KNUST Researcher Explores How AI Could Improve Brain Scan Analysis in Ghana, Shares Findings
- Dr Prince Ebenezer Adjei of KNUST travelled to Germany to explore how AI tools perform when analysing brain scans from Ghanaian patients
- The biomedical engineer worked with roughly 1,600 head CT scans from Ghana to train AI models without requiring specialists to manually label every image
- Dr Adjei believes the technology could eventually help doctors interpret brain scans within minutes, particularly in hospitals with few radiologists
PAY ATTENTION: Mark YEN.com.gh as a preferred source, and our content will appear higher in your Google feed!
A Ghanaian biomedical engineer at the Kwame Nkrumah University of Science and Technology (KNUST), Dr Prince Ebenezer Adjei, is researching whether artificial intelligence can strengthen brain scan analysis and stroke diagnosis across Ghanaian hospitals.
Source: UGC
The researcher recently completed a visit to the German Cancer Research Centre (DKFZ) in Heidelberg, Germany, where he worked alongside scientists to understand how AI-powered medical imaging tools hold up when processing scans from Ghanaian patients.
KNUST published details of his work on Monday, November 9, 2026.
Testing AI on Ghanaian brain scans
At DKFZ, Dr Adjei applied a technique called self-supervised learning to roughly 1,600 unlabelled head CT scans sourced from Ghana.
The approach trains AI systems to recognise patterns in medical images without specialists having to manually categorise each one, which could significantly reduce the workload on already stretched health professionals.
He tested two imaging tools, nnFoundation and nnU-Net, on tasks such as identifying brain damage caused by acute stroke and estimating patients' brain age.
The preliminary results revealed a nuanced picture: extra AI training did not guarantee across-the-board improvements.
While nnFoundation outperformed nnU-Net in certain tasks, it did not do so when detecting stroke damage or estimating brain age from Ghanaian scans.
Dr Adjei, who leads the Global Health AI Lab at the Kumasi Centre for Collaborative Research in Tropical Medicine (KCCR) and lectures in KNUST's Biomedical Engineering programme, noted that the quality of training data shaped outcomes considerably.
Clearer, more accurate examples boosted the models' performance, while pinpointing smaller strokes remained a persistent challenge.
The path to AI-assisted stroke care
Dr Adjei envisions AI tools that could interpret brain scans in a matter of minutes and flag potential stroke cases to doctors, a capability that would be particularly valuable in facilities where access to radiologists is limited.
He cautioned, however, that the technology needs rigorous testing across multiple Ghanaian hospitals before it can be responsibly integrated into everyday clinical practice.
He also emphasised the need for robust data protection frameworks, supportive health policies, and proper training for medical staff, insisting that physicians must retain final authority over patient decisions.
The research was made possible through the inaugural Global Research Opportunities for the South (GROW) programme, which is funded by Germany's Federal Ministry for Economic Cooperation and Development.
Going forward, Dr Adjei's team is continuing its collaboration with DKFZ and Spectra Health Imaging Limited, with a focus on building a larger collection of high-quality Ghanaian brain scans.
The partnership aims to grow local capacity and advance AI tools developed with African data for African patients.
The X post by KNUST is below.
KNUST wins award to build health system
Earlier, YEN.com.gh reported on KNUST’s $5.1 million Takeda grant to launch PharmAdapt, a programme that will prepare Ghanaian pharmacists to address health risks linked to climate change.
The initiative aims to establish national standards for climate-responsive care, with the potential to extend its impact across West Africa. KNUST is one of four global projects sharing $22.2 million under Takeda’s 2026 Global CSR Programme.
Don't miss out! Get your daily dose of sports news straight to your phone. Join YEN's Sports News channel on WhatsApp now!
Source: YEN.com.gh
