Loading Curie Briefs...
Getting the latest healthcare news for you
Getting the latest healthcare news for you
AI-powered imaging may catch skin cancer before you can even see it. A new study found that an AI-assisted optical coherence tomography tool successfully identified subclinical basal cell carcinomas (BCCs) — lesions with no visible signs — in high-risk patients with a positive predictive value of up to 94.4%. Researchers say the approach could enable earlier, less invasive treatment, though questions about natural history and treatment thresholds remain.
AI-powered imaging may catch skin cancer before you can even see it. A prospective feasibility study published in JAMA Dermatology tested an AI-assisted imaging technique — line-field confocal optical coherence tomography (LC-OCT) — to detect subclinical basal cell carcinomas (BCCs) in high-risk patients. BCC is the world's most common skin cancer, and catching it early means less invasive treatment, especially on the face where it can cause functional and cosmetic damage.
The study enrolled 150 patients (mean age 72.9 years) with at least two established BCC risk factors at a German hospital. Researchers scanned clinically normal-looking facial skin and flagged suspicious lesions for biopsy. The AI system identified 18 subclinical lesions as BCC, with 15 histologically confirmed — and a sensitivity analysis bumped the positive predictive value to 94.4%.
By the Numbers:
Why it matters: These "subclinical BCCs" represent a newly described disease category with unknown natural history — some may progress, others may not. Experts caution that while early detection is promising, frameworks for treatment thresholds and long-term outcome data are still needed before widespread screening can be recommended.
