Paediatric Oncology — ML for Hodgkin Lymphoma in Low- and Middle-Income Settings
Machine learning on affordable blood tests to replace expensive imaging in paediatric Hodgkin lymphoma.
A long-running collaboration with Prof. Jennifer Geel at the Charlotte Maxeke Johannesburg Academic Hospital (University of the Witwatersrand) and Prof. Monika Metzger at St Jude, focused on improving paediatric Hodgkin lymphoma care in low- and middle-income countries.
Machine learning on affordable blood tests to predict interim treatment response — work motivated by the prohibitive cost of mid-treatment imaging (PET, CT) in resource-constrained settings. Using longitudinal blood-test data we built models that anticipate PET response and could safely stratify patients away from costly imaging. Published in Journal of Clinical Oncology Global Oncology, Paediatric Blood and Cancer, and presented at the International Society of Paediatric Oncology (SIOP) where the work won a Global Health Network Award (Top 3 of 1,600 abstracts) in 2023.
Related work on national harmonised treatment protocols and outcomes for children and adolescents with classical Hodgkin lymphoma in South Africa.