Clinical field studies validating the use of AI-based digital diagnostics continue in East Africa with new major grant support

A project led by FIMM-affiliated researchers has received significant funding for validating and implementing artificial intelligence for diagnostics of cancer and infectious diseases in East Africa. Such methods will have a huge societal impact globally, especially in resource-limited settings where AI can help mitigate a critical shortage of pathologist and other experts.

A research team led by Johan Lundin (Karolinska Institutet and FIMM), Nina Linder (FIMM and Uppsala University),  Andreas Mårtensson (Uppsala University), Harrison Kaingu (Kinondo Kwetu Hospital, Kenya) and Billy Ngasala (Muhimbili University of Health and Allied Sciences, Tanzania) has received a research grant of 10 million SEK divided over three years from the Erling-Persson Foundation and a 4.5 million SEK grant from Vetenskapsrådet (VR, the Swedish Research Council).

The group has been developing applications that enable artificial intelligence (AI) supported automated diagnostics at the point-of care for several years. Their approach is based on digitizing the sample with portable microscopy scanners at the point-of-care followed by image transfer and remote diagnosis utilising both AI-analysis and human experts.  

“With these methods, biological samples can be analyzed with an accuracy comparable to a highly trained expert, at a fraction of the cost and time”, explains MD, PhD, Docent Nina Linder, who is leading the project related activities at FIMM.

The group has been carrying out field studies in several resource-poor African countries since year 2017. The results of these proof-of-concept studies demonstrate that both cervical cancer screening, malaria as well as soil-transmitted parasite diagnostics is technically and diagnostically feasible. Based on the work, several articles and one doctoral dissertation has already been published.

With the support from the Erling-Persson Family Foundation and VR the team will now be able to further assess the feasibility and clinical value of the methods. During 2022 – 2024 they plan to validate and implement the approach within routine healthcare and disease control programs in East Africa.

The group will conduct the field trials in Kenya and Tanzania in collaboration with the local clinicians and stakeholders. The studies will focus on validating the cervical cancer screening approach and evaluating the diagnostic performance of the AI-based methods to detect malaria and soil-transmitted parasites by comparing the results with those achieved with conventional methods.

Importantly, the research team is especially focusing on vulnerable patient groups, such as children and women with HIV.

 “Improved access to lifesaving diagnostics of cancer and infectious diseases is a global health priority. By taking advantage of recent technology innovations using AI this could be an achievable goal”, Dr. Linder says.

“This means a significant step towards a more equal and sustainable access to high-quality diagnostics in resource-poor countries.”

Dr. Linder is also active in health policy discussions and leads one of the World Health Organizations’s “AI for Health” Topic groups, focusing on point-of-care diagnostics. She illustrates:

"This WHO topic group will bring together researchers focused on the use of AI-based digital diagnostics at the point-of-care and allow joint efforts in standardization, development of guidelines and assessment of diagnostic accuracy of the novel methods ."


More information:

Press release from Karolinska Institutet AI delivers cervical cancer screening to rural areas of Kenya

Lancet: WHO and ITU establish benchmarking process for artificial intelligence in health