A major leap forward in Diabetic Retinopathy (DR) screening

Diabetic retinopathy is a complication of diabetes and is one of the leading causes of blindness among adults. If left untreated, it can lead to sight loss with the devastating impact on individuals and their families. By promptly identifying and treating the disease, these effects can be reduced or be avoided completely.

It is estimated that 30 to 50% of people with diabetes have a form of DR. However, its development often goes unnoticed until it reaches an advanced stage and vision loss occurs.

Early detection and systematic screening for DR is therefore paramount and should be an essential component of diabetes care, integrated within routine diabetes care. However, that can only be realistically implemented if the screening device is efficient enough in terms of workflow, processing time and accuracy.

Evolucare Imaging OphtAI: fast, accurate DR screening

Evolucare Imaging OphtAI (EIO) was designed to significantly facilitate the organization of DR screening and the follow up of patients through time.

EIO is a comprehensive modular solution which results from three years of collaboration between artificial intelligence experts, prominent eye care professionals and a leader in medical workflow software development. It relies on an Artificial Intelligence (AI) algorithm called OphtAI and Evolucare core imaging workflow solution. EIO allows to manage the patient through the entire examination workflow, from admission to invoicing, and detects if the patient suffers from DR.

Large scale screening is made easier

Hosted securely in the cloud, any type of eye care establishment can access EIO: public hospitals, private clinics, reading centers, small practices, permanent or temporary screening organizations. Through secure communication channels, any of these institutions can send their retina images for analysis and benefit from the powerful engine of EIO in less than 3 seconds.

Decentralized organizations can centralize their data and provide consistent analyses throughout their network of points of care. If connected to an EHR or EMR, they will be able to build a structured and consistent history for the patient.

An all-win solution

 

  • Larger patient reach: fosters large scale, mobile screening
  • Lower screening cost
  • Increased prevention: early detection of DR
  • Useful information for diagnosis with lesions mapping
  • Time saving
  • Focus on actual cases where expertise is needed
  • Low to no footprint
  • Modular solution
  • Adaptable to any site
  • Easy and secure web access
  • Early detection of DR
  • Improved care : systematic, regular screening now affordable
  • Consistent history of exam
  • Time saving

About our Artificial Intelligence algorithm: OphtAI

The algorithm, OphtAI, is a deep-learning AI which provides recommendation on the referability of the patient and the on the DR grading. In less than 3 seconds, OpthAI will analyze the retina image and provide an assessment of the quality of the image, the mapping of the aneurysms and an assessment of the referability of the patient. Depending of his/her authority, the reader will also be able to complete the report with the patient clinical context, other symptoms he detected and the recommended course of action.

The AI was trained on close to 350 000 retina images provided by an actual telemedicine eye care network managed by AP-HP and tested with several databases, including Messidor1. In addition, presence of DR was annotated by three ophthalmologists and the ground truth was defined by a majority vote. ROC curves for the referable DR detection analysis are reported in the graph below.

  1. Methods for Evaluating Segmentation and Indexing techniques Dedicated to Retinal Ophthalmology

Receiver-operating characteristics for OphtAI

With a specificity of 87.0%, OphtAI has a sensitivity of 99.0%.

With a sensitivity of 96.8%, OphtAI has a specificity of 90.2%.

With these operating points, OphtAI provides unmatched efficiency for DR screening: less pathological patients are sent home undetected, less healthy patients are erroneously referred to an ophthalmologist..