Innovative AI project launched to tackle skin disease diagnosis in Africa

In a significant development aimed at addressing dermatological healthcare challenges in Africa, researchers from the University of Basel, under the leadership of Professor Alexander Navarini, have launched the PASSION project, an initiative designed to harness artificial intelligence (AI) for skin disease diagnosis in regions with sparse medical resources. The project, officially known as Pediatric AI Skin Support In Outreach Nations, seeks to improve diagnostic capabilities in regions where dermatological specialists are exceedingly scarce.

The PASSION project comes in response to a critical shortage of dermatologists across many African countries. According to data, the ratio of dermatology specialists falls significantly short of the World Health Organization’s recommended one per 50,000 people, with less than one specialist available per million people in these areas. This inadequacy is most pronounced in rural zones, where an estimated 87% of children remain without treatment for various skin conditions.

An important component of the PASSION project is the creation of a comprehensive database that particularly represents skin diseases on dark skin tones. AI, which relies heavily on visual data for diagnostic purposes, often lacks a diverse set of images, with the majority of existing images being predominantly of light skin. This has created a gap in accurately diagnosing skin conditions on darker skin tones. To bridge this gap, researchers focused on collecting images of common skin diseases such as eczema, fungal infections, scabies, and superficial skin infections caused by streptococci and staphylococci.

Local dermatologists from Madagascar, Malawi, and Guinea contributed to the compilation of this extensive image repository, capturing over 4,200 images from approximately 1,300 patients, two-thirds of whom are under the age of 18. These images, taken between 2020 and 2023 with patient consent, have been meticulously annotated to include vital information such as diagnosis, patient age, gender, and affected body part, before being entered into an anonymized database.

The vision driving this initiative, as articulated by Professor Navarini, is to enable patients to self-diagnose by uploading images of their skin conditions via smartphones. The AI system would then analyse the images and provide treatment recommendations. Such a system, if proven effective, could be used in triage situations and initiate treatment, involving human dermatologists only when necessary.

Currently, the method is undergoing phased testing in Madagascar, where researchers aim to achieve a diagnostic accuracy rate exceeding 80%. As explained by Philippe Gottfrois, a doctoral student and lead author on the study, the goal is to eventually offer this diagnostic tool with thorough scientific oversight.

Future plans for the PASSION project include expanding the database to include neglected tropical skin diseases, hoping that AI technology will help alleviate the stark disparities in dermatological care across these underserved regions. The project’s findings and objectives were presented at the MICCAI 2024 conference in Marrakesh, highlighting the innovative approaches being taken to improve health outcomes in Africa.

Source: Noah Wire Services