MUNICH, Germany – Aigora provided dermoscopic image data used to train the AI algorithm behind Dermalyser, an AI-based decision support tool for melanoma detection. The tool’s performance has now been evaluated in a prospective real-life clinical trial, published in the British Journal of Dermatology.
The study was conducted at 36 primary care centres in Sweden and included 228 patients with skin lesions of concern. Primary care physicians photographed suspicious lesions with a dermoscope attached to a smartphone, and the app assessed whether there was evidence of melanoma. The tool showed high diagnostic accuracy, with an AUROC of 0.960 for all melanomas and 0.988 for invasive melanomas. Its negative predictive value of 99.5% suggests it could help physicians avoid unnecessary excisions and referrals of benign lesions.
Most studies of AI for melanoma detection have been retrospective, and few have tested the technology prospectively in real clinical settings, particularly in primary care, where most skin lesions are first assessed. The trial adds this evidence and highlights the role of high-quality training data in developing reliable medical AI.
Read the publication in the British Journal of Dermatology
About Aigora
Aigora is a real-world data partner for life sciences and medtech companies. Through its network of healthcare partner sites across Europe and beyond, Aigora discovers, sources, curates and delivers multimodal, regulatory-grade datasets tailored to specific research and development needs.