Researchers at Florida Atlantic University combined Raman spectroscopy with machine learning to distinguish skin cancer from normal tissue with about 81 to 84 percent accuracy in a preliminary study, the university said on Tuesday.

The strongest machine-learning models correctly classified normal skin, basal cell carcinoma and squamous cell carcinoma about 81 to 84 percent of the time, according to results published in the Proceedings of SPIE as part of Advanced Chemical Microscopy for Life Science and Translational Medicine 2026.

The team used a mobile Raman system with a 785-nanometre diode laser and handheld probe to generate nearly 1,000 spectra from more than 50 clinical samples examined outside the body. Raman spectroscopy analyses how light scatters off molecules, providing a chemical fingerprint of tissue without removing or preparing it.

Skin cancer is the most common cancer in the United States. Nearly 1.5 million new cases were diagnosed globally in 2024, including nearly 340,000 melanomas, and 5.4 million nonmelanoma cases are diagnosed annually in the US, the university said. Biopsy followed by microscopic examination remains the standard for diagnosis, but biopsies are invasive, costly and sometimes performed on lesions that prove benign.

The promise of this technology is that it could give clinicians another way to look beneath the surface of a skin lesion without immediately having to remove tissue.

The quotation is from Andrew Terentis, senior author, professor and chair of FAU’s Department of Chemistry and Biochemistry. K-nearest neighbours and support vector machine classifiers achieved the highest overall test accuracy, at about 84 percent. The support vector machine recorded 78.7 percent sensitivity and 88.6 percent specificity, while a shallow neural network achieved 80.8 percent accuracy and the highest receiver operating characteristic area under the curve, at 0.910.

The models distinguished cancerous tissue from normal skin more reliably than they separated basal cell carcinoma from squamous cell carcinoma, whose molecular signatures overlapped more. Cancerous samples showed stronger protein-related signals while normal tissue showed stronger lipid-related signals, the researchers found.

The team said its findings were preliminary and larger studies, further optimisation and approaches such as deep neural networks would be needed before clinical use. “Ultimately, we want to develop technology that is not only accurate, but also practical, portable and accessible enough to become a useful tool in the clinical setting,” Terentis said.

Study co-authors are former FAU graduate student Venkata Dhulipalla, dermatology affiliate professor John Strasswimmer, and undergraduate students Phuong Nguyen, Lizzie Klein and Max McCain.