Passion Dermatology: à pilot study for an intelligent self-assessment educational system

The alarming shortage of dermatologists in the world poses a real problem in terms of access to medical expertise for the treatment of skin diseases.The PASSION Dermatology project, led by the University of Basel in Switzerland, seeks to provide a solution by combining artificial intelligence with teledermatology.The aim is to develop a Machine Learning algorithm capable of recognising skin lesions with different types of pigmentation, using data from several European, African and Asian countries to form a federated learning system.For Madagascar, which joined the project in 2020, the data is collected by a team of dermatologists via a collaborative teledermatology platform.The team has sent a total of 323 photos of atopic dermatitis/eczema, dermatophytosis, impetigo and scabies to the server.Currently, a local study is being carried out to improve the collection system in order to speed up the volume of data with a view to reaching 1,000 photos per year for model training with local photos for these 5 most common skin diseases in Madagascar.Using the algorithm developed in Switzerland and India, the model was evaluated using a confusion matrix to give a precision/recall performance result of 0.92/0.93(n =2429).In addition, the Madagascar technical team is looking to improve this model and is currently developing an algorithm that will produce an intelligent computer-guided medical self-learning system, a new challenge for the African continent.