Dr. Josep Malvehy, from the Dermatology Department at Hospital Clínic de Barcelona, delivered an extensive presentation on the role of artificial intelligence (AI) in dermatology at the conference held in Thessaloniki in 2024.
His lecture focused on how AI is transforming dermatology, particularly in the detection and classification of skin lesions, with a notable emphasis on its application in melanoma diagnosis.
The presentation began with an introduction to AI, providing an overview of its impact across various fields in healthcare, including dermatology, radiology, ophthalmology, and pathology. Dr. Malvehy highlighted the rapid growth of AI in healthcare, citing projections that the global AI healthcare market will expand from USD 14.6 billion in 2023 to USD 102.7 billion by 2028. This set the stage for a discussion on the increasingly vital role of AI technologies, particularly in imaging-based diagnosis for dermatology.
Dr. Malvehy emphasized the importance of supervised learning in dermatology. He outlined the process by which AI algorithms are trained using large datasets of dermoscopic images labeled as benign or malignant, allowing the system to recognize patterns and classify lesions with high accuracy. He then delved into the evolution from basic algorithms to advanced techniques like convolutional neural networks (CNNs).
A key focus of the presentation was on AI's potential to outperform dermatologists in specific diagnostic tasks. Dr. Malvehy referred to several studies demonstrating that AI, particularly CNNs, have shown superior or equivalent performance to clinicians in melanoma detection. He discussed how these systems are trained to analyze dermoscopic and clinical images, ultimately supporting dermatologists in making more accurate diagnoses, thereby improving patient outcomes.
However, Dr. Malvehy also addressed the limitations and challenges associated with the integration of AI in clinical practice. He underscored concerns about the potential biases in AI systems, which may arise from imbalanced training datasets, as well as the ethical implications of relying on AI for critical diagnostic decisions. He advocated for the careful regulation of AI tools, particularly concerning data privacy, transparency, and the need for rigorous validation before these tools are widely implemented in healthcare settings (AI Thyessaloniki).
The presentation also mentioned the collaborative efforts needed to advance AI in dermatology. Dr. Malvehy highlighted the importance of interdisciplinary research, bringing together clinicians, computer scientists, and engineers to develop robust AI systems. He cited ongoing projects, such as the iToBoS initiative, which seeks to integrate multimodal data—including clinical, genomic, and imaging data—to create more personalized diagnostic tools for melanoma detection.
In conclusion, Dr. Malvehy expressed optimism about the future of AI in dermatology. He emphasized the need for continuous innovation and research to refine these technologies and ensure they are safe, reliable, and beneficial to both clinicians and patients. By the end of the session, participants had gained a comprehensive understanding of how AI is shaping the future of dermatological diagnosis, particularly in the early detection of melanoma.
Find out more at iToBoS at EADV Dermoscopy Course.