Pr3vent: Revolutionizing Newborn Eye Screening with Machine Learning

• Analytics & Modeling - Computer Vision Software
• Analytics & Modeling - Machine Learning

• Cement
• Education

• Quality Assurance

• Clinical Image Analysis
• Computer Vision

• Data Science Services
• Training

Pr3vent, a Silicon Valley-based diagnostic company, was faced with the challenge of improving patient diagnosis and eye screening availability through computer-aided diagnosis. The company aimed to scale doctors’ expertise through AI, with the goal of reducing the per-screen cost for better accessibility to 4M infants in the US alone while increasing diagnosis accuracy. The challenge was to utilize the power of AI to combat preventable vision loss in infants. Due to the scarcity of trained doctors who can diagnose eye diseases by a newborn’s retina, the team’s vision was to marry Deep Learning and data to scale the expertise of ophthalmologists who can, to cut per-screen cost, increase accuracy, and improve screening availability. The solution needed to be highly accurate in detecting pathology in a newborn’s retina, to receive FDA approval. This required Pr3vent to accurately label a database of 350K fundus and retina images by a team of experienced ophthalmologists, build an AI-driven image analysis and anomaly detection engine, and develop an application for ophthalmologists to handle retina images.

Pr3vent is a Silicon Valley-based diagnostic company that specializes in building AI/ML-powered eye screening solutions. The company's mission is to detect and prevent ophthalmic conditions in infants. Pr3vent has exclusive access to a database of more than 350K retina and fundus images of newborns and has attracted some of the best ophthalmologists in the US to label the training dataset. The company's vision is to utilize the power of AI to combat preventable vision loss in infants by scaling the expertise of ophthalmologists, reducing per-screen cost, and improving screening availability.

To address the challenge, Pr3vent teamed up with Provectus to deliver an FDA-compliant eye screening solution. The solution was an ML-powered disease screening platform that processes, analyzes, and labels a wide range of medical images to detect pathology. It consisted of three components — to manually label and store images, to build and train ML models, and the app for physicians to check the results. The first step was to build an image labeling infrastructure that facilitated and accelerated labeling of medical images, enabling doctors to process up to 72 eye screens per minute. The second step was to build an FDA-ready AI infrastructure with fully auditable labeling, dataset management, model training, model evaluation, model release management, model inferencing, prediction explainability, and model monitoring components integrated into an end-to-end AI platform for healthcare. The final step was to develop a disease detection and diagnosis application for ophthalmologists to check how accurately pathology has been detected and classified by a given ML model.

• The ML-powered eye screening solution provided by Pr3vent has revolutionized the way newborn eye screening is conducted. The solution has not only made the process more efficient but also more accurate. With the ability to detect pathology in a newborn’s retina and fundus images with high accuracy, the solution is ready for FDA approval. If approved, Pr3vent’s solution could scale newborn screening in the United States and globally. By reducing cost per screening by 10x and by scaling ophthalmologists’ expertise through automation, it makes timely eye screening accessible to no less than 4 million babies in the US alone. This means that ophthalmologists can now focus on patients who require immediate assistance, instead of looking through dozens of normal images of healthy individuals. Given that up to 97% of eye conditions can be treated in the first few months of life, Pr3vent may significantly reduce the number of infants and adults suffering from vision impairment and vision loss.

• Achieved an accuracy of ~96% in detecting pathology in newborn eye retina

• Achieved a 10x reduction in cost per screening

• Made eye screening accessible to more than 4M infants in the US alone