal diseases. It shows changes in retinal anatomy, although fluorescein angiography and indocyanine green angiography, requiring intravenous dye injection, are still needed in retinal and choroidal vasculature analysis in the initial diagnosis of the disease. Several new noninvasive imaging methods are emerging in retinal and choroidal vessel visualization, such as scanning laser Doppler flowmetry,14 laser speckle photography,15 and laser speckle flowgraphy.16 However, they are not able to show the blood flow in the normal vessels of pathologic eyes. Optical coherence tomography angiography is a promising method to
Fig. 6. Representative picture of retinal vascular superficial and deep network interconnected by vertical anastomosis.
distinguish the retino/choridal vessel anatomy. The pioneering studies by Miura et al17 and Hong et al18 reported a new tool with the possibility of choroidal neovascularization imaging with some limitations in distinguishing it from the choroidal normal vasculature. In their study, Miura et al, in fact, described Doppler optical coherence angiography with the capability to detect only some parts of the choroidal or retinal vasculature, announcing more sensitive technique in development. Hong et al described a new tool, but the vast amount of information given could be difficult to use by clinicians. Hong et al had necessity to use a sophisticated data browser. This device is custom made and not available to the clinicians. The beautiful images obtained cannot be easily made in everyday clinic. Recently, Jia et al19 reported a new method to visualize well-defined choroidal neovascularization toward the retino/choroidal normal vessel by OCT angiography. Optical coherence tomography angiography is a new analytical method that allows high-resolution automated imaging of the retinal microvasculature without intravenous dye injection. This new technology is not invasive in contrast to fluorescein angiography, the gold standard method of vasculature imaging. Optical coherence tomography angiography is able to detect the intraluminal flow, independent of time and dye injection. In OCT angiography, the brightness signals the flow velocity. As reported by Leitgeb et al20 in OCT angiography, bright signals mean presence of blood flow. Fluorescein angiography does not allow visualization of vessel morphology; it only indirectly indicates a vascular permeability defect. Optical coherence tomography angiography intricately shows the silhouette of vessels in pathologic eyes, with none of the leakage, pooling, or staining representative of fluorescein angiography. Several anatomical studies have reported the retinal vessel distribution, organized in 2 distinct vessel networks: 1) the superficial network, ophthalmoscopically
2202 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 11
Fig. 7. A stack of OCT angiography parallel multiple images were captured in sequence from the superficial to the deep network (from left to right). The red circle highlights a capillary region detail and the yellow arrow show the capillary flow that moves from the superficial to the deep network.
observable with large and medium diameter vessels in correspondence to the retinal nerve fiber layer and 2) the deep network, small size capillary complex located in the outer plexiform layer.13,21,22 Optical coherence tomography angiography is able to separately detect the superficial vascular network in the ganglion cell layer and the deep vascular network in the outer plexiform layer. These two networks have different aspects, not seen with standard angiographies. In fluorescein angiography, the superficial and inner vascular networks overlap, and a selective evaluation is not possible. In healthy eyes, the superficial network consists of larger vessels than the deeperderived vessels from the superior and inferior arcades that terminate in the foveal avascular zone. The deep network shows a smaller interconnected net of dense vessels. In agreement with the description recently by Spaide et al,23 the importance of OCT angiography is mainly due to the capability to visualize all retinal vasculature layers without dye injection and to highlight at least two vascular plexus. One of the most important limitations of OCT angiography is the small area of the scan, involving only the macular region. Most likely, a full-field area of analysis providing more details could produce more information in the future. At the moment, commercially available systems does not show structures below retinal pigment epithelium. In conclusion, OCT angiography makes vessel analysis an option in everyday clinical practice without dye injection. The future possibilities of retinal details assessed by OCT angiography are unique from fluorescein angiography data, as vessel anatomy is not masked by leakage, staining, or pooling. The images of healthy and pathologic eyes may be compared, and the lack of leakage allows for better visualization of capillary abnormalities, preretinal new vessels, and CNV net (Type I or Type II). Furthermore, the lack of wall staining allows for the visualization of an abnormal blood column and the assessment of vessel wall details. Capillaries in
the macula, deep capillaries, and intraretinal potential shunts can only be finely assessed with OCT angiography. In the future, we plan to compare healthy eyes, visualized with OCT angiography imaging, with agerelated maculopathy, macular ischemia, and other diseases, to assess capillary anomalies and help delineate correct diagnoses. Key words: deep vasculature network, OCT angiography, retinal vascularization, spilt-spectrum amplitude-decorrelation angiography with optical coherence tomography, superficial vasculature network. References 1. Alvis D. Happy 50th birthday [letter]. Ophthalmology 2009; 116:2259. 2. Marmor MF, Ravin JG. Fluorescein angiography: insight and serendipity a half century ago. Arch Ophthalmol 2011;129: 943–948. 3. Lipson BK, Yannuzzi LA. Complications of intravenous fluorescein injections. Int Ophthalmol Clin 1989;29:200–205. 4. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science 1991;254:1178–1181. 5. Makita S, Hong Y, Yamanari M, et al. Optical coherence angiography. Opt Express 2006;14:7821–7840. 6. Fingler J, Schwartz D, Yang C, Fraser SE. Mobility and transverse flow visualization using phase variance contrast with spectral domain optical coherence tomography. Opt Express 2007;15:12636–12653. 7. Kim DY, Fingler J, Werner JS, et al. In vivo volumetric imaging of human retinal circulation with phase-variance optical coherence tomography. Biomed Opt Express 2011;2:1504– 1513. 8. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude decorrelation angiography with optical coherence tomography. Opt Express 2012;20:4710–4725. 9. Hogan M, Alvarado J, Weddell JE. Histology of the Human Eye—An Atlas and Textbook. Philadelphia, PA: WB Saunders; 1971. 10. Druault A. Appareil de la Vision. Traité d’Anatomie Humaine. Poirier et Charpy, 1911;1:1018. 11. Redslob E. Anatomie du Globe Oculaire. Traité d’Ophtalmologie. Paris, France: Masson, édit, 1939;5:382. 12. Redslob E. Traite d’Ophtalmologie: Société Française d’Ophtalmologie. Masson et CIE, 1939;1:465–472.
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL 13. Duke-Elder S. The Anatomy of Visual System. London, United Kingdom. 1961;2:372–376. 14. Michelson G, Schmauss B, Langhans MJ, et al. Principle, validity, and reliability of scanning laser Doppler flowmetry. J Glaucoma 1996;5:99–105. 15. Briers JD, Fercher AF. Retinal blood-flow visualization by means of laser speckle photography. Invest Ophthalmol Vis Sci 1982;22:255–259. 16. Tamaki Y, Araie M, Kawamoto E, et al. Noncontact, two dimensional measurement of retinal microcirculation using laser speckle phenomenon. Invest Ophthalmol Vis Sci 1994; 35:3825–3834. 17. Miura M, Makita S, Iwasaki T, Yasuno Y. Three-dimensional visualization of ocular vascular pathology by optical coherence angiography in vivo. Invest Ophthalmol Vis Sci 2011;52: 2689–2695. 18. Hong YJ, Miura M, Makita S, et al. Noninvasive investigation of deep vascular pathologies of exudative macular diseases by
19.
20.
21.
22.
23.
2203
high-penetration optical coherence angiography. Invest Ophthalmol Vis Sci 2013;54:3621–3631. Jia Y, Bailey ST, Wilson DJ, et al. Quantitative optical coherence tomography angiography of choroidal neovascularization in age-related macular degeneration. Ophthalmology 2014; 121:1435–1444. Leitgeb RA, Werkmeister RM, Blatter C, Schmetterer L. Doppler optical coherence tomography. Prog Retin Eye Res 2014;41:26–43. Paques M, Tadayoni R, Sercombe R, et al. Structural and hemodynamic analysis of the mouse retinal microcirculation. Invest Ophthalmol Vis Sci 2003;44:4960–4967. Tick S, Rossant F, Ghorbel I, et al. Foveal shape and structure in a normal population. Invest Ophthalmol Vis Sci 2011;52: 5105–5110. Spaide RF, Klancnik JM Jr., Cooney MJ. Retinal vascular layers imaged by fluorescein angiography and optical coherence tomography angiography. JAMA Ophthalmol 2015;133:45–50.
T
nous dye injection can cause some side effects.3 Furthermore, in a diseased eye, the fluorescein leaks radially through the damaged fenestrations of the vessels, not allowing an evaluation of anatomical details of the vascular morphology. The main problem with fluorescein angiography is that it cannot dissociate the two (superficial and deep) retinal networks. With the introduction of optical coherence tomography (OCT) in 1991, clinical practice saw remarkable development.4 The high-resolution spectral domain OCT provides information comparable with a histological examination. Despite the rapid evolution of imaging, even the most modern OCT devices have not been able to provide an adequate microvasculature visualization of chorioretinal diseases. This limitation frequently requires the diagnosis to be completed with fluorescein angiography or indocyanine green angiography to investigate retinovascular disorders, such as diabetic
he present body of knowledge on retinal vascularization morphology is derived from the works of many investigators, dating from 1930 to 1960. These studies show that there are two (superficial and deep) retinal networks located in the inner retina. After the 1960s, more interest was shown in fluorescein angiography. Since the pioneer studies by Novotny and Alvis more than 50 years ago, fluorescein angiography of the retina has been considered to be the best imaging modality to assess and study the retinal vascular network.1,2 Although fluorescein angiography is able to
detect significant microvascular details, the intraveFrom the Italian Macula Center, Rome, Italy. None of the authors have any financial/conflicting interests to disclose. M. C. Savastano and B. Lumbroso contributed equally to this study. Reprint requests: Maria Cristina Savastano, MD, PhD, via Angelo Brofferio, 7 00195 Rome- Italy; e-mail: [email protected]
2196
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
retinopathy, age-related macular degeneration, and vascular occlusions. Innovative and promising OCT angiography is able to assess retinal vessels without intravenous dye injection. The first report of OCT angiography was made by Makita et al.5 They used Doppler shifts of the bulk motion for axial displacement between adjacent A-lines and fast correlation-based algorithm for axial shifts between neighboring images. Fingler et al6 and Kim et al7 described the retinal microvasculature for the first time using phasevariance OCT. In this new approach, the identification of the vessels is based on areas of motion between consecutive B-scans that are compared with static areas. In the retina, the regions showing motion correspond to the vasculature because the vessels are distinct from other retinal tissues that are static. The variance of motion-corrected phase captured with multiple B-scans at the same position, and repeated over the entire volumetric scan, generated a threedimensional phase-variance OCT image of the retinal vasculature. Recently, a new analytical algorithm, split-spectrum amplitude-decorrelation angiography with optical coherence tomography, was described by Jia et al.8 Split-spectrum amplitude-decorrelation angiography with optical coherence tomography is able to improve the signal-to-noise ratio of flow detection using the spectral bands separately and then averaged, showing the microvascular network and allowing the automated removal of motion errors. The images observed with the split-spectrum amplitude-decorrelation angiography with optical coherence tomography represent the blood flow. The classical histological and anatomical studies of retinal vessels report the clear distinction between the superficial and deep vascularization. These important studies were performed before fluorescein angiography allowed a fast and unique in vivo study. All classical histology publications highlight that there are two parallel vascular networks at the inner retinal level.9 The superficial network consists of about 75 mm caliber vessels in a network, lying in the nerve fiber layer. The deep network is represented by a dense and complex system of smaller vessels (about 20 mm), lying in the outer plexiform layer. The superficial network is interconnected to the deep network through small vessels in vertical course.10–13 Druault10 and Redslob11 report that the retinal capillary system is formed by two superposed vascular networks that are in communication with each other. The first has a wide mesh (75 mm) and is located in the nerve fiber layer and in the ganglion cell layer. It originates directly from arterioles and, itself, is the source of venule orig-
2197
ination. The second, the outer network, has a small mesh (20 mm) and is located in the ganglion layer. Its farthest loops reach the outer plexiform layer. The outer network originates from the first net or is connected to it by vertical vessels (arterioles). In the same year, Redslob12 reported that retinal vessels, arteries, and veins have a sinuous course inside the nerve fiber layer under the inner limiting membrane. Thin nerve fibers innervate the vessel walls. A first vascular network is located in the inner plexiform layer and a second in the outer plexiform, at the edge of the outer nuclear layer. The two networks are connected by vertical arterioles. The inner vascular net is composed of a wide mesh. This mesh is much tighter in the outer network; the two networks merge in venules that follow the course of the arterioles. Anatomical studies by Duke-Elder13 described the retinal artery emerging from the main branches of the central artery in the superficial parts of the nerve fiber layer, interdigitating in their arrangement with the corresponding veins. From these vessels, two capillary plexuses are given off: 1) the superficial (inner) capillary network, lying in the more superficial parts of the nerve fiber layer, arranged in a two-dimensional pattern and 2) the deep (outer) capillary network, which forms a denser and more complex pattern than the other, lying in the boundary plane between the inner nuclear layer and the outer plexiform layer. Retinal vessel assessment by fluorescein angiography delivers vascular anatomy information, merging the retinal vascular structures, the superficial and deep retinal networks. Fluorescein angiography cannot dissociate the different networks that form the complex retinal vascularization. Because fluorescein imaging cannot dissociate the two parallel vascular networks at an inner retina level, most authors ceased stressing the existence of this important anatomical feature after the 1960s. The aim of our study was to evaluate the retinal vascular anatomy by splitspectrum amplitude-decorrelation angiography with optical coherence tomography (OCT angiography) in healthy eyes by analyzing the superficial network and deep network morphologic features.
Methods This study adhered to the tenets of the Declaration of Helsinki, and written informed consent was obtained from all participants after a detailed description of our study and its goals. A total of 52 eyes from 26 healthy volunteers (from 35 to 48 years; mean age: 41.94 years; SD: ±4.13) were included in the study. Inclusion criteria were no other eye pathologies (i.e.,
2198 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
Fig. 1. B-scan of healthy eye showing the superficial vascular network distribution (120 mm caliper) and deep vascular network arrangement (60 mm caliper).
uveitis, glaucoma, and so forth), emmetropic eyes (spherical equivalent ±1.00 diopter [D]), and no diopter opacities, to obtain good image quality. The exclusion criterion was inadequate cooperation to obtain satisfactory images. All recruited eyes underwent ocular fundus imaging evaluation (spectral domain OCT RTVue XR AVANTI, version 2014.2.0.13; Optovue Inc, Fremont, CA), producing 5 consecutive B-scan acquisitions of 3 · 3 mm. The OCT angiography protocol acquisition consists of a 216 · 216 a-scan
Fig. 2. The superficial network was defined at a 60 mm thickness with an inner limiting membrane reference (6 mm offset), and the deep network has been identified at 30 mm with an inner plexiform layer reference (60 mm offset). Scan projections on white paper were used to highlight major vessels and secondary vessels and to draw them, simplifying the network to its main feature.
2015 VOLUME 35 NUMBER 11
repeated 5 times on the same position. The acquisition is divided into 2 steps: the first consists of 2 horizontal acquisitions (the operator may choose the best one), followed by 2 vertical acquisitions. The aim of the double acquisition is to reduce the motion artifacts. The grid area may be customized between 3 · 3 mm and 6 · 6 mm, although the pixel number in the analysis area is the same. The immediate evidence is the lack of resolution in the larger grid, whereas the 3 · 3 and 4 · 4 grids have more pixels · mm. Results According to the anatomical features, OCT angiography was able to detect the superficial network in the ganglion cell layer and the deep network in the outer plexiform layer in all assessed eyes. The anastomotic vessels run from the superficial to the deeper network. To define the reference plane for the different networks visualization by OCT angiography, selected parameters were chosen. The requirement to establish
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
references was to compare different operator analyses. The considered parameters were outline (inner limiting membrane, inner plexiform layer, retinal pigment epithelium, retinal pigment epithelium reference), thickness (layer analysis selected), and offset (distance from the outline). Figure 1 shows the B-scan OCT visualization of the superficial network with a mean representative caliper measurement (white arrowhead) of 120 mm. The evaluation of this network by OCT angiography was selected at a 60 mm thickness from ILM, to define all superficial vessels (Figure 2). The deep network distribution area, lying in the outer plexiform layer, was detected by the caliper (black arrowhead). The reference was set at a 30 mm thickness scan to obtain
2199
the entire visualization of the deep network (Figures 3 and 4). As reported by Hogan et al,9 the deep network has small capillaries with a mean diameter of 20 mm. The references outline are set at 30 mm to incorporate all the capillaries of the deep network as they described. Fluorescein angiography cannot differentiate between the superficial and deep network, merging vascular anatomy information of all retinal vascular structures (Figure 5). Superficial Network Features In healthy eyes, the superficial network showed a defined silhouette morphology with a linear and continuous white shape against a black background. The wall thickness was homogenous along the entire
Fig. 3. Four different samples of the superficial network OCT angiography representation and the respective references on B-scan. In all eyes, OCT angiography shows the superficial network harmonic distribution around the foveal avascular zone with a spider web feature. B-scans show depth and shape of the layer under study.
2200 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 11
Fig. 4. Same eyes described in Figure 3, with scans taken at deep network level. Optical coherence tomography angiography shows small complex fans with multiple interconnections with irregular features around the foveal avascular zone. B-scans show depth and shape of the layer under study.
scan; the vessel course was regularly distributed, originating from the arched branches, superior and inferior, around the foveal avascular zone. The vessel flow representation (decorrelation) was homogenous for the visualized scan. The superficial network may be compared with a spider web. Around the foveal avascular zone, the vessel design shows continuous regularly spaced macular web branches, centered on the fovea, and homogeneously distributed (Figure 6). Deep Network Features The deep network in healthy eyes had a regular distribution around the foveal avascular zone, with more complex multiple tiny radial and horizontal
interconnections. The vessel flow was dense throughout the entire visualized scan, and the vasculature texture was homogeneous. The vascular network showed multiple vertical anastomoses connecting the superficial to the deeper network. From the inferior end of each vertical anastomosis, originate horizontal fan-like capillaries that interconnect to form a continuous web, parallel to the superficial network. Figure 7 shows a stack of parallel images captured in sequence through the superficial and deep network. In all the images, the persistence of the flow between the superficial and deep network can be followed showing anatomical capillary interconnections. Moreover, it is possible to follow the capillary flow from the superficial network to the deep network.
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL
2201
Fig. 5. Fluorescein angiography and OCT angiography superficial and deep network representations of the same healthy eye. The fluorescein angiography shows the overlapping information of the superficial and deep networks. The OCT angiography revealed the possibility to individually view the two networks.
Discussion Optical coherence tomography has become an essential tool in the management of retinal diseases. It shows changes in retinal anatomy, although fluorescein angiography and indocyanine green angiography, requiring intravenous dye injection, are still needed in retinal and choroidal vasculature analysis in the initial diagnosis of the disease. Several new noninvasive imaging methods are emerging in retinal and choroidal vessel visualization, such as scanning laser Doppler flowmetry,14 laser speckle photography,15 and laser speckle flowgraphy.16 However, they are not able to show the blood flow in the normal vessels of pathologic eyes. Optical coherence tomography angiography is a promising method to
Fig. 6. Representative picture of retinal vascular superficial and deep network interconnected by vertical anastomosis.
distinguish the retino/choridal vessel anatomy. The pioneering studies by Miura et al17 and Hong et al18 reported a new tool with the possibility of choroidal neovascularization imaging with some limitations in distinguishing it from the choroidal normal vasculature. In their study, Miura et al, in fact, described Doppler optical coherence angiography with the capability to detect only some parts of the choroidal or retinal vasculature, announcing more sensitive technique in development. Hong et al described a new tool, but the vast amount of information given could be difficult to use by clinicians. Hong et al had necessity to use a sophisticated data browser. This device is custom made and not available to the clinicians. The beautiful images obtained cannot be easily made in everyday clinic. Recently, Jia et al19 reported a new method to visualize well-defined choroidal neovascularization toward the retino/choroidal normal vessel by OCT angiography. Optical coherence tomography angiography is a new analytical method that allows high-resolution automated imaging of the retinal microvasculature without intravenous dye injection. This new technology is not invasive in contrast to fluorescein angiography, the gold standard method of vasculature imaging. Optical coherence tomography angiography is able to detect the intraluminal flow, independent of time and dye injection. In OCT angiography, the brightness signals the flow velocity. As reported by Leitgeb et al20 in OCT angiography, bright signals mean presence of blood flow. Fluorescein angiography does not allow visualization of vessel morphology; it only indirectly indicates a vascular permeability defect. Optical coherence tomography angiography intricately shows the silhouette of vessels in pathologic eyes, with none of the leakage, pooling, or staining representative of fluorescein angiography. Several anatomical studies have reported the retinal vessel distribution, organized in 2 distinct vessel networks: 1) the superficial network, ophthalmoscopically
2202 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 11
Fig. 7. A stack of OCT angiography parallel multiple images were captured in sequence from the superficial to the deep network (from left to right). The red circle highlights a capillary region detail and the yellow arrow show the capillary flow that moves from the superficial to the deep network.
observable with large and medium diameter vessels in correspondence to the retinal nerve fiber layer and 2) the deep network, small size capillary complex located in the outer plexiform layer.13,21,22 Optical coherence tomography angiography is able to separately detect the superficial vascular network in the ganglion cell layer and the deep vascular network in the outer plexiform layer. These two networks have different aspects, not seen with standard angiographies. In fluorescein angiography, the superficial and inner vascular networks overlap, and a selective evaluation is not possible. In healthy eyes, the superficial network consists of larger vessels than the deeperderived vessels from the superior and inferior arcades that terminate in the foveal avascular zone. The deep network shows a smaller interconnected net of dense vessels. In agreement with the description recently by Spaide et al,23 the importance of OCT angiography is mainly due to the capability to visualize all retinal vasculature layers without dye injection and to highlight at least two vascular plexus. One of the most important limitations of OCT angiography is the small area of the scan, involving only the macular region. Most likely, a full-field area of analysis providing more details could produce more information in the future. At the moment, commercially available systems does not show structures below retinal pigment epithelium. In conclusion, OCT angiography makes vessel analysis an option in everyday clinical practice without dye injection. The future possibilities of retinal details assessed by OCT angiography are unique from fluorescein angiography data, as vessel anatomy is not masked by leakage, staining, or pooling. The images of healthy and pathologic eyes may be compared, and the lack of leakage allows for better visualization of capillary abnormalities, preretinal new vessels, and CNV net (Type I or Type II). Furthermore, the lack of wall staining allows for the visualization of an abnormal blood column and the assessment of vessel wall details. Capillaries in
the macula, deep capillaries, and intraretinal potential shunts can only be finely assessed with OCT angiography. In the future, we plan to compare healthy eyes, visualized with OCT angiography imaging, with agerelated maculopathy, macular ischemia, and other diseases, to assess capillary anomalies and help delineate correct diagnoses. Key words: deep vasculature network, OCT angiography, retinal vascularization, spilt-spectrum amplitude-decorrelation angiography with optical coherence tomography, superficial vasculature network. References 1. Alvis D. Happy 50th birthday [letter]. Ophthalmology 2009; 116:2259. 2. Marmor MF, Ravin JG. Fluorescein angiography: insight and serendipity a half century ago. Arch Ophthalmol 2011;129: 943–948. 3. Lipson BK, Yannuzzi LA. Complications of intravenous fluorescein injections. Int Ophthalmol Clin 1989;29:200–205. 4. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography. Science 1991;254:1178–1181. 5. Makita S, Hong Y, Yamanari M, et al. Optical coherence angiography. Opt Express 2006;14:7821–7840. 6. Fingler J, Schwartz D, Yang C, Fraser SE. Mobility and transverse flow visualization using phase variance contrast with spectral domain optical coherence tomography. Opt Express 2007;15:12636–12653. 7. Kim DY, Fingler J, Werner JS, et al. In vivo volumetric imaging of human retinal circulation with phase-variance optical coherence tomography. Biomed Opt Express 2011;2:1504– 1513. 8. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude decorrelation angiography with optical coherence tomography. Opt Express 2012;20:4710–4725. 9. Hogan M, Alvarado J, Weddell JE. Histology of the Human Eye—An Atlas and Textbook. Philadelphia, PA: WB Saunders; 1971. 10. Druault A. Appareil de la Vision. Traité d’Anatomie Humaine. Poirier et Charpy, 1911;1:1018. 11. Redslob E. Anatomie du Globe Oculaire. Traité d’Ophtalmologie. Paris, France: Masson, édit, 1939;5:382. 12. Redslob E. Traite d’Ophtalmologie: Société Française d’Ophtalmologie. Masson et CIE, 1939;1:465–472.
RETINAL VASCULARIZATION BY OCT ANGIOGRAPHY SAVASTANO ET AL 13. Duke-Elder S. The Anatomy of Visual System. London, United Kingdom. 1961;2:372–376. 14. Michelson G, Schmauss B, Langhans MJ, et al. Principle, validity, and reliability of scanning laser Doppler flowmetry. J Glaucoma 1996;5:99–105. 15. Briers JD, Fercher AF. Retinal blood-flow visualization by means of laser speckle photography. Invest Ophthalmol Vis Sci 1982;22:255–259. 16. Tamaki Y, Araie M, Kawamoto E, et al. Noncontact, two dimensional measurement of retinal microcirculation using laser speckle phenomenon. Invest Ophthalmol Vis Sci 1994; 35:3825–3834. 17. Miura M, Makita S, Iwasaki T, Yasuno Y. Three-dimensional visualization of ocular vascular pathology by optical coherence angiography in vivo. Invest Ophthalmol Vis Sci 2011;52: 2689–2695. 18. Hong YJ, Miura M, Makita S, et al. Noninvasive investigation of deep vascular pathologies of exudative macular diseases by
19.
20.
21.
22.
23.
2203
high-penetration optical coherence angiography. Invest Ophthalmol Vis Sci 2013;54:3621–3631. Jia Y, Bailey ST, Wilson DJ, et al. Quantitative optical coherence tomography angiography of choroidal neovascularization in age-related macular degeneration. Ophthalmology 2014; 121:1435–1444. Leitgeb RA, Werkmeister RM, Blatter C, Schmetterer L. Doppler optical coherence tomography. Prog Retin Eye Res 2014;41:26–43. Paques M, Tadayoni R, Sercombe R, et al. Structural and hemodynamic analysis of the mouse retinal microcirculation. Invest Ophthalmol Vis Sci 2003;44:4960–4967. Tick S, Rossant F, Ghorbel I, et al. Foveal shape and structure in a normal population. Invest Ophthalmol Vis Sci 2011;52: 5105–5110. Spaide RF, Klancnik JM Jr., Cooney MJ. Retinal vascular layers imaged by fluorescein angiography and optical coherence tomography angiography. JAMA Ophthalmol 2015;133:45–50.
