The collaboration amongst the three institutions would see the utilization of existing classified skin lesion datasets, first from the University of Nigeria Teaching Hospital (UNTH) and the Lagos University Teaching Hospital (LUTH), in validating and training an A.I model.
Of the many promises of artificial intelligence, the potential to augment the capabilities of doctors by expanding their ability to provide more personalized and cost-effective care to patients, stands out most and especially, for healthcare in resource-limited environments.
For dermatology patients of African descent globally, this is the vision and promise of this tripartite collaboration initiated by Uburu Health -a Nigerian health tech startup building the data infrastructure for BigPharma and A.I R&D in Africa.
As expressed by Christoph Sadee, a biomedical informatics expert at Stanford University's Gevaert Lab,"AI for augmenting human capabilities has leaped forward in recent months with the emergence of ChatGPT and similar foundational models. This technology has the potential to revolutionize medical treatment in underserved communities and even in nations with insufficient access to healthcare. A prevailing challenge, however, is that such models are trained on datasets unrepresentative of these communities, which results in significantly poorer performance.."
The collaboration amongst the three institutions would see the utilization of existing classified skin lesion datasets, first from the University of Nigeria Teaching Hospital (UNTH) and the Lagos University Teaching Hospital (LUTH), in validating and training an A.I model, and then scale collaboration with more institutions in Nigeria and across Africa.
Prof Olivier Gevaert at Stanford's Biomedical informatics center says, "..a collaborative study between Stanford University and the Nigerian Institutions, to train a fair, unbiased foundational model, leveraging our pre-existing partnership; aims to: first, assemble a diverse skin lesion image dataset; second, jointly train a foundational model using Stanford's and newly gathered data; and third, test this model within Nigeria and African-American communities in the US.."
The model will be deployed and clinical trial tested for real-life patients, under supervision by dermatologists in Nigeria at the University of Nigeria Centre for Excellence for Clinical Trials (UNNCECT). A smartphone interface will be created to query a server hosting the model. Patients will then rate its usefulness, while dermatologists evaluate its correctness and accuracy. The findings will be published along with the collected data for future AI researchers.
Sign up for free AllAfrica Newsletters
Get the latest in African news delivered straight to your inbox
Prof Ifeoma J. Okoye, Director at UNNCECT and advisory board member at Uburu Health, reaffirms the potential impact this collaboration would bring to Nigerian patients who would benefit from faster, cost-effective and personalized dermatology care from their doctors. Prof Okoye welcomes the collaboration and urges for more research initiatives that drive value to African patients while addressing the global paucity of African datasets in A.I R&D.
Dr Nkiru Onodugo and Dr Ayesha Akinkugbe, who co-lead local experts and consultant dermatologists at UNTH and LUTH respectively, expressed excitement at the potential impact. Considering the paucity of dermatology specialists in Nigeria with a current ratio of one (1) dermatologist to 2.4Million Nigerians, a positive outcome of this collaboration would augment existing dermatology service delivery via better informed referrals and tele-dermatology. The latter, driving growth of Nigeria's (and Africas) digital health ecosystem.
