RESEARCH TRIANGLE PARK – Researchers from RTI International and the University of North Carolina Chapel Hill released a new application that uses machine learning technology to detect trachomatous trichiasis.
Trachomatous trichiasis, according to RTI, is the painful end stage of the neglected tropical disease (NTD), trachoma, that can lead to blindness if not promptly diagnosed and treated with surgery.
January 30 marks World NTD Day, which Rebecca Flueckiger, associate director of Monitoring, Evaluation, Research, Learning and Adapting (MERLA) at RTI, told WRAL TechWire is a day that is “intended to raise awareness on neglected tropical diseases and the progress to eliminate them globally.”
WRAL TechWire spoke with Flueckiger about the new technology, which the organization found to be 92% accurate, and about trachoma. A lightly edited transcript of that conversation appears below.
WRAL TechWire (TW): Tell us more about trachoma and World NTD Day.
Rebecca Flueckiger, RTI (Flueckiger): These viral, parasitic, and bacterial diseases cause severe disability among more than 1.7 billion people around the world. World NTD Day is an opportunity to celebrate the amazing achievements toward controlling and eliminating these diseases – to date, 43 countries have eliminated at least one NTD – and highlights the need to keep the momentum going. To address these diseases we must work together, secure resources and facilitate political leadership.
TW: What’s the latest available data say with regard to trachoma prevalence, in the United States and globally?
Flueckiger: Trachoma is the leading infectious cause of blindness worldwide and is currently a public health problem in 44 countries. Trachoma used to exist in the United States. However, it disappeared from North America because of improved living conditions. Trachomatous trichiasis (TT) is the painful end stage of trachoma, that can lead to blindness if not promptly diagnosed and treated with surgery. It can result in scarring that turns the eyelid inward and leads the eyelashes to scratch the eye. Trachoma is responsible for blindness or visual impairment of an estimated 1.9 million people. Most individuals with trichiasis live in rural areas of low- and middle-income countries where health care access is limited and few, if any, ophthalmologists are available. The odds of women having trichiasis is 2-4 times higher than men, and the social factors associated with it perpetuate a devastating cycle of poverty among afflicted families.
Trachoma is targeted for global elimination as a public health problem by the year 2030 and there is an effective strategy for fighting trachoma, called SAFE (Surgery, Antibiotics, Facial cleanliness, and Environmental improvement). Our aim is to provide an accurate and user-friendly app that can be scaled and used in a variety of settings. It is our mission to ensure that all those who develop trichiasis can be identified and informed of surgical services to manage this painful condition and prevent progression to blindness.
TW: Describe the research conducted at RTI and UNC CH that led to the application.
Flueckiger: Trachoma programs rely on highly trained doctors and community members to go door-to-door to identify the disease, and as we get closer to elimination, the resources needed to find and treat each case increase.
Dr. Emily Gower, associate professor of epidemiology at UNC-CH, approached RTI with an idea to develop an easier, more cost-effective way to identify individuals impacted by trachoma. Given RTI’s and UNC-CH’s long history working with neglected tropical diseases, the opportunity to collaborate is always exciting. Dr. Gower and her team are known for their innovative and applied technologies for advancing efforts to eliminate NTDs, having previously spearheaded the development of a high-fidelity, low-cost surgical simulator to improve training for surgeons performing trichiasis surgery (HEAD START) .
We teamed up with Dr. Gower, who led this work with financial support from RTI and UNC-CH through the RTI University Scholars Program. Dr. Juan Prieto and Ms. Hina Shah from UNC-CH joined the research team, bringing extensive experience using machine learning methods to develop technology for low-resourced settings. The team includes researchers with a range of expertise and experiences from both UNC-CH and RTI, who leveraged machine learning technology to develop an algorithm that assesses high-quality photos of eyelids to detect trichiasis. We then integrated this algorithm into a new smartphone app called the “TT Screener.”
TW: Why a smartphone app — and how does it work?
Flueckiger: We decided to develop a smartphone app because mobile smartphones are readily available and easy to use in trachoma endemic countries.
The “TT Screener” technology is stored on the phone and does not require consistent access to a mobile network to run, which means it can be used in many of these remote communities where mobile networks or Wi-Fi might not be available.
The simple app interface prompts the user to enter a patient identifier and then opens the phone’s camera, prompting the user to take a picture of the patient’s eyelid. After the image is captured, the app runs the algorithm to assess presence or absence of trichiasis.
The current processing time to return a screening result following image capture is 13 seconds. The algorithm currently has an estimated 92% accuracy rate. Additionally, through our prior research and early testing on this project, we have shown that lay community members without prior experience with a smartphone can easily be trained to use smartphones to capture good quality pictures.
TW: 92% seems high — how high is this really, comparatively? What would increase to higher success rate?
Flueckiger: Disease identification is a key component for reducing the burden of NTDs. Many diseases require screening throughout the community, which necessitates training community members how to identify the disease in question. This process can be costly, and often does not yield optimal outcomes. For example, evaluation of trichiasis case finders shows that the accuracy of appropriate case detection can be as low as 15-30%. So, yes 92% is very high and could greatly impact the trajectory of trachoma elimination.
TW: What’s RTI’s role?
Flueckiger: RTI leads USAID’s Act to End NTDs | East program, which supports the national governments of 12 countries to reach their control and elimination goals through proven, cost-effective public health interventions. Through this large-scale program, we work regularly with ministries of health to understand the unique challenges they face in eliminating trachoma, and this experience has helped inform the practical design of the app and how it would realistically be used in practice. In addition to funding from the RTI University Scholars Program for Dr. Gower’s time on the study, RTI also contributed funding as part of a $3 million research investment to support global NTD control and elimination efforts and fill gaps in knowledge.
TW: How has the global ecosystem changed with regard to infectious disease research following SARS-CoV-2?
Flueckiger: The SARS-CoV-2 pandemic has definitely put a spotlight on infectious disease research and has highlighted the great disparities worldwide when it comes to access to health screening tools and services.
We have also seen a shift in how we implement NTD programs – for example, early in the pandemic we had to stop surveys for trachoma temporarily because safety protocols weren’t yet in place, and diagnosis for trachoma requires close face-to-face interaction. We’ve since pivoted and developed a loupes face shield, which pairs the special magnifying glasses used for trachoma diagnosis with face shields to keep people safe from COVID-19 while also continuing efforts to end NTDs. And now the TT Screener app has the potential to reduce the need for as many outsiders to come into communities to diagnose trichiasis, which could further help mitigate the spread of COVID-19. All in all, the pandemic has reminded us that our world is small, we are all human and all connected. We must work together and be able to quickly adapt.