High resolution iridocorneal angle imaging system by axicon lens assisted gonioscopy

Direct visualization and assessment of the iridocorneal angle (ICA) region with high resolution is important for the clinical evaluation of glaucoma. However, the current clinical imaging systems for ICA do not provide sufficient structural details due to their poor resolution. The key challenges in achieving high quality ICA imaging are its location in the anterior region of the eye and the occurrence of total internal reflection due to refractive index difference between cornea and air. Here, we report an indirect axicon assisted gonioscopy imaging probe with white light illumination. The illustrated results with this probe shows significantly improved visualization of structures in the ICA including TM region, compared to the current available tools. It could reveal critical details of ICA and expected to aid management by providing information that is complementary to angle photography and gonioscopy.

without using any fluorescence stains. However, multiphoton sources (lasers) are very expensive and are not used in routine clinical procedures. In 2015, McNabb et al. demonstrated complete 360° circumferential gonioscopic optical coherence tomography imaging of the iridocorneal angle 8 . However, high resolution structural information of the TM region was not possible by this method.
In this article, we report the first demonstration of an integrated gonioscopy concept named as axicon assisted gonioscopy. The proposed concept and method uses gonioscopy imaging approach by integrating Bessel Beam Microscopy (BBM) concept 19,20 . The proposed concept is explained through optical simulation which is followed by experimental demonstration using the developed system. The imaged ICA region of porcine eyes indicates the efficiency of the system and its clinical significance. A Comparison of ICA images taken by the newly developed system with gonioscopy and EyeCam, demonstrate improved resolution in documenting the TM, pigmentation and ciliary body base.

Results
Optical design and simulation of gonioscopic system. Figure 1 shows the optical setup of the proposed simple indirect white light gonioscopic imaging system (the schematic diagram is provided in the Supplementary  Fig. S4). Here, the indirect gonioscopy imaging system is designed using a conventional gonioscopic contact lens (Hoskins-Barkan Goniotomy lens, adult 11 mm lens, Ocular instruments, WA, United States of America), as the imaging head. For optical simulation, a 3D model of gonioscopic lens is made using Zemax software by using the design specifications of Hoskins-Barkan Goniotomy lens. Figure 2 shows the 3D model of gonioscopic lens and eye. In Fig. 2, only the chief ray is illustrated. The parameters for the eye model and ray tracing of the imaging system are provided in the Supplementary material.
The results of 3D Huygen point spread function (PSF) and the PSF grid for gonioscopy imaging system integrated with and without axicon lens unit are shown in Figs 3 and 4. We performed the simulations with two point sources separated by 3 μ m. For the gonioscopy imaging system integrated with axicon lens unit, improvement in resolution is evident from Huygen PSF, which showed a distinct separation of the 3 μ m point sources. However, it is also evident from Fig. 3 that when axicon is not used there is clear overlap of the two point sources. The difference between the gonioscopy imaging system with and without the axicon lens unit was further compared with result of PSF grid simulation, as shown in Fig. 4.
For theoretical modelling, we used a plano convex lens (F = 25 mm) in combination with a gonio lens as the imaging head. However, for performing the experiments, we utilized LSM03 (4.6X, VIS OCT scan lens, Thorlabs, USA) instead of plano convex lens. Therefore, we further performed an experimental comparison of both systems by imaging a USAF chart through gonio lens. The result is shown in Fig. 4c. The USAF resolution chart 1951, group 7 element 4 (2.76 μ m) was resolved for gonioscopy system integrated with the axicon lens unit. On the other hand, the gonioscopic system without the axicon lens unit only resolved group 6 element 2 (16 μ m) (Fig. 4c). Thus, it is clearly evident from these results that the performance of gonioscopy system in terms of spatial resolution is improved with the integration of axicon lens unit. ICA image of demo eye. Functionality of the imaging system was first tested on an ocular imaging eye model (OEM1-7, Ocular Instruments Inc., Bellevue, WA, USA). Gonioscopic lens was placed onto the eye model interfacing with coupling gel (Vidisic gel). The result is shown in Fig. 5a. Figure 5b shows the image of the demo eye model. TM region of porcine eye visualized by the developed probe. The ICA of the porcine eyes imaged with our system is shown in Fig. 6. Figure 6a is the acquired raw image. Raw image has structures slightly false colored since the light source used to illuminate has yellow tint and not perfectly white. The processed image is shown in Fig. 6b. It is evident that the different regions in the ICA of the porcine eye are distinctly visible in the image. Close to the iris, there is a simple organization of ciliary body base circumferentially distinct in a discontinuous bundle like pattern. In this zone, fibers are longer and reach the sclera and the scleral spur. It is also evident that the ciliary body is highly pigmented. This observation is corroborated by the earlier studies which indicated higher pigmentations in porcine ciliary body as compared to ciliary body of human eyes 21 .
The TM region is very distinct in the image (Fig. 6b). However, TM meshwork structures are further highlighted by performing edge detection on the image. This result is shown in Fig. 6c. Pigmentation in the angular region is also clearly evident in Fig. 6b. It is important to note that the imaging was performed on enucleated porcine eyes, and we therefore assume the image quality will be even better for live animal due to less or no corneal haze.
Comparison with other direct ICA imaging methods. Figure 7 shows the comparison of ICA of Porcine eyes imaged using the newly developed system with the state of the art imaging systems such as EyeCam and Gonioscopy. The amount of clouding (haze) of the cornea is known to affect the image quality drastically. Due to this reason, imaging was performed on clear and mildly cloudy porcine eyes with all the imaging systems, for comparative purposes. It is evident from the Fig. 7 that, irrespective of the different imaging systems used, the structures of the ICA imaged for clear eyes were more distinct than cloudy ones. The details of the EyeCam images were better than gonioscopic images. Wide field camera of the EyeCam also allows to get a larger view of iridocorneal angle region as compared to gonioscopy images (Fig. 7). However, it is clear ICA imaged with our system is much better than EyeCam and gonioscopy. Notably, even for cloudy eyes our system showed better image clarity (Fig. 7j). Moreover, fine structural details of pigmentation, TM, ciliary body can be visualized by our system. Further, we also compared the results with an integrated flexible handheld wide angle probe which was recently developed in our research centre for imaging and documenting ICA 17 (Fig. 7g,h). Though the probe is flexible and can image large area using an endoscopic CCD camera as compared to other imaging systems, the resolution is poor compared to the proposed new axicon assisted gonioscopy. All the above mentioned results confirms the improved quality of ICA imaging which can be achieved by our system as compared to the state of the art systems as well as those reported in literatures. Illustration of pigmentation in the ICA with clarity. Assessment of pigmentation in the TM is of major significance to glaucoma diagnosis [22][23][24] . (Porcine eyes are generally known to have higher pigmentation than human eyes 25 .) The pigmentation pattern of an individual angle is dynamic over time. Heavy pigmentation of the TM should indicate pigment dispersion or exfoliation syndrome. In addition, a line of pigment deposition anterior to the Schwalbe line is often present in exfoliation syndrome (Sampaolesi line). Moreover, studies have also revealed that pigmentation can affect or introduce challenges for glaucoma treatment such as laser trabeculoplasty 26 . Other conditions that cause increased anterior chamber angle pigmentation include malignant melanoma, trauma, surgery, inflammation, angle closure, and hyphema.
However, there are only few relevant studies which statistically analyzed the amount of pigmentation in the TM region [27][28][29] . Currently available imaging systems can neither provide new insight nor quantify the pigment level of TM 27 . For this reason, we have demonstrated the potential of our system by illustrating better quality porcine eye images with different pattern of pigmentation. The result is shown in Fig. 8. It is evident from Fig. 8, that the pigmentation have better contrast. We believe the high contrast imaging of ICA region would be useful to perform further statistical studies to get quantitative information regarding pigmentation.

Discussion
In vision sciences research, porcine eyes were often used as an ex vivo animal model because of similarity in its morphology to the human eye 30,31 . In addition, porcine eyes are also generally used as animal model eyes for validation in a variety of studies such as R & D on glaucoma 32 , in neuroretinal studies 33 , cataract surgery 34 , corneal transplant studies 35 and in aberrometry studies 36 . In our study, we have used the porcine eye as the test eye samples because of its resemblance to the human eye and due to easy availability.
ICA imaging systems must have resolution in the range of 1 to 5 μ m 16 to resolve TM region. However, the existing clinical white light gonioscopic systems do not have this above mentioned spatial resolution to resolve the structural details of TM. This key drawback limits their efficiency to visualize critical details of ICA.
In summary, we have demonstrated for the first time to the best of our knowledge, an axicon assisted white light indirect gonioscopic system which can image the ICA with spatial resolution down to 3 μ m. This imaging system is thus capable of direct imaging of ICA with better structural clarity as compared to current imaging system such as gonioscopy, OCT and EyeCam. Our gonioscopic system may pave the way for future ICA imaging.

Methods
Preparation of Sample. Enucleated porcine eyes were obtained from abattoir. To avoid corneal haze which would reduce image quality, porcine eyes were imaged within 5 to 6 hours of death. The ex vivo samples were   transported on ice (approximately − 4 °C) to the laboratory to maintain their freshness. Before imaging, the extraocular tissues such as conjunctiva and lacrimal gland were removed from the samples. Each sample was fixed on to a custom eye holder, which was mounted on a translational stage with micrometer accuracy. A sample size of 50 porcine eyes were used for the study. The experiment was conducted in accordance to the regulations of Agri-Food & Veterinary Authority of Singapore and Nanyang Technological University's regulations on biosafety.
Gonioscopy. Hoskins-Barkan Goniotomy lens (adult 11 mm lens, Ocular instruments, WA, USA) was used in this study. The thickness of the gonioscopic lens is 1 cm. The gonioscopic lens was fixed to an optical post and secured to optical table. A thin layer of ophthalmic gel (Vidisic Gel, Bausch & Lomb, NY, USA) was applied to the gonioscopic lens before placing it on the eye, to provide lubrication and to reduce the refractive index mismatch. The porcine eye (with cornea in centre) was then placed gently on to the developed gonioscopic lens based imaging system to image the TM region in the ICA.
Additionally using Latina 5 Bar SLT, (single mirror, Ocular instruments, WA, USA) static gonioscopy was also performed on porcine eyes under white light. Vidisic gel was used as the lubricant. Gonioscopy was performed on clear as well as mildly cloudy porcine eyes to compare the image quality. The images were recorded using a colour CCD camera.
EyeCam. EyeCam (Clarity Medical systems, Pleasanton, CA) imaging was also performed on porcine samples (n = 4). The images were taken by a trained technician at Singapore Eye Research Institute, who is specialized on EyeCam imaging and had basic knowledge of angle anatomy and structures. Using foot pedal, illumination was adjusted to avoid overexposure. Clear still images of the ICA were then recorded with the probe being placed at the opposite limbal region.
In-house developed handheld imaging probe for iridocorneal angle. The flexible handheld probe has an imaging sensor located at the central axis of the probe and has a variable resolution at different depths which is optimized for recording the ICA of the eye 17 . The imaging probe was placed near the limbal region of the porcine cornea to image the opposite ICA through vidisic gel. The imaging probe has LEDs integrated to provide illumination.
Optical Setup. In the gonioscopic setup shown in Fig. 1, a white light source (Correct Shimadzu FA-150EN Fiber Illuminator, Japan) was used for illumination. (3D model and the schematic diagram is also provided in the Supplementary Fig. S4.) The imaging head is a gonioscopic lens (Hoskins-Barkan Goniotomy adult 11 mm lens; Ocular Instruments, WA, USA). Through the gonioscopic lens, ICA region of the eye was imaged by a microscopic objective lens (LSM03 -4.6X, VIS OCT scan lens, Thorlabs, USA). The long working distance of the objective lens (25 mm) and it's large field of view enabled the integration of the microscopic lens with a gonioscopic lens. The light from the microscopic objective was then collected via a tube lens (ITL 200, Thorlabs). A convex lens (LB-1844-A, Thorlabs) was placed at a focal length distance from the focal plane of the tube lens. An axicon lens was placed in close proximity of convex lens. The combination of convex lens and axicon lens was known to transform the wavefront of a point source into a Bessel beam, which has the special property to propagate without diffracting. A beam passing through axicon lens was captured by the CCD camera (EXi Aqua Bio-Imaging Microscopy Camera, QImaging, Canada). Lateral resolution of the imaging system is around 3 μ m. Further, optical ray tracing of the imaging system is detailed in the Supplementary Fig. 3. Additionally information about basic BBM configuration is further detailed in the Supplementary material. Acquired images were post processed using ImageJ software. In order to enhance the image quality, noise reduction filtering and thresholding was applied to the captured images. Additionally, edge detection was also performed to highlight the structures in the images.