According to the World Health Organization (WHO), nearly 200 million people in Africa and the Americas are at risk of contracting the debilitating disease onchocerciasis, or river blindness. The parasitic infection spreads by blackfly bites, causing extreme skin irritation.
Treatments are available, but they can be logistically challenging to implement for remote populations or children. Left untreated, the infection can cause permanent blindness. To effectively eliminate the disease, a visionary solution is needed.
But clinical trials necessary to develop new treatments are time-consuming. Previously, only a handful of experts were trained to carry out the specialized—and manual—analysis of hundreds of complex images. Meanwhile, an estimated 20 million people are already affected by the disease. Capgemini wanted to help move the process forward, faster, using the power of artificial intelligence (AI).
Utilizing the scaling capabilities of Amazon Web Services (AWS), Capgemini hosted a hackathon with almost 1,000 participants. Teams created AI models that could review tissue samples provided by the Institute of Medical Microbiology at University Hospital Bonn, directed by Prof. Dr. med. Achim Hoerauf. The winning model generated close to human-level performance, or 90 percent accuracy, which could potentially shave months off the research process, reduce time-to-market for future treatments, and help more people before the disease advances.
Capgemini and University Hospital Bonn teams continue to develop the AI system, exposing the system to more images—and a greater variety of image quality—so the AI can learn what anomalies to ignore as much as what to look for. Within a few years, if teams are successful in minimizing disagreement between AI and human analysts, scientists predict the AI could eventually perform a majority of the review, with human experts verifying its work.
There is more to be done before the fight against river blindness is won, but with innovative technology doing some of the heavy lifting, a cure could be in sight sooner than anticipated.
