By Azfa Razzaq

Ophtha Automatic Eye Screening or Ophtha AES is the only Sri Lankan project that is funded by the Ministry of Economy, Trade and Industry of Japan for 3 years continuously. The project has improved the security of user data  through end-to-end systems via user validation, encryption and decryption technologies. AES is a successful multi-party collaboration between Sri Lanka and Japan. This case study is compiled using the in-house experience gathered while developing the Ophtha AES.

What is Automatic Eye Screening?

Screening is a way of identifying healthy people who may have an increased risk of a particular condition. Vision screening, also called an eye test, is a brief exam that looks for potential vision problems and eye disorders. An ophthalmologist (a medical doctor who specializes in eye health) would perform an eye screening to identify different diseases related to vision impairment. Here, a retinal image of the eye (taken by a fundus camera), is examined thoroughly to detect eye conditions like Diabetic Retinopathy, Glaucoma, AMD, hypertensive retinopathy etc. 

Automatic Eye Screening is a way of identifying any eye disease using a software application, without the help of an eye specialist. The system examines the retinal image  and automatically generates reports related to the eye condition. The patient could use the generated report and get necessary treatments based on the outputs of the report.

Diabetic Retinopathy

Diabetic Retinopathy (DR) is one of the fastest growing causes of blindness in the world which occurs due to a complication of Diabetes. Approximately 463 million people in the world suffer from Diabetes and each one of them has a risk of going blind due to DR. It is caused by damage to the blood vessels of the light-sensitive tissue at the back of the eye (retina). The longer you have Diabetes and the less controlled your blood sugar is, the more likely you are to develop this eye complication.

In the beginning, Diabetic Retinopathy may exhibit no symptoms or only mild vision problems. However, if left untreated it can eventually cause complete blindness. The condition can develop in anyone who has been diagnosed with type 1 or type 2 Diabetes. Diabetic Retinopathy affects up to 80% of patients who suffer from Diabetes for over 20 years. Each year in the United States, Diabetic Retinopathy accounts for 12% of all new cases of blindness. It is also the leading cause of blindness in people aged 20 to 64. Of these cases, at least 90% of new cases could be reduced with proper treatment and monitoring of the eyes. The key to prevent blindness is regular screening (once a year as recommended by medical officers worldwide) because this condition is asymptomatic, i.e. the condition develops without any symptoms at first and the patient may get to a point of vision loss at the final stage of the disease which is irreversible.

DR can be diagnosed by visual examination of retinal images to detect most common pathological features such as microaneurysms, hemorrhage, exudates and neovascularization. This process is  manual and time consuming and the outcomes depend on the examiner’s perspective. Thus, there’s a potential bias in the examiner’s decision.Once the pathological signs are identified, these examinations can be used to identify  different disease stages such as  “No DR”, “Moderate DR’’, “Proliferative DR” etc.

Why AES for Diabetic Retinopathy?

With the increasing number of Diabetic patients in the world, the need for eye screening is growing worldwide. A shortage of ophthalmologists who are expertized in eye screening has become a major problem worldwide. In a US study, it was estimated that by 2029, around 20% of the American population will be over the age of sixty-five. Therefore, it has  been presumed that the demand for ophthalmology care in the next two decades will be higher than ever before because diseases like cataracts, glaucoma, macular degeneration and DR are all age-related eye diseases.

Furthermore, it is expected that the number of DR patients will rise up to 191 million by 2030. Even in rural areas of Sri Lanka, a shortage of doctors who are specialized in eye screening can be seen. Also, eye screening and having an appointment with a specialized doctor to diagnose the disease condition can be very costly. Even if the patient consults a doctor, the doctor’s decision can be inconsistent. I.e there can be a variability or disagreement about the different stages of DR (eg. No disease, moderate, severe etc.) from different doctors. Because of these issues, about half of people suffer from vision loss even before they are diagnosed with the disease condition. Hence AES with a very high accuracy is a much needed cost-effective solution to overcome this global situation.

Objective of Our Eye Screening Application

Effective Solutions developed the Ophtha AES application to detect DR during the period 2017 to 2020. The main objective of the system is to detect different stages of DR and to identify different pathological features of the disease. Even Though, the AES system was initially built for the portable fundus camera made in Japan, it can and will still be used for DR identification of retinal images taken by  third party fundus cameras as well.

How Does Our Eye Screening Application Work ?

Ophtha AES is an automatic eye screening application which mainly focuses on detecting DR. It is a cloud-based system where anybody can upload the retinal images of the eye/s to the system from anywhere in the globe and get the full report of the disease condition in a very short period of time. Image processing techniques of AI are used for the segmentation of different pathological features of the retinal image and deep-learning techniques in artificial neural networking have been used to identify different stages of DR.

What Distinguishes Our Product from the Other Automatic Eye Screening Applications ?

Even Though, several companies have done researches  on Automatic Diabetic retinopathy detection, these companies have only focussed on identifying the severity of DR using retinal images, i.e. the different stages of disease progress from “no DR to proliferative DR”.What distinguishes Ophtha AES software from the others is that our system not only detects the severity of the disease, but also detects pathological features of DR and marks their specific locations in the retinal image. For this, the system is trained with over 11,800 retinal images at several stages. The following images show how the pathological features are identified by the system. 

Our Achievements

By considering consultant ophthalmologists’ opinion as the golden standard in measuring accuracy, sensitivity and specificity, Ophtha AES has a sensitivity of 75% and specificity of 95% in pathological feature identification. Hence, the system has an accuracy of 97% in pathological feature identification. Also in identification of different stages of DR, currently Ophtha AES is 82% accurate. Ophtha AES was awarded as the first runner up in “Sirasa Kotiyak Watina Adahasak” competition and won the Spiralation 2021. The product research paper also was presented in the Conference of Healthcare in Fourth Industrial Revolution 2018.

Future Plans

Currently, the system is used to detect Diabetic Retinopathy condition only and the team is hoping to develop the system to identify other serious eye complications like Glaucoma and AMD in future. Also, it is in the pipeline to expand this solution worldwide and around different areas of Sri Lanka where there is limited access to eye screening through Ophtha Innovations, with affordable pricing and convenient access. 
You can now experience the Beta release of Ophtha AES via Home (ophthaes.xyz)