To discuss the transition from selected-lesion classification to whole-body imaging in skin cancer surveillance using AI-assisted three-dimensional total-body photography.
Approach:
Literature Review: The review focuses on recent evidence regarding AI-assisted three-dimensional total-body photography (3D TBP) for skin cancer surveillance, including automated triage, lesion detection, risk prediction, and longitudinal tracking.
Key Findings:
AI studies primarily focus on lesion classification using cropped images, which do not align with real clinical screening needs.
Three-dimensional total-body photography captures a broader skin surface, enhancing lesion detection and monitoring.
Current applications of AI in 3D TBP include automated triage, multimodal risk prediction, and longitudinal tracking.
Evidence supporting the effectiveness of AI-assisted 3D TBP is limited, with challenges in skin tone reporting, workflow integration, and equity.
Interpretation:
Clinical value should be assessed by the ability of AI-assisted 3D TBP to enhance patient-level surveillance over time.
Limitations:
The evidence base for AI-assisted 3D TBP is limited and requires stronger prospective and longitudinal studies.
Key adoption issues include dataset transparency, false-positive and false-negative rates, cost, and equity.
Conclusion:
Research is needed to establish the clinical utility of AI-assisted 3D TBP in routine practice.
