Optical Coherence Tomography as a Diagnostic Tool for Retinal Pathologies in Avian Ophthalmology

Abstract

Purpose.: Optical coherence tomography (OCT) is an established diagnostic tool for retinal pathologies in human eyes and has been adapted to small animal models. However, there have been only a few attempts to use OCT for examination of avian eyes, and little is known about structural details of healthy or pathologically affected retinas in living birds.

Methods.: We used SD-OCT (high-resolution spectral domain OCT) to investigate eyes of various avian species including birds of prey. The birds were anesthetized by isoflurane application during OCT examination. Eyes of a common buzzard ( Buteo buteo ) could be used for a comparative analysis of OCT images and histologic/immunohistochemical examinations.

Results.: We investigated 45 wild and domestic birds (25 different species, 40 g–7.7 kg body mass) without and with diverse pathologic indications (e.g., body or head trauma). Animals were generally and ophthalmologically examined, and the diagnostic findings of direct ophthalmoscopy and OCT were compared. The OCT examination revealed an increased number of animals with clinical findings and allowed a detailed assessment of structural changes in retinal and choroidal tissue compared to simple direct ophthalmoscopy. Common findings were retinal and choroidal degeneration, retinal detachment, choroidal schisis, drusen, and drusenoid changes. Histologic and immunohistochemical analysis of retinal tissue confirmed the findings of the OCT examination.

Conclusions.: Spectral domain OCT of eyes in living birds is applicable and useful as a diagnostic tool in veterinary clinical practices and for vision research in general. Optical coherence tomography improves the quality of the common assessment methods in avian ophthalmology, and expands the diagnostic possibilities with respect to identification and prognosis of diseases. This will be particularly important for hereditary retinal defects, especially of precious breeding individuals, or estimation of treatment success in traumatized wild birds with the aim of release back into the wild.

Introduction

All vertebrate eyes and retinas are basically similar, but those of birds possess some unique characteristics. ^1,2 Avian retinas are cone dominated (with a few exceptions in nocturnal birds); they are avascular but have developed a specialized nutritive structure, the pecten oculi. Several avian species possess one or two foveal structures: a fovea centralis and a fovea temporalis or area-like structures, respectively. ^1,3 Whereas a number of histologic investigations have been conducted to characterize avian retinal structures, only a few attempts have been made to examine eyes of living birds. ^4–6 Recent technological advances in optical coherence tomography (OCT) in human ophthalmology make it possible to extend the application of this method to animal vision research, ⁷ especially to monitor pathologic changes in retinal tissue. ⁸  

Birds, with few exceptions, are incredibly visual animals. ² Outstanding visual acuity and visual performance are of great importance, particularly for raptorial species and wild birds. Avian ophthalmology is an advancing field of avian medicine. This results from advancements in exotic pet husbandry; in wild animal protection; in welfare-oriented animal husbandry in zoos and falconry; and additionally in the focus on health control of precious breeding animals, especially of birds of prey. Animals from all of these categories are presented to the avian veterinarian with a different prevalence of ocular diseases and are subject to different management and financial criteria. Avian veterinarians are frequently confronted with ocular diseases of their patients, which could be the result of systemic infectious diseases (bacterial, mycotic, viral), noninfectious diseases (traumata, neoplasias, nutritional, toxic), or of unknown etiology. ⁹ Wild bird disorders of the posterior eye segment are often trauma-related hemorrhages (most frequently arising from the pecten), ¹⁰ and retinal edema, retinal separation, and focal or extensive retinal and choriodal degeneration have been described. ^11–13  

Although other diagnostic tools have been applied to avian species (i.e., electroretinography, ocular ultrasound, computed tomography, magnetic resonance imaging), ^11,14 direct and indirect ophthalmoscopy still remain the basic and routine methods to investigate the status of the posterior eye segment. However, the ophthalmological and diagnostic information is limited, particularly if the inter- and intraspecific variation of almost 9000 avian species is taken in consideration. On the other hand, a detailed analysis of the retinal pathology and the degree of chorioretinal damage and neurodegeneration is necessary in order to make a decision on the appropriate medical treatment, the extent of visual impairment, and the prognosis (including the success of releasing the birds back into the wild). Therefore, we applied high-resolution spectral domain OCT (SD-OCT) as an innovative ophthalmological assessment method and compared the results with observations of direct ophthalmoscopy and histologic examinations. 

Materials and Methods

Bird Handling and Anesthesia

All general clinical examinations and the handling and anesthesia of birds to execute OCT were performed in accordance with applicable European laws (European Communities Council Directive 86/609/EEC) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and after approval by the local authorities (Faculty of Veterinary Medicine of the University of Leipzig, Germany, and Landesdirektion Leipzig, Germany). 

For this study, 45 wild and falconry birds were examined (Table 1). The wild birds were admitted to the Clinic of Birds and Reptiles as wildlife casualties either with suspicion of generalized trauma or as orphaned juveniles. The few falconry birds, which were admitted to the clinic for periodical health control examinations, were included in this study with the owner's consent. A detailed anamnesis was carried out, and each bird received a full-body general examination and a standard ocular examination. Direct ophthalmoscopy was achieved with the use of a simple monocular ophthalmoscope (Heine Beta 200; Heine Optotechnik, Herrsching, Germany). Further examinations with 30-, 60-, and/or 90-diopter lenses (adapted to the special needs of different bird species) were carried out by the same examiner. For OCT examination, the subjects were anesthetized using a mask and standard concentrations of isoflurane (induction 5%, maintenance 1.5%–2.5%) and oxygen (induction 2 L/kg, maintenance 0.8–1 L/kg flow). During OCT examination, the birds were intubated. Anesthetic monitoring was carried out by an experienced avian veterinarian and included the basic life parameters (heart rate, respiratory rate, body temperature, ventilation). The complete anesthetic duration and the OCT examination took approximately 5 to 30 minutes per animal. Birds were held in an upright position to avoid internal organ and air sac compression. 

Table 1

 

Survey of All Examined Bird Species

Ordo	Familia	Species	Age	Weight, kg	Activity	Animal Husbandry
I	Accipitridae (accipiters)	Milvus migrans (black kite)	Adult	0.81	Diurnal	Wild
		Milvus milvus (red kite)	Juvenile/adult*	0.83–1.1	Diurnal	Wild
		Accipiter gentilis (goshawk)	Juvenile	0.85	Diurnal	Wild
		Accipiter nisus (sparrowhawk)	Adult	0.21	Diurnal	Wild
		Haliaaetus albicilla (white-tailed sea eagle)	Adult	3.60	Diurnal	Wild
		B. buteo (common buzzard)	All adult*	0.77–0.82	Diurnal	Zoo†
	Falconidae (falcons)	F. tinninculus (kestrel)	Juvenile/adult*	0.08–0.22	Diurnal	Wild
		F. peregrines (peregrine falcon)	All juvenile*	0.55–0.88	Diurnal	Wild/falconer
II	Anatidae (ducks)	C. olor (mute swan)	Adult	7.72	Diurnal	Wild
		Anas cyanoptera (cinnamon teal)	Juvenile	0.32	Diurnal	Zoo†
		Anas platyrynchos domesticus (Indian runner duck)	Adult	1.53–1.9	Diurnal	Zoo†
III	Corvidae (corvids)	C. corone cornix (hooded crow)	Juvenile	0.25	Diurnal	Wild
		C. corone corone (carrion crow)	Juvenile/adult*	0.25–0.55	Diurnal	Wild
	Turdidae (thrushes)	Turdus merula (common blackbird)	Juvenile	0.06	Diurnal	Wild
IV	Charadriidae (plovers)	Vanellus vanellus (northern lapwing)	Adult	0.20	Diurnal/nocturnal	Wild
V	Columbidae (pigeons)	Columba livia domestica (pigeon)	Adult	0.45	Diurnal	Domestic
VI	Strigidae (typical owls)	Bubo bubo omissus (eagle owl)	All subadult*	1.75–1.9	Nocturnal/crepuscular	Falconer
		S. aluco (tawny owl)	Adult	0.49	Nocturnal	Zoo†
		A. otus (long-eared owl)	All adult*	0.44–0.46	Nocturnal	Zoo†
		Bubo scandiaca (snow owl)	Adult	1.24	Diurnal/nocturnal/ crepuscular	Falconer
	Tytonidae (barn owl)	Tyto alba (barn owl)	Adult	0.48	Nocturnal/crepuscular	Zoo†
VII	Apodidae (swifts)	A. apus (common swift)	Adult	0.04	Diurnal/crepuscular	Wild
VIII	Upupidae (hoopoe)	Upupa epops (hoopoe)	Adult	0.05	Diurnal	Wild
IX	Picidae (woodpeckers)	Dendrocopos major (great spotted woodpecker)	Juvenile	0.07	Diurnal	Wild
X	Rheidae (rheas)	Rhea americana (rhea)	Juvenile	4.80	Diurnal	Domestic

 Ordo/Order: I, falconiformes (birds of prey); II, anseriformes (waterfowl); III, passeriformes (passerine); IV, charadriiformes (plover-like waders); V, columbiformes (plover and pigeons); VI, strigiformes (owls); VII, apodiformes (swifts); VIII, coraciformes (raven-like birds); IX, piciformes (woodpeckers and allies); X, rheiformes (ratites).

*  Examination of more than one bird.

†  Length of captivity unknown.

Spectral Domain Optical Coherence Tomography

Images were taken with OCT with use of the Spectralis SD-OCT (Heidelberg Engineering GmbH, Heidelberg, Germany), including its active live eye tracking function. The laser light projected to the retina was reflected differently by the layers of retinal tissue; therefore the scan enabled histology-like examination of retinal structures in vivo. Two simultaneous laser beams presented retinal scan and overview image registered in real time and shown live, which aided the measurement procedure while the whole fundus was examined and areas of interest were simultaneously scanned. Within the device, the infrared image was taken with a wavelength of 815 nm, and the OCT scan was acquired with a superluminescent diode with a mean wavelength of 870 nm (850–920 nm). 

For every bird, an OCT volume scan with a field size of 15° (temporal-nasal) × 5° (superior-inferior) was acquired. The OCT scans were acquired in high speed mode. This means that every B-scan contained 384 A-scans, corresponding to a distance between the single A-scans of 0.039°. A total of 131 B-scans were acquired within the specified field, resulting in a distance between individual B-scans of 0.038°. In some test subjects, the field size was increased while the scan settings were kept the same to cover an area of specific interest. In these cases the distances between the B-scans and between the single A-scans were the same as in the smaller field. 

The Spectralis system acquires 40,000 A-scans per second and applies a real-time noise reduction system to increase the signal-to-noise ratio of the acquired scans. In general, noise is inherent within OCT scans; and with the use of averaging techniques, for example, per scanning location, image processing within this SD-OCT allowed the signal to be enhanced. At each scan location, for every B-scan, 10 single B-scans were acquired and averaged to reduce the speckle noise in the images. The implemented noise reduction enabled fine detail to be visible in the bird eyes. 

Confocal scanning laser ophthalmoscopy, integrated into this SD-OCT system, presented infrared reference fundus images of each region examined. This detailed outline is comparable to the view obtained on manual funduscopy techniques. Capturing such images alongside the OCT scan allows perspectives of the retina to be examined retrospectively, which is not possible in standard ophthalmoscopy. The simultaneous fundus image enabled accurate registration of the scan location. The eye tracker system aided accurate image capture despite small eye movements in the bird by employing statistical methods based on landmarks in the fundus, which were brought into alignment during the scanning procedure. 

An OCT scan usually achieves a scan depth in human retinal tissue of 1.9 mm with 496 pixels per A-scan. Therefore, the scan has a digital axial resolution of 3.9 μm (optical resolution of 7 μm). 

Histology and Immunohistochemistry

One common buzzard ( Buteo buteo ) with an OCT examination had to be euthanized for other veterinary medical reasons, and the use of additional eye material from euthanized birds was in accordance with all applicable laws and approved by the local authorities (see above). Removed eyes were fixed in 4% paraformaldehyde (PFA) containing phosphate-buffered solution (PBS) for 48 hours. Fixed eyes were circumferentially opened at the limbus, and retinal pieces were isolated and embedded in 3% agarose gel for sectioning with a vibratome slicer (Microm HM 650 V Vibration microtome; Thermo Scientific, Pittsburgh, PA). Retinal sections of 30-μm thickness were histologically stained with hematoxylin or hematoxylin/eosin. For immunohistochemistry, the slices were preincubated in PBS with 5% normal donkey serum plus 1% dimethyl sulfoxide and 0.3% Triton X-100 (both Carl ROTH GmbH + Co. KG, Karlsruhe, Germany) to block nonspecific binding of antibodies. Immunostaining with primary antibodies was performed with mouse monoclonal anti-glutamine synthetase (10 μg/mL, MAB302; EMD Millipore Corporation, Billerica, MA) to label Müller glia cells and rabbit polyclonal anti-G_αt1 (4 μg/mL, K-20; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) to detect rod photoreceptors. Biotin-conjugated peanut agglutinin (PNA) from Arachis hypogaea (5 μg/mL; Sigma-Aldrich, St. Louis, MO) was used to visualize cone photoreceptors. Secondary antibodies conjugated to carbocyanin 2 (Cy2) or Cy3 (7 μg/mL), respectively, and Cy3-conjugated streptavidin (9 μg/mL; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) were used to detect primary labeling. The cell nuclei were stained with Hoechst 33258 (1:1000; Molecular Probes, Eugene, OR). The retinal slices were mounted with Immu-Mount (Thermo Scientific) on glass slides and analyzed with a confocal laser scanning microscope (LSM 510 Meta; Zeiss, Oberkochen, Germany). All materials not indicated separately were purchased from Sigma-Aldrich (Deisenhofen, Germany). 

Results

We investigated 45 wild birds/falconry birds (25 different species, 0.04–7.7 kg body weight; Table 1) with standard ophthalmological examination and specifically compared the diagnostic findings of direct ophthalmoscopy and SD-OCT. No anesthetic complications or behavioral changes were observed during or after the OCT examination. For reasons of completeness and comparability of all examinations, the presented study includes only the examinations of the right eyes. With high-resolution OCT we were able to visualize retinal layers and special retinal structures such as the pecten oculi and the fovea centralis and fovea temporalis in the avian eye. A direct comparison of different methods (such as immunohistochemistry, histology, and OCT) to visualize retinal tissue revealed analogous retinal layer structures and similar, but not identical, retinal layer dimensions (Figs. 1A–C). In some cases, the OCT images actually resolved some of the outermost retinal structures as shown in Figures 1D and 1E. Additionally, in some birds we were able to visualize the choroid and the cartilaginous lamina and external fibrous layer of the avian sclera (see Fig. 3). An observation in several examined bird eyes was a retraction of the vitreous bodies several hundred micrometers from the inner retinal surface and irregular margins possessed by the hyaloid. 

Figure 1

 

Histologic and immunohistochemical staining and OCT images to visualize retinal layer structure of the common buzzard ( B. buteo ) in corresponding healthy retinal areas. (A) Immunohistochemical staining with anti-glutamine synthetase antibodies (green) labels parts of Müller cells in the NFL and from the INL to the ONL. PNA lectin (red) labels mainly photoreceptor cells. Cell nuclei were visualized by Hoechst 33258 (blue) staining. (B) Light microscopic image of histologic staining with hematoxylin/eosin reveals the well-organized retinal layers. The pigmented cell processes of retinal pigment epithelial cells surrounding photoreceptor outer segments appear brown (bottom). (C) Two-dimensional OCT image of a healthy retinal area of a living common buzzard. The layer structure of the retina corresponds to the histologic preparations. Note the diffuse appearance of the outermost structures (oS/RPE). (D, E) OCT image of a healthy area at the border of the fovea centralis. It is noteworthy that several separate layers are discernible in the outermost structures from the ELM onward. NFL, nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; ELM, external limiting membrane; iS, inner photoreceptor segments; oS, outer photoreceptor segments; RPE/BM, retinal pigment epithelium/Bruch's membrane; Chor, choroid.

Figure 2

 

Infrared funduscopy ([A, C, E, G]; left column) and 2D-OCT images ([B, D, F, H]; right column) from eyes of four avian species with different abnormalities in the outer retinal layers and in the subretinal space. (A, B) F. tinnunculus (common kestrel): Infrared funduscopy revealed dark spots, whereas the OCT showed no obvious pathologic abnormalities. (C, D) Milvus milvus (red kite): The dark spot in the funduscopy appeared in the OCT image as a drusenoid change with a dark central cavity (white arrow in [D]) in the area of the photoreceptor segments and the subretinal space. The asterisk in (C) indicates the central fovea. (E, F) F. peregrinus (peregrine falcon): In the funduscopy, several small dark dots were observed near the pecten occuli; they appear as diffuse drusenoid alterations spreading from the subretinal space toward the external limiting membrane. (G, H) Bubo scandiacus (snowy owl): Many small speckled white dots were found in the fundus of the snowy owl. The drusenoid changes with a dark central cavity in the subretinal space deform the external limiting membrane. The small white arrows in the funduscopy images indicate the orientation of the corresponding 2D-OCT pictures. The bold white arrows in the OCT images (D, F, H) point to the retinal pathologic changes enlarged in the insets.

Figure 3

 

Infrared funduscopy ([A, C, E, G]; left column) and 2D-OCT images ([B, D, F, H]; right column) from four individual birds (three different avian species) with different retinal and choroidal pathologies. (A, B) A. otus (long-eared owl): Diffuse distributed light gray deposits and dots of different sizes were observed in the funduscopy. OCT image revealed drusenoid deposits with a dark central cavity (vacuole-like) in the outer nuclear layer and in the choroid (arrowhead and inset). (C, D) C. corone (carrion crow): The funduscopy shows a locally restricted pathologically altered fundus. The OCT image revealed a massive widening of the choroid (arrow) and degenerative changes in the subretinal space and the outer retinal layers (arrowhead). (E, F) A. otus (long-eared owl): The funduscopy shows pathologically altered areas. The OCT image revealed a widening of the choroid (arrow) and explicit degenerative changes in the subretinal space and the outer retinal layers (arrowhead). (G, H) Tyto alba (barn owl): The funduscopy revealed an inhomogeneous appearance of various fundus areas. The OCT clearly shows a degenerative retinal alteration of the entire retina with a schisis-like appearance. The retinal layer structure is totally disorganized, with edema-like enlargements (arrowhead). The white arrows in the funduscopy show the orientation of the OCT image.

Using direct ophthalmoscopy, only 5 animals in the examined group of 45 birds presented pathologic changes of the right eye. In contrast, we identified abnormalities or pathologic changes in a total of 17 birds by employing OCT technology to examine fundus and retinal structures. In 3 of 17 birds, we observed changes of the fundus pigmentation and assessed these for nonpathologic abnormalities. In the remaining 14 cases we observed various pathologic changes. A summary of characteristic and representative abnormalities and pathologic changes of eyes from 10 exemplary birds is given in Table 2 and corresponds to the pictures presented in the figures. The remaining 7 birds, not specified in Table 2, had the following similar abnormalities or pathologic changes: nonpathologic changes of the fundus pigmentation ( Falco peregrinus , peregrine falcon; Corvus corone cornix , hooded crow); drusen/drusenoid textures ( Falco tinninculus , kestrel; Cycgnus olor , mute swan; Strix aluco , tawny owl; Asio otus , long-eared owl); or retinal pigment epithelium detachment ( Apus apus , common swift). In this study, foveal areas were coincidentally not directly affected by the pathologic changes. 

Table 2

 

Comparison of Pathologic Findings Examined by Direct Ophthalmoscopy or OCT

No.	Species, Age	Direct Ophthalmoscopy O.D.	OCT Infrared Scan O.D.	OCT 2D and Volume Scan O.D.
1	F. tinnunculus (kestrel), 35 d	n.a.d.	Pigmented round dots in the temporal fundus, inferior to the fovea temporalis	n.a.d.
2	Milvus milvus (red kite), adult	n.a.d.	Pigmented round dot	1 drusenoid texture in upper half nasal fundus area
3	F. peregrines (peregrine falcon), juvenile	n.a.d.	Drusenoid textures in the nasal fundus, next to the pecten oculi and inferior to the fovea centralis	Drusenoid textures in the most posterior layer
4	Bubo scandiacus (snowy owl), adult (falconry)	n.a.d.	Multiple, widespread drusen mostly in the nasal fundus, next to the pecten oculi	Drusenoid textures/vacuoles in the most posterior layers of the retina
5	A. otus (long-eared owl), adult, animal No.1	n.a.d.	Widespread irregular, yellow hypopigmented “deposits” (O.S.: 1 distinct drusen in the temporal fundus, next to the pecten oculi)	Drusenoid textures/vacuoles in the most posterior layer of the retina, partial intraretinal schisis, some blurry hyperpigmented round areas and dark, possibly fluid-filled areas in the chorioidea (O.S.: a large drusenoid texture in the most posterior layer of the retina extending deep into the anterior layers)
6	C. corone (carrion crow), juvenile	Extensive peripectenal bleeding	Extensive “discoloration” or blurry pigmentation in the nasal fundus area	Choriodal schisis underneath the retinal layers in the affected area; inner layers largely unaffected
7	A. otus (long-eared owl), adult, animal No. 2	Centrally degenerative, hypopigmented area with irregular presence of choroidal blood vessels	Centrally degenerative, hypopigmented area with irregular presence of choroidal blood vessels	Partial intrachoroidal schisis, degeneration of photoreceptor segments, and pathologically altered subretinal space
8	Tyto alba (barn owl), adult	Degenerative changes	Extensive retinal degeneration	Intraretinal schisis, with associated edema and atrophy
9	B. buteo (common buzzard), adult, animal No. 1	Retinal degeneration, possible retinal detachment	Extensive retinal degeneration in the superior, nasal fundus area, nasosuperior to fovea centralis	Retinal detachment and degeneration of the posterior layers, especially photoreceptor segments
10	B. buteo (common buzzard), adult, animal No. 2	Retinal degeneration, possible old bleeding	Retinal degeneration in the superior, temporal fundus area, superior to fovea temporalis	Complete retinal degeneration affecting all layers, retracted vitreous body with irregular margins

 n.a.d., no abnormalities detected; O.S., oculus sinister (left eye); O.D., oculus dexter (right eye).

The most frequently observed pathologic findings in our study cohort were changes to the fundus pigmentation, drusen/drusenoid textures, and vacuoles (Fig. 2). Furthermore, we observed complete or partial choroidal schisis with suspicion of intrachoroidal edema, intraretinal schisis with edema and atrophy, retinal degeneration of the outer layers, and a retinal detachment (Figs. 2–4). In three animals, fundus pigmentation was not associated with pathologic changes of the retinal structure (Figs. 2A, 2B). Drusenoid patterns were found to be a single spot (Fig. 2C), a locally restricted group of dots (Fig. 2E), or evenly distributed in a large retinal area of the fundus (Fig. 2G). The drusenoid textures seem to emanate mostly from the subretinal space and the photoreceptor segment layer and in some cases penetrate the external limiting membrane and the outer retinal layers. The drusenoid textures possessed two different appearances: (1) an irregular shape and sharp borders, with a black cavity in the center (Figs. 2D, 2H), and (2) diffuse margins and a homogeneous content without cavities or inclusions (Fig. 2F). 

Figure 4

 

(A, B) Infrared funduscopy and 2D-OCT image from a common buzzard ( B. buteo ) with local retinal detachment and pathologically choroidal alterations. Whereas the funduscopy shows an expanded pathologically altered area, the OCT image revealed an exclusive degeneration of photoreceptor segments in the subretinal space of a retinal detachment.

Other birds presented with more obvious and serious pathologic changes. Such changes to the fundus pigmentation were observed in infrared funduscopy as well as in the OCT scan (birds numbered 6–10 in Table 2) as shown in Figures 3, 4, and 5. The infrared fundus images give a good overview of the local dimensions of pathologic changes in the fundus areas. However, no information about the affected choriodal or retinal layers or about the degeneration of retinal pigment epithelium cells, photoreceptors, or retinal neurons can be obtained from the funduscopy. This information, however, is revealed by the two-dimensional (2D) OCT scan; and we could therefore clearly identify choriodal or retinal schisis with edema-like structures, retinal detachment, or different forms of neuroretinal degeneration (Figs. 3–5). 

Figure 5

 

Comparison of histologic, immunohistochemical, and OCT examinations of a pathologic buzzard retina. (A) Gross morphology, funduscopy, and OCT images from a single buzzard right eye with pathologic changes. The eye cup was circumferentially opened at the limbus, and a pathologically altered area is observable in the superior–temporal periphery (arrow, left picture). The enlarged inset shows the circumscribed pathologic area (dotted line) and the location of the two foveas (fovea temporalis and fovea centralis, asterisks). Shape and dimension of the pathologically changed area are clearly visible in the infrared funduscopy (middle right: white arrow shows the orientation of the 2D-OCT image; asterisks indicate the fovea temporalis). The 2D-OCT image (right) revealed a massive degeneration of all retinal layers in the affected areas and obvious structural alterations of the choroid. (B) Histologic cross section with comparable orientation as seen in the OCT image. Hematoxylin staining shows the transition zone (dotted line) between the healthy retinal part on the right and the degenerated retina. (C) Immunohistochemical labeling of a consecutive retinal cross section with anti-glutamine synthetase antibodies (Müller cells, green), anti-G_αt1 antibodies (rod photoreceptors, red), and Hoechst 33258 (blue) staining of all cell nuclei. Note the loss of immunoreactivity in the degenerated retinal areas. (D) Microscopic images of hematoxylin-stained retinal sections from degenerated and healthy retina illustrate the massive loss of retinal cells. Pigmented cells migrate into the degenerated retina (arrows). (E) Enlarged microscopic images of immunohistochemical stainings from degenerated and healthy retina as seen in (C) confirm the severe retinal degeneration and the loss of photoreceptors and glutamine synthetase immunoreactivity in Müller cells in the affected region, respectively.

One of the common buzzards was euthanized for veterinary reasons (not because of bad prognosis regarding ophthalmological injuries), and we used the eyes for a comparison of the OCT images with histologic and immunohistochemical examinations (Fig. 5). The infrared funduscopy image revealed the margins and dimensions of the pathologic changes in fundus pigmentation, whereas only the 2D OCT scan provided insights into the degenerative changes of the neuroretinal tissue and the affected retinal and choroidal layers (Fig. 5A). Additionally, the histologic examination confirmed the findings of the OCT images. Here, the neurodegeneration presented with clear-cut margins; and in the pathologically changed retinal tissue, nearly all retinal layers were affected by neurodegeneration and a dramatic cell loss (Fig. 5B). Furthermore, pigmented cells migrated into the retinal tissue, presumably originating from the retinal pigment epithelium (Fig. 5D). Additionally, immunohistologic examination revealed a massive photoreceptor cell loss and changes in the expression pattern of the glial-specific enzyme glutamine synthetase in Müller glial cells (Figs. 5C, 5E). Nearly all neuronal cells of the outer retinal layers were lost in the degenerated retinal tissue, and the Müller glial cells were immunonegative for glutamine synthetase. 

Discussion

High-resolution SD-OCT is a modern diagnostic tool in human ophthalmology, and we used this method to examine avian eyes in veterinary clinical practice. We compared the results of the OCT examination with findings of simple direct ophthalmoscopy. The SD-OCT examination and even the infrared funduscopy revealed an increased number of animals with clinical findings and allowed a detailed assessment of structural changes in retinal and choroidal tissue, far beyond simple direct ophthalmoscopy. Therefore, SD-OCT improves the quality of the common assessment methods in ornitho-ophthalmology. Spectral domain OCT expands the diagnostic possibilities in estimating pathologies and prognosis and helps to evaluate hereditary or acquired retinal defects, especially in precious breeding animals. In addition to advanced investigation of pathologic changes, OCT enables the examiner to monitor all special retinal characteristics such as the foveal structures (see Supplementary Material) and the pecten oculi. ⁶  

We were able to investigate many avian species including diurnal and nocturnal birds of prey, with different body and eye sizes, different body weights, and different ages, by means of the OCT imaging technique. During the OCT examination we used a standard protocol for avian anesthesia, while others have investigated birds of prey without anesthesia. ⁶ The pupil of birds is controlled by striated muscles, and the commonly used mydriatics for humans and mammals do not work in birds; therefore, we did not use specific mydriatics to achieve or control pupil dilation in the examined birds. We used a standard isoflurane anesthesia until it reached a surgical plane; it can therefore be assumed that pupil dilation was maximal, since all muscles including the ocular muscles were fully relaxed. 

Furthermore, we used a commercially available SD-OCT device without further modifications for our examinations, whereas others have used an individually modified OCT setup. ⁶ Our approach might enable an experienced avian veterinary staff to perform OCT examinations. 

The high-resolution cross-sectional images of the retinal and choroidal structures provided by the SD-OCT facilitate a detailed characterization of ophthalmological peculiarities. Some of the changes of the posterior inner eye detected with our examinations could be identified as innocuous changes with no need for therapeutic treatment, for example, pigmentations. Others could be identified as more or less harmful, causing retinal tissue damage, with different consequences for the visual capability of the animal—for example, identified drusen and drusenoid changes. The appearance of either vacuole-like or diffuse drusenoid changes in the avian eyes was similar to that of drusen/drusenoid changes observed in human retina pathologies. ^15,16 Bilateral multiple punctate chorioretinal lesions in the ventral aspect of the fundus in owls have already been reported. ¹² The affected retinal structures could be estimated by the OCT images, but the consequences for the pathologic progress or the prognosis are uncertain. Additional histologic examinations are necessary to confirm assumptions about the different structures, appearances, and pathologic consequences of the drusenoid changes. As known from ophthalmology, drusenoid pathologic changes might be induced through environmental effects (e.g., UV light), caused by nutrition deficits, affected by breeding conditions, or related to age-dependent changes and/or genetic aspects. However, all the birds examined belonged to wild avian species and were not domestic (exceptions are the rhea and the pigeon; see Table 1). Even the animals from the zoo included in this study mostly originated from the wild. In free-living or captive birds it is difficult to estimate the age and the actual diet. Although we cannot be sure, it may be assumed that light and environmental conditions, diet, or age-related changes do not play a dramatic role in causing retinal pathologies in free-living birds. For animals growing up or living in a zoo or for falconry birds, one can assume that to some extent diet, light, and husbandry conditions do not necessarily reflect the optimal natural habitat and therefore that these could play an important role in retinal pathologies. However, in this study it would have been speculative to correlate the retinal abnormalities found with environmental conditions or possible age-related changes. 

Some animals (n = 5) in our cohort presenting with large areas of pigment alteration were diagnosed by direct ophthalmoscopy. However, this method did not allow for a determination of which retinal layer or tissue was really affected. Spectral domain OCT images revealed a detailed picture of affected tissues and allowed an accurate account of the extent of the neuronal damage and thus of the consequences for visual performance for the patients. Thus, we could distinguish between choroidal or retinal changes and could estimate the degree of degeneration of neuronal tissue; therefore, the examiner obtained the relevant information to estimate the visual impairment and the success of a possible release into the wild. 

In addition to the cross-sectional images of the retina, the infrared funduscopy mode of the SD-OCT gave an excellent overview of large areas of the posterior eye for orientation. Direct or indirect ophthalmoscopy is largely subjective and is influenced by the experience of the ophthalmologist. The infrared funduscopy images presented the opportunity to perform a more objective evaluation for various examiners, since each one could study the same findings retrospectively. Similarly, this might be possible with fundus pictures taken by a fundus camera, but there are significant limitations of that method (e.g., eye size, mydriasis achievement, room light quality, flash disturbance of the animal). Additionally, an effective eye tracking system enables repeated examinations of the same retinal areas during follow-up monitoring or over the course of therapeutic treatment. In other studies, orientation and accurate reorientation were found to be challenging because the avian retina is avascular and lacks blood vessels to serve as “landmarks.” ⁶ The built-in eye tracking in the OCT system proved to be very reliable and effective. In every case (i.e., every examined retina even with an inconspicuous fundus), the eye tracking system was able to continuously measure the scan outlined and therefore able to align all the recorded images and produce a 3D composition of all B-scans regardless of brief interruptions in the scanning procedure caused by movements of the animal or the animal holder. The eye tracking system employs statistical methods based on landmarks in the fundus that are brought to alignment during the scanning procedure. In the avascular bird retina, different gray values of nerve fiber structures or the underlying choroid were sufficient for alignment during the scan of retinal areas. 

In comparison to examinations of human eyes, in some birds it was even possible to image the entire choroid and the two layers (cartilage and fibrous) of the scleral tissue. Currently, we do not know if the depth of the OCT images depends on species differences (e.g., larger image depths were observed in owl species), on pigmentation and/or dimension of the pigment epithelium, on retinal thickness, on the anterior visual apparatus, or on special conditions/adjustments and technical parameters of the OCT examination/system itself. 

A comparison of OCT images and histologic and immunohistologic sections of retinal tissue revealed good accordance of retinal layers and structures. It is noteworthy that the measured dimensions of the retinal layers or other anatomical structures in the OCT images are not equal to the dimensions in the histologic sections (note the slightly different scale bars in Figs. 1A–C). This was also observed in an earlier investigation of the monkey retina, in which the histologic dimensions were slightly larger than those assessed in the OCT images. ¹⁷ Possible reasons for a discrepancy in the optical reflection properties and the anatomical dimensions have to be carefully investigated in separate series of experiments, because it is obvious that OCT image dimensions do not correlate exactly with the anatomical structures of the retinal and adjacent tissues (e.g., influences of resolution limitations, coherence length and point spread function in OCT imaging systems).¹⁸ Similar problems arise for the designation of several lines of the outer and outermost structures revealed by high-resolution OCT images in some birds (see Fig. 1). Recent publications have questioned and scrutinized the designation of the four outer lines in OCT images of the human retina. ^18,19 For instance, one of the hyperreflective OCT bands (correlating in birds with the most hyperreflective band in Figs. 1D, 1E) represents the ellipsoid portion of the inner segments of photoreceptors rather than the junction between the inner and outer photosegments. ¹⁸ However, the designation of OCT layers is still controversial. Therefore, these cellular and subcellular structures should be examined at an ultrastructural level and should be correlated with their optical properties. Especially in different bird species, the anatomical structures of the retina vary considerably (e.g., length of photoreceptor inner and outer segments and their proportion of myoid and ellipsoid parts, photoreceptor dimensions in temporal and central foveal and nonfoveal regions, microvilli length of Müller cells, height of RPE cell processes ensheathing the photoreceptor outer segments). Additionally, in many bird species, the existence of double cones and the number, distribution, and optical properties (e.g., absorption wavelength profile) of oil droplets in many photoreceptors might influence the reflecting properties of the outer retinal parts. 

In general, using the built-in OCT algorithm to quantify the retinal layer dimensions of bird eyes provides incorrect data because the OCT system is optimized for usage in human eyes. The analysis software uses optical and statistical parameters of the human eye, and the calibration is based on optical formulas of Gullstrand's model eye or similar historical data (e.g., eye and tissue dimensions, refractive indexes of the optical relevant tissues). 

However, histologic and immunohistochemical analysis of retinal tissue from a pathologic eye of a common buzzard confirmed the findings of the OCT examination and underlined the degree and the dimensions of the neurodegeneration. It is noteworthy that the loss of glutamine synthetase immunoreactivity in Müller glial cells occurs exactly in the area of neuronal cell loss. Glutamine synthetase is a glia-specific enzyme and is involved in the so-called glutamate-glutamine transmitter recycling system. ²⁰ Glutamate released by active neurons is specifically taken up by Müller glia cells and is converted into glutamine by this enzyme. An explanation for the reduced glutamine synthetase immunoreactivity in Müller cells might be the reduced need to recycle glutamate for neurotransmission caused by the massive neuronal cell loss as observed under various pathologic conditions in animal models. ²¹  

The results of our investigation clearly show that OCT diagnostics provide more detailed and accurate information about the pathologies of the posterior eye, with a higher incidence of detecting pathologic changes in comparison with direct ophthalmoscopy. Furthermore, we conclude that SD-OCT is an excellent tool for the in vivo evaluation of developmental or pathologic alterations of avian retinas, as well as for studies on structural differences between species in living birds. Thus, although the OCT setup is an expensive device and special education is needed, and although OCT might be applicable only in specialized facilities, we are sure that it will become an indispensable tool for up-to-date ornitho-ophthalmology and for vision research. 

Supplementary Materials

mov S1 

Acknowledgments

Supported by funding from the German Federal Ministry of Education and Research (BMBF, PtJ-Bio, 0315883 [CK, NK, MF]), from the Deutsche Forschungsgemeinschaft (GRK 1097/1 and 1097/2 [AR]), and the Medical Faculty of the University of Leipzig (grant [WV]). 

Disclosure: F.G. Rauscher, None; P. Azmanis, None; N. Körber, None; C. Koch, None; J. Hübel, None; W. Vetterlein, None; B. Werner, None; J. Thielebein, None; J. Dawczynski, None; P. Wiedemann, None; A. Reichenbach, None; M. Francke, None; M.-E. Krautwald-Junghanns, None 

Footnotes

 FGR and PA contributed equally to the work presented here and should therefore be regarded as equivalent authors.