Diabetic Retinopathy (DR) is a widespread disease and a leading cause of preventable blindness in the adult working population.
In 2020, the number of adults worldwide with DR was estimated to be 103.12 million1. In addition, there is a continuous increase in type II diabetes which is often accompanied by DR, with only 40% of diabetic patients obtaining their annual recommended screening2.
Whilst this diabetic population continuously grows, ophthalmology is facing a workforce crisis, with the number of ophthalmologists retiring each year superior to the new entering the field, rendering impossible to meet the DR screening needs of diabetic and DR patients.
Therefore, there is a need to improve the screening of DR both for public health and economic benefit. Many other challenges need to be addressed among which the fact that many countries have poor resources for DR screening and the patient adherence is low.
But artificial intelligence (AI) can help in a robust way. Deep learning-based methods show high accuracy in screening and are more robust to human error. AI, together with telemedicine, can also help when it comes to increasing patient adherence.
Fundación Ver Salud, together with RetinAI, developed an algorithm, LuxIA, to optimize the screening of DR from primary care consultations and thus speed up its diagnosis and treatment.
In this way, clinicians have better access to automated retina screening methodology for diagnosis and treatment, aimed ultimately at reducing the incidence of blindness caused by DR.
Evaluate the performance of the LuxIA algorithm in screening patients for mtmDR using a single 45 degree color fundus images from Topcon machines.
Fundus data was collected from five sites across Spain.
Clinical data collection was done through the RetinAI Discovery platform to ensure the reviewers were blinded to each other's grades, as well as patient demographic information. Electronic case report forms were used to collect patient data, image quality information, as well as DR grades.
RetinAI Discovery is a CE-marked medical device according to the Medical Devices Regulation (EU) 2017/745
and the AI models are CE-marked devices according to Medical Devices Directive 93/42/EEC
RetinAI Discovery® is a 510(k) FDA Cleared medical device in US.
RetinAI Discovery® and Retinai® are both trademarks of RetinAI Medical AG.
Discovery and AI Models are CE-Marked medical devices since April 2020 & December 2020 respectively.