- Top of page
- Material and Methods
Several imaging methods exist for the evaluation of corneal morphology. The most frequently used method is slit-lamp biomicroscopy, but visualization of deeper corneal structures can be limited by light absorption and scattering through the cornea, particularly in case of corneal edema. The application of new technologies including high-frequency ultrasound (ultrasound biomicroscopy [UBM]), confocal microscopy, Scheimpflug imaging, and optical coherence tomography (OCT) has led to improved resolution and complementary evaluation of corneal conditions.[1-3] OCT, developed in 1990, was initially dedicated to evaluation of the retina in human ophthalmology. The use of OCT for the examination of the anterior segment of the eye, a practice started in the early 2000s, has opened a new field in the clinical and experimental approaches to evaluating these structures.[1-3] Anterior segment spectral domain OCT (SD-OCT) has been the subject of more than 500 scientific articles over the last 5 years. The application of SD-OCT for medical and surgical purposes has been specifically described for the structural analysis of tear meniscus, normal and pathological cornea, and iridocorneal angle, as well as evaluation of the anterior chamber, iris, and lens.[5-16] OCT has also been compared with slit-lamp examination and with Scheimpflug imaging for corneal endothelial evaluation.
Spectral domain optical coherence tomography function is similar to ultrasonography with a few major differences. Ultrasonography uses ultrasound waves emitted by a probe in contact with the tissue to be studied, whereas SD-OCT uses infrared (IR) light emitted at a distance from the cornea, making contactless image acquisition possible. While optical transparency is not required for ultrasound, the use of IR light in OCT requires transparent media. The light passing through different ocular media experiences interference, which is compared with light reflected on a reference mirror at the same working distance. Scanning the reference mirror through a range of distances allows generation of an axial image (A-scan). A series of axial sections are combined to produce a composite image, similar to the standard two-dimension (B-scan) image produced in ultrasonography. SD-OCT images have an axial resolution of 2–4 μm and a lateral resolution of 20–25 μm. As a comparison, the axial resolution of 50-MHz ultrasonography is 50 μm. Additional information about technical features of SD-OCT can be found elsewhere.[1, 4]
Heavy and costly SD-OCT devices were initially limited to specialized human ophthalmology centers or research centers. But now, lighter and more affordable models have become available. The use of SD-OCT for the morphological analysis of the cornea of dogs and cats in healthy and pathological conditions has yet to be described. Our study aims to evaluate spectral domain OCT for corneal imaging in dogs and cats in clinical practice conditions.
- Top of page
- Material and Methods
The goal of this study was to test the feasibility of using SD-OCT to image the cornea in cats and dogs in common practice situations. Our results confirm that this technique is applicable to the evaluation of the cornea in healthy and pathological conditions. For most cases studied, SD-OCT evaluation provides reliable and accurate additional information to the standard examination using a slit lamp for corneal disorders. Through the study of many different conditions, we were able to underline diagnosis- and image-related advantages and limitations of SD-OCT.
This study is one of the first evaluations of SD-OCT for corneal diseases in private practice conditions. This means, we have evaluated the pathological conditions that came through our clinic during the time of the study. Although a wide range of conditions has been examined, many other conditions such as extended endothelial dystrophies or eosinophilic keratitis were not evaluated because we were not presented with those conditions during the study.
The SD-OCT imaging resolution was excellent at all depths, and we were able to measure the central cornea thickness with a resolution of 5 μm. This measurement required the device to be positioned axially in the center of the cornea. The results obtained in the ‘pachymetry’ mode were compatible with available data.[18, 19] However, statistical analysis of these data was not the aim of this study and has not been performed. The influence of the corneal curvature on the pachymetry value is known in humans, but has not been measured in this study. As is the case with humans, the central trajectory of IR illumination was accompanied by a central reflection that could disturb tissue analysis. This reflection was not systematic and was related to the perpendicular position of the incident light in relation to the corneal surface.
In all pathological cases studied, qualitative and quantitative analyses of the lesions were possible using SD-OCT. In the case of chronic superficial corneal ulceration, we were able to detect both epithelial detachment and hyperplasia, as well as the increased reflectivity of the anterior stroma. The analysis of bacterial keratitis highlighted the presence and the intensity of cellular infiltration, corneal edema, and tissue destruction. Measurements made on the different compartments of the cornea allowed us to monitor the development of these lesions over time. Our observations were similar to those performed in human ophthalmology for which SD-OCT is now a useful additional technique in the evaluation of bacterial keratitis.[21, 22] However, in some cases, the hyper-reflectivity of lesions limited the assessment of deep corneal areas, and as a result, evaluation of the entire stroma was not always possible.
Images of corneal dystrophy and degeneration were correlated with available histologic data by identifying areas of high reflectivity in the anterior stroma. We were able to measure the thickness of these high-density regions in vivo using SD-OCT. These observations were similar to human data produced from analysis of presumed calcified lesions.[24, 25] Precise localization in the subepithelial stroma, specific reflectivity, and evaluation of the deposits thickness provided a noninvasive confirmation of clinical diagnosis. The deepest stromal lesions, Florida spots, were also characterized by the presence of an infiltrating component without any epithelial modification. SD-OCT imaging of major stromal changes such as complete corneal edema and bullous keratopathy produced spectacular pictures showing profound changes in the stromal architecture that could not be evaluated by slit-lamp examination. What we call feline bullous keratopathy is named, in human ophthalmology, corneal hydrops, a condition associated with endothelial rupture or detachment. SD-OCT examination failed to show endothelial structures in bullous keratopathy because it was not technically possible to have the entire corneal thickness on the same scan, so results could, therefore, not be compared with human data. Similarly, when imaging endothelial lesions, the SD-OCT examination provided an accurate evaluation as long as the corneal thickness was <1300 μm, and thus, absorption was minimal. This limitation can be further investigated by imaging endothelial dystrophies and degenerations, which were not included in our study. This limitation could also be handled by the use of UBM that uses high-frequency ultrasound, which does not depend on corneal transparency. UBM can scan extremely thick corneas but with a lower resolution than OCT.
We also used SD-OCT for evaluation and follow-up of surgical procedures. The preoperative analysis of corneal sequestra in cats was not reliable. Our aim was to anticipate the choice of the surgical technique and of the prognosis by preoperative evaluation. In two cases, the size and position of the lesion were accurately measured, while in other cases, the reflectivity of the pigmented lesion produced a posterior shadow that made analysis of the deep stroma impossible. However, this lack of information about the depth of the sequestrum did not change either the treatment choice or the outcome of the surgical procedure. OCT was used after the keratectomy to measure the corneal thickness and to detect residual hyper-reflective areas. The shadow effect was also observed in the evaluation of corneal foreign bodies. Serial scans were required to overcome this difficulty and accurately quantify the depth of penetration.
Preoperative and post-operative evaluation during superficial keratectomies and conjunctival or biomaterial (such as porcine intestine submucosa) grafts was possible with SD-OCT. The thin graft material was transparent enough to allow unobstructed imaging of underlying structures and the measurement of their dimensions. These observations were similar to those made in humans where OCT is used frequently in the operative scope of corneal surgery to aid in choice of treatment strategy.[2, 5, 7, 11, 25, 28] Currently, more data are needed to support this kind of protocol, but we believe that increasingly widespread use of OCT will improve treatment strategies in the future. For example, the ability to accurately evaluate corneal thickness before or during the surgical procedure will allow the veterinary surgeon to use novel surgical tools like lasers or corneal cross-linking, which should only be used when minimal residual corneal thickness can be guaranteed.
The analysis of corneal wounds proved more difficult. Perforation areas were not always visible in SD-OCT because they required the beam to be oriented accurately along the axis of the wound. In addition, the IR absorption due to corneal edema limited the imaging of deep structures in most of the cases. In these conditions, the use of SD-OCT adds little to no value to the slit-lamp examination.
The SD-OCT images of healthy corneas in carnivores were comparable to images of the human cornea, where different layers are easily distinguishable. These results correlate strongly with a previous study performed on a smaller group. However, the SD-OCT image of Descemet's membrane and the endothelium fades as the thickness of the cornea increases.
The use of SD-OCT for corneal evaluation in normal and pathological conditions in dogs and cats opens an investigative field that is comparable to the existing field in human ophthalmology. In our study, the advantages of this technique were found in the ability of giving a qualitative and quantitative evaluation of all cornea layers, in vivo, in real time. The rapid acquisition time and the absence of contact with the ocular surface make the method insensitive to eye movements and compatible with the fragility of the cornea to be examined. Furthermore, the production of quantitative images and measurements allows us to monitor clinical situations in an objective manner, particularly in the context of microbial keratitis.
However, we encountered three types of difficulties in the application of SD-OCT to corneal evaluation. First, focusing must be very precise, and it was very difficult to use the device on conscious animals. This fact led us to use SD-OCT on sedated animals only, which added the necessity of permission, from the owners, for repeated sedations for follow-up imaging sessions. Additionally, due to the fine focus requirements, small lesions were not easily detected during the examination, as was the case for the corneal wounds and foreign bodies. Serial scans were required to obtain useable evaluation images in such cases. Finally, for certain disorders, opacity of corneal structures obscures image acquisition and interpretation, and thus, accurate analysis of corneal sequestra was difficult using SD-OCT. In these cases, SD-OCT presented the same limitations as slit-lamp examination. For highly edematous lesions, examination was limited by corneal thickness for which it was not possible, with our device, to scan the entire cornea. The use of UBM, which does not depend on corneal transparency but has a lower resolution than OCT, could help to evaluate such corneal conditions.
We were driven to complete this study because of the possibility of obtaining equipment that is compatible with veterinary practice. Here, we demonstrate the successful use of the SD-OCT technique for the imaging and evaluation of canine and feline cornea in clinical conditions for a wide range of corneal diseases. This technique is not intended to replace careful slit-lamp examination. The results and images presented here show that SD-OCT optical analysis has a resolution comparable to low-magnification histologic images, and the images obtained are in agreement with available clinical and histologic data in the literature. In most of the cases, images provided us with quantitative information that completed the slit-lamp examination. The major advantage of this technique is real-time, in vivo, contactless evaluation of animal corneal structures, and SD-OCT corneal evaluation in pathological and surgical conditions is very promising diagnostic tool for therapeutic decision-making and for follow-up of corneal healing.