Smartphones could become a more important tool for detecting eye disease, particularly in regions where access to ophthalmologists and specialist imaging equipment is limited. An international research team led by University Hospital Bonn (UKB) and the University of Bonn has reviewed the rapidly developing field of smartphone-based retinal imaging and concludes that the technology is already sufficiently mature to support selected screening and triage applications.
The researchers analysed 30 technical solutions, 274 scientific publications and other relevant sources. Their review, published in Progress in Retinal and Eye Research, covers not only image quality and diagnostic performance but also telemedicine, artificial intelligence, data protection, regulation and the potential role of non-physician healthcare workers. The findings point towards a model in which retinal imaging becomes less dependent on specialised ophthalmology facilities. But the researchers also stress that considerable work remains before smartphone-based fundus imaging can be implemented at scale.
Turning a phone into a retinal camera
Fundus photography produces images of the back of the eye, including the retina, blood vessels and optic disc. Conventional imaging generally requires dedicated equipment. Smartphone-based systems instead combine a phone's camera with optical adapters and illumination, creating smaller and more mobile alternatives. According to the review, some of these systems can already deliver image quality and a field of view suitable for particular screening and triage tasks. Evidence is especially relevant for diabetic retinopathy, glaucoma-related changes to the optic disc and selected applications involving retinopathy of prematurity.
That could be particularly valuable in areas where specialist capacity is scarce. Appropriately trained healthcare workers could capture retinal images locally and transmit them to specialist centres for remote assessment. The smartphone would therefore function as part of a telemedicine pathway rather than as a replacement for an ophthalmologist. This distinction is important. Smartphone imaging can potentially help identify people who need further assessment, but a photograph alone does not constitute a comprehensive eye examination.
AI could support image capture and assessment
Artificial intelligence could make this decentralised model more practical. An AI system could determine whether an image is of sufficient quality while it is being captured and check whether the required retinal area is visible. Algorithms could also select the most useful images from a sequence, combine multiple images and assist in detecting abnormalities. Such automation could become particularly relevant in large screening programmes, where the number of images can quickly exceed the capacity of specialists to assess them manually.
The concept is not purely theoretical. Smartphone-based eye screening is already being investigated in real-world settings. A recent study in rural India, previously reported by us, used a low-cost smartphone imaging system at 19 rural eye camps. Remote referral decisions based on smartphone images agreed with in-person assessments in 96 percent of cases. The researchers emphasised that the technology was intended for screening and triage rather than replacing comprehensive ophthalmological examinations.
The eye as a diagnostic window
Smartphone eye photography could eventually extend beyond diagnosing eye disease itself. Another study demonstrated how images of the eye could potentially help screen children for anaemia. Researchers from Purdue University, Rwanda Biomedical Center and the University of Rwanda captured more than 12,000 smartphone images from 565 children aged five to 15. Rather than photographing the retina, they photographed the conjunctiva: the inside of the eyelid and the white part of the eye. Machine learning and radiomics were then used to detect patterns associated with anaemia.
Notably, the approach used grayscale images instead of relying on colour. The researchers analysed subtle structural patterns in blood vessels, reducing some of the problems caused by differences in lighting and smartphone cameras. As with smartphone retinal imaging, the aim was not to replace established diagnostics. Conventional anaemia diagnosis requires blood testing, while the smartphone approach could potentially identify children who should be referred for further investigation.
Together, these developments illustrate how the combination of widely available cameras, specialised optics and increasingly sophisticated image analysis could turn smartphones into accessible screening instruments.
Evidence and standards still needed
Widespread implementation nevertheless presents significant challenges. Smartphone-based retinal imaging systems differ in image quality, field of view and compatibility with different phones. Regulation, light and device safety, cybersecurity and protection of medical data also need to be addressed. Crucially, it is not yet clear whether large-scale smartphone screening programmes are economically sustainable or ultimately improve treatment outcomes. The researchers therefore call for larger implementation and cost-effectiveness studies as well as minimum reporting standards.
The potential lies not simply in replacing an expensive camera with a cheaper one. Smartphone retinal imaging could enable a different care pathway: images captured closer to patients, assessed remotely and increasingly supported by AI. Whether that model can translate technical accessibility into earlier diagnosis and better outcomes will now require evidence from large-scale clinical implementation.
References
University of Bonn (research)
Add ICT&health on Google
Show more content from ICT&health in Google Search.
