AI technology is increasingly being deployed in regions that lack sufficient medical expertise, according to Professor Lundin, who is currently collaborating with health authorities in Kenya and Tanzania. “At the moment, AI is being used where it’s least needed,” he notes, highlighting a critical imbalance in the distribution of AI resources. In these sub-Saharan countries, the scarcity of pathologists is stark; there are fewer than one pathologist per million people, compared to Sweden’s 30 to 40 per million.
Professor Lundin aims to leverage AI and digital technology to enhance access to image-based diagnostics in these underserved regions. The infrastructure for digitizing medical samples is already being adapted from the mobile phone industry, allowing data to be transmitted efficiently across locations. “The one assessing the microscope images could be sitting in another city, or even country,” he explains.
In an additional initiative, researchers are tackling cervical cancer, which is a leading cause of cancer-related deaths among women in many low-income countries. While the Nordic region has successfully implemented screening and vaccination efforts to prevent the disease, the World Health Organization (WHO) aims to have 70 percent of women of screening age tested globally by 2030. To meet this target, approximately 400 million more women need to be screened, a goal that may be unattainable without the assistance of automated methods.
At Kenya’s Kinondo Hospital, a team has begun screening over 3,000 women, with an additional 600 screened in primary care settings in Tanzania. A nurse collects swabs from the ectocervix, which are then placed on microscope slides for preparation and digitization. AI analyzes these samples, and the findings are verified remotely by an experienced pathologist. The researchers reported that AI’s accuracy matched that of human experts, and they have also identified the presence of the virus that can lead to cervical cancer.
Professor Lundin noted the challenges posed by regions where 25 to 30 percent of women are positive for the virus. Implementing standard screening methods can be overwhelming, making AI-mediated sample analysis a crucial intermediary to alleviate pressure on healthcare systems. Furthermore, intestinal parasites, a concern affecting around 1.5 billion people globally and particularly prevalent among children, are also receiving attention. AI has shown promise in enhancing the efficiency of diagnosing these neglected diseases.
In a study involving 2,500 schoolchildren, AI significantly reduced the time required to analyze faecal samples for worm eggs from 10-15 minutes per slide to just seconds. Notably, AI detected eggs in over 10 percent of cases that human microscopists had overlooked. “It’s like looking for a needle in a haystack,” Professor Lundin remarked. However, he cautioned that sample preparation can vary between laboratories, affecting AI’s effectiveness.
To address this issue, the AI models must be adapted to local conditions through standardized routines and quality controls. Professor Lundin recommends initially manually quality-assuring the first 50 to 100 samples to ensure accuracy. But he is optimistic about the future, suggesting that advancements may eliminate the need for traditional staining techniques currently used to aid human analysis.
He emphasized that responsibly deployed AI has the potential to reduce global healthcare inequality. “We’ve seen a great deal of support for these developments in low-resource environments, especially given the lack of experts,” he stated. “AI can’t just be high-tech for rich countries, as the greatest need is in low-resource countries.”
While AI has primarily been employed to detect anomalies in medical imaging, it is also being utilized to analyze data and uncover relevant medical patterns. For instance, Helga Westerlind, an epidemiology docent at Karolinska Institutet, is developing AI models to identify patient subgroups that respond distinctly to specific drugs. In her current project focusing on patients with rheumatoid arthritis, she highlights the inefficiencies in prescribing the immunosuppressant methotrexate, which often proves ineffective for about a third of patients within a year.
These initiatives illustrate the transformative potential of AI in global health, particularly in regions facing critical shortages of medical professionals. By adapting technology to local needs, researchers hope to enhance diagnostic capabilities and ultimately save lives. The integration of AI in healthcare may not only address immediate challenges but also pave the way for a more equitable future in medical access and treatment.
See also
NVIDIA’s GB200 NVL72 Achieves 10x Speed Boost for Mixture-of-Experts AI Models
Europe’s AI Data Centers Face Urgent Sovereignty, Energy, and Environmental Struggles
AI-Powered eSignatures Revolutionize Healthcare Approvals, Boost Compliance and Efficiency
Amazon Unveils Agent-Driven AI Strategy, Analysts Boost Price Targets to $340
Quantum-AI Convergence Looms: Altucher Warns of Major Tech Disruption Ahead
