In the rapidly expanding field of health technology, Health Tech World award-winner, Delft Imaging, stands out as a pioneer in leveraging advanced medical imaging and artificial intelligence to address some of the world’s most pressing health challenges. Founded in the early 2000s, with a focus on tuberculosis (TB) screening, Delft Imaging has evolved significantly over the decades, now reaching around 80 developing countries worldwide.Today its technology has directly enabled 20 million people to access TB screening, saving over 20,000 lives by identifying cases which otherwise would have gone undetected. Having been endorsed by the World Health Organization (WHO) in 2021, Delft Imaging was the first company to develop Computer-aided Detection (CAD) software that uses AI to detect TB from X-ray images, and its CAD4TB software is the most deployed on ground today, with CAD for TB detection being endorsed by the WHO in 2021.It is now turning its attention to maternal health, with 300,000 women losing their lives during pregnancy every year—over 90 per cent of which are in sub-saharan Africa.After it was named Health Technology Company of the Year at the HTW Awards in May, director Florent Geerts, reflected on the successes so far and shared his vision for global equity in healthcare. HTW: Delft Imaging has been around since the 20th century, can you tell me about the original vision and motivation behind Delft Imaging and how this has developed over the years?Florent Geerts (FG): Our vision is to bring equity in healthcare by ensuring everyone, regardless of where they live, has access to quality healthcare. We do this through a combination of medical imaging, such as X-rays and ultrasounds, and artificial intelligence, to aid in the detection of certain diseases and health burdens. We operate in about 80 countries, all of which are emerging.We can trace our roots to the first Dutch radiology equipment factory (ENRAF) and De Oude Delft (established in the early 20th century), but the current organisation was founded in the early 2000s, by Guido Geerts who was deeply motivated by addressing the challenges of TB. Now, we really want to focus on healthcare burdens that shouldn’t exist. While no one dies of TB in the Netherlands anymore, it still claims 1.5 million lives globally. Additionally, about 300,000 women die during pregnancy each year, 90 per cent of them in Sub-Saharan Africa. Our overarching motivation is to address these burdens.HTW: You touched on your move into the maternal health space there, can you tell me more about how this has come about and what it hopes to achieve?FG: We’ve been researching this area for about eight years now. Initially, we aimed to develop our own low-cost ultrasound device, which we allocated a lot of resources towards, but faced some challenges. Instead, we partner with ultrasound manufacturers while developing AI to analyse the ultrasound images. Our AI can currently identify gestational age, foetal presentation and placenta localisation. We are working on algorithms to identify more relevant parameters that can lead to severe maternal and antenatal complications in resource-constrained settings. We focus on settings with high maternal mortality rates, such as Sierra Leone, where we are working with primary health centres in rural areas to assess the experience of our users and the implementation of the solution.In all honesty, we are still figuring out the business model behind it. We invest heavily in research and development because we believe in the potential impact, but we also need to create a business model that can sustain itself for the future. We focus on impact first and always trust that the financial aspect of it will follow.HTW: What do you know about the human impact that this technology has had?FG: On the TB front, we have numerous publications confirming the accuracy of our software. We’ve shown that our technology can reduce programme costs by about 50 per cent while at the same time, screening two and a half times as many people. We’ve screened around 20 million people, and our calculations suggest that we’ve directly saved approximately 20,000 lives by identifying TB cases that would otherwise have been missed.HTW: With that in mind, what do you feel have been your biggest successes as a company?FG: One of our milestones was the WHO endorsement of our AI for TB screening and X-ray technology in 2021. But also the logistics of implementing these projects successfully in countries with limited road infrastructure, high temperatures, and high humidity has been another major achievement. For example, in Mozambique, we installed 54 X-ray systems with AI software at the same time. In Pakistan, we rolled out around 100 portable X-ray systems and 60 mobile ones last year. In Nigeria, we’re working on deploying 200 units. These efforts have made a substantial impact in countries with large populations, such as Nigeria. These are really some of our biggest successes.HTW: And what about the challenges?FG: I think these experiences also speak to the challenges of it. We really try to understand the specific contexts and obstacles our clients face, because they are not challenges we have in the Netherlands anymore, such as operating systems in extreme conditions with high temperatures and dust getting into the system. HTW: How has your work developed along with advances in health technology?FG: When we started with the development of AI back in 1996, it was not a hot topic like it is today. Over the years, we can say we single-handedly revolutionised screening and triage globally, with WHO endorsement in 2021 for using AI for TB. Many AI applications remain small-scale, but we have screened about 20 million people around the world, often in areas without robust internet access. While AI brings exponential change, we remain cautious. There’s a tendency to overestimate AI’s capabilities, but it doesn’t work for everything. However, AI has brought exponential change and the software has reached a level where its performance matches expert human readers. This allows us to deploy effective screening solutions in places where it would be impossible to station thousands of radiologists. This is something which technology has enabled that was not accessible before. HTW: What is your vision for the future of Delft Imaging?FG: In the future, we are looking at expanding AI applications to other diseases. For instance, we’re exploring AI for detecting silicosis, which is prevalent among miners and correlates with TB, as well as pneumonia and lung nodules. We’re also partnering on AI for retinal imaging to detect conditions like glaucoma and diabetic retinopathy. Our guiding principle is to eliminate burdens that shouldn’t exist, so it is this which really guides our thinking. Photo Courtesy of Delft Imaging
Sarah Sinclair
Delft Imaging: Making equitable healthcare a reality for millions around the globe
