Curie Brief
Turn on cookies to sign in
Signing in saves your progress to your Curie account. We can only do that with cookies on — turn them on to continue.
News
Researchers at Florida Atlantic University are combining Raman spectroscopy — a technique that reads the molecular "fingerprint" of tissue — with machine learning to detect skin cancer without a biopsy. In a study of 50+ clinical samples, the best-performing AI models hit around 84% accuracy in distinguishing cancerous from normal skin. The team sees this as a potential tool to help clinicians decide which lesions actually need a biopsy.
Skin cancer is the most common cancer in the U.S., with 5.4 million nonmelanoma cases diagnosed annually — and biopsies, while the gold standard, are invasive, costly, and often performed on lesions that turn out to be benign. Researchers at Florida Atlantic University (FAU) are working on a smarter first step: using Raman spectroscopy, which analyzes how light scatters off tissue molecules to create a chemical "fingerprint," paired with machine learning to identify cancerous tissue non-invasively.
In their study, the team analyzed over 50 clinical skin samples — including basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and normal skin — generating nearly 1,000 Raman spectra. They found that cancerous tissue tended to show stronger protein-related signals, while normal skin showed stronger lipid-related signals. BCC and SCC were harder to tell apart from each other than from normal skin.
By the Numbers:
Why it matters: These are early-stage results, but the technology could one day give clinicians a rapid, portable, non-invasive way to triage skin lesions — reducing unnecessary biopsies and improving the patient experience. Larger studies and deeper neural network models are next on the roadmap.
