Comprehensive Summary
The following study evaluated a real-time artificial intelligence (AI) assistant integrated with Line-field Confocal Optical Coherence Tomography (LC-OCT) to enhance basal cell carcinoma (BCC) diagnosis. LC-OCT is a high-resolution, non-invasive imaging modality capable of visualizing skin architecture at cellular depth, but its clinical adoption has been limited by the expertise required. In this multicenter, retrospective study, 43 dermatologists of varying experience levels participated in a randomized double-round quiz involving 200 equivocal BCC lesions from the four European hospitals. Diagnoses were made using clinical and dermoscopic (C&D) images, LC-OCT alone, and AI-assisted LC-OCT. The AI model, trained on over 685,000 annotated images, provided real-time decision support with interpretable heatmaps highlighting regions of diagnostic concern. Results demonstrated that AI-assisted LC-OCT improved overall diagnostic accuracy from 87.5% to 93.0%, with significant gains in sensitivity (+10.5%) and specificity (+2.5%), outperforming unaided LC-OCT and bridging a two-year expertise gap for less experienced clinicians.
Outcomes and Implications
This study establishes AI-assisted LC-OCT as a transformative advance in non-invasive skin cancer diagnostics. By enabling real-time “digital biopsies,” the system improves diagnostic confidence, enhances accuracy across experience levels, and minimizes the need for traditional histopathological confirmation. The technology shows particular promise in democratizing access to advanced imaging modalities, making LC-OCT practical for wider clinical use. Beyond diagnostic support, AI integration could facilitate automated surgical margin mapping, reducing re-excision rates and improving cosmetic outcomes. If validated prospectively, this approach may soon become a standard adjunct in dermatologic oncology, offering faster, biopsy-free diagnosis of BCC and potentially other skin cancers. The findings underscore how AI-human collaboration, rather than replacement, can enhance precision medicine, optimize workflows, and improve patient care outcomes.