Unveiling leprosy through machines: a review of artificial intelligence in a neglected tropical disease

Leprosy persists as a major public health challenge in many areas of the world, with nearly 200,000 new cases reported annually despite the success of multidrug therapy.Timely diagnosis remains pivotal to preventing disability and interrupting transmission; however, dependence on clinical acumen and variable diagnostic infrastructure continues to impede early detection.Recent advances in artificial intelligence (AI) herald transformative potential across diagnostic, classification, monitoring, and epidemiological dimensions.Convolutional neural networks and hybrid deep learning architectures have demonstrated diagnostic accuracies exceeding 90% in differentiating leprosy from phenotypically similar dermatoses, while explainable AI frameworks enhance interpretability and clinician confidence.Machine learning algorithms leveraging registry and questionnaire-based data enable reliable classification of paucibacillary and multibacillary forms, facilitating community-level triage.Integration of biochemical, spectroscopic, and geospatial analytics further supports therapeutic monitoring and targeted surveillance.Persistent challenges include limited dataset diversity, insufficient external validation, and unresolved ethical issues surrounding data governance, bias, and privacy.Future directions lie in federated learning, multimodal integration, and patient-centric digital platforms.The fusion of computational precision with human compassion may ultimately redefine early detection and accelerate global leprosy elimination.