This open-source device is a shot in the arm for diabetic retinopathy diagnosis

Did you know that close to 30% of diabetics — both Type 1 and Type 2 — suffer from diabetic retinopathy or damage to blood vessels in the retina that can result in loss of vision? Half the diabetic population is unaware of this risk. With the disease taking on epidemic proportions in India — it affects over 65 million Indians and the number is estimated to rise to 109 million by 2035, according to this 2015 report — this is a cause for concern.
The prohibitive cost of retinal imaging devices — in the range of $10,000-$25,000 — makes them inaccessible people in rural areas or developing countries, a major hindrance in diagnosis of diabetic retinopathy. One of five diabetics in India suffer from diabetic retinopathy.

A Hyderabad-based team of doctors and engineers from Srujana Innovation Centre at the L V Prasad Eye Institute (LVPEI), has come up with a low-cost, open source device that could address this problem. The device, called an Open Indirect Ophthalmoscope (OIO), uses a Raspberry-Pi board and can be used by clinicians to capture good-quality retinal images for diagnosis of diabetic retinopathy.

The device: Affordable and 3D-printing enabled

An ophthalmoscope is a device used to examine the interior surface at the back of the eye. This examination is essential for diagnosis and early detection of everything from glaucoma to diabetic retinopathy and also helps in detecting brain tumors.

As with most medical equipment, ophthalmoscopes tend to be expensive, making them inaccessible to a majority of the population, especially in developing countries.
The OIO designed by the Srujana team will cost under $400 to build, making it easy to reproduce and, thereby, be more accessible to people from developing countries.

They started work on the project in April 2016 with just the aim of seeing the retina with a Raspberry Pi camera. After getting good results, they made a more compact version of the device and housed it in a compact box. The OIO is fitted with a rechargeable battery backup (any smartphone charger can be used to charge it) and can run continuously for at least six hours.
The OIO was designed to be open-sourced, as this allows it to be developed, debugged and rethought by others across the globe. “The idea behind making it an open-source design is to make it more accessible to people and doctors. Anyone can get the parts and assemble the device themselves,” says Sandeep Vempati, product engineer of OIO.

Also, other than the electronics and optics, most of the components required for building the device can be 3D printed. The schematics and specifications for making the OIO have been made available on the internet for people to be able to reproduce.

How it works

Once the pupils of the patients’ eyes have been dilated using a medicated eye drop, the device is placed on their face across their eyes at the correct distance, determined by their vision. After this, retinal images are taken by the camera module with the help of a 3-Watt LED. At present, the OIO is capable of taking a nearly 40-degree retinal picture.
The images thus taken are saved on the device and can also be uploaded to a computer for diagnosis. They are then run through a machine learning algorithm to help grade the diabetic retinopathy on a score of severity from zero to four.

What next

The team has already developed and deployed two OIO devices at the L V Prasad Eye Institute, and is currently testing and evaluating the sensitivity and specificity of the devices. They are also developing a device that can be used for undilated examination as that is more convenient for the patient.