Singapore Eye Lesion Analyser (SELENA+) is a deep-learning AI software system that can detect potential threatening eye conditions accurately and efficiently. It can detect abnormal fundus images, and in particular, signs of diabetic retinopathy disease. The use of SELENA+ to augment the Singapore Integrated Diabetic Retinopathy Programme (SiDRP) will potentially reduce the workload by up to 50 per cent, with patient results made available in minutes instead of hours or days. This enables resources to be optimised to support other areas such as formulating personalised care plans for patients.
SELENA+ at a glance
Diabetic patients have retinal photographs taken once a year to detect signs of diabetic eye disease, a condition which can lead to blindness if not treated early. Over 100,000 patients undergo eye scans each year under the Singapore Integrated Diabetic Retinopathy Screening programme. This number is expected to double by 2050. Instead of a human being assessing these scans, SELENA+ can do the job.
SELENA+ analyses retinal photographs using a deep-learning system to detect three major eye conditions:
• diabetic eye disease
• age-related macular degeneration
In studies conducted to date, the AI-powered image reader has proven to be faster and as accurate compared to human graders. For patients, this means earlier, more targeted treatment, cheaper medical bill and better quality of life. The use of the AI system also allows clinicians to reach out to patients earlier, make better decisions supported by AI and manage patient condition more efficiently.
Building on SELENA+’s capabilities, a predictive risk assessment model for cardiovascular diseases will be developed based on the analysis of retinal images. Changes in the blood vessels in the retina can provide useful information on the individual’s cardiovascular status and risks. This will help identify patients with chronic conditions who are at high risk of getting a heart attack or a stroke.
Beyond SELENA+, enhancing the ability to predict cardiovascular diseases
AI to identify and predict your risk of cardiovascular disease
We will build on SELENA+’s capabilities to create an AI model that can be applied to retinal images.
This is a multi-prong approach which includes
- identifying imaging biomarkers from eye scans
- developing predictive algorithms from existing health data
- creating an AI engine to analyse genomic data, medical scans, clinical data and lifestyle information to create a personalised risk score for individuals
AI to create personalised treatment plansBased on the risk score, AI will be used to support clinicians in developing a personalised care plan for at-risk patients and monitoring their progress in the type of medication they have to take and the lifestyle changes they would need to make.
AI to empower patients to better manage their chronic diseases
Through the use of a wearable device or a mobile application, AI will give patients real-time friendly nudges to comply with their care plan. Some examples:
- Remind them throughout the day of the lifestyle changes they need to make
- Prompt them to adhere to medication schedules
- Urge them to participate in the necessary screening tests