Article Text
Abstract
Digital fundus imaging is used extensively in the diagnosis, monitoring and management of many retinal diseases. Access to fundus photography is often limited by patient morbidity, high equipment cost and shortage of trained personnel. Advancements in telemedicine methods and the development of portable fundus cameras have increased the accessibility of retinal imaging, but most of these approaches rely on separate computers for viewing and transmission of fundus images. We describe a novel portable handheld smartphone-based retinal camera capable of capturing high-quality, wide field fundus images. The use of the mobile phone platform creates a fully embedded system capable of acquisition, storage and analysis of fundus images that can be directly transmitted from the phone via the wireless telecommunication system for remote evaluation.
- Diagnostic tests/Investigation
- Imaging
- Retina
- Telemedicine
Statistics from Altmetric.com
Introduction
Since the late 19th century, when retinal imaging was first described, there has been steady technical improvement in imaging the fundus of the eye.1 Recently, this has included the development of digital imaging, non-mydriatic imaging systems and optical coherence topography imaging systems.2 These technological advancements have enabled clinicians to use a number of different imaging methods to detect, diagnose and monitor retinal diseases.
One limitation of most of the current imaging systems is the bulky and stationary nature of the equipment. Current gold standard tabletop fundus cameras require patients to be seated upright, which can be difficult in sick and hospitalised patients.3 Portable cameras have been introduced for fundus imaging,4 ,5 but these devices are often costly and typically must be connected to a computer for image processing, storage and visualisation. To overcome these limitations, we have developed a mobile phone-based retinal camera that leverages the compact size, high-resolution camera, large data storage capacity and wireless data transfer capabilities of current mobile devices to capture diagnostic retinal images. The camera includes a custom-designed mobile phone attachment that houses optics capable of capturing a retinal field-of-view of approximately 55°, which we call the Ocular CellScope. The device allows a completely portable and inexpensive solution for retinal imaging that could be used in the hospital setting and also for community vision screening.
Materials and methods
The Ocular CellScope is comprised of a mobile phone, a housing that contains the illumination and collections optics, and an integrated phone holder that ensures alignment of the optics with the camera on the phone (figure 1A). The acrylonitrile butadiene styrene plastic housing was designed using computer-aided design software and constructed using a 3D printer for use with an iPhone 4S mobile phone (figure 1B). The phone requires no modification, and the Ocular CellScope attachment is configured such that the user can easily slide the phone in and out of the holder. The rubber cup rests on the orbital rim of the sitting or supine subject, providing user-controlled stabilisation.
Imaging
The retina is imaged through a 54-dioptre ophthalmic lens. The ophthalmic lens forms an intermediate image that is relayed by a 20 mm focal length achromatic lens to the 8-megapixel camera of an iPhone 4S. The auto-focus mechanism of the iPhone camera is used to correct for the variability in axial length and refractive error in subject eyes. Images are captured using the native camera application. The Ocular CellScope in its present form does not have a built-in fixation method, requiring the subject to use the contralateral eye for fixation.
Illumination
Illumination of the retina is provided by three white light-emitting diodes (LEDs) controlled independently from the iPhone by an on/off dimming circuit powered by two 9-volt batteries (figure 1A). A 25 mm focal length collector lens is placed in front of the LEDs followed by a ground glass diffuser, a plastic linear polariser, a matte board annular mask with a 7.0 mm inner diameter and a 15 mm outer diameter, and a 50 mm focal length condenser lens. A polarising wire grid beam splitter is placed at a 45° angle to reflect the illumination towards the eye. The properties of the beam splitter are such that it preferentially reflects light that is polarised parallel to its vertical axis, and transmits light that is polarised parallel to its horizontal axis. This allows vertically polarised light to be reflected towards the ophthalmic lens, which then forms an image of the annulus with a 4.5 mm inner diameter and a 9.6 mm outer diameter near the pupil of the eye. The image of the annulus blocks light at the central region of the pupil to decrease reflection off the central cornea. The unmasked light passes at the periphery of the pupil and defocuses to form diffuse, uniform illumination on the retina. As the light reflects off the retinal surface, it is depolarised, creating light in the vertical and horizontal axes relative to the beam splitter. Light that is parallel to the horizontal axis of the beam splitter is preferentially transmitted through the beam splitter. The polarised light then travels through a second polarizer that is oriented parallel to the horizontal axis of the beam splitter (perpendicular to the first plastic polarizer), and through the relay lens to the iPhone camera. This cross-polarisation technique is important for limiting reflection artefacts from the ophthalmic lens and surface of the eye.
Results
The mobile phone-based retinal camera presented here enables the capture of fundus images using off-the-shelf components coupled to an unmodified, commercially available mobile phone. When used through a dilated pupil, the device captures a field-of-view of approximately 55° in a single fundus image. The image is captured on 2652×2448 pixels of the camera sensor, resulting in approximately 48 pixels per retinal degree. This surpasses the minimum image resolution requirement of 30 pixels per degree described by the UK Nation Health Service for diabetic retinopathy screening.5 If used continuously, the device's battery life is approximately 2–4 h in the current design, with the LED light source as the primary power consumer. Based on initial testing with the device, the time required to capture 10 photographs from a single subject is approximately 5 min.
Images captured by the mobile phone-based retinal camera could be stitched together using i2k Retina software (DualAlign LLC, Clifton Park, New York, USA) to create a composite image that captures a larger field-of-view of the retina (figure 2A). This mosaic compared well with a mosaic created using images taken with a TRC-50EX retinal camera (Topcon Medical Systems, Oakland, New Jersey, USA) when contrast and exposure were similarly scaled (figure 2B,C). To assess the potential for the device as a telemedicine tool, diagnostic quality images of diabetic retinopathy (figure 3A) and active cytomegalovirus retinitis (figure 3B) were captured from dilated patients in Thailand and transmitted directly from the mobile phone to a secure server. These images were of sufficient quality to enable the remote ophthalmologist in the USA to accurately provide a real-time diagnosis of the retinal diseases.
As demonstrated by the blue artefact in figures 2 and 3, the cross-polarisation technique reduced, but did not eliminate the reflection from the back surface of the ophthalmic lens. In addition, the cross-polarisation technique used by the device has been shown to increase the visibility of the choroid, optic disc, and blood vessels and also accentuate nerve fibre layer defects by reducing nerve fibre layer reflectivity.6 ,7 However, cross-polarisation also decreases specular reflected light from the internal limiting membrane that can be helpful for photography of certain retinal pathology.6
Discussion
Mobile devices are beginning to play a central role as medical diagnostic tools. Taking advantage of the portability, data storage capacity and wireless connectivity of mobile phones, it is conceivable that a mobile phone-based retinal camera could soon play an important role in hospitals and clinics. One example is the iExaminer (Welch Allyn, Skaneateles Falls, New York, USA), which can capture retinal images with a mobile phone, but over a limited field-of-view (25° maximum). We have used off-the-shelf components to develop and demonstrate a retinal camera based on a mobile phone that is capable of wide field imaging, enabling convenient and high-resolution diagnostic imaging for a broad set of applications (table 1). The readily available Bluetooth and Wi-Fi connectivity available on contemporary smartphones could allow seamless integration into an electronic medical record. Given the relatively low cost of the Ocular CellScope, the device may be most promising for use in resource-poor settings, where it could improve the accessibility of retinal screening programmes.
Acknowledgments
The authors would like to thank Neil Switz and Clay Reber for helpful discussions, and David Clay, Clay Reber, Ragi Maamari and Salim Saikaly for assistance with data collection. Collaboration was conducted with Somsanguan Ausayakhun, MD in Chiang Mai, Thailand.
Footnotes
-
Contributors RNM, JDK, DAF and TPM designed the Ocular CellScope. RNM constructed the device. RNM, JDK and TPM performed image acquisition. RNM, JDK, DAF and TPM wrote the paper.
-
Funding University of California, Berkeley Blum Center for Developing Economies and That Man May See, San Francisco, CA provided seed funding for this work. They had no involvement in the device design; the collection, analysis and interpretation data; in the writing of the report; and in the decision to submit the paper for publication.
-
Competing interests RNM reports a grant from Research to Prevent Blindness during the conduct of the study. RNM and the University of California have pending intellectual property related to the camera described in this manuscript. At the current time this intellectual property has no financial value. DAF reports a grant from the UC Berkeley Blum Center for Developing Economies during the conduct of the study. DAF and the University of California have pending intellectual property related to the camera described in this manuscript. At the current time this IP has no financial value. He is also a cofounder of CellScope, which is a medical device start-up using mobile phones for disease screening and diagnosis. CellScope was not involved in this research. TPM reports a grant from That Man May See during the conduct of the study. TPM and the University of California have pending intellectual property related to the camera described in this manuscript. At the current time this intellectual property has no financial value.
-
Ethics approval The Committee on Human Research, University of California, San Francisco.
-
Provenance and peer review Not commissioned; externally peer reviewed.