Access to cervical cancer screening in low- and middle-income countries can be improved with the help of AI, according to a new study from Karolinska Institutet, Uppsala University and the University of Helsinki, published in The BMJ. However, the technology requires support from functioning healthcare systems to be effective.
In many low- and middle-income countries, women lack access to screening for life-threatening diseases such as cervical cancer. Now, researchers from Sweden and Finland show that artificial intelligence (AI) can be used to analyse cell samples from the cervix at the primary healthcare level in resource-limited settings.
Projects in Kenya and Tanzania
The study is based on experiences from clinical projects in rural areas in Kenya and Tanzania, where researchers tested an AI-based system for cervical cancer screening. Cell samples from 3,000 women were digitised on site and analysed using AI, followed by review by medical experts. Local healthcare workers were trained to use the system, and the project was integrated into existing healthcare structures.
"Only a third of women worldwide have been screened for cervical cancer, even though serious disease is preventable," says the study's last author, Johan Lundin , professor at the Department of Global Public Health at Karolinska Institutet. "AI enables screening in low-resource settings where access to pathologists and laboratories is limited, and can shorten the time from sampling to diagnosis."
However, the researchers emphasise that the technology is not a universal solution. Without a functioning electricity supply, access to reagents, trained personnel and the trust of the local community, AI-based diagnostics risks becoming ineffective and widening health gaps rather than closing them. They suggest that AI-based diagnostics should be integrated into national health programmes and that the technology should be developed to work offline and with limited resources.
Must be adapted to local conditions
The researchers also emphasise the importance of assessing cell-based AI analysis in combination with molecular HPV testing in order to find the most effective way of screening for cervical cancer in resource-limited settings.
"The technology only works if it is adapted to local conditions," says Nina Linder, first author of the paper and guest professor of AI-based diagnostics at the Department of Women's and Children's Health at Uppsala University. "For AI to be truly beneficial, investments are needed throughout the entire healthcare chain. Our study provides important lessons on how AI can be implemented responsibly."
The study was conducted in collaboration with Kinondo Kwetu Hospital in Kenya and Muhimbili University of Health and Allied Sciences in Tanzania. It was funded by the Erling-Persson Foundation, the Knut and Alice Wallenberg Foundation and the Swedish Research Council, among others. Johan Lundin is a co-founder of Aiforia Technologies, but the company has no products for cervical cancer screening and did not contribute to the study.
Publication
"AI-supported diagnostic innovations for impact in global women's health" , Nina Linder, Dinnah Nyirenda, Andreas Mårtensson, Harrison Kaingu, Billy Ngasala, Johan Lundin, The BMJ, online 10 October 2025, doi: 10.1136/bmj-2025-086009.
