AI for Skin Cancer Detection: A Smarter Way Forward?

The global incidence of skin cancer is on the rise, yet there’s a critical shortage of dermatologists, specialized nurses, and telemedicine experts. Can technology provide a solution? According to Prof. Gordon Wishart, Founder, CEO, and Chief Medical Officer of Check4Cancer, artificial intelligence (AI) might be the answer we need to revolutionize cancer pathways. 

During his keynote address at MedTech Malta 2024, Prof. Wishart emphasized, “By using AI in cancer and cancer pathways, we can speed up many of the pitfalls that slow down cancer screening or diagnostic pathways.” This vision aligns with Check4Cancer’s mission to use AI not just for detection but to enhance decision-making and optimize the patient and clinician experience.

Building on Proven Success

Check4Cancer’s approach isn’t new to leveraging data for clinical innovation. Prof. Wishart previously developed the PREDICT tool, aiding clinicians in treatment decisions post-breast cancer surgery. Now in its third iteration, PREDICT has demonstrated the importance of high-quality data, a principle guiding the development of Check4Cancer’s new AI model for skin lesion classification. 

Prof. Wishart highlighted that AI’s effectiveness relies heavily on data quality. For example, recent studies have shown that AI can boost breast cancer detection during mammography screening by up to 13%—a success attributed to the robust quality assurance in breast screening programs. “The message I want to get across is it’s not just about the quantity of data you use for AI or decision aids; it’s the quality of the data,” he noted.

The Proposed AI Model

Check4Cancer’s teledermatology pathway already shows promise. Currently, patients with suspicious lesions undergo assessment at nurse-led clinics, where high-quality images are captured and reviewed remotely. This system has impressive results: 80% of patients are reassured without needing further tests, and among those biopsied, 96% of melanomas are detected early (stage 1 or 2), significantly improving survival rates. 

Scaling this success requires innovation. Check4Cancer’s AI model integrates images captured via an app—whether in clinics or at home—with clinical data to assist healthcare professionals. This AI-driven decision aid aims to enhance the accuracy and efficiency of skin cancer detection, helping telemedicine reporters match the diagnostic prowess of seasoned clinicians.

Demonstrated Effectiveness

The AI model uses over 70,000 images from nearly 20,000 patients, developed through a partnership with the University of Essex and funded by Innovate UK. Testing has shown remarkable accuracy: the model’s sensitivity for detecting skin cancer using images alone is 98% but, the specificity increases from 63% to 80% when combined with clinical data. This specificity enhancement is crucial to avoid unnecessary face-to-face assessments. 

Compared to competitors and even human reporting, Check4Cancer’s model performs exceptionally. Benchmarking against NHS data and Cochrane reviews, the AI’s 98% cancer detection rate at 80% specificity surpasses many human assessments, suggesting AI could play a pivotal role in future diagnostics.

Future Prospects and Challenges

Check4Cancer’s AI model is more than just a diagnostic tool; it’s a step toward transforming global skin cancer care. Future plans include expanding regulatory approvals (UK, EU, USA) and addressing healthcare equity, particularly improving performance for darker skin tones. Prof. Wishart emphasized the importance of explainability, ensuring healthcare professionals understand and trust AI outputs. 

The company also aims to integrate its AI model with a skin cancer app, empowering patients to capture high-quality images at home. This vision of seamless collaboration between industry and academia exemplifies the potential of AI in healthcare: smarter, faster, and more equitable.

Takeaway

AI’s role in cancer detection represents a significant advancement, but as Prof. Wishart underscored, success hinges on data quality and strategic implementation. By combining high-quality clinical information with cutting-edge AI, Check4Cancer is setting a new standard in early detection—one that could redefine how we combat the growing challenge of skin cancer worldwide.

Secure Your Spot at MedTech Malta 2025—Early Rates End Soon!

Have you booked your tickets yet? Don’t miss out on early rates before prices go up. Join the MedTech community in Valletta, Malta, from November 12-14, and be part of the discussions that matter. Reserve your spot today!