A new AI-powered tool, PanDerm, has boosted skin cancer diagnostic accuracy by 11% for dermatologists, according to researchers from Monash University and the University of Queensland. The tool analyzes multiple image types, including dermoscopy, pathology slides, close-up photos, and full-body images, simultaneously. In testing, it also improved diagnostic accuracy by 16.5% for non-specialists, offering valuable support in primary care and rural clinics with limited access to dermatologists. Trained on more than 2 million images from 11 institutions globally, PanDerm excels even with limited labeled data, making it highly adaptable to diverse clinical settings and imaging resources.
PanDerm Trained on Diverse Images to Mimic Dermatologist Reasoning
Unlike traditional AI models limited to a single task, PanDerm is multimodal, enabling it to process various image types just like a human dermatologist would. Its development was led by a team of AI researchers and clinicians across several countries and institutions, including Princess Alexandra Hospital, Alfred Health, and the University of Florence. First author Siyuan Yan explained that training the system across diverse imaging methods allows it to interpret complex skin conditions more holistically. The model supports multiple tasks: mole counting, lesion tracking, skin type classification, cancer risk prediction, and full lesion segmentation. Its strength lies in synthesizing cross-modal data to enhance diagnostic reliability.
This AI tool is especially powerful in detecting early changes in lesions that might indicate malignancy, which even trained professionals might miss. The team emphasized that PanDerm performed robustly even with minimal labeled datasets, a major advantage in settings where annotated medical data is scarce. By working in tandem with clinicians, PanDerm is designed not to replace human judgment but to strengthen existing diagnostic workflows, helping doctors catch warning signs earlier and provide consistent care regardless of location or resource constraints.
Real-World Applications and the Road to Clinical Integration
PanDerm has shown promise across multiple healthcare environments, particularly in resource-limited or high-throughput clinics, where timely diagnosis is crucial. Professor Victoria Mar of Alfred Health highlighted its potential in monitoring lesion changes over time and assessing metastatic risk, making it a key tool for managing high-risk melanoma patients. As an AI model, it’s not limited to static diagnosis, it’s capable of identifying trends and subtle lesion evolutions that may otherwise go unnoticed.
Its real-world utility stems from its compatibility with diverse diagnostic routines. Dermatologists often rely on a combination of close-ups, dermoscopic images, and pathology results to reach a decision. PanDerm processes all of them. It also accommodates the growing demand for teledermatology and remote diagnostics, making it relevant in a global context where 70% of the population faces some form of skin condition.
Currently, PanDerm is undergoing evaluation for broader implementation, with researchers focused on verifying its performance across demographic groups. The team is developing standardized assessment protocols to ensure equity in diagnostic outcomes, especially in multicultural and underserved communities. Collaborating institutions include NVIDIA’s AI Technology Centre in Singapore and Spain’s Hospital General Universitario de Alicante, signaling strong international interest in advancing AI-driven skin diagnostics.
AI Tool Revolutionizing Skin Cancer Diagnosis Globally
PanDerm represents a breakthrough in AI-supported dermatology, enabling faster, more accurate diagnosis of skin diseases using multi-modal imaging. With a foundation of over 2 million images and global clinical partnerships, it has the potential to standardize and democratize skin cancer detection across healthcare systems. Its ability to assist in early detection and patient monitoring positions it as a next-generation diagnostic tool. As evaluation progresses, PanDerm could soon become a critical asset in combating melanoma and other skin conditions, particularly in regions with limited specialist access. Its success may reshape how AI augments care in dermatology worldwide.
About The Author
